4Th October 2008 Surgical Oncology

Home , Mast cell sarcoma

Association of Veterinary Soft Tissue Surgeons Autumn Scientific Meeting 3rd – 4th October 2008 Surgical Oncology

The AVSTS would like to thank the following sponsors for generously supporting this meeting:

PROGRAMME

FRIDAY 3rd OCTOBER

9.009.30 Registration & Coffee

9.30‐10.15 Surgical Oncology – What Is It and Where Is It Going? Nick Bacon

10.15‐11.00 Surgical Margins and Getting the Pathologist to Nick Bacon & Evaluate Them. Tim Scase

11.00‐11.30 Grading Soft Tissue Sarcomas and Mast Cell Tumours Tim Scase – why Pathologists keep changing the systems.

11.3012.00 Coffee

12.00‐12.30 Soft Tissue Sarcomas – Anything New Worth Nick Bacon Knowing?

12.30‐1.00 Soft Tissue Sarcomas – Anything Else Worth Jonathan Bray Knowing?

1.002.00 Lunch

2.00‐2.45 Radiation Therapy for Soft Tissue Sarcomas: Susan North What Radiotherapists need to know from the Surgeons, Challenging locations and Outcomes of Incomplete Resection with Post‐ Operative Radiotherapy

2.45‐3.15 Soft Tissue Sarcoma Panel Discussion: Nick Bacon, Susan North, Jonathan Bray

3.153.45 Tea

3.45‐4.30 Canine Histiocytic Disorders: an Immunological and Steven Baines Oncological Perspective.

4.30‐4.45 Discussion

4.45‐6.00 AVSTS Committee meeting

6.007.00 Tour of Castle Caves

7.308.00 Drinks (in Bar)

8.00 Dinner

SATURDAY 4th OCTOBER

9.3010.00 Coffee

10.00‐10.30 Mast Cell Tumours – Anything New Worth Knowing? Nick Bacon 10.30‐11.00 Chemotherapy, New Molecular Targets for Diagnosis Richard Elders and Therapy in Mast Cell Tumours 11.00‐11.15 Discussion

11.1511.45 Coffee

11.45‐12.30 Maxillofacial Tumours in Humans Andy Burns

12.301.30 Lunch

1.30 Abstracts Session: 1.30‐1.45 Intra‐ and Post‐ Operative Electrochemotherapy Ron Lowe (ECT) in the Management of Canine and Feline Tumours 1.45‐2.00 Janos Butinar Reconstruction of Major Dorsal Nasal Defect induced by Intranasal Radiation with a Forehead Transposition Flap

2.00‐2.30 Use of Intra‐Pleural Chemotherapy for Management Ana Lara of Malignant Pleural Effusion

2.453.15 Tea 3.15‐3.45 Urinary Bladder and Urethral Tumours Nick Bacon

3.45‐4.15 Options for Prostatic Carcinoma Henry L’Eplattenier

4.15‐4.30 Discussion

CONTENTS Page

Surgical Oncology – What Is It and Where Is It Going? Nick Bacon 4

A Suregon’s Perspective on Margins Nick Bacon 11

Surgical Margins and Getting the Pathologist to Evaluate Tim Scase 15 Them: The Pathologist’s View

Grading Soft Tissue Sarcomas and Mast Cell Tumours – Tim Scase 19 why Pathologists keep changing the systems.

Soft Tissue Sarcomas – Anything New Worth Knowing? Nick Bacon 24

Soft Tissue Sarcomas – Anything Else Worth Knowing? Jonathan Bray 29

Radiation Therapy for Soft Tissue Sarcomas: Susan North 30 What Radiotherapists need to know from the Surgeons, Challenging locations and Outcomes of Incomplete Resection with Post‐Operative Radiotherapy

Histiocytes, Histiocytoses and Dendritic Cells: Stephen Baines 35 A Review of the Histiocytic Diseases of the Dog

Intra‐ and Post‐ Operative Electrochemotherapy (ECT) in Ron Lowe 56 the Management of Canine and Feline Tumours

Reconstruction of Major Dorsal Nasal Defect induced by Janos Butinar 60 Intranasal Radiation with a Forehead Transposition Flap

Mast Cell Tumours – Anything New Worth Knowing? Nick Bacon 61

Chemotherapy, New Molecular Targets for Diagnosis and Richard Elders 66 Therapy in Mast Cell Tumours

Use of Intra‐Pleural Chemotherapy for Management of Ana Lara 70 Malignant Pleural Effusion

Urinary Bladder and Urethral Tumours Nick Bacon 75

Treatment Options for Prostatic Carcinoma (PCA) Henry L’Eplattenier 81

What is Surgical Oncology and Where is it Going??

Nicholas Bacon MA VetMB CertVR CertSAS DipECVS MRCVS European Specialist in Small Animal Surgery Secretary of Veterinary Society of Surgical Oncology University of Florida Assistant Professor of Small Animal Surgery

It is a fact that 60% of human patients who are presently cured from cancer are cured by surgical resection alone and it is likely that the figure in veterinary patients is similar. In humans there are 2 broad but distinct groups of oncologic surgeons: the generalist, and the anatomic‐site‐specific. The general surgical oncologists are able to operate on most solid tumours, and have minimal experience or practice in benign disease. Those that are anatomic site specific retain the right to treat patients with complex benign disorders related to their area of interest, e.g. orthopaedic or musculoskeletal oncologists, who will treat an array of benign bony pathologies such as bone cysts, or osteomata. In humans, the perceived trend is towards growth of the latter group; however, generalists will likely persist given the nature and needs of health‐care globally.

Veterinary surgeons are familiar with the concept of general surgery and so veterinary surgeons with an interest in surgical oncology are typically involved in general oncologic surgery (benign and malignant) as the case load and client expectations do not require anatomic‐specific specialists. There is no formally recognised subspecialty in veterinary surgical oncology (as there is in medical and radiation oncology), despite over 50% of companion animals over 10 years old dying of cancer. This figure does not include those that are cured of cancer, or those that have cancer, but die from a co‐existing disease.

Potential for surgical oncology training is improving with certificates, diplomas, and residencies in surgery and post‐residency fellowships in surgical oncology growing in number and opportunity but it is important to combine any training with exposure to radiation and medical oncology – major advances in these fields are dramatically changing the face of cancer treatment and the role of surgery. Cancers considered unresectable with curative intent may be brought to potentially curable surgical resection with neo‐adjuvant strategies. Likewise long‐term outcome for patients following local cure of a solid tumour can be significantly improved with appropriate adjuvant strategies.

The founding father of human surgical oncology in the UK was Dr Stanford Cade. In 1940 he wrote:

“Successful treatment depends on three main factors: a sound knowledge of the disease; a wise selection of the method of treatment; and accurate and skilful technique.” 4

Knowledge of the oncologic condition, rather than simply knowledge of how to perform a surgical procedure is vitally important to successfully manage oncology cases and advise clients. Many people hold the belief that surgical oncology is becoming ever more aggressive and the patient and client considerations are becoming less important – people worry that ‘because we can’ is replacing ‘whether we should’. This may seem true in some select cases, but what is most likely happening is that through better pre‐operative planning, there is an improved understanding of the behaviour and distribution of the tumour, and so ‘big’ surgeries are being performed to effect a cure, where previously surgery would have been less aggressive and more likely to be incomplete. These larger surgeries are supported by improvements in pain management, anaesthesia, availability of blood products, and better understanding of adjunctive therapies.

At the same time, there is a growing body of research to challenge previous surgical dogmas about veterinary surgical oncology. An example is the rule concerning 3cm margins, which results in many patients being over‐treated by having large resections for low‐grade masses with small tumour volumes. In the future it is likely many patients will receive significantly LESS surgery than historically, with the same outcome in terms of local control.

Second to an improved understanding of tumour behaviour is the role of adjuvant therapies. Whenever and wherever possible multimodality cancer therapy needs to be considered. Not only does this include medical, surgical or radiation therapies, but also immunotherapy and potentially investigational therapies, such as veterinary clinical trials. If possible, cases should be presented regularly between members of the team – all members of the group will learn something from discussion of the treatment options. Some of these discussions may be with the client present in an attempt to identify the right path for them and their family.

The surgeon must have sufficient skill to succeed at the goal of surgery – typically to achieve local control of the tumour and to reduce the incidence of local recurrence to as low a level as possible. As much information as possible should be gained on local, regional or distant spread, in particular draining lymph node involvement and pulmonary metastasis. The surgeon must also ensure all of this is achieved with an acceptable quality of life, which leads to the last consideration of surgical oncology ‐ “when not to operate”. Many scenarios exist. Possibilities include: 1. Surgery resulting in significant residual local disease with insufficient benefit to the patient 2. Significant co‐morbidities existing, which are a higher priority than the tumour 3. Surgery itself is associated with severe or grave morbidity, subjectively worse than the symptoms of cancer 4. The patient is not expected to be discharged from the hospital, regardless of surgical outcome 5. Owner wishes hugely exceed realistic expectations 6. Owner wishes supersede patient’s best interests

These principles can be applied to something as simple as a lipoma. There are five reasonable scenarios to remove a lipoma – • rapid growth; • change of texture/feel; • owner concern; • causing clinical signs due to physical presence; • dog bothering the mass.

If at least one of these is satisfied, the surgeon is justified operating on the patient. It goes without saying that surgeons’ wishes do not feature on any of these lists, but sometimes surgeons’ wishes are thinly disguised as something else, especially if it is a surgery you are ‘desperate’ to try, or feel would be a suitable challenge.

An appreciation of tumour biology helps create a clear understanding of biopsy techniques and principles. Creating differential lists and obtaining good imaging studies prior to biopsy will guide the surgeon on where and how to biopsy, the benefits of each biopsy type and even if biopsy is indicated. An awareness of tumour margins, compartments and likely surgical plans will also improve the quality of the biopsy, and minimise potential complications associated with it.

A large proportion of oncologic surgery are cutaneous and subcutaneous resections, and familiarity, interest, experience and confidence in reconstructive surgery is paramount. Drains are not frequently employed unless extremely confident about clean tumour margins. Likewise, use of random local flaps or axial pattern flaps needs to follow a definitive curative –intent surgery, where there is little to no risk of residual disease. A knowledge of some of the more unusual tissue flaps is beneficial as the location and extent of wide tumour excisions can create defects in difficult locations – some creativity may be needed. A good example is oro‐nasal fistulae of the caudal mucoperiosteum/soft palate following maxillectomies. Reconstructive options such as bipedicle advancement flaps, temporal muscle flaps, and angularis oris APFs are all options. Cadaver work before electing to perform these flaps would be sensible.

A close working relationship between the surgeon and the pathologist is vital for good practice and to provide a high quality service. If possible, spending some time (even one day) with a pathologist to observe tissue orientation, trimming, mapping, and margin evaluation is invaluable to learn how to get a greater return from your surgical pathology. Watching the process of a receiving a formalinised specimen evolve to a macro‐ and microscopic report with clinical comments is an education to many and helps explain some of the uncertainties, frustrations and pitfalls that arise. There is an adage that ‘pathology cannot lie’. This is true, but is only relevant if: 1. the tissue submitted is the tissue of interest; and 2. the history, clinical findings and other pertinent information are available to the pathologist, otherwise it can be quite misleading.

‘Mass on leg’ or ‘spleen’ are simply not sufficient, and will result in a similarly

non‐committal pathology report.

Tumour Biology

Most solid tumours are graded (I‐III) in terms of their histopathologic characteristics – necrosis, differentiation, mitosis, invasion, lymphatic or vascular invasion etc... The grading information is then used to help predict biologic behaviour, but the converse does not always hold true. Occasionally histologically grade I tumours will be metastatic or locally problematic, a good example being the HiLo oral fibrosarcoma (FSA) ‐ histologically low grade, biologically high grade oral FSA. When creating a treatment plan for any patient, it is important to appreciate the interplay of grade, behaviour, and anatomic location, as the relative importance of each may differ between cases.

An understanding of local/regional metastasis, draining lymph nodes, propensity to metastasize and patterns of distant spread are vital in helping plan and offer a “best guess” of an outcome. You will never regret knowing what a mass is before surgery and this will help plan appropriate staging, interpretation of tests, and ultimately surgical planning.

Whereas solid carcinomas frequently metastasize to lymph nodes first, this would be unusual with sarcomas where secondary pulmonary metastasis is much more likely. The process of tumour metastasis is actually very inefficient, and each step is associated with a large loss of viable cells. Steps involved in spreading from primary to secondary sites include

1. cells must enter lymphatic or vascular vessels (intravasation) 2. cells disseminate to all body tissues via lymphatic or venous system 3. cells lodge and attach to vessel wall 4. cells migrate through the vessel wall (extravasation) 5. cells enter dormant state (esp. true for melanoma or osteosarcoma) 6. cells initiate new blood vessels developing locally (angiogenensis) 7. cells undergo progressive growth

These processes are all complicated and subject to specific interactions between cancer cells, their adhesion molecules, and the host organ vasculature, and endothelial surfaces. Many owners have fears that biopsying the mass before surgery will accelerate spread, and appreciating the complexity of metastasis and the steps involved will help to allay many owners’ fears that the risk is small, and countless millions of cells are often shed from these malignancies hourly even before biopsy.

Advanced / CrossSectional Imaging

CT and MRI are playing an increasing role in tumour staging, surgical planning, radiation planning, and outcome monitoring. It should be remembered that they can also lead to surgery being ruled out due to distribution and invasion of

tumour, which may not have been apparent with radiographs and/or ultrasound. Examples where CT/MRI is invaluable include large oral malignancies, chest wall resections, proximal limb sarcomas, and caval invasion with adrenal tumours or thymomas. New lymph‐node‐specific MR contrast agents (ferumoxtran‐10) becoming available may help in the non‐invasive detection of nodal metastases in normal‐sized nodes.

Micrometastasis of Solid Epithelial Tumours / Microscopic Staging

Improved surgical techniques and advances in diagnostic imaging have improved prognosis in many epithelial tumours by reducing mortality and lowering morbidity. Notable examples are anal sac tumours, liver tumours, urogenital tumours, and oral carcinomas. The fate of the patient is increasingly dependent on stage (i.e. TNM), and this usually determines whether systemic adjuvant chemotherapy is recommended to prevent or delay secondary metastasis. Early dissemination of neoplasia however is often not detectable by even high‐resolution cross sectional imaging or post‐operatively by conventional histopathology. With our reliance on the staging system, and the recognition that lymphatic or vascular invasion is of concern, improved staging techniques that are able to detect minimal cancer are of increasing importance.

In veterinary medicine, we often use ultrasound for staging of carcinomas of the perineum and urogenital region, and palpation and possible fine needle aspiration for tumours of the head and neck. Ultrasound relies on published estimates of size to predict normality (N0 vs. N1), with huge scope for error, and palpation of a deep node in the neck is very insensitive. In many human solid carcinomas, peripheral blood and bone marrow are also now being investigated, using sensitive immunocytochemical, immunohistochemical and molecular assays, which identify disseminated microscopic tumour cells remote from the primary tumour. The processes from which tumour cells reside within bone marrow or lymph nodes, and then become disseminated are not clear. This is part of the phenomenon of “tumour cell dormancy”, which refers to the time from tumour cell dissemination to the development of clinically overt metastasis.

One thing we know however in both humans and animals is that regional lymph node metastasis is one of the most important risk factors for disease recurrence in patients with completely resected solid tumours. There is great interest in the human literature regarding improving the identification of very small metastatic deposits (1 to 3 cells) in draining lymph nodes. Removal of a few sentinel lymph nodes with more intensive histologic analysis has indicated that as many as 30% of previously “negative” regional nodes by traditional histologic analysis may actually harbour occult metastasis or isolated tumour cells.

Studies in humans have demonstrated that detecting small tumour deposits in the nodes of patients with histopathologically ‘node‐negative’ disease is of prognostic significance in cancers of the breast, colon, stomach, oesophagus, prostate, pancreas, non‐small‐cell lung cancer, and melanoma. It is only logical to believe that veterinary species also have microscopic deposits of cells within

draining lymph nodes, and these too maybe predictors of biological behaviour.

Work currently underway at the University of Florida is exploring that question in dogs, by prospectively submitting draining lymph nodes from solid carcinomas for both histopathology and immunohistochemisty. Nodes are being excised as a staging tool whether or not they appear enlarged on imaging, or are palpably enlarged. Early findings include sublumbar nodes diagnosed ‘normal’ by ultrasound containing heavy tumour burdens from the urogenital tumours they drain, plus detecting deposits of carcinoma cells at the margins of normal sized lymph nodes. In humans, finding tumour cells in afferent vessels or sinuses is considered a precursor to nodal metastasis. Once the tumour cells escape from the sinus, they invade the parenchyma and start to induce stoma cell reactions. Consequently parenchymal involvement is a significantly worse prognosis than subcapsular or marginal involvement.

Research is being pursued in human breast, kidney and prostate cancer regarding the impact of finding circulating tumour cells in the peripheral blood (so‐called ‘blood staging’). Laboratory techniques can isolate as few as a handful of cells per 5ml of blood, and can potentially be used to stratify patients into different stages, and therefore treatments. Identifying circulating tumour cells promotes debate into their significance ‐ are these true metastatic deposits, or are they just cancer cells in circulation (part of the metastasis pathway) which are non viable, with no ability to attach, invade or undergo progressive growth? i.e. these cells could be a hallmark of more aggressive malignancies or indicate a poorer prognosis, but not themselves ultimately lead to death from disseminated disease.

The consequence of increased efforts to look for microscopic tumour metastasis is ‘stage migration’, i.e. it becomes progressively more difficult to diagnose a cancer patient as TxN0M0, as increasingly they are diagnosed as TxN1M0, TxN0M1 or TxN1M1 as small tumour deposits are found which would previously have been overlooked. Stage migration makes analysis of contemporary cancer patients difficult as more and more detail is learnt from new, sophisticated diagnostic testing, resulting in the ‘established’ knowledge of prognosis (and recommendations for treatment) from historical cases becoming harder to compare. It means only tentative comparisons must be made between new and old drug protocols, and new and old surgical studies.

Although stage migration does not impact the total patient population, it can have profound impact on curability by stage. For example – ultrasound and needle aspirate identifies a lymph node as negative. If the node is excised and malignant cells are subsequently found in the node by IHC then the patient might move from stage I to stage II, or N0 to N1 depending on the cancer. However stage II (N1) would also be the same stage as dogs with huge lymph nodes effaced and replaced with metastasis, and the former has an excellent prognosis compared to the latter. Removing dogs with possible adverse prognostic factors from stage I also means the curability of stage I patients increases. In actual fact the prognosis in both groups is likely improved (“Will Rogers” effect). If a new drug/surgery/therapy targeting stage I or II is introduced at the same time

as this stage migration, then this will bias assumptions about its value, as it may appear to perform significantly better compared to previous reports or historic controls.

Where to next…..?

Performing bigger surgeries in the future to reduce the risk of local recurrence is only really justified if the current recurrence rate for a particular solid malignancy is high (>25%). This must also be balanced against the cost, morbidity and reduced function of the larger surgery. Certainly before that is done the role of adjunctive radiation could be explored in more detail in our patients. As mentioned previously there is likely to be new research on REDUCING margins of normal tissue to achieve the SAME rate of recurrence, rather than INCREASING margins to LOWER the recurrence rate. It may be that smaller but carefully quantified margins of tumour‐free tissue may be as adequate as wide excisions. In veterinary patients with short life spans, many of our patients will die from other disease before dying from recurrence or tumour‐ related disease and that is an important consideration in our decision‐making. There should not be a blind quest for clean margins in every patient. It is not the over‐riding goal of surgical oncology and treatment should be tailor‐made for the individual and their situation.

Other areas of expansion in surgical oncology are likely to include sentinel node studies, lymph node mapping, and non‐invasive lymph node characterisation. Surgeons need to be familiar with new and emerging therapies, for example anti‐ angiogenic drugs, which shut off the blood supply to tumours leading to an elimination or reduction of tumour volume. Development of local therapies to reduce local recurrence (with a smaller surgery) are also likely to evolve, to include interventional oncology, intra‐arterial chemotherapy, chemo‐ embolisation, and isolated limb perfusion.

Radiation will evolve, for example stereotactic radiosurgery can deliver a highly conformed dose of radiation to a target in a single treatment, with little to no peripheral damage. Tissues which have been treated to date include brain, nose, skull, skeleton, urethra and prostate. The availability of radiation will increase as will the understanding and appreciation of its roles in local cancer control.

A goal of future systemic treatment is the definition of specific categories of patients or cancers that will benefit from systemic therapy, and those that will not, and eliminate the recommendation of universal non‐specific toxic chemotherapy for some cancers. One example is appendicular osteosarcoma. Two similar patients can present with volumetrically‐equivalent distal radial osteosarcoma. There can be no nodal disease, and bone scans and thoracic radiographs can be clean at time of amputation. One patient could receive chemotherapy and suffer lung metastases at 150 days. The other could receive chemotherapy and be tumour free at 600 days. Arguably the first patient could be spared the time, expense, and toxicity of chemotherapy for a comparatively small increase in survival over amputation alone.

A Surgeon’s Perspective on Margins

As far as the surgeon, owner and patient are concerned, an awareness of surgical margins is vital to effect local control and so maximise the chance of a long disease‐free interval. Not all surgeries are curative‐intent in nature, but even those that anticipate likely leaving tumour behind will benefit from understanding how to minimise the volume of residual tumour, potentially to a level too low to recur in the patient’s lifetime.

Surgical margins are defined by: (a) the tissue plane through which dissection and excision is done; and importantly, (b) the actual (or perceived) residual neoplastic disease in the wound bed. Classification of the margin achieved around the mass will help with judging the effectiveness of a surgical procedure in achieving local control of a tumour. Classification makes comparisons between different techniques possible, and rapidly communicates the intent of the surgery.

Expanding tumours are typically histologically circumscribed but un‐ encapsulated, with a poorly defined surrounding reactive zone, consisting of some or all of; • a vascular response (new blood vessels); • mesenchymal response (to physical presence of the tumour and abnormal local tissue forces); • inflammatory response (to necrosis / haemorrhage / degranulation of mast cells).

This reactive zone may be several millimetres in width in smaller low grade tumours, but several centimetres in high grade tumours. The reactive zone creates a visible and palpable ‘edge’ to the tumour mimicking a fascial plane, but microscopically or immuno‐histochemically it is not ‘all or nothing’, instead being a three‐dimensional ‘halo’ of malignancy.

Curative oncological excision to resect this zone is then further complicated by tumour site, anatomical constraints, maintenance of local function, biological characteristics of the tumour, and length and cost of surgery and reconstruction. Following excision, all cut edges of the tissue (skin margin and deep margin) should be stained with India Ink, left to dry, and then placed into 10% formalin to fix, at a ratio at least 1 part mass: 10 parts formalin. On processing and interpretation, identifying tumour cells in ‘inked’ tissue represents tumour at the wound edge, and so an incomplete excision. Larger tissue masses may need to be

‘bread‐loafed’ prior to fixing, as the formalin cannot diffuse to a depth of greater than 1cm tissue. Alternatively some surgeons prefer to tie small suture knots at locations where concern exists over the width of the normal tissue margin, to encourage the pathologist to examine these areas more critically.

Our understanding of margins is based largely on Enneking’s pioneering work in musculoskeletal tumours in humans, whereby he classified them as intralesional (intracapsular), marginal, wide or radical.

• An intracapsular margin is achieved by piecemeal removal (‘debulking’) of a lesion from within the capsule. This is also used if the capsule is accidentally entered during dissection as the surgical field is now contaminated. Gross and/or microscopic disease remains. Examples include incisional biopsy, curettage of bone lesions, infiltrative lipomas, and some hepatic masses.

• A marginal margin is achieved by an extracapsular dissection through the reactive zone around the mass. Classically these are termed ‘shell‐ outs’ and involve peeling the mass out from its tissue bed and off local attachments. Both benign and malignant lesions may have extracapsular microextensions of disease, microsatellites (in the reactive zone, e.g. mast cell disease), and ‘skip’ metastases of high‐grade lesions (in normal tissue of the same compartment e.g. soft tissue sarcomas). These both have implications for marginal excisions in terms of potential for local recurrence.

• A wide margin is achieved by en bloc removal of the lesion, its capsule and the surrounding reactive zone but always working in normal uncontaminated tissue within the compartment of the lesion. Non‐ neoplastic, non‐reactive intracompartmental normal tissue is left at the margins and there is the possibility of ‘skip’ metastases arising in the remaining portion of the compartment (e.g. synovial cell sarcomas, osteosarcoma).

• A radical margin removes the lesion, reactive zone, and all the tissue of the associated compartment. There is no potential for residual neoplasm locally. The typical example is amputation, along with variants such as hemi‐pelvectomy.

It is important to realize the surgery is defined by the margin it achieves, not the surgery itself, i.e. a hemipelvectomy is typically considered a radical surgery if extracompartmental, but if the mass is proximal and the reactive zone around the mass is entered it would be classified as a marginal excision. The margin is also defined by the least margin at any portion. A dissection that is 95% wide and 5% marginal is a marginal excision, i.e. excising 3cm of skin around a mast cell tumour will still be classified as marginal if the pseudocapsule is entered on the deep surface, the commonest site for incomplete excision.

This classification gives us an indicator of the completeness of surgical excision, but how generous this margin of normal tissue around the reactive zone of a solid tumour should be has not been defined for most veterinary tumour types or grades. There are however three fundamental approaches to delivering a surgical margin;

• the Metric Approach, whereby the mass is visualized as a homogenous tissue in three dimensions and is resected following direct measurement beyond the visible or palpable edge of the mass. The metric approach is most commonly employed for small to medium sized cutaneous or subcutaneous masses. Examples include margins of 2cm for grade 2 mast cell tumours, and 3‐5cm for soft tissue sarcomas or feline vaccine sarcomas.

• the Barrier Approach relies on conceptualizing the mass being constrained within anatomical boundaries and dissecting up to an uninvolved barrier, or a barrier with great functional significance (e.g. sciatic nerve, vena cava, spinal cord). Many procedures are in fact a combination, i.e.

• the Metric/Barrier Hybrid, especially in areas where there is little non‐ functional tissue e.g. head, neck, thorax. Excision of cutaneous/subcutaneous masses often also relies on this hybrid approach, with a superficial metric skin margin, and a deep barrier margin of an uninvolved fascial plane. This approach often leads to a histopathological report of several cms clearance laterally, but only 1‐2 mm clearance deep.

Histopathology reports are often most easily interpreted by the surgeon who performed the surgery, and adequate communication with the pathologist who reviews the sample is needed. Wherever possible as much information should be given to the pathologist at time of sample submission, including tumour location, orientation, any areas in the tissue block of concern, what sutures and inking represent, and if possible thin masses such as bladder, mucosa, nerve could be fixed to cardboard to minimize shrinkage and curling artefact. The pathologist should be contacted beforehand to clarify the colour of India ink they prefer (red is notoriously hard to interpret!) and when inking rules such as thin application, allowing to dry, avoiding tissue plane slippage and no ‘double‐dip’ should be followed.

The list of what we don’t know about tumour margins is significantly longer than what we do know. We reach for the generic ‘3cm’ distance for most malignancies, often with little consideration of tumour biology, natural history, clinical behaviour, and anatomic constraints. There is not enough long‐term follow‐ up in the veterinary literature to guide future decision‐making in terms of ‘how much is enough’. The 3cm rule has been challenged recently in regards to mast cell tumours and soft tissue sarcomas, and that is described in the relevant lecture notes. For many tumour types however we are best‐guessing based on historical anecdote, i.e. discrete intestinal malignancies, liver tumours, anal sac

tumours, bladder masses, the feasibility of partial splenectomies, nephrectomies or lung lobectomies, effaced lymph nodes and local recurrence, significance of capsular invasion in thyroid carcinomas, rib resections, acetabulectomy versus coxo‐femoral disarticulation for femoral head and neck bone tumours, canine versus feline nosectomies for SCC.

The concept of appropriate margins is also linked to the quality of surrounding ‘normal’ tissue. It may be more important to focus on a bone margin for bone‐ originating tumours than a soft‐tissue margin. This is demonstrated by limb sparing surgery where a wide bony margin is married to a marginal soft‐tissue margin, and yet local recurrence is rarely seen unless the capsule is inadvertently broached or a biopsy tract left behind. In chest wall resections for rib tumours, the focus of surgery has always been to take a healthy rib cranial and caudal to the mass and a wide 3‐5cm rib margin dorsally and ventrally. It is assumed one fascial plane of soft‐tissue overlying the tumour will be an equivalent barrier to local spread. In addition to tissue type, few publications address the issue of tumour grading or size in decision making on width of margins. Increasingly histopathologists are being asked to grade tumours, as our knowledge of its impact on prognosis and adjuvant therapy increases. This grading will in future also impact the initial surgical plan.

The impact of previous biopsy on surgical planning and width of margins is often underestimated. Whereas biopsy is the cornerstone of treatment, biopsies are not benign and the benefit needs to be weighed against the local risks. Location of skin puncture, direction of tract, direction and extent of regional or dependent haemorrhage potentially carrying neoplastic cells, adherence of local tissues from fibrosis post‐biopsy and the consequence disrupting a virgin capsule all need consideration. This is especially true in percutaneous biopsies of thyroid masses, lung masses, deep chest wall sarcomas, splenic lesions or urogenital masses.

Despite accurate pre‐surgical imaging and an appropriate surgical plan, you can still end up with unexpected ‘dirty’ margins. So what do you do? Whereas large studies help us predict the likely behaviour in 25, 50 or 100 patients, the problem rests with our poor ability to predict the behaviour of one individual’s tumour. Of particular significance and concern is the observed phenomenon of local recurrence after wide tumour negative margins have been achieved, or following wide resection of scars containing too few tumour cells to even be detected histologically. Additionally, a causative relationship between local recurrence and systemic metastasis has yet to be elucidated, and so the importance of avoiding local recurrence at all costs is not always clear.

It is suggested that patient prognosis is dictated primarily by a multitude of independent biologic factors such as tumour grade, size and depth of invasion. Local recurrence is thought to serve more as a marker for, rather than a cause of diminished survival in patients with malignancies. As our individual experience increases, and our understanding of tumour biology improves, we hope we will soon be able to make decisions based more on the nature of recurrence rather than the fact that it has occurred at all.

Surgical Margins and Getting the Pathologist to Evaluate them: The Pathologist’s View

Tim J. Scase BVM&S, PhD, DipACVP

The evaluation of surgical margins as an indicator of the effectiveness of surgical excision can often present a challenge to clinicians, pathologists and histology technicians.

The evaluation of surgical margins represents a trade off between what is most practical and cost effective for an individual case, because the entire tissue cannot be easily examined. For instance if the entire specimen of a 3x2x1cm elliptical piece of skin, was cut into 5 micron sections, it would take approximately 4000 histological sections to examine the entire tissue. Unless you have both a very rich client and a very patient pathologist, this is not going to happen!

Consequently any practical evaluation of surgical margins involves some pragmatism and degree of judgement as to which of the many orientations or planes of sections are the most likely to provide the information. In human medicine, it has been recognised that it is very important to have a pathologist involved in the evaluation and dissection of a gross specimen right from its initial receipt into the lab. This should enable the most appropriate sections to be taken for subsequent microscopic evaluation.

In human medicine there are set guidelines for how tissue specimens should be examined and processed (provided by, for instance, the Royal College of Pathologists in the UK and the College of American Pathologists in the US). Even in human medicine there is a degree of controversy about how best to take the margins from any given tissue/tumour type, and indeed for some tissue the guidelines are much less stringent than for others. For instance the breast tumour guidelines are stringent whilst the CNS tumour guidelines are very vague indeed. In some areas (i.e. for some endocrine tissues) there are no guidelines at all!

However, even where these guidelines are in common usage, the application varies. For instance a recent study (Apple SK, 2006) looked at pathology reports from 91 consecutive breast cancer cases over a 2 year period from 50 hospitals in the US. Of these, there were problems with both the surgeons submitting the tissues and the pathologists reviewing the slides. For instance, only 65% of surgeons had marked the orientation of the tissues (medial, lateral, dorsal, etc) and only 30% of the tissues had been submitted in toto. Only 18% of the pathology reports stated how the margins were taken, and only 76% had margins reported by the pathologists. This suggests that it is not just in

veterinary medicine where we have these types of problems!

The American College of Veterinary Pathologists is currently forming a working party on surgical biopsy margin evaluation, although they have yet to publish their findings or discussions. Until this happens, there will continue to be no agreed guidelines for the pathological evaluation of surgical biopsies in veterinary medicine.

Lessons can be learnt and appropriated from the medics, and therefore, for the remainder of the discussion, the Royal College of Pathologists guidelines for the examination of surgical margins will be used. The guidelines can be divided into different sections, from samples being taken by the surgeon to final microscopic evaluation by the pathologist.

1) Surgeon

a) The surgeon needs to provide some sort of description of the case, signalment (human version of species, sex, breed, age!) and history, previous treatment, etc.

b) Tissue orientation needs to be marked. It doesn't matter too much how this is done (e.g. suture tags, staples, inks). As long as the meaning of the markings are written on the submission form, and are readily and individually identifiable, this is usually straightforward.

c) If large tissues are submitted, then it is preferable to have rapid fixation over tissue orientation, as areas of inadequate fixation can prevent meaningful tumour grading or application of immunohistochemical markers at a later date. The tissue should be incised such that the skin surface (where there is one) is left intact. Keeping an anatomical surface intact should still enable the tissue to be orientated following fixation.

d) If the tissue needs to be submitted in formalin, then the tissue should be dunked/painted in some type of ink. For fatty tissues (ie: mammary gland, skin with a lot of subcutaneous adipose tissue) dunking it in alcohol first, drying it and then painting it will increase the amount of ink that will stick on. Many different inks are available from proprietary surgical paints through to Indian ink, alcian blue or dyed gelatine. For Indian ink, 'fix' the ink by dunking it in 10% acetic acid prior to putting into formalin to stop it being washed away. For the pathologists at the microscope, prior inking of the tissues makes a MASSIVE difference to how easy it is identify the 'real' surgical margins. It is surprising how tissues can get flipped around in cassettes without anyone realising, and therefore without inking there is great the potential for both false positives and negatives to arise.

2) Receiving pathologist.

a) The specimen is weighed (not routinely done in veterinary medicine) and measured (routinely done in veterinary medicine).

b) The specimen is then sliced at 3‐5mm intervals throughout its entirety. Cassettes are then labelled and the sections placed into them so that they can be re‐orientated at the microscope. The slicing of the specimen can be done in a number of different ways depending on the specimen:

i) Skin ellipses ‐ slices taken transversely across the specimen, such that the narrowest margin will be included. ii) Large specimens ‐ slices taken from medial to lateral; from internal to external or for large round specimens using a cruciate cut +/‐ shaved margins.

c) An alternative method for taking margins is to shave off slices from around the tissue periphery and embed those separately in labelled cassettes. This enables more of the marginal tissue to be examined in fewer slides, but prevents accurate measurement of the tumour‐tissue edge distance. This distance has prognostic value in some tumours. For instance this measurement in ductular carcinoma in situ in human breast cancer, is highly associated with local tumour recurrence (P=0.006; (Apple SK, 2006). Alternatively, in intestinal resections, coronal margins of intestinal segment ends are not taken when a tumour is >30mm away from the segment end, as they are so rarely involved when the tumour is that far away.

d) Another alternative method is for the surgeon to provide further 'cavity shaves' of tissue from the surgical bed, which are embedded separately.

e) If the tissue is small (i.e. <1cm in widest diameter), the entire tissue is blocked up for histological sectioning.

f) In general this will lead to a lot of tissue sections ‐ some breast cancer papers quoting a median of 13 sections per case, with a range of 3‐47.

In most veterinary diagnostic laboratories examination of margins may not be as thorough as this; indeed, in the majority of commercial diagnostic practices, trimming of gross tissues is performed by a technician. The 'standard' method for tissues where it can be done, is to take ‘cruciate’ sections (North, East, South, West) such that margins can be assessed around the tissue. However, this process does not work particularly well in all cases, and particularly where there are large specimens, or large specimens from anatomically complex sites (ie: head/neck). Therefore, peripheral 'shaves' may be more appropriate in these cases, aided by pre‐fixation inking.

In my opinion, the most cost effective method of margin assessment in veterinary medicine is for the surgeon to take 'cavity shaves' from the tissue bed that is left behind after tumour removal, and for these to be submitted separately. This takes away some of the variation in the different ways that tumours are handled in different labs, and gives your pathologists something to concentrate their efforts on! In addition, when done in combination with the standard ‘cruciate’ sections through the main tumour mass, useful measurements can still be obtained for the tissue margin/tumour margin distance.

It is also possible, however, that as it stands, veterinary pathologists are misdirecting their efforts into providing great histopathological descriptions and comments. In human medicine, the histological descriptions are small if present at all, with the vast majority of the pathology report received by the clinician describing the gross lesion, with the remainder made up of tick boxes of different tumour types, grades and features: particular breast cancer type; nuclear grade; degree of necrosis, total size of tumour; aggregated grade; presence or absence of vascular invasion; etc.

As it is much more difficult to trim in more tissue from a case that has already been cut into pieces, the most effective way to obtain meaningful information from your veterinary pathologist on surgical margins is to choose a good lab and to talk to the pathologists there about your requirements at the time of sample submission.

Further Reading:

The Royal College of Veterinary Pathologists’ tissue preparation guidelines can be found at: http://www.rcpath.org/index.asp?PageID=254

Apple, SK (2006) Variability in Gross and Microscopic Pathology Reporting in Excisional Biopsies of Breast Cancer Tissue The Breast Journal 12 pp145‐149

Abstract: Accurate and complete information in pathology reporting is essential since most breast cancer treatment decisions are based on pathologic findings. The College of American Pathologists (CAP) has guidelines for breast cancer reporting; however, pathology reports remain variable. Data were collected on 91 consecutive breast cancer excisional biopsies from "outside slide review" (OSR) cases for a 2-year period to determine the variability in pathology reports in gross and microscopic examinations from 50 different outside community and university hospitals located primarily in the southwestern United States. From the gross pathology report, the following items were analyzed: measurement and weight of specimens, orientation provided by surgeons, number of blocks submitted, designation of margins, and whether margins were indicated as "shaved" or "perpendicular" in relation to the breast tissue at the time of grossing. From the final diagnoses, the following items were analyzed: type and size of tumor, and surgical margins. The results show that 100% of the reports documented the measurement of specimen size, and 30% documented the specimen weight. Surgeons provided orientation of the breast specimens in 65% of cases. Surgical margins were inked in 58%, while only 18% described how margins were submitted (either shaved or perpendicular to the mass). Only 30% of specimens were submitted in toto, 1% were submitted with an unknown amount of tissue, and 69% were submitted in representative sections with an average of 13 blocks for lumpectomies. In the final diagnoses, all reports had documentation of the tumor type and size of the invasive cancer; 26% of the final diagnoses had ductal carcinoma in situ (DCIS) and just 5% of those reports documented the size of the DCIS. The surgical margin status was reported in only 76% of the final diagnoses. This study shows that the pathology reports were heterogeneous with respect to reporting gross and microscopic final diagnoses from the variable hospitals.

Grading Soft Tissue Sarcomas and Mast Cell Tumours: Why Pathologists keep Changing the Systems!

Tim J. Scase BVM&S, PhD, DipACVP

Tumour grading systems are by their very nature only developed on the basis of current knowledge and when that knowledge base changes, when managements options change or when larger or more well designed studies are performed, the grading systems may often be shown to be incomplete, inadequate or inaccurate. Hence any grading system has to be considered a ‘work in a progress’ and it should come as no surprise that the schemes change!

Soft Tissue Sarcomas (STS) The current histological grading scheme for STS is the same as that used by human pathologists (Trojani et al, 1984), although the number of publications that have confirmed that the grading scheme is prognostic in dogs is small (Kuntz et al, 1997; Ettinger et al, 2006). Because of this very small number of studies, containing relatively small numbers of cases, it is not possible to be totally confident that the human STS grading scheme is applicable in dogs; indeed in the Ettinger study, intermediate grade STS had a shorter median survival time than high grade STS! In addition, there are no studies to my knowledge using this general scheme for STS in cats. Furthermore, as both studies were carried out at referral institutions, the prognostic value of the scheme in primary care institutions has not been evaluated.

Only limited numbers of STS types are included in these studies, on the assumption that some STS types will behave differently to the majority of STS types. For instance, haemangiosarcoma and histiocytic sarcomas are often not included in these studies as they are expected to have a higher rate of metastasis/local aggressiveness than STS as a whole.

The histological grading scheme is based on identification of the following histological parameters: • tumour differentiation • mitotic rate over 10 / HPF • extent of tumour necrosis.

Scores are assigned for each category and summed to give a final score. The higher the score, the higher the grade and the worse the prognosis.

In practice, although the extent of necrosis and mitotic index are generally straightforward to assess, the degree of tumour differentiation can be very subjective and indeed varies from tumour type to tumour type. This issue has

been addressed with guidelines available on how to score the differentiation for each human STS type. This has yet to be done for veterinary STS types and therefore a degree of subjectivity still comes into grading these tumours at the moment.

This is further complicated by the ongoing process of refining the list of STS types that are recognised in dogs and cats, through the development of molecular and immunohistochemical markers specific for a particular cell of origin. The most recent example of this is the reclassification of a large group of common STS in dogs, under the umbrella heading of ‘perivascular wall tumours’. This particular study (Avallone et al, 2007) used a variety of immunohistochemical markers and histological features to subclassify those tumours that have previously been lumped together as ‘peripheral nerve sheath tumours’ or ‘haemangiopericytomas’. Using the human WHO STS classification scheme as a guide, a number of different tumours were identified that had unique immunohistochemical and histological appearances (hemangiopericytoma, myopericytoma, angioleiomyoma, angiofibroma, etc). This type of study will likely have consequences for STS grading, as these individual STS types could behave differently.

Canine Mast Cell Tumours The canine tumour that has received the most attention for development of grading systems is the cutaneous mast cell tumour. Arguably the most problematic feature of MCTs in terms of clinical management is their diverse biological behaviour, which can range from essentially benign to highly malignant and metastatic with many stages between.

1. Cytology of MCTs

Although MCTs can be identified readily by cytological examination of stained smears of fine needle aspirates, no grading schemes exist for MCT based on cytology alone, although the variation in granule content and nuclear morphology gives some indication of the degree of differentiation of the tumour cells. It would seem surprising that no cytological grading scheme has been developed, particularly as the nuclear morphology is used in the Bostock (1973) histological grading schemes.

2. Histopathology of MCTs

The basic pathology of MCTs is well documented, as is the concept that histological features such as cellular differentiation, mitotic rate and tumour invasion may be used to “grade’ the tumour and thus predict the biological behaviour and prognosis of an individual tumour.

There are 2 histological grading schemes published (Patnaik et al, 1984; Bostock, 1973) for canine cutaneous MCT, both of which were developed using a relatively small number of retrospectively collected MCT biopsy samples. The most frequently used is the Patnaik scheme; however, there are numerous

problems in standardising the histological grading criteria set out in the original Patnaik paper, as exemplified by recent studies showing that different veterinary pathologists will assign different histological grades to the same MCT (Northrup et al, 2005). Strict application of Patnaik’s grading technique precludes the grading of the large numbers of MCTs that originate from within the subcutis as the grading scheme – as written –refers only to dermal MCTs. A recent paper (Newman et al, 2007) suggests that subcutaneous MCTs behave like ‘intermediate grade’ MCTs, i.e. the majority are benign.

Less subjective methods of predicting individual MCT behaviour, based on quantification of morphological or staining features of the tumours have been studied. Markers of cell proliferation have been widely investigated in human medicine as potential tools for predicting prognosis, response to therapy and tumour behaviour.

For instance, the mitotic index (MI; mitotic figures/10 high‐power fields) in regions of the tumor with the highest overall mitotic activity is predictive for survival and is associated with Patnaik grade (Romansik et al, 2007). The median survival time for dogs with a MI < or =5 was 70 months compared to 2 months for those with a MI >5 , and this was independent of histological grade. For grade III tumors with a MI >5, the median survival was <2 months, vs. median value not reached, for those grade III tumours with a MI<5.

Staining for argyrophilic nucleolar organizing regions (AgNORs) has been used in a wide range of tumours in both human and veterinary medicine as a marker of tumour kinetics and tumour metabolic activity. AgNORs are nucleolar organizing region‐associated proteins such as nucleolin and nucleoplasmin, that bind silver ions, and can be visualized by light microscopy using a silver‐based histochemical stain. The number of AgNOR dots per neoplastic nucleus is inversely proportional to the doubling time. A number of studies have shown that AgNOR counts in MCT are associated with prognosis (Webster et al, 2007; Scase et al, 2006; Simoes et al, 1994) but that they are not independent of histological grade. There is some controversy, given the difficulties of standardising AgNOR staining and counting between laboratories, as to whether they add significantly useful information about the likely tumour behaviour above that provided by the histological grade and Ki67 index.

Ki67 is an antigen that is expressed during the cell cycle that can be detected by immunohistochemistry and has recently been shown to be highly associated with MCT prognosis (Scase et al, 2006). Importantly Ki67 has been shown to be an independent prognostic factor, independent of histological grade. A follow‐on paper (Maglennon, in press) details 163 dogs with intermediate grade MCT, for which Ki67 staining was performed. This confirmed the very high discriminatory power of Ki67 (binary cut off of 1.8%; hazard ratio 19; p<0.0001) to identify those dogs that had long survival times (median survival not reached) from those with short survival times (median survival 292 days). Thus, Ki67 immunohistochemistry is a promising non‐subjective, quantitative prognostic marker to accurately predict MCT behaviour. There remains however, some problems in defining the exact counting methodology that should be

employed when performing the counts, as each of the different research groups use slightly different methods!

KIT expression in both normal and neoplastic canine mast cells has been demonstrated by immunohistochemistry (Webster et al, 2004). In some neoplastic mast cells and all normal mast cells, the positive staining is restricted to the cell membrane, while in many MCTs, the staining is localized in the cytoplasm, often around the nucleus. This aberrant cytoplasmic KIT expression in MCT is associated with c‐kit mutations and is associated with a poor clinical outcome.

The majority of c‐kit mutations are internal tandem duplications in the juxtamembrane domain, compared with human c‐kit mutations which are mostly single codon mutations in the kinase domain (Riva et al, 2005). The frequency of these mutations in MCTs varies between 9‐33%, with higher grades associated with more frequent c‐kit mutations (30‐50% of grade II/III MCTs contain them). In addition, approximately 80% of MCT with c‐kit mutations had aberrant immunohistochemical staining

A number of other studies have been performed in attempts to identify prognostic markers for MCT. Assessment of mast cell nuclear characteristics by morphometric analysis demonstrates an association with MCT grade, but is not an independent prognostic marker. Other studies examining expression of an inhibitor of apoptosis, (survivin), and p53 do not show significant association with tumour prognosis. Similarly, DNA aneuploidy is associated with clinical stage I versus non‐clinical stage I, but it is not associated with survival. Canine MCT express both matrix metalloproteinases (MMP) 2 and 9, and high grade MCT exhibit higher levels of proenzyme MMP‐9 expression than intermediate grade MCT.

A number of diagnostic centres are now offering a service to 'grade' MCTs using a combination of histopathology, Ki67 staining, AgNOR counts and KIT immunohistochemistry. However, it the majority of cases, a combination of histopathology and Ki67 staining is likely to be sufficient to accurately assign the MCT grade in the vast majority of cases.

Ultimately, I believe that the MCT grading system will be simplified into low and high grades, mostly based on histological appearance of the tumour and Ki67 index. The low grade MCTs will be benign, and indeed may not be truly neoplastic, and high grade MCTs, will be malignant, will locally recur and have significant metastatic potential. There is some agreement between veterinary pathologists that this will be proposed in the not‐too‐distant‐future. This will then do away with the clinical problem of what to do with the intermediate grade MCTs!

References:

Avallone G, Helmbold P, Caniatti M, Stefanello D, Nayak RC, Roccabianca P (2007) The spectrum of canine cutaneous perivascular wall tumors: morphologic, phenotypic and clinical characterization Veterinary Pathology 44:5 pp 607‐620

Ettinger SN, Scase TJ, Oberthaler KT, Craft DM, McKnight JA, Leibman NF, Charney SC, Bergman PJ (2006) Association of argyrophilic nucleolar organizing regions, Ki‐67, and proliferating cell nuclear antigen scores with histologic grade and survival in dogs with soft tissue sarcomas: 60 cases (1996‐2002). Journal of the American Veterinary Medical Association 228:7 pp 1053‐1062

Kuntz CA, Dernell WS, Powers BE, Devitt C, Straw RC, Withrow SJ (1997) Prognostic factors for surgical treatment of soft‐tissue sarcomas in dogs: 75 cases (1986‐1996). Journal of the American Veterinary Medical Association 211:9 pp 1147‐1151.

Newman SJ, Mrkonjich L, Walker KK, Rohrbach BW (2007) Canine subcutaneous mast cell tumour: diagnosis and prognosis. Journal of Comparative Pathology 136:4 pp 231‐239.

Northrup NC, Howerth EW, Harmon BG, Brown CA, Carmicheal KP, Garcia AP, Latimer KS, Munday JS, Rakich PM, Richey LJ, Stedman NL, Gieger TL (2005) Variation among pathologists in the histologic grading of canine cutaneous mast cell tumors with uniform use of a single grading reference. Journal of Veterinary Diagnostic Investigation 17:6 pp 561‐564

Riva F, Brizzola S, Stefanello D, Crema S, Turin L (2005) A study of mutations in the ckit gene of 32 dogs with mastocytoma. Journal of Veterinary Diagnostic Investigation 17:4 pp 385‐388

Romansik EM, Reilly CM, Kass PH, Moore PF, London CA. (2007) Mitotic index is predictive for survival for canine cutaneous mast cell tumors. Veterinary Pathology 44:3 pp 335‐341

Scase TJ, Edwards D, Miller J, Henley W, Smith K, Blunden A, Murphy S (2006) Canine mast cell tumors: correlation of apoptosis and proliferation markers with prognosis. Journal of Veterinary Internal Medicine 20:1 pp 151‐158

Simoes JP, Schoning P, Butine M (1994) Prognosis of canine mast cell tumors: a comparison of three methods. Veterinary Pathology 31:6 pp 637‐647

Trojani M, Contesso G, Coindre JM, Rouesse J, Bui NB, de Mascarel A, Goussot JF, David M, Bonichon F, Lagarde C (1984) Soft‐tissue sarcomas of adults; study of pathological prognostic variables and definition of a histopathological grading system. International Journal of Cancer 33:1 pp 37‐42

Webster JD, Kiupel M, Kaneene JB, Miller R, Yuzbasiyan‐Gurkan V (2004) The use of KIT and tryptase expression patterns as prognostic tools for canine cutaneous mast cell tumors. Veterinary Pathology 41:4 pp 371‐377

Webster JD, Yuzbasiyan‐Gurkan V, Miller RA, Kaneene JB, Kiupel M (2007) Cellular proliferation in canine cutaneous mast cell tumors: associations with c‐KIT and its role in prognostication. Veterinary Pathology 44:3pp 298‐308 23

Soft Tissue Sarcomas – Anything New Worth Knowing??

Soft tissue sarcoma (STS) is a catch‐all classification referring to tumours that arise from the embryonic mesoderm and as such can occur anywhere in the body. Consequently there is huge variability in histopathological subtypes of STS. Generally they are classified according to cellular lineages (morphology on H&E staining, or immunohistochemisty (IHC)) such as fibrosarcoma, peripheral nerve sheath tumour, myxosarcoma, liposarcoma, or leiomyosarcoma, but sometimes these distinctions are not clear and the generic terms “soft tissue sarcoma” or “spindle cell sarcoma” are employed. Some tumours of mesoderm behave in a much more aggressive and less predictable fashion and these tend not to be included in STS; examples include lymphangiosarcoma, rhabdomyosarcoma, synovial cell sarcoma, haemangiosarcoma, chondrosarcoma, and osteosarcoma. Surgical resection is the principal treatment for primary localised disease as STS are relatively chemo‐insensitive and radiotherapy is more of value in a curative context as an adjunct to surgery. The great variation in anatomic location, factored with variable size and grade, can present significant problems when making a treatment plan.

STS are graded into low (I), intermediate (II) and high (III) grade tumours taking into account histological features such as mitotic rate, extent of necrosis and cellular differentiation. The difficulty of attributing a “T” stage to STS is that whereas the system may be accurate for tumours of the upper limb, chest wall and flank, it may be less accurate when the tumours involve the head and neck, retroperitoneum or distal limbs, when specific anatomical restraints at these sites may prevent wide curative‐intent excision. In these locations the amount of tumour excised/excisable might prove to be more important than grade. In the distal limbs however, volume of residual disease and grade interplay, with some volumes of tumour being too small to recur, whilst others may recur in 2 months, and some may take 3 years or more.

Diagnostic investigation for STS has 3 aims: • to accurately define the tumour ahead of treatment; • to define the anatomical relations of the primary tumour for planning (surgery and/or radiation), and; • to identify the presence or absence of metastatic disease.

Radiographs may yield some information regarding local behaviour, but may only confirm the mass is of soft tissue density. Ultrasound (esp Doppler) can be

useful but cross sectional imaging (CT or MRI) is the imaging of choice and is typically supportive of the diagnosis. Whereas MRI is traditionally regarded as superior for soft tissue detail, CT (esp contrast CT) offers a fast, simple, and accurate solution for all but the most complex STS. It is cheaper, as useful as MRI, and it has the advantage that imaging the thorax for metastatic disease can easily be performed at the same time. In fact, when both are available, CT is the more commonly used technique for assessing STS and planning surgery and/or radiation.

In terms of biopsy, fine needle aspirates have an important role in ruling out more likely subcutaneous differentials for example mast cell tumours, lipomas or inflammatory lesions, all of which exfoliate cells well. Enough mesenchymal cells may not be aspirated from a STS to make a diagnosis. If an aspirate of a SQ mass fails to yield many cells on the slide, your index of suspicion for a STS should be raised, and prompt a core biopsy. Percutaneous core biopsies (e.g. Trucut) are the best technique for achieving a safe and accurate diagnosis and can easily be performed with local anaesthetic alone or with sedation in nervous patients. Trucut biopsies will reliably differentiate benign from malignant disease and in most cases will also give a good indication of grade. The simplicity and accuracy of core biopsy for STS means incisional biopsies are infrequently indicated and come with the added concerns of location and direction of scar, and tumour dissemination from post‐incisional biopsy haematoma.

Surgical resection is the most effective treatment for STS. The aim of curative‐ intent surgery is to widely excise the primary tumour (dogma: 3cm and/or a fascial plane) and achieve negative histopathological margins. The requirement for a 3cm margin to maximize local control of STS was challenged by Banks and Straw in 2004. In a prospective study, 14 dogs with 15 subcutaneous STS were treated using a standardized protocol. Histopathological evaluation was also standardized and the dogs were followed for a minimum of 12 months. A lateral surgical margin of > 10mm and a single fascial plane deep (or > 10mm of deep tissue) yielded a local disease control rate of 100% with a 93% one year disease‐ free interval. Local recurrence occurred in one dog with a grade III STS over the lateral elbow, despite achieving a fascial plane as the deep surgical margin. This case generated discussion that suggested a margin of 11‐30mm may be inadequate for high grade STS and that adjunctive radiotherapy should be considered when the margin assessment has been labelled as “close,” even if a deep fascial plane of normal tissue has been excised. It may be this recurrence is attributable to skip metastasis of high‐grade sarcomas.

Banks’ research into what is ‘an appropriate margin for STS’ is relevant for daily practice, given that these masses can arise more‐or‐less anywhere, there is a great variety of biological behaviours seen between the grades, and that the advice to ‘widely excise the primary tumour’ is much more complex than for other malignancies which present more uniformly (e.g. lung, thyroid, intestine, mammary). The Enneking classification of intralesional, marginal, wide or radical is valuable in the flank or upper limb, but becomes problematic in the distal limb. Large wide resections of the upper limb or flank are possible and may not need complex reconstructive techniques, unless in awkward

anatomical locations, for example around the perineum. This is because not only is there an abundance of skin on the flanks of most patients, but also many STS arise deep within the body wall and primary skin closure can be achieved by preserving the superficial subdermal plexus, but still getting sufficient tissue resection deep. Radiation has a role if incomplete margins are found, if further surgery is declined, or if the mass is close to important unresectable structures.

This is a different situation in the distal limbs however (typically at/below the stifle, and at/below the elbow) where a wide surgical margin of skin is usually only achievable using free skin grafts or random flaps, and options such as amputation or marginal resection and radiation are considered. Both these options have drawbacks in terms of altered function, morbidity and cost.

Treatment options for a histologically diagnosed extremity sarcoma include: 1. Amputation; 2. Curative‐intent excision with reconstruction, plus radiation if incomplete margins; 3. Pre‐operative radiation (48Gy) with sterilisation of tissue around tumour and planned marginal excision of mass; 4. Planned marginal excision of mass then post‐operative radiation (57Gy) to sterilise tumour bed and periphery; 5. Skin‐sparing surgery – planned marginal excision to microscopic disease, including fascia where possible.

Inadvertent shell‐out procedures on masses assumed to be benign, that subsequently turn out to be malignant represent an unfortunately common but easily avoidable clinical scenario. The difference between planned and unplanned marginal excisions is:

• In planned marginal excisions, typically in the distal extremities, the surgeon is aware they are incompletely excising a malignancy, anticipate local recurrence, and so attempt wherever possible to work outside the pseudocapsule to excise the infiltration of the tumour into the surrounding tissues. Wide margins are not possible without reconstructive techniques (which have been discussed and declined by the owner) and so the aim is to remove as much tumour as possible, but still close the skin over the wound. This may mean removing tumour off fascia, or even excising fascia beneath the mass and this is made easier if an exsanguinating Esmarch tourniquet is used. At some point during the surgery the pseudocapsule is entered (normally when peeling tumour off skin) and so the surgery is classified as marginal.

• An unplanned marginal excision is a true shell‐out working in the pseudocapsule. Fascia is not taken, and small macroscopic disease may remain.

Treatment options for a histologically diagnosed extremity sarcoma removed inadvertently (i.e. without previous biopsy, unplanned excision) include:

1. Amputation; 2. Post‐operative radiation (57Gy) to sterilise tumour bed; 3. Re‐excision of scar with local closure techniques; 4. Wait‐and‐See.

In both human and veterinary surgical oncology, universally accepted guidelines on treatment of solid tumours have been difficult to establish. A large stumbling block in creating guidelines on treatment of STS is the indecision and uncertainty over what to do with an incomplete surgical margin and its relevance on local recurrence and overall survival.

The impact of leaving residual tumour cells in the wound bed of an excised STS is a many‐fold increase in the rate of local recurrence (Kuntz). Several studies have reported the results of adjunctive radiotherapy in the management of STS with surgically incomplete margins and its success has generally been measured by its ability to provide consistently long overall survival times. The effect on local tumour control however, has been variable, with recurrence rates ranging from 17‐60%.

Retrospective investigation into the results of surgery alone (‘primary re‐ excision’) for the treatment of STS after incomplete resection studied 41 dogs that had undergone aggressive scar revision (attempted wide (1‐3cm) margins / 1 fascial plane deep) for incompletely excised STS. Complete margins were obtained after re‐excision in 90% (37/41) of all the cases with mean margin widths of 2.7cm on the proximal portion of limbs and 1.4cm on the distal portion of limbs. Local tumour recurrence occurred in 15% (6/39) of the dogs at a median time to recurrence of 142 days. Correlations for recurrence were not identified and of particular note, in 4 out of the 6 recurrences, there was no evidence of tumour in the resected tissue and the scar was excised with clean margins. All the masses that recurred were grade 2, no recurrence was seen below the stifle or elbow, and only 2/6 dogs were ultimately euthanized for reasons related to the tumour.

Comparisons of the local control rates achieved with re‐excision of the scar compared to radiation therapy of the scar suggest that the outcomes of surgery, when possible, may be equivalent to those achieved with adjunctive radiation therapy. In fact, when the costs and relative morbidity of radiation are factored in, an attempt at surgical excision alone may be a more desirable first line approach. The question then arises, what if the site of recurrence does not afford an extensive re‐excision, or if the owner declines further treatment after a tumour positive margin is confirmed? Veterinary opinion suggests that local recurrence rate will be high if ‘wait‐and‐see’ is employed but limited data has been published to support this assumption.

Two recent studies have evaluated the local recurrence rate of canine STS of the distal limbs treated by marginal excision alone (wait‐and‐see):

• The first (Cavanaugh et al, 2007) followed 26 dogs with tumours of the distal antebrachium or pes, all of whom were determined to have tumour positive margins after marginal excision by a variety of surgeons,

some planned, some unplanned. All grades of STS were represented and follow‐up intervals were long (median, 781 days, minimum 594 days). The rate of local recurrence was 37% (10/27 tumours) with only 12% (3/26) of the dogs being euthanized for problems relating to local disease. Interestingly, of the 10 recurrences, 4 were untreated and all 4 dogs died of tumour‐related disease. The other 6 underwent further conservative surgery (intra‐lesional, marginal) and only 2 died from tumour‐related disease, suggesting that managing local recurrences by repeat marginal excisions may be an acceptable solution in some circumstances. This study had long follow‐up, all excisions had dirty margins, and all grades were represented.

• The second study (Stefanello 2008) was a retrospective review of planned marginal excisions of low‐grade distal extremity soft tissue sarcomas by two experienced surgeons. Using this technique, 32% surgeries has clean margins, 34% were ‘clean but close’ and 34% incomplete/dirty margins. Recurrence rate for this study was 11%. Follow up ranged from 210‐2202 days. In this study, minimum follow‐up time was shorter, and all tumours were effectively grade 1. 34% had clean surgical margins, and the rate of local recurrence was too low to accurately predict factors influencing recurrence.

Both these studies challenge the dogma that local recurrence rate following incomplete resection of STS is high, common, inevitable or guaranteed, depending on your text source. As is the goal of any retrospective study, these two raise questions rather than provide answers. The question which needs to be answered now is not how many recur, but what can we do to predict which are likely to recur. Ettinger and Scase have shown that AgNORs and possibly Ki67 could be used over and above grade to help predict overall survival in STS, and for a starter, this work needs to be repeated in the context of local recurrence with a similar population of dogs with similar wounds in terms of tumour burden.

Future directions and treatments??

It would be ideal to identify those at risk of local recurrence or metastasis BEFORE surgery based on imaging and better predicting biological behaviour. We could then tailor the “dose” of surgery and treatment accordingly, taking into account the animal’s age, status and prognosis. If we follow human advances, then intra‐arterial chemotherapy, interventional chemo‐embolisation or isolated limb‐perfusion (melphalan/TNF∝) treatments may be considered in the future.

The bigger picture here is that STS in humans are relatively uncommon, accounting for <1% of all cancers, with 8500 new cases being diagnosed a year in the US, and 13000 in the EU. Although relatively rare in the dog, STS are seen commonly in secondary and tertiary referral practices. Scope for translational trials (what would be called “pre‐clinical trials” in the human setting) exists and is likely to benefit both parties.

Soft Tissue Sarcomas – Anything Else Worth Knowing??

Jonathan Bray MVSc CertSAS MACVSc DipECVS MRCVS RCVS Recognised Specialist in Small Animal European Specialist in Small Animal Surgery Davies Veterinary Specialists

Radiation Therapy for Soft Tissue Sarcomas (STS): Indications and Applications

Susan M. North BSc(Hons), PhD, DVM DipACVIM (Medical Oncology) DipECVIM (Internal Medicine and Medical Oncology) Dip ACVR(Radiation Oncology) American, European and RCVS Specialist in Medical Oncology Specialist in Radiation Oncology

Introduction

Traditionally STS has been considered the territory of the soft tissue surgeon because if these tumours can be excised with good margins then further treatment is not usually required. Also, as sarcomas are chemoresistant there is little indication for chemotherapy in the management of STS, although in cases of high‐grade sarcomas some oncologists do recommend doxorubicin chemotherapy in combination with surgical excision and/or radiotherapy. This remains controversial and debatable.

The expanding role of radiation in the management of STS, histiocytic sarcomas and invasive haemangiosarcomas continues to be appreciated as better and more powerful Linacs are available for veterinary patients along with the planning systems required for complex treatments. This means that it is important that the oncological surgeon and the radiation oncologist work together as a team.

Principles of management • The same in both dogs and cats, although there are relevant species differences

Canine STS • Aetiology: unknown • Incidence in dogs is interestingly greater than in man as they account for ~15% of canine malignancies • Predominantly on the extremities

Feline STS • Naturally‐occurring sarcomas are seen less frequently than in the dog • Distribution more frequently on the head than extremities

Feline vaccine‐associated sarcomas (VAS), now known as injection‐site sarcomas (ISS) • Unique to the cat • Initially thought to be exclusively associated with adjuvant containing vaccines

• Local inflammation leads to uncontrolled fibroblast proliferation and oncogenesis

Surgery

For patients with STS amenable to surgical excision with wide margins, that is the treatment of choice. However, even when surgical excision is achieved recurrence is still a possibility. Kuntz et al (1997) reported 85% local control with a median time to recurrence of 360 days, Posterino et al (1988) 79% control for 2 years. External beam radiation is indicated in patients with tumours not amenable to complete surgical excision and there are a number of different ways radiation can be utilised.

Radiation

• Adjuvant treatment after cytoreductive surgery 1. Hypofractionated 2. Hyperfractionated • Neo‐adjuvant (pre‐operative) • Sole treatment‐palliative • Intra‐operative

A. Adjuvant Radiotherapy 1. Most common use of radiation in the management of STS 2. Indications are tumours that have been incompletely resected, usually on distal extremities 3. Most effective on microscopic or <1cm macroscopic disease • Higher growth fraction • Recurrence uncommon for low‐intermediate‐grade tumours over 3 years • A higher rate of recurrence is expected with high grade tumours 4. Important for the radiation oncologist to know size and exact location of the tumour if the patient is referred after surgery. • Send a pre‐operative photograph of the tumour so the exact location of tumour relative to scar can be determined • A detailed surgical report with the surgeons opinion as to extent of tumour left behind and the most suspicious areas for potential regrowth • Haemoclips and a post‐operative radiograph are also helpful • If CT or MRI scans were carried out prior to surgery these are also helpful 5. If the tumour is large, consider radiotherapy before surgery (see below). 6. Grafts and flaps make radiation oncologists nervous especially if carried out by non‐oncological surgeons 7. Refer early for radiation. If the margin is unclean or suspicious and further surgery would be difficult, i.e. you cannot go back and get good

margins then refer for radiation. Optimal time to start treatment is once surgical healing has been achieved (~14 days after surgery). 8. Remember when planning cytoreductive surgery the radiation oncologist cannot irradiate 360 degrees on a distal extremity, as this will damage lymphatic drainage from the paw. 9. The majority of STS referred for adjuvant radiotherapy are located on the limbs, head and neck, hip, tail base and close to the anus. For those tumours located on the flank and lateral thoracic wall radiotherapy is usually not an option. 10. Feline sarcomas that have been incompletely resected from the interscapular region are not good candidates for radiation because of the difficulty in planning with potential toxicity to lungs and spinal cord.

Radiation Protocols: Pros and Cons

There are a number of treatment protocols; these broadly fall into 2 categories: • Hypofractionated • Hyperfractionated (Definitive)

Hypofractionated

4 treatments once weekly, usually 8‐9 Gy/fraction, therefore maximum total dose 32‐36Gy.

Advantages: i. Only 4 treatments ii. Less expensive Disadvantages: i More potential side effects, especially late effects ii. Total dose is low for curative intent

Hyperfractionated

These treatments are more in line with human protocols. There are a number of protocols used e.g. MWF (3Gy/fraction to either 48 or 54Gy), or M‐F (2Gy/fraction up to 60Gy). See: McKnight et al (2000), Forest et al (2000)

Advantages: i. Better long‐term control of tumours such as STS ii. Reduced rate of late effects of radiation Disadvantages: i. More treatments and therefore more expensive ii. Requires a lot of work from the oncological team

B. NeoAdjuvant Radiation (PreOperative) Why consider this in the first place?

Surgery‐fails at the edges of the tumour Radiation‐fails at the centre of the tumour due to hypoxia

Advantages: • Surgical cure may be achieved, i.e. a tumour that was considered not operable can be made operable • Less morbidity from surgery, reduction of surgical dose can mean that a client who initially declined surgery may decide to go ahead after radiation; this mostly applies to tumours of the maxilla and mandible.

Disadvantages: • Wait for the effects of radiation • Wound healing after radiotherapy?

What is the goal of pre‐operative radiotherapy? • Sterilise the margins • Shrink the tumour or free up deep margins • Plan to make an inoperable tumour operable • Usually wait 2‐4 weeks after treatment to maximise the effects of radiation

Indications: • Tumours considered inoperable by an experienced oncological surgeon (or client declines surgery) • No metastatic disease identified • Undertaken on the understanding that the tumour may not become surgical • Therefore the surgeon and radiation oncologist agree from the beginning on the plan aided by advanced imaging

Tumours amenable to this approach • STS and other invasive sarcomas • Thyroid carcinomas • Large oral tumours • Mast cell tumours?

C. Sole Treatment 1. Problem: most sarcomas grow slowly 2. Typical growth fraction is ~10% 3. Sometimes can control tumour‐slow growth 4. One study: ~50% control at doses of 50Gy in 5Gy fractions 5. Palliative therapy

D. IntraOperative Radiation

• Indicated when gross tumour is left behind at surgery • Typically one dose of 9Gy is given • Normal wound healing allowed before giving additional treatments • Goal is to prevent recurrence of aggressive tumour before starting conventional radiation

Radiotherapy and the feline patient with ISS • Side effects to other organs must be considered • Tumours incompletely resected from the interscapular region are a problem because: o Spinal cord o Lung Early consultation with a radiation oncologist is recommended.

References:

Forrest LJ, Chun R, Adams WM, et al. (2000) Postoperative radiotherapy for canine soft tissue sarcoma Journal of Veterinary Medicine 14: 578‐582

Kuntz CA, Dernell WS, Powers BE, et al. (1997) Prognostic factors for surgical treatment of soft‐tissue sarcomas in dogs: 75 cases (1986‐1996) Journal of American Veterinary Medical Association 211: 1147‐1151

McKnight JA, Mauldin N, McEntree MC, et al. (2000) Radiation treatment for incompletely resected STS in dogs Journal of American Veterinary Medical Association 217: 205‐210

Posterino NC, Berg RJ, Bowers BE, et al. (1988) Prognostic variables for canine haemangiopericytoma: 50 cases (1979‐1984) Journal of American Animal Hospital Association 24: 501‐509

Histiocytes, Histiocytoses and Dendritic Cells: A Review of the Histiocytic Diseases of the Dog

Stephen J Baines MA VetMB PhD CertVR CertSAS DipECVS MRCVS RCVS Recognised Specialist in Small Animal Surgery European Specialist in Small Animal Surgery Royal Veterinary College Lecturer in Small Animal Surgery

Introduction

The dog is affected by a number of proliferative diseases involving cells with histiocytic morphology. This is a pleiomorphic group of diseases and differentiation between them has been made primarily on light microscopy of cytological or histological samples and the clinical behaviour of the disease. The development of appropriate immunological reagents has made it possible to characterise these diseases more fully and to provide a more rational differentiation between them. In addition, the finding that many of these disorders arise from cells of the dendritic cell lineage may help understanding the pathogenesis of these diseases and guide a more logical approach to their treatment.

1. Canine Cutaneous Histiocytoma

1.1. Incidence and clinical features Canine cutaneous histiocytoma (CCH) is a benign tumour of the skin [1‐4]. It is a relatively common skin tumour in the dog, accounting for 3‐19% of all skin tumours [2, 5‐7].

CCH characteristically affects young dogs, with approximately 50% of tumours occurring in dogs under two years [8‐10]. There is a considerable drop in prevalence after two years, although older dogs can also be affected [3]. Pure‐ bred dogs, particularly boxers, cocker spaniels, Great Danes, Shetland sheepdogs, dachshunds, English bulldogs, Dobermans, schnauzers, Shar Peis and Scottish terriers are reported to be predisposed [1, 3, 5, 10] whilst Poodles have a significantly lower risk of developing CCH [10]. Other authors have discounted a breed predilection for solitary lesions [11]. There is no sex predisposition [3, 9].

The tumours are usually solitary, but can be multiple [12‐16]. The multiple forms represent 0.5‐4% of all histiocytomas [9, 10] and are seen predominantly in Shar‐Pei and Shar‐Pei‐mixed breed dogs [11].

Lesions are often found on the head [10], especially on the muzzle, pinna or periocular region [17], the extremities and the scrotum. Other sites are the skin of the distal limbs, neck, trunk and tail. Only 1.5‐4% of tumours are found in a perigenital location and none are found on mucosae [9, 10]. In rare cases, the

regional lymph nodes may be enlarged [18]. This lymphadenopathy generally resolves with regression of the cutaneous nodule [11].

1.2. Pathological features The lesions are small, firm, dome or button‐shaped cutaneous nodules which seldom grow more than 2 cm in diameter but occasionally reach 4 cm [8‐10]. The surface is covered with thin, shiny or ulcerated epidermis, through which sparse hairs project. Erythema of this semi‐glabrous surface has prompted the synonym “strawberry tumour”. Upon cutting, the tumour is pliable, though rather tough and resilient. The cut surface of fresh tissue bulges slightly and is nearly white with varying shades of yellow. The appearance is generally homogenous, although it may be finely nodular and stippled or mottled red. Deep margins are well defined but not encapsulated. These margins describe arcs varying from nearly flat to a crude semicircle extending into the subcutis.

The histological appearance of this tumour has been well described [8‐10, 19, 20]. With haematoxylin and eosin staining, canine cutaneous histiocytoma comprises uniform sheets of cells that infiltrate the dermis and subcutis, displacing collagen fibres and adnexal structures. The cells are closely packed and arranged randomly in the deeper dermis, but near the surface may be lined up in rows, perpendicular to the dermo‐epidermal junction. The tumours have a base‐narrow and top‐heavy configuration and the borders of the tumour merge gradually with the surrounding tissue.

Large round to ovoid cells with large pale‐staining nuclei and abundant pale, poorly demarcated cytoplasm are found. The nuclei are usually round or oval although they may be oblong, reneiform, C‐shaped, wrinkled or shapeless. The nuclei are often eccentric and vary considerably in size with occasional binucleate forms. Typically there is a high mitotic index, averaging 3 ‐ 8 mitotic figures per high power field [9], or 2.4 ‐ 8.7 (average 5.5) per 1000 tumour cells [10]. There may be some difficulty in distinguishing canine cutaneous histiocytoma from small mastocytoma or ‘histiocytic’ lymphoma. Differentiation from masT‐cell tumour requires the use of Giemsa or toluidine blue stains [21] and immunohistochemistry on fixed tissue, using antibodies recognising CD3 and CD79a, may allow a definitive diagnosis of lymphoma to be made.

Hyperkeratosis and epidermal hyperplasia are present in a few tumours. In others, the epidermis is thinned with flattened rete pegs, accompanied by vesicles and pustules. More commonly, tumours show epidermal ulceration and necrosis, with inflammatory cells mixed with the superficial tumour cells. Epidermal invasion by histiocytes is noted in some cases and nests of histiocytes may resemble Pautrier’s microabscesses, as seen in epidermotropic lymphoma [22]. Some lesions show focal areas of necrosis and lymphocytic infiltration, which may be arranged as solid focal aggregates or in palisading rows. With significant lymphocytic infiltrate, tumour cells may be individualised and rounded with more intensely eosinophilic cytoplasm. Most tumours contain scattered masT‐cells and fibroblasts.

The pattern of lymphocytic infiltrate has been divided into four groups, ranging from no lymphocytes (Group I) through nodular infiltrates at the periphery (Group II) and centrally (Group III) to a diffuse infiltrate which outnumbers the histiocytes (Group IV) [23]. Foci of necrosis within the tumour in association with the lymphoid infiltrate were absent from group I but present in at least 50% of the other groups, with maximal occurrence (84%) in Group III. Ulceration of the overlying epithelium was present in the majority of dogs and increased progressively with increasing group number to reach 100% in Group IV. Similar findings were reported by Magnol et al (1982).

Most of the features identified histologically are recognisable on cytological specimens prepared by aspiration or imprinting [20]. However, the diagnosis may be in doubt if insufficienT‐cells are obtained (which is more likely for this round cell tumour than any other) or if a high proportion of lymphocytes or neutrophils are present [20].

In cases where there is epidermal invasion or a large lymphocyte infiltration, differentiation from epitheliotropic or non‐epitheliotropic cutaneous lymphoma may be difficult on morphological grounds. In addition, if multiple lesions are present, differentiation from cutaneous histiocytosis may be difficult [24]. However, epidermotropism and epidermal invasion are not common features of cutaneous histiocytosis.

Esterase staining of tumour impression smears demonstrates cells with diffuse cytoplasmic staining for α‐naphthyl acetate esterase [25]. The cells stain with lysozyme and α‐1‐antitrypsin in a proportion of cases [26, 27]. The cells do not stain with acid phosphatase, α‐1‐antichymotrypsin, S‐100 or neuron‐specific enolase [19, 27].

1.3. Clinical behaviour and treatment Typically, there is rapid growth over one to four weeks and most tumours regress spontaneously one to three months after recognition [5, 21]. Multiple histiocytomas usually persist for a longer period, up to nine months in some cases [11]. Metastasis is rare, but has been documented in a few cases [18]. Some tumours may recur at the previous surgical site [1], although only three of 196 (1.5%) excised histiocytomas recurred at the site of excision in one survey [10]. Development of another tumour at a different site is also rare, representing three of 196 tumours in the above survey. Regression of a CCH over a two week period, following biopsy, has been reported [10]. In a review of 23 tumours, Moore and others (1996) reported a higher incidence of multiple tumours (32%) and recurrence (9%).

Clinical management consists of observation without treatment or surgical excision. Topical application of glucocorticoid in dimethyl sulphoxide has been recommended for those lesions where surgical excision would be difficult [3]. However, if regression is mediated by cytotoxic T‐cells, administration of an immunosuppressive drug would appear to be counterproductive. More recent

experimental work has shown that corticosteroids induce dendritic cell apoptosis [28] so this therapy would appear to have a rational basis for the treatment of any proliferative disorder of dendritic cells. Surgical excision may be indicated if the cytological appearance is equivocal, for grossly ulcerated or pruritic lesions, or for lesions which do not show regression. Persistence of a focus of inflammation, particularly where there is a foreign body reaction to implanted hair shafts, may give the appearance that the tumour has not undergone regression.

1.4. Immunophenotype Immunophenotypic analysis of CCH shows they express a number of markers normally found on canine LCs, namely CD1a, CD1b, CD1c, CD11a, CD11c, CD18, CD45 and MHC‐II [18, 24, 29]. In addition, most histiocytomas express additional markers that are not present on normal canine LCs, namely CD11b, CD44, CD49d and ICAM‐1, and express CD1b uniformly and diffusely [18]. These markers are upregulated on activated human LCs and murine LCs [30], indicating that CCH cells may have an activated LC phenotype [18, 24]. The lymphocytic infiltrate consists primarily of CD8+ cytotoxic T‐cells, with smaller numbers of CD4+ T‐cells and B‐cells [18]. Expression of E‐cadherin by the CCH cells, but not the infiltrating lymphocytes, further supports the Langerhans cell origin, since LCs are the only non‐epithelial cells to express this adhesion molecule [31].

1.5. Other remarks The aetiology of the disease is unknown. An infectious cause has been proposed based on behavioural and epizootological features [5, 9, 23, 32, 33] and the appearance of various intracellular structures on electron microscopy [19, 25]. However, transmission studies and cell culture have failed to show an infectious agent [10]. It has been suggested that cutaneous histiocytoma represents a peculiar focal inflammatory lesion [21]. However, the monomorphic cell population, numerous mitotic figures, rapid growth rate, dermal invasion and occasional metastases suggest that it is a true neoplasm rather than an inflammatory reaction to an unknown agent [9].

Further evidence in support of the Langerhans cell origin of the neoplastic cells is the strong stimulatory potential of CCH cells in the allogeneic mixed lymphocyte reaction, one of the hallmark features of DCs in vitro [34].

It is suggested that the phenomenon of regression reflects killing of the tumour cells by the infiltrating lymphocytes and that lesions consisting solely of histiocytic cells are early lesions, with increasing numbers of lymphocytes invading the tumour with time [23]. However, no correlation between the degree of lymphocyte invasion and duration of the tumour has been determined since the historical data is generally incomplete and unreliable [23, 35]. Progressive foci of acute coagulation necrosis and other signs of degeneration in the tumour cells have been noted in tumours with a greater lymphocytic

infiltrate [23]. The presence of CD8+ lymphocytes in the tumour is consistent with killing by cytotoxic T‐lymphocytes, but no in vitro proof of lymphocyte‐ mediated killing is available.

Those histiocytomas with a plasma cell infiltrate and no T‐lymphocytes may show delayed regression and may persist for up to 12 months. This behaviour is reported to be more common in Shar Peis [11, 24]. Observations of spontaneous regression in vivo support the hypothesis of an anti‐tumour response. However, most tumours are excised, either because of their clinical similarity to cutaneous lymphoma or masT‐cell tumour, or because they ulcerate and cause the animal irritation, and the true incidence of regression is not clear.

Regression of the tumours, as measured by increasing lymphocytic infiltrate [23] was associated with upregulation of IL‐2, TNF‐α, IFN‐γ and iNOS mRNA expression [36]. An initial infiltration of the tumours by CD4+ T‐cells, followed by expression of Th1 cytokines and recruitment of anti‐tumour CD8+ effector T‐ cells was suggested as the principal mechanism for tumour regression.

Regression of the tumour is associated with a change in the pattern of MHC‐II expression from the cytoplasm to the cell surface [37] and an increased expression of MHC‐I, MHC‐II and ICAM‐1 and a lower expression of E‐cadherin [38].

Some animals develop multiple widespread cutaneous histiocytomas in which the lesions are almost confluent in affected regions. This has been termed progressive Langerhans cell histiocytosis [39]. In one case report, there was no response to immunosuppressive doses of prednisolone, but the lesions all regressed after 7 weeks’ therapy with griseofulvin. However, lymphadenopathy and hepatosplenomegaly developed following cessation of therapy. Other authors have seen similar animals and in these cases, rapid internal spread is seen and animals have all been euthanised (Moore, personal communication).

2. Cutaneous Histiocytosis

2.1. Incidence and clinical features This is a rare, benign histiocytic proliferative disorder of dogs [40‐42]. Collies, Shetland sheepdogs and Golden retrievers may be predisposed [43, 44]. No age or sex predilection is apparent and affected dogs range from 2 to 13 years old [11, 40, 45]. No seasonal incidence is reported [44]

The lesions consist of multiple haired or alopecic cutaneous nodules or plaques, 1 to 5 cm in diameter, in the dermis, which may extend into the panniculus. They occur primarily on the face, neck, back and trunk, but may occur anywhere on the body, and can involve the nasal mucosa. Multiple lesions often occur in clusters, but may be more generalised. The severity of the disease varies between individuals, with some animals having a few, intermittent lesions and others having repeated bouts of up to 50 lesions. Some dogs have lesions limited to the nasal planum and nasal mucosa, giving the appearance of an enlarged 39

‘clown‐nose’ appearance [40, 43]. Nasal lesions may be accompanied by stertorous breathing [44]. The lesions are non‐pruritic, erythematous and occasionally ulcerated. The lesions often wax and wane and appear in new sites. Lymphadenopathy and systemic involvement have not been reported at presentation, but one dog developed lymph node involvement with effacement of the normal lymph node architecture by atypical histiocytic cells [44].

Haematology and biochemistry changes which might be referable to the disease comprise mild regenerative anaemia in one dog [44]. No pathogenic organisms have been cultured on aerobic, anaerobic, mycobacterial and fungal cultures [44]

2.2. Pathological features Histological examination reveals nodular or diffuse dermal or subcutaneous infiltration by cytologically normal histiocytes, and the lesions have been confused with CCH [40]. Occasional erythrophagocytosis is observed and, rarely, the histiocytes may appear vacuolated and full of lipid. Foci of neutrophils, lymphocytes and eosinophils may be present, although plasma cells are uncommon [1, 11, 43]. In some cases, small lymphocytes may represent up to 50% of the infiltrating cell population [45]. A consistent finding is the lack of epitheliotropic invasion of the epidermis and the outer follicular root sheath [45]. There is little tendency to infiltrate the upper dermis, buT‐cells are frequently present in the panniculus. The mitotic index is generally lower than in CCH. In many cases, the lesions resemble those of systemic histiocytosis, except that vessel wall invasion by histiocytes is not prominent and the lesions are restricted to the skin and mucous membranes [42, 45]. Cultures, special stains and electron microscopy have not revealed an aetiological agent [11, 40, 46]. The cells stain positively for non‐specific esterase and negatively for acid phosphatase, as do the cells of CCH [40].

2.3. Clinical behaviour and treatment Few reports have evaluated the response of cutaneous histiocytosis to treatment. The clinical course of this disorder may involve lesions that wax and wane and spontaneous resolution of lesions without treatment has been reported [11, 41, 45]. Antibiotics are usually ineffective in treating cutaneous histiocytosis [11, 40, 41]. Surgical excision is often followed by recurrence at the surgical site or at other locations [40, 46].

Treatment has been attempted with immunosuppressive doses of glucocorticoids [40], cytotoxic drugs (e.g. azathioprine) [40, 45, 47] or the immunosuppressive agents cyclosporin A and leflunamide [24, 45] with variable results. Other agents, including tetracycline/doxycycline, niacinamide, vitamin E and essential fatty acids, either alone or in various combinations have also been used [3, 44]. When evaluating the response of cutaneous histiocytosis, it should be remembered that some lesions regress without treatment and the apparent success of therapy in some cases may simply reflect the phenomenon of spontaneous regression.

It is reported that approximately 50 – 82% of animals respond to immunosuppressive doses of corticosteroids [11, 44, 45]. Systemic cyclosporin or leflunomide, or intralesional corticosteroids have been used successfully in those animals which showed a poor response to systemic corticosteroids [11, 45]. Following cessation of therapy, some dogs remain in remission for an indefinite period of time. Development of new lesions requires re‐ administration of these drugs until regression is observed. Other patients develop new lesions as soon as therapy is withdrawn, and need continuous therapy with immunosuppressive drugs [45].

In one multi‐centre retrospective case series of 32 dogs, a variety of treatments were used and all dogs had complete resolution of dermatological lesions with initial treatment with a median time to resolution of 45 days. After resolution, 17 dogs received maintenance therapy with tetracycline/niacinamide either alone (7 dogs) or with other medications including safflower oil, essential fatty acids and vitamin E and 5 dogs received therapy with drugs including prednisone, azathioprine, cyclosporine and ketoconazole. Nine dogs (28%) experienced a recurrence, either a single episode (2 dogs) or multiple episodes (7 dogs) with a median time to recurrence of 130 days. The recurrence rate for lesions involving the nasal planum/nares (66%) was greater than for dogs without nasal involvement (33%). The presence of previous inflammatory skin disease did not affect the rate of recurrence. Five of these dogs were receiving maintenance therapy at the time of recurrence and 4 dogs had recurrence when the maintenance therapy was tapered. Short courses of corticosteroids with or without new immunomodulatory drugs were used to induce remission again and, of the 6 animals available for follow‐up, none had lesions and 4 were on maintenance therapy. One dog developed histiocytic sarcoma of the mandibular lymph node, was treated with lomustine for 6 months and remained in remission for both diseases for 2 years [44].

2.4. Immunophenotype The cells express CD1a, CD1b and CD1c, consistent with a dendritic cell origin [45]. Immunophenotypic analysis suggests that the neoplastic cells arise from dermal dendritic cells since they express Thy‐1, which is not expressed by epidermal Langerhans cells, and CD4, which is an activation marker for dermal dendritic cells, [24, 45]. In addition, variable expression of CD44, CD49d, CD50 and CD54 by the histiocytic cells is present, which indicates that these cells exist in a similar activated state as cutaneous histiocytoma [24, 45].

A lymphocytic infiltrate, consisting primarily of CD8+ T‐cells, is usually present, and represents up to 50% of the infiltrating cells [24, 45]. Smaller aggregations of CD21+ B‐cells are observed in many cases. It has been speculated that these infiltrating lymphocytes may represent effector cells responsible for inducing remission of the lesions, or alternatively lymphocytes whose secreted cytokine profile stimulates the proliferation of the histiocytic cells within the lesions [24, 45].

2.5. Other remarks Cutaneous histiocytosis has been described as a reactive dermatosis arising in the context of disordered immune regulation [24, 45]. It is suggested that the histiocytes may be defective in antigen presentation, leading to prolongation of the initial antigen‐dependent encounter, or alternatively that the T‐lymphocytes secrete cytokines which induce histiocyte proliferation. An increase in pro‐ inflammatory cytokines (TNF‐α, IL‐6, IL‐12 and IFN‐γ) have been demonstrated in lesional skin [48]. This provides a rational basis on which to explain the therapeutic efficiency of cyclosporin and leflunamide, which are potent inhibitors of T‐cell activation.

3. Systemic Histiocytosis

3.1. Incidence and clinical features This is a rare histiocytic proliferative disorder involving the skin and other organs [1, 26, 42, 49‐53]. It has been described in closely related, predominantly male, Bernese Mountain Dogs, aged from two to eight years old. An autosomal recessive mode of inheritance has been suggested [26, 51]. The disorder has also been recognised in other breeds. Some dogs develop the disease at three to four months old and the clinical course of the disease may be more acute [42].

The skin lesions consist of poorly‐circumscribed, firm papules, plaques or nodules up to 4 cm in diameter. The overlying epidermis is either smooth and sparsely haired or ulcerated and crusted. Lesions may be present anywhere on the body, particularly the muzzle, nasal planum, eyelids and scrotum [26]. The distribution of lesions is similar to that of cutaneous histiocytosis, but is often more widespread [45]. Lesions may also involve the lymph nodes, eye, nasal cavity, liver, spleen, bone marrow and lungs [11, 26, 45, 50, 54].

Mild to severe lymphadenopathy is generally present but gross hepatosplenomegaly is uncommon. Anorexia, weight loss and depression may be noted. Involvement of the nasal mucosa occurs in a proportion of dogs, resulting in increased respiratory noise and nasal discharge. Involvement of the eye and periocular tissues may manifest as conjunctival hyperaemia, chemosis, corneal oedema, episcleral thickening, anterior and posterior uveitis, glaucoma, exophthalmos and retinal detachment [26, 51, 53]. Ocular signs may be present before cutaneous lesions [51] or may be the sole manifestation of the disease [53]. Mild to moderate anaemia, monocytosis and lymphopoenia have been described in a proportion of dogs [26]. Hypercalcaemia has been reported in a small number of animals [45].

Most animals have several clinical episodes of gradually increasing severity, with intervening asymptomatic periods, although the disease is rapidly fatal in some dogs. In severe disease, lesions become persistent and do not respond to immunosuppressive doses of corticosteroids [26]. Ultimately, most dogs are euthanised because of the chronic, debilitating nature of the disease [26, 51, 53].

3.2. Pathological features The histopathologic features of the cutaneous and subcutaneous lesions in cutaneous and systemic histiocytosis are similar, if not identical [45]. The histological picture is of a distinctive histiocytic infiltrate in many tissues (predominantly skin, lymph node, spleen, liver, kidney, bone marrow, orbital tissues and skeletal muscles of the head), accompanied by lymphocytes and occasionally by neutrophils and eosinophils. A superficial and deep dermal and subcutaneous infiltrate is present, organised into a perivascular, nodular or diffuse pattern. The infiltrate may invade blood vessels leading to thrombosis and ischaemic necrosis. Typically the cells are large, cytologically normal histiocytes with indented nuclei and abundant cytoplasm. Multinucleated giant cells are rare and there is a low mitotic index.

Affected lymph nodes are characterised by numerous histiocytes in the subcapsular and medullary sinuses, the paracortex and along the trabeculae [45]. Cells from two of six tumours were positive for non‐specific esterase and acid phosphatase [26].

3.3. Clinical behaviour and treatment Treatment with glucocorticoids and cytotoxic drugs is generally ineffective, although a few dogs will show a response [11, 26, 45, 53]. Some authors report apparent success with immunosuppressive doses of glucocorticoids in forms of systemic histiocytosis that are limited to the skin. However, such forms of systemic histiocytosis would be difficult to differentiate from cutaneous histiocytosis, which is often corticosteroid‐responsive. There are anecdotal reports of a response to bovine thymosin fraction 5 in two dogs [26], similar to the response of human Histiocytosis‐X to thymic extract [55]. More recently, success with immunosuppressive doses of cyclosporin or leflunamide, which are potent inhibitors of T‐cell activation, has been reported [11, 24, 45]. Therapeutic success with doxorubicin was reported in one dog [45]. Ocular lesions may respond to topical corticosteroids or cyclosporin [53]. The majority of animals require either long‐term therapy to avoid immediate relapses, or intermittent therapy following the development of new lesions [45].

3.4. Immunophenotype Immunophenotypic studies reveal that the cells have an activated dermal dendritic cell phenotype, identical to those of cutaneous histiocytosis [24, 45]. It is therefore easy to understand the consistent involvement of the lymph nodes in this disease, since this is the target for migration of activated dendritic cells. However, the lack of lymph node involvement in cutaneous histiocytosis, whose cells have the same phenotype, indicates that there must be other factors responsible for determining migratory behaviour.

3.5. Other remarks It has been suggested that, on the basis of immunophenotypic studies, cutaneous histiocytosis and systemic histiocytosis represent opposite ends of a spectrum of histiocytic proliferative disease, and that it is their tissue tropism that differentiates them. Although a range of clinical behaviours may be exhibited by each disease, cutaneous histiocytosis is likely to remain skin‐limited and systemic histiocytosis is likely to be more systemically widespread. The pleiomorphic appearance of the lesional cells of systemic histiocytosis under electron microscopy and the presence of cells with an intermediate appearance suggest the cells represent a continuum of histiocytic differentiation [24, 45].

It is further suggested that both cutaneous and systemic histiocytosis are proliferative diseases arising in the context of disordered immune regulation arising from a defective interaction between dendritic cells and T‐lymphocytes. The end result of this interaction is chronic proliferation of the histiocytic cells, possibly mediated via T‐lymphocyte derived cytokines [24, 45]. Regression of the lesions is not reported to be associated with the appearance of CD8+ T‐cells, as it is for cutaneous histiocytoma [11, 18]. It is unknown whether the T‐cells are recruited secondarily to the lesions, perhaps because of the cytokine microenvironment, or whether the lymphocytes are the primary proliferating population, which attract dendritic cells to the lesions [11].

4. Malignant Histiocytosis (Disseminated Histiocytic Sarcoma)

4.1. Incidence and clinical features Malignant histiocytosis is a rare malignant neoplasm of dogs [56‐73]. The disease has been reported in many breeds, notably the Bernese Mountain Dog, Rottweiler, Golden Retriever and Flat Coat Retriever. The disease typically affects middle‐aged to older animals, primarily males. It is familial in the Golden Retriever and has been reported in closely‐related Bernese Mountain Dogs, predominantly males [58].

Primary lesions occur in the spleen, liver, lung, lymph node and bone marrow [74]. Cutaneous lesions are rare, but consist of multiple firm dermal or subcutaneous nodules, arising anywhere on the body, and may be hairless or ulcerated [65]. A gingival lesion was reported in one dog [71]. The clinical course of the disease is more rapid than the other histiocytoses and is inevitably fatal. Animals usually present with anorexia, lethargy, weight loss and pyrexia. Clinical signs referable to other organs are common and include hepatosplenomegaly, lymphadenopathy, lameness and ocular signs (e.g. hyphaema), pulmonary signs (e.g. dyspnoea, coughing), gastrointestinal signs (e.g. diarrhoea) or neurological signs (e.g. paraparesis, paraplegia or tetraplegia, inco‐ordination and seizures) [56‐59, 64, 70, 72‐75].

Clinicopathological abnormalities may include regenerative or non‐regenerative anaemia, thrombocytopaenia, neutrophilia, monocytosis, lymphopaenia, leucopaenia, pancytopaenia, hypoalbuminaemia, hyperglobulinaemia,

hyperferritinaemia, hypercalcaemia and elevations in serum alanine transaminase and alkaline phosphatase [56, 57, 59, 67, 70, 76, 77]. A bone marrow aspirate may reveal infiltration by atypical histiocytes, which may exhibit erythrophagocytosis, and reduced or increased cellularity [56], reduced myeloid: erythroid ratio [57] and megakaryocytic hyperplasia [57]. A histiocytic leukaemia has been reported as a rare occurrence [58, 72, 76]. A confirmed CD1c+ CD11c+ CD11d‐ dendritic cell leukaemia, with lesions in the lung, liver, lymph node and bone marrow, has also been reported [78]. Radiography may reveal nodular or diffuse hepatomegaly or splenomegaly, abdominal masses or lymphadenopathy, ascites, nodular pulmonary masses or mediastinal masses, lobar consolidation, tracheobronchial, mediastinal and sternal lymphadenopathy, pleural effusion and a marked pulmonary interstitial pattern, and osseous lytic lesions [56, 59, 64, 79‐81]. Ultrasonography may reveal discrete hypoechoic regions within the spleen [70].

4.2. Pathological features Histologically, the infiltrates in all organs are composed of large, pleiomorphic mononuclear cells and multinucleated giant cells within a fine fibrovascular stroma. Marked cytological atypia and frequent bizarre mitotic figures are present. Within the skin, there is nodular to diffuse deep dermal and subcutaneous infiltration by atypical histiocytes, accompanied by lymphocytes and neutrophils. Histiocytes frequently show phagocytosis of erythrocytes, neutrophils and other histiocytes [1, 56, 58, 61]. Occasional areas of necrosis and haemorrhage are present. Histiocytes may also be present in the liver, spleen, lymph nodes, myocardium, skeletal muscles, gastrointestinal tract, pancreas, mesenteric lymph nodes, kidneys, adrenal glands, salivary glands, axial and appendicular skeleton, epidural space and brain [56, 58, 59, 65].

The histiocytic cells are positive for lysozyme, non‐specific esterase, acid phosphatase, desmin and cathepsin B in all cases, and α‐1‐trypsin in some [58, 61, 70, 82, 83].

4.3. Clinical behaviour and treatment This disease has a rapid clinical progression, despite therapeutic intervention, and rapidly progresses to death or euthanasia. In one series of 20 dogs, 1 died, 6 were euthanised at the owners request, 7 were euthanised because of a very poor general condition and 5 were lost to follow‐up [74]. Response to therapy with corticosteroids or cytotoxic agents has generally been poor and of short duration [57, 63, 65, 84]. More recently, the use of lomustine has been reported, although median survival times remain in the 3‐6 month range [85]. In another study of dogs treated with lomustine, the median survival time of 59 dogs was 106 days, although 3 dogs with minimal residual disease at the initiation of chemotherapy lived for 433 days or more [86]. There are a few reports of responses to chemotherapy with doxorubicin, liposomal doxorubicin and paclitaxel [87, 88]

One study reported some success using a human MHC‐non‐restricted cytotoxic T‐cell line, TALL‐104 [89]. This cell line has potent tumoricidal activity and 45

induced a complete remission of between 4 and 16 months in four dogs. However, it is not clear that these dogs had malignant histiocytosis.

4.4. Immunophenotype Immunophenotypic analysis confirms their Langerhans cell phenotype [24]. The lesional histiocytes express CD1, CD11c and MHC‐II and do not express CD4 and Thy‐1, thus differentiating them from the reactive histiocytoses of dermal dendritic cell origin.

4.5. Other remarks It has been suggested that systemic histiocytosis and malignant histiocytosis represent opposite ends of a spectrum of histiocytic proliferative disease [58, 90]. This is supported by the observation that animals affected with both diseases may come from the same family lines [59]. However, the diseases are distinct clinically and pathologically. Systemic histiocytosis occurs in younger dogs (mean age of onset four years old) and is characterised by a prolonged, fluctuating clinical course. The skin and peripheral lymph nodes are consistently involved. The infiltrating histiocytes lack cytological atypia and multinucleate giant cells are rare. Malignant histiocytosis occurs in older dogs (mean age of onset seven years old) and has a rapidly progressive clinical course. Infiltrates occur frequently in the lung, lymph nodes and liver, and the skin and eyes are rarely involved. The infiltrates consist of extremely pleiomorphic mononuclear cells and multinucleate giant cells, which frequently manifest cytological atypia. In addition, animals with an intermediate stage have not been described, and immunophenotyping suggests they have a different cell of origin [24, 58]. In the absence of imunophenotypic studies, confusion with the other histiocytoses, the large cell form of cutaneous lymphoma and anaplastic pulmonary carcinoma is possible [24, 59].

5. Localised Histiocytic Sarcoma

5.1. Incidence and clinical features Localised histiocytic sarcoma (LHS) is a solitary, rapidly‐growing soft tissue mass with a moderate to high metastatic rate [91]. It has been reported in many breeds, particularly Flat Coated Retrievers, Bernese Mountain Dogs, Labrador retrievers and Rottweilers [91]. Most affected dogs are 6 to 11 years old.

Primary lesions are commonly found in the subcutaneous and underlying tissues, often on the extremities and particularly near joints, particularly the stifle and elbow. Other predilection sites include the spleen, liver, lung, brain, vertebra, oral cavity, nasal cavity, eye and bone marrow [73, 91‐96]. Subcutaneous lesions often invade the deep dermis, skeletal muscle and fascia and periarticular tumours often invade the joint capsule, ligaments, tendons and muscle. Animals with ocular histiocytic sarcoma may show pulmonary masses (20%) and lymphadenopathy (8%), but systemic signs are uncommon [96].

A number of abnormalities may be identified on ophthalmic examination, including corneal oedema, conjunctival hyperaemia, increased intra‐ocular pressure, buphthalmos, decreased vision or blindness, an intra‐ocular mass, hyphaema, a painful eye, a red eye, episcleral swelling, exophthalmos and retinal detachment. Uveitis and secondary glaucoma are common.

Metastasis to the regional lymph nodes may be seen in up to 60% of cases and distant metastasis may be seen in 38‐60% of dogs [74, 92]. If organs systems other than the skin, e.g. spleen, are involved then either a more aggressive form of the disease (haemophagocytic histiocytic sarcoma) or disseminated histiocytic sarcoma should be suspected.

Clinical signs are generally referable to the presence of a mass and depend on the organ system involved. Most dogs have a palpable cutaneous or subcutaneous mass, but are not systemically ill, unlike dogs with DHS [74, 92]. Many dogs with a mass on the extremity present with lameness initially and the mass is identified on physical examination or diagnostic imaging. Anaemia and hypoproteinaemia may be seen in dogs with splenic lesions [97]. Hypocholesterolaemia is uncommon in histiocytic sarcoma, compared to haemophagocytic histiocytic sarcoma [98].

5.2. Pathological features The typical histological appearance is a highly cellular, invasive mass of pleiomorphic histiocytic cells which effaces the normal tissue architecture and shows multiple areas of necrosis. Cellular morphology and arrangement of cells within the lesion varies between lesions and between areas within a single lesion. Two patterns of cellular morphology are reported, one comprising individualised round cells with abundant cytoplasm and other plump spindle cells with long cytoplasmic processes. Often both cell types are present within a lesion [74]. The neoplastic cells demonstrate anisocytosis, with abundant eosinophilic and finely granular to foamy cytoplasm, and anisokaryosis, with large centrally‐located nuclei and prominent nucleoli of varying sizes. Multinucleated giant cells and mitotic figures, including bizarre forms, are common [73, 74, 92, 99]. Phagocytosis by the tumour cells is occasionally seen. A variable reactive leucocytic infiltrate is seen, with neutrophils commonly present along with small lymphocytes arranged as dispersed cells or small aggregates [74]. Tumour cells are generally positive for vimentin, but fewer cells label positively for lysozyme [73, 92].

5.3. Clinical behaviour and treatment Local management of non‐metastasised LHS comprises aggressive surgical excision, with 3cm lateral margins and at least one fascial plane deep, as recommended for all soft tissue sarcomas. For distal limb tumours, adequate local control may require amputation [100].

The role of radiotherapy for gross disease or in the adjuvant setting has not been well investigated, although other round cell tumours are radiosensitive. Limited efficacy in a small number of cases has been reported [74, 85, 100], including

a lesion on the tongue [74]. Post‐operative chemotherapy should be considered, given the aggressive biological behaviour and high metastatic rate [74, 85, 100], and single agent therapy with doxorubicin or lomustine is recommended although there are few reports of this [85, 86].

In one multicentre retrospective study, 59 dogs with incompletely resected or metastatic histiocytic sarcoma (23 of which were tested and found to express CD18) were treated with lomustine at 60‐90mg/m2 every 3‐4 weeks, for a median of 4 doses. Overall median survival time was 106 days, although all 3 dogs with minimal residual disease lived 433 days or more before relapse [86]. In this study, thrombocytopaenia and hypoalbuminaemia were associated with a poor outcome (survival < 1 month).

Dogs with LHS have a more favourable prognosis than dogs with DHS and euthanasia should not be considered automatically, although the prognosis remains guarded. An accurate prognosis for dogs with LHS of the skin and subcutis is not available. In one case series of 13 dogs with LHS affecting the extremities, wide surgical excision or amputation alone provided local control in five dogs and no metastasis was recorded, but the length of follow‐up is not stated and follow‐up data was only available for 5 of the dogs [74].

In one single centre retrospective study of 37 flat‐coated retrievers with histiocytic sarcomas of varying clinical stages, the overall median survival time was 123 days, with a median survival of 167 days for those dogs receiving any non‐palliative therapy compared to 17 days for those dogs receiving palliative therapy [85]. Dogs given chemotherapy had a median survival of 185 days compared to 34 days for dogs that were not. Dogs receiving radiotherapy survived a median of 182 days compared to 60 days for dogs that did not. A protocol involving radiotherapy and CCNU induced minimal toxicity and provided a median survival of 208 days compared with 68 days for all other dogs. In this series, diagnosis was based on histological criteria and CD18 immunohistochemistry was attempted, but was unsuccessful. It is not clear whether this was for technical reasons or whether these tumours were not CD18+ histiocytic sarcomas.

The prognosis for dogs with synovial LHS is poor. In one case series, the median survival time was 5.3 months for all dogs and 6 months for dogs undergoing amputation, with or without chemotherapy [100]. In this series the metastatic rate was 91%.

The prognosis for dogs with LHS affecting internal organs, such as the spleen is poor, with a median survival time of 1 month and a 1 year survival rate of 0‐20%. Factors associated with a poor prognosis in splenic LHS include lymphoid:fibrohistiocytic ratio, mitotic index and histologic grade.

5.4. Immunophenotype There is no difference between the immunophenotype of localised and disseminated histiocytic sarcoma [74]. Expression of CD45 and CD11a/CD18 confirms their leucocytic origin. Tumour cells express CD1a, CD1b, CD1c, CD11c, CD44, ICAM‐1 and MHC‐II, consistent with a myeloid dendritic cell origin. [101, 102]. There is inconsistent expression of CD11b, CD45RA, CD49d, and ICAM‐3. Thy‐1 expression was not expressed by the majority of tumour tissue and was limited to a perivascular subpopulation of tumour cells in 8 out of 30 cases [74].

Lack of expression of E‐cadherin allows differentiation from cutaneous histiocytoma [31, 74] and lack of expression of CD4 allows differentiation from cutaneous and systemic histiocytosis [18, 45, 74]

Reactive lymphocytic infiltrates of mainly CD8+, TCRαβ+ T‐cells, with smaller numbers of CD21+ B‐cells.CD4+ T‐cells and TCRγδ+ T‐cells were rare [74].

In one study, 35 tumours previously diagnosed as synovial cell tumours were re‐ evaluated by histology and immunohistochemistry and 18 of these were re‐ classified as histiocytic sarcomas on the basis of CD18 expression [100].

5.5. Other remarks Localised and disseminated histiocytic sarcoma may represent two different stages on a continuum of the same disease. There is no known histological or immunohistochemical method for differentiating between localised and disseminated histiocytic sarcoma.

6. Haemophagocytic Histiocytic Sarcoma

6.1. Incidence and clinical features A particular subtype of histiocytic sarcoma with clinical, morphological and immunophenotypic features different from localised or disseminated histiocytic sarcoma has recently been described [98]. Breeds commonly affected include Bernese Mountain Dog, Golden Retriever, Rottweiler, and Labrador Retriever, with a range of ages from 2 to 13 years old. No sex predilection is present.

Clinical signs include lethargy, inappetance, weight loss and pale mucous membranes. Jaundice was present in a few animals and clinical signs were present from 2 to 32 weeks. Splenomegaly was present in all animals examined with concurrent hepatomegaly in a many animals. Many animals were initially presented for evaluation of immune‐mediated haemolytic anaemia. Thrombocytopaenia and a mild prolongation of PTT were seen in the majority of animals. Other common abnormalities included hypoproteinaemia, hypoalbuminaemia, hypocholesterolaemia and hyperbilirubinaemia. Coombs tests, performed in 8 dogs, were negative.

Gross lesions were most common in the liver and spleen. Splenic lesions were multifocal tan to red masses of 3‐5cm in diameter, accompanied by discoloured green‐tan to white infarcts in some cases. Hepatomegaly was diffuse with tan to red discolouration, a prominent reticular pattern and a textured capsular surface, but masses were not observed. Multiple white nodules were also identified in the lungs (12% cases) and the kidneys.

6.2. Pathological features Microscopic lesions most consistently involved the spleen, liver, bone marrow and lung. The splenic red pulp was expanded by histiocytes which frequently obliterated the white pulp. Infarction and thrombosis of the red pulp was also seen. Phagocytosis of red cells, red cell precursors and granulocytes was seen along with the presence of haemosiderin deposits within histiocytes. Foci of extra‐medullary haematopoiesis were often seen close to histiocytic infiltrates. A small infiltrate of lymphocytes and plasma cells were also seen.

There was considerable variation in the morphology of the histiocytes, some of which resembled normal splenic red pulp macrophages and some of which displayed gross cytological atypia. Multinucleated giant cells were frequent and mitotic figures were common. A histiocytic infiltrate was present in the liver, lung and bone marrow in all dogs examined. Foci of extra‐medullary haematopoiesis were also seen in the liver. Microscopic lesions were identified in lymph nodes in approximately half the animals.

Splenic and bone marrow aspirates revealed well‐differentiated haemophagocytic histiocytes in approximately half the animals, which may have confused the diagnosis of HHS. However, the large number of macrophages and their presence in aggregates was a useful feature to allow differentiation from the changes seen in immune‐mediated haemolytic anaemia

6.3. Clinical behaviour and treatment All animals with haemophagocytic histiocytic sarcoma were euthanised and no description of the clinical behaviour, response the therapy or prognosis can be given.

6.4. Immunophenotype Neoplastic histiocytes in formalin‐fixed specimens express CD11d, CD18 and CD45. Positive labelling with CD11d allowed small numbers of histiocytes invading vessels in liver and lung to be identified, which might have been missed on examination of standard H&E sections [98].

In frozen sections, neoplastic histiocytes express CD45 and MHC‐II, with more variable expression of CD1c, CD80, CD86 and Thy‐1. CD11b and CD11c expression was less common than CD11d expression. CD45RA and CD4, an activation marker of macrophages and DCs, were not expressed. This

indicates that HHS is a proliferative disorder of macrophages, which may be differentiated from HS by expression of CD11d and CD4.

Infiltrating lymphocytes comprised mainly a mixed population of CD4+ and CD8+ T‐cells, with a preponderance of TCRαβ, with smaller numbers of CD21+ CD70a B‐cells and CD79a+ plasma cells.

6.5. Other remarks HHS affects the same breeds as localised and disseminated HS, but may be differentiated by immunophenotypic studies. HHS is a proliferative disorder of splenic red pulp and bone marrow macrophages, rather than dendritic cells. However, mixed disorders have been identified, i.e. animals with haemophagocytic histiocytic sarcoma occurring in the context of histiocytic sarcoma of DC origin [98].

Summary

The canine histiocytic diseases are a group of related disorders which continue to present a considerable diagnostic challenge as it may be difficult to make a definitive diagnosis or differentiate between members of this group on the basis of clinical features and routine histological examination. The key to understanding this heterogeneous group of disorders and making a definitive diagnosis lies in establishing the immunophenotype of the proliferative cells. However, this technique requires specialised immunological techniques and access to fresh rather than fixed tissue. Many advances have been made in recent years in the understanding of this group of disorders and these have helped to provide a rational classification of these disorders as well as to guide appropriate therapeutic strategies.

References:

1. Goldschmidt, M.H. and F.S. Shofer, Canine cutaneous histiocytoma., in Skin tumours of the dog and cat., M.H. Goldschmidt and F.S. Shofer, Editors. 1992, Pergamon Press: Oxford. p. 222‐230. 2. Walder, E.J. and T.L. Gross, Neoplastic diseases of the skin., in Veterinary Dermatopathology: A macroscopic and microscopic evaluation of canine and feline skin disease., T.L. Gross, P.J. Ihrke, and E.J. Walder, Editors. 1992, Mosby Year Book: St Louis. p. 465‐473. 3. Scott, D.W., W.H. Miller, and C.E. Griffin, Muller and Kirk's Small Animal Dermatology. 5th ed. 1995, Philadelphia, PA: W B Saunders. 1076‐1078. 4. Vail, D.M. and S.J. Withrow, Tumors of the skin and subcutaneous tissues., in Small Animal Clinical Oncology, S.J. Withrow and E.G. MacEwen, Editors. 1996, WB Saunders: Philadelphia, PA. p. 167‐191. 5. Conroy, J.D., Canine skin tumours. Journal of the American Animal Hospital Association, 1983. 19: p. 91‐114. 6. Rothwell, T.L., et al., Skin neoplasms of dogs in Sydney. Australian Veterinary Journal, 1987. 64: p. 161‐4.

7. Bostock, D.E., Neoplasms of the skin and subcutaneous tissues in dogs and cats. British Veterinary Journal, 1986. 142: p. 1‐19. 8. Mulligan, R.M., Neoplastic diseases of dogs. II. Mast cell sarcoma, lymphosarcoma, histiocytoma. Archives of Pathology, 1948. 46: p. 477‐492. 9. Howard, E.B. and S.W. Nielsen, Cutaneous histiocytomas of dogs. National Cancer Institute Monographs, 1969. 32: p. 321‐327. 10. Taylor, D.O.N., C.R. Dorn, and O.H. Luis, Morphologic and biologic characteristics of the canine cutaneous histiocytoma. Cancer Research, 1969. 29: p. 83‐92. 11. Affolter, V.K. and P.F. Moore, Canine cutaneous histiocytic disease., in Current Veterinary Therapy: Small Animal Practice XIII, J.D. Bonagura, Editor. 2000, W. B. Saunders: Philadelphia. p. 589‐591. 12. Frye, F.L., J. Carney, and J.‐P.E. Cucuel, Generalized eruptive histiocytoma in a dog. Journal of the American Veterinary Medical Association, 1969. 155: p. 1465‐1466. 13. Howard, E.B., Eruptive histiocytoma in a dog. Journal of the American Veterinary Medical Association, 1970. 156: p. 140‐143. 14. Garma, A.A. and E. Fromer, Generalized cutaneous histiocytoma in a dog (a case report). Veterinary Medicine Small Animal Clinician, 1979. 74: p. 1269‐70. 15. Bender, W.M. and G.H. Muller, Multiple, resolving, cutaneous histiocytoma in a dog. Journal of the American Veterinary Medical Association, 1989. 194: p. 535‐537. 16. Kumar, N., et al., Multiple canine cutaneous histiocytoma in an Alsatian bitch. Indian Journal of Veterinary Surgery, 1993. 14: p. 98‐99. 17. Gelatt, K.N., Histiocytoma of the eyelid of a dog. Veterinary Medicine Small Anim Clinican, 1975. 70: p. 305. 18. Moore, P.F., et al., Canine cutaneous histiocytoma is an epidermotropic Langerhans cell histiocytosis that expresses CD1 and specific Beta2 integrin molecules. American Journal of Pathology, 1996. 148: p. 1699‐1708. 19. Kelly, D.F., Canine cutaneous histiocytoma: A light and electron microscopic study. Veterinary Pathology, 1970. 7: p. 12‐27. 20. Duncan, J.R. and K.W. Prasse, Cytology of canine cutaneous round cell tumours. Veterinary Pathology, 1979. 16: p. 673‐679. 21. Pulley, L.T. and A.A. Stannard, Tumors of the skin and soft tissue., in Tumors in domestic animals, J.E. Moulton, Editor. 1990, University of California Press: Berkeley, CA. p. 23‐87. 22. Jones, T.S., R.D. Hunt, and N.W. King, Veterinary Pathology. 6th ed. 1997, Baltimore, MD: Williams & Wilkins. 867‐869. 23. Cockerell, G.L. and D.O. Slauson, Patterns of lymphoid infiltrate in the canine cutaneous histiocytoma. Journal of Comparative Pathology, 1979. 89: p. 193‐203. 24. Moore, P.F., et al., The use of immunological reagents in defining the pathogenesis of skin diseases. Advances in Veterinary Dermatology, 1998. 3: p. 77‐94. 25. Glick, A.D., M. Holscher, and G.R. Campbell, Canine cutaneous histiocytoma: Ultrastructural and cytochemical observations. Veterinary Pathology, 1976. 13: p. 374‐ 380. 26. Moore, P.F., Systemic histiocytosis of Bernese Mountain dogs. Veterinary Pathology, 1984. 21: p. 554‐63. 27. Sandusky, G.E., W.W. Carlton, and K.A. Wightman, Diagnostic immunohistochemistry of canine round cell tumours. Veterinary Pathology, 1987. 24: p. 495‐499. 28. Brokaw, J.J., et al., Glucocorticoidinduced apoptosis of dendritic cells in the rat tracheal mucosa. American Journal of Respiratory Cell and Molecular Biology, 1998. 19: p. 598‐ 605. 29. Marchal, T., et al., Immunophenotypic and ultrastructural evidence of the Langerhans cell origin of the canine cutaneous histiocytoma. Acta Anatomica, 1995. 153: p. 189‐202. 30. Aiba, S., et al., Upregulation of a4 integrin on activated Langerhans cells: analysis of adhesion molecules on Langerhans cells relating to their migration from skin to draining lymph nodes. Journal of Investigative Dermatology, 1993. 100: p. 143‐147. 31. Baines, S.J., E.F. McInnes, and I. McConnell, Ecadherin expression in canine cutaneous histiocytoma. Veterinary Record, 2008. 162: p. 509‐513. 32. Theilen, G.H. and B.R. Madewell, Tumours of the skin, in Veterinary Cancer Medicine, G.H. Theilen and B.R. Madewell, Editors. 1979, Lea & Febiger: Philadelphia, PA. p. 123‐191. 33. Quimby, H.H., Cutaneous histiocytosis vs leptospirosis. Journal of the American Veterinary Medical Association, 1986. 188: p. 1142‐1143.

34. Baines, S.J., et al., CCH cells are potent stimulators in the allogeneic mixed lymphocyte reaction. Veterinary Immunology Immunopathology, 2007. 119: p. 316‐321. 35. Magnol, J.P., P. Devauchelle, and S. Achache, Canine histiocytoma bibliographical study and personal observations. Revue de Medecine Veterinaire, 1982. 133: p. 391‐401. 36. Kaim, U., et al., The regression of a canine Langerhans cell tumour is associated with increased expression of IL2, TNFalpha, IFNgamma and iNOS mRNA. Immunology, 2006. 118: p. 472‐482. 37. Kipar, A., et al., Expression of major histocompatibility complex class II antigen in neoplastic cells of canine cutaneous histiocytoma. Veterinary Immunology and Immunopathology, 1998. 62: p. 1‐13. 38. Baines, S.J., et al., Maturation states of dendritic cells in canine cutaenous histiocytoma. Veterinary Dermatology, 2000. 11: p. 9‐10. 39. Nagata, M., et al., Progressive Langerhans' cell histiocytosis in a puppy. Veterinary Dermatology, 2000. 11: p. 241‐6. 40. Calderwood Mays, M.B. and J.A. Bergeron, Cutaneous histiocytosis in dogs. Journal of the American Veterinary Medical Association, 1986. 188: p. 377‐381. 41. Thornton, R.N. and C.J. Tisdall, Multiple cutaneous histiocytosis in 2 dogs. New Zealand Veterinary Journal, 1988. 36: p. 192‐193. 42. Scott, D.W., Canine cutaneous histiocytoses, in Current Veterinary Therapy: Small Animal Practice X, R.W. Kirk, Editor. 1989, W B Saunders: Philadelphia. p. 625‐626. 43. Gross, T.L., P.J. Ihrke, and E.J. Walder, Noninfectious nodular and diffuse granulomatous and pyogranulomatous diseases of the dermis., in Veterinary Dermatopathology: A macroscopic and microscopic evaluation of canine and feline skin disease., T.L. Gross, P.J. Ihrke, and E.J. Walder, Editors. 1992, Mosby Year Book: St Louis, MO. p. 190‐206. 44. Palmeiro, B.S., et al., Cutaenous reactive histiocytosis in dogs: a retrospective evaluation of 32 cases. Veterinary Dermatology, 2007. 18: p. 332‐340. 45. Affolter, V.K. and P.F. Moore, Canine cutaneous and systemic histiocytosis: reactive histiocytosis of dermal dendritic cells. American Journal of Dermatopathology, 2000. 22: p. 40‐48. 46. Panic, R., Sterile pyogranulomatous and granulomatous disorders of dogs and cats., in Current Veterinary Therapy: Small Animal Practice XI., R.W. Kirk, Bonagura, J. D., Editor. 1992, WB Saunders: Philadelphia, PA. p. 536‐539. 47. Collins, B.K., et al., Idiopathic granulomatousdisease with ocular adnexal and cutaneous involvement in a dog. Journal of the American Veterinary Medical Association, 1992. 201: p. 313‐316. 48. Affolter, V.K., et al., Canine reactive histiocytosis: dysregulation of the immune system, characterisation of the cytokine milieu. Veterinary Dermatology, 2003. 14: p. 216. 49. Moore, P.F., Utilization of cytoplasmic lysozyme immunoreactivity as a histiocytic marker in canine histiocytic disorders. Veterinary Pathology, 1986. 23: p. 757‐762. 50. Scott, D.W., D.K. Angarano, and M.M. Suter, Systemic histiocytosis in two dogs. Canine Practice, 1987. 14: p. 7‐12. 51. Scherlie, P.H., et al., Ocular manifestation of systemic histiocytosis in a dog. Journal of the American Veterinary Medical Association, 1992. 201: p. 1229‐1232. 52. Brearley, M.J., et al., Systemic histiocytosis in a Bernese Mountain Dog. Journal of Small Animal Practice, 1994. 35: p. 271‐274. 53. Paterson, S., P. Boydell, and R. Pike, Systemic histiocytosis in the Bernese Mountain dog. Journal of Small Animal Practice, 1995. 36: p. 233‐236. 54. Affolter, V.K. and P.F. Moore. Histiocytes in skin disease. in Proceedings of the 4th World Congress of Veterinary Dermatology. 2000. San Francisco, CA. 55. Osband, M.E., et al., Histiocytosis X. Demonstration of abnormal immunity, Tcell histamine H2 receptor deficiency and successful treatment with thymic extract. New England Journal of Medicine, 1981. 304: p. 146‐153. 56. Scott, D.W., et al., Lymphoreticular neoplasia in a dog resembling malignant histiocytosis (histiocytic medullary reticulosis) in man. Cornell Veterinarian, 1979. 69: p. 176‐197. 57. Wellman, M.L., et al., Malignant histiocytosis in four dogs. Journal of the American Veterinary Medical Association, 1985. 187: p. 919‐921. 58. Moore, P.F. and A. Rosin, Malignant histiocytosis of Bernese Mountain Dogs. Veterinary Pathology, 1986. 23: p. 1‐10.

59. Rosin, A., P.F. Moore, and R. Dubielzig, Malignant histiocytosis in Bernese Mountain Dogs. Journal of the American Veterinary Medical Association, 1986. 188: p. 1041‐1045. 60. Tisdall, R.N., R.N. Thornton, and B.M. Veal, Malignant histiocytosis in a Bernese Mountain Dog. New Zealand Veterinary Journal, 1987. 36: p. p43. 61. Hayden, D.W., et al., Disseminated malignant histiocytosis in a Golden Retriever: Clinicopathologic, ultrastructural and immunohistochemical findings. Veterinary Pathology, 1993. 30: p. 256‐264. 62. Kohn, B., et al., Malignant histiocytosis of the dog 26 cases (19891992). Kleintierpraxis, 1993. 38: p. 409‐416. 63. Peaston, E.A., R.J. Munn, and B.R. Madewell, Malignant histiocytosis. Journal of Veterinary Internal Medicine, 1993. 7: p. 101‐103. 64. Schmidt, M.L., et al., Clinical and radiographic manifestations of canine malignant histiocytosis. Veterinary Quarterly, 1993. 14: p. 117‐120. 65. Uno, Y., et al., Malignant histiocytosis with multiple skin lesions in a dog. Journal of Veterinary Medical Science, 1993. 55: p. 1059‐1061. 66. Marholdt, F. and A. Besch, Malignant histiocytosis in a Bernese Mountain Dog. Praktische Tierarzt, 1994. 75: p. 690‐691. 67. Newlands, C.E., D.M. Houston, and D.Y. Vasconceles, Hyperferritinaemia associated with malignant histiocytosis in a dog. Journal of the American Veterinary Medical Association, 1994. 205: p. 849‐851. 68. Freeman, L., et al., Malignant histiocytosis. Journal of Veterinary Internal Medicine, 1995. 9: p. 171‐173. 69. MacEwen, E.G., Malignant histiocytosis., in Small Animal Clinical Oncology, S.J. Withrow, MacEwen, E. G., Editor. 1996, WB Saunders: Philadelphia, PA. p. 505‐509. 70. Ramsey, I.K., et al., Malignant histiocytosis in three Bernese Mountain Dogs. Veterinary Record, 1996. 138: p. 440‐444. 71. Carioto, L., Malignant histiocytosis in a Bernese Mountain Dog presenting as a mandibular mass. Canadian Veterinary Journal, 1997. 38: p. 105‐107. 72. Schouben, Y., D. Heripret, and C. Fournel, Malignant leukaemia histiocytosis in the Bernese Mountain Dog. Pratique Medicale et Chirurgicale de l'Animal de Compagnie, 1998. 33: p. 317‐322. 73. Chandra, A.M.S. and P.E. Ginn, Primary malignant histiocytosis of the brain in a dog. Journal of Comparative Pathology, 1999. 121: p. 77‐82. 74. Affolter, V.K. and P.F. Moore, Localized and disseminated histiocytic sarcoma of dendritic cell origin in dogs. Veterinary Pathology, 2002. 39: p. 74‐83. 75. Suzuki, M., et al., A comparative pathological study on granulomatous meningoencephalomyelitis and central malignant histiocytosis in dogs. Journal of Veterinary Medical Science, 2003. 65: p. 1319‐1324. 76. Madewell, B.R. and B.F. Feldman, Characterisation of anaemias associated with neoplasia in small animals. Journal of the American Veterinary Medical Association, 1980. 176: p. 419‐425. 77. Uehlinger, P., et al., Hypercalcemia in dogs A retrospective study of 46 cases. Schweizer Archiv Fur Tierheilkunde, 1998. 140: p. 188‐197. 78. Allison, R.W., et al., Dendritic cell leukaemia in a Golden Retriever. Veterinary Clinical Pathology, 2008. 37: p. 190‐197. 79. Berry, C.R., et al., Radiographic diagnosis: malignant histiocytosis. Veterinary Radiology, 1989. 30: p. 142‐144. 80. Lord, P.F. and R.R. Dubielzig, Radiographic diagnosis: Malignant histiocytosis. Veterinary Radiology, 1991. 32: p. 17‐18. 81. Schmidt, M.L., G. Rutteman, and P. Wolvekamp, Canine malignant histiocytosis (MH): clinical and radiographic findings. Tijdschrift voor Diergeneeskunde, 1992. 117: p. S43‐ 44. 82. Moore, P.F., Characterization of cytoplasmic lysozyme immunoreactivity as a histiocytic marker in normal canine tissues. Veterinary Pathology, 1986. 23: p. 763‐769. 83. Thoolen, R., et al., Malignant fibrous histiocytomas in dogs and cats an immunohistochemical study. Research In Veterinary Science, 1992. 53: p. 198‐204. 84. Shaiken, L.C., S.M. Evans, and M.H. Goldschmidt, Radiographic findings in canine malignant histiocytosis. Veterinary Radiology, 1991. 32: p. 237‐242.

85. Fidel, J., et al. Retrospective analysis of flatcoated retrievers with histiocytic sarcomas. in Proceedings of the 12th Annual Congress of the European Society for Veterinary Internal Medicine. 2002. Munich. 86. Skorupski, K.A., et al., CCNU for the treatment of dogs with histiocytic sarcoma. Journal of Veterinary Internal Medicine, 2007. 21: p. 121‐126. 87. Poirier, V.J., A.E. Hershey, and K.E. Burgess, Efficacy and toxicity of paclitaxel (Taxol) for the treatment of canine malignant tumours. Journal of Veterinary Internal Medicine, 2004. 18: p. 219‐222. 88. Vail, D., et al., Preclinical trial of doxorubicin entrapped in sterically stabilised liposomes in dogs with spontaneously arising malignant tumours. Cancer Chemotherapy Pharmacology, 1997. 39: p. 410‐416. 89. Visonneau, S., et al., Successful treatment of canine malignant histiocytosis with the human major histocompatibility complex nonrestricted cytotoxic Tcell line TALL104. Clinical Cancer Research, 1997. 3: p. 1789‐1797. 90. Padgett, G.A., et al., Inheritance of histiocytosis in Bernese Mountain dogs. Journal of Small Animal Practice, 1995. 36: p. 93‐8. 91. Affolter, V.K. and P.F. Moore, Localised and disseminated histiocytic sarcoma of dendritic cell origin in dogs. Veterinary Pathology, 2002. 39: p. 74‐83. 92. Bass, M., et al., Localized histiocytic sarcoma in a dog: an uncommon diagnosis in forelimb lameness. Veterinary and Comparative Orthopaedics and Traumatology, 2004. 17: p. 48‐ 52. 93. Bettini, G., L. Mandrioli, and B. Brunetti, Canine splenic pathology: a retrospective study of 109 surgical samples with special emphasis on fibrohistiocytic nodules. European Journal of Veterinary Pathology, 2001. 7: p. 101‐108. 94. Cruz‐Arambulo, R., R. Wrigley, and B. Powers, Sonographic features of histiocytic neoplasms in the canine abdomen. Veterinary Radiology & Ultrasound, 2004. 45: p. 554‐ 558. 95. Ramirez, S., J.P. Douglass, and I.D. Robertson, Ultrasonographic features of canine abdominal malignant histiocytosis. Veterinary Radiology & Ultrasound, 2002. 43: p. 167‐ 170. 96. Naranjo, C., R.R. Dubielzig, and K.R. Friedrishs, Canine ocular histiocytic sarcoma. Veterinary Ophthalmology, 2007. 10: p. 179‐185. 97. Dobson, J., et al., Histiocytic sarcoma of the spleen in flatcoated retrievers with regenerative anaemia and hypoproteinaemia. Veterinary Record, 2006. 158: p. 825‐829. 98. Moore, P.F., V.K. Affolter, and W. Vernau, Canine hemophagocytic histiocytic sarcoma: A proliferative disorder of CD11d+macrophages. Veterinary Pathology, 2006. 43: p. 632‐ 645. 99. Fant, P., et al., Primary gastric histiocytic sarcoma in a dog A case report. Journal of Veterinary Medicine Series a‐Physiology Pathology Clinical Medicine, 2004. 51: p. 358‐ 362. 100. Craig, L.E., M.E. Julian, and J.D. Ferracone, The diagnosis and prognosis of synovial tumours in dogs: 35 cases. Veterinary Pathology, 2002. 39: p. 66‐73. 101. Affolter, V.K., Activation of human dendritic cells through CD40 crosslinking. Journal of Experimental Medicine, 1994. 180: p. 1263‐1272. 102. Morris, J.S., et al., Immunohistochemical and histopathologic features of 14 malignant fibrous histiocytomas from flatcoated retrievers. Veterinary Pathology, 2002. 39: p. 473‐ 479.

Intra‐ and Post‐operative electrochemotherapy (ECT) in the management of canine and feline tumours.

Ron Lowe PetCancerVet, Knaresborough, North Yorks.

Electrochemotherapy (ECT) comprises enhancing the anti‐tumour effectiveness of a cytotoxic drug by means of electroporation. For a detailed explanation and full bibliography see Cemazar et al 2008.

Electroporation involves subjecting cells to very brief, high voltage, square wave electrical pulses. This has been shown in vitro to cause the formation of holes in the cell membrane large enough for macromolecules (such as DNA) to enter the cell but such that most cells survive. It can, therefore, be used to allow the entry of large molecular weight cytotoxic drugs such as bleomycin, where conventional systemic administration would fail to achieve cytotoxic intracellular drug levels. With bleomycin, it can be shown in vitro that intracellular levels can be increased by a factor of 700 times or more. Other drugs have been used: cis‐platin is used clinically by some workers; other drugs have been used experimentally but with limited enhancement of effectiveness.

The electrical pulses used are typically 1200 Volts/cm for 0.1 milliseconds each and eight pulses are needed. The rate of pulse delivery is of little significance in vitro; in vivo, rates lower than 1 per second (1Hz) and higher than 1kHz cause marked muscle contractions. Older delivery systems used 1 Hz but more recent ones deliver at 5kHz so that the chain of 8 pulses is delivered in 0.0016 seconds.

In patients, the pulses are delivered using a variety of electrode systems: surface electrodes wires, plates or needles. The author has used only needle electrodes based on the Gehl array: 2 parallel rows of needles. The gap between the rows is 6 mm and the length of the needles is either 1 cm or 2.5 cm. The needles are mounted in a handle and are inserted into the tissue sequentially to cover the entire volume to be treated.

The cytotoxic drug can be administered either locally or intravenously. The author prefers the intravenous route on grounds of operator safety and maximal distribution through the tissue. Bleomycin achieves maximum tissue level 8 minutes after i/v injection (this is based on extremely limited experimental data from a single human patient but is generally quoted) and a high level is then maintained for about 20 minutes. The ECT should therefore be applied during this period.

Experimental studies have shown that ECT can be applied to normal tissues with no significant damage occurring so normal tissue in tumour margins can be treated with a high degree of selectivity for tumour cells. This phenomenon has not been adequately explained. Besides direct cytotoxic effects, ECT can be shown to have marked haemodynamic effects in treated tissues, again not fully

explained. The author's experience shows that some necrosis of apparently normal tissue in tumour margins can occur in some cases suggesting that perhaps if the load of tumour cells within the visually normal tissue is relatively high, toxins released during death of tumour cells may lead to damage to normal tissue. Some patients receiving ECT to margins show no necrosis suggesting a low tumour load in the area.

In theory, all tumours should be sensitive to ECT but clinical experience suggests this may not be the case. Mast cell tumours, squamous cell carcinomas and soft tissue sarcomas have a high response rate. Unlike the situation in humans, canine melanomas seem poorly responsive.

In humans, ECT has been used mainly in palliative treatment of cutaneous metastases but in veterinary studies, primary tumours have been treated with intention to cure. Standard Operating Procedures have been published for human use (3) and at least one hospital in the UK is now using the procedure.

ECT can be applied clinically in 3 settings: as sole therapy, intra‐operatively and post‐operatively. Its use as sole therapy is not within the range of this abstract. Spugnini et al (3) have used ECT intra‐operatively and followed this with post‐ operative ECT (in the same cases) but this has not been confirmed as beneficial by other clinicians. The author has used single treatments in most patients whether intra‐ or post‐operatively.

Intra‐operative ECT facilitates local control of a tumour whilst reducing the extent of tissue loss. A tumour can be debulked and the margins treated with ECT prior to suturing/reconstruction. Healing is not directly affected by ECT though, if necrosis occurs in the margins (see above) there may be some degree of breakdown. Post‐operative ECT has been used chiefly where surgical margins have been shown to be inadequate on histopathology (although this measure of adequacy is currently being questioned).

Clinical Data

The patients were 32 dogs and 4 cats. The tumours involved were:

Mast Cell Tumours 16

Fibrosarcomas 4 Haemangiopericytomas 3 Soft Tissue Sarcomas 3 Spindle Cell Tumours 4 TOTAL Sarcomas 14

Squamous Cell Carcinomas 4

Undefined Carcinoma 1

Lymphoma 1

Mean follow‐up period, or survival time from treatment, was 296 days. At last report date 33 patients were alive (8 with recurrent tumour), 3 were dead with no tumour recurrence and 1 lost to follow‐up.

Post‐therapy wound healing was delayed in 13 cases and in 3 cases was incomplete at long term follow‐up.

As all of these patients had tumours of the distal limb or face, either amputation or disfiguring surgery would have been the conventional approach.

The complete remission rate in the cases was 28/36 (77.8%).

Side‐effects of therapy were local swelling, lameness if the limb was affected and local necrosis (see above). The swelling was very noticeably worse in cats and caused compromise to circulation in the paws in 2 treated cats receiving intra‐ operative ECT. No side‐effects attributable to bleomycin were seen.

Discussion

As a method of targeting local therapy whilst reducing surgical morbidity, ECT seems to have value. It has advantages over radiotherapy in that it is cheaper, repeatable, more targeted on tumour cells and easier to apply during surgery. It is limited in depth of treatment to the length of the electrodes used and there are theoretical limits on this.

The canine cases in this report show a high remission rate. Pre‐treatment assessment of and access to margins are factors affecting outcome.

The use of ECT in cats needs more consideration. Intra‐operative use in tumours of the paw appears to carry significant morbidity; perhaps such tumours might be better treated post‐operatively, although only one such case occurs in the series.

Client responses to results of ECT have been almost exclusively positive. There is a desire to preserve the anatomical integrity of pets if at all possible.

A widening of use of ECT would allow more rapid accumulation of case numbers for validation of the method.

References:

1) M. Cemazar, Y. Tamzali, G. Sersa, N. Tozon, L.M. Mir, D. Miklavcic, R. Lowe and J. Teissie (2008) Electrochemotherapy in Veterinary Oncology Journal of Veterinary Internal Medicine 22 pp 826–831

2) Michel Marty, Gregor Sersa, Jean Remi Garbay, Julie Gehl, Christopher G. Collins, Marko Snoj, Valerie Billard, Poul F. Geertsen, John O. Larkin, Damijan Miklavcie, Ivan Pavlovic, Snezna M. Paulin‐Kosir, Maja Cemazar, Nassim Morsli, Declan M. Soden, Zvonimir Rudolf, Caroline Robert, Gerald C. O’Sullivan, Lluis M. Mir (2006) Electrochemotherapy – An easy, highly effective and safe treatment of cutaneous and subcutaneous metastases: Results of ESOPE (European Standard Operating Procedures of Electrochemotherapy) study. European Journal of Cancer Supplements 4 pp 3–13

3) Enrico P. Spugnini, Alfonso Baldi, Bruno Vincenzi, Franco Bongiorni, Corrado Bellelli, Gennaro Citro, Alessandro Porrello (2007) Intraoperative versus postoperative electrochemotherapy in high grade soft tissue sarcomas: a preliminary study in a spontaneous feline model Cancer Chemotherapy and Pharmacology 59:3 pp 375‐81

Reconstruction of Major Dorsal Nasal Defect induced by Intranasal Radiation with a Forehead Transposition Flap

J. Butinar1, J.P. de Vos2, L. Van Kuijk2

A 6 year old female mongrel dog was presented with a histopathologically confirmed undifferentiated nasal carcinoma. CT‐images showed extension of the tumour over both nasal cavities with septum destruction, but no involvement of the cribriform plate and frontal sinuses. Staging procedure showed no evidence of metastatic disease.

Intranasal high‐dose‐rate brachytherapy (192Iridium) was performed using an afterloading system (Microselectron HDR; Nucletron) and MRI‐based dose distribution calculation by the planning system PLATO (version 14.2.4.; Nucletron). This hypofractionated twice weekly protocol, consisting of eight 4Gy fractions, resulted in a complete remission.

Subsequently to the maxillary fistula, multiple oral‐nasal fistulae and finally bone lacking complete palatine defect as well as large bone and skin dorsal nasal defect developed within 1 year post‐radiation. After several unsuccessful surgical attempts palatine defect was covered with a custom made removable Vitallium frame prosthesis with an attached methylmetacrylate obturator.

Dorsal nasal defect was reconstructed with a transposition flap from the forehead. The idea for flap geometry was primarily taken from ancient Indian surgeon Sushruta (The Healing Hand, Man an Wound in the Ancient World ‐ Guido Majno) as well as from gulwing flap from Grabb&Smith plastic surgery textbook.

Results: 48 months after initial therapy the tumour is still in complete remission, frame prosthesis with obturator and dorsal nasal reconstruction are fuctional and cosmetically effiecient.

1 Animal Hospital Postojna, Postojna, Slovenia 2 De Ottenhorst, clinic for companion animal medicine Terneuzen, The Netherlands

E‐mail: [email protected]

Mast cell tumours – anything new worth knowing??

Principles of surgical oncology are driven by ‘what is it, where is it, how bad is it and how can we treat it?’; i.e. diagnosis, staging and therapeutic options. This tenet is particularly appropriate in the management of mast cell tumours (MCTs) and it is important to remember that more MCTs are cured by surgery alone than by any other single or multi‐modality plan. This is achieved in the main by planning the appropriate ‘surgical dose’ in order to minimize the need for repeat surgery or post‐operative adjunctive therapies.

Fundamental to completely excising MCTs is an understanding of the principles of histological invasion, tumour reactive zones, and decisive curative‐intent surgery. Mast cell tumours are typically histologically circumscribed but un‐ encapsulated, with a poorly defined surrounding reactive zone, consisting of some or all of; • a vascular response (new blood vessels), • mesenchymal response (to physical presence of the tumour and abnormal local tissue forces) and, • inflammatory response (to necrosis / haemorrhage / degranulation).

This reactive zone may be several millimetres in width in smaller low grade tumours, but over a centimetre in high grade degranulating tumours. Within the reactive zone of MCTs will be normal inflammatory mast cells in addition to neoplastic mast cells, and occasionally local extensions of tumour called ‘satellites’, distinct from the main tumour mass. This surrounding zone of predominantly inflammatory cells admixed with neoplastic cells is successfully excised if a surgical margin of ‘normal’ tissue around the borders of the grossly visible cutaneous (or subcutaneous) mass is taken.

Historically a 3cm skin margin laterally around the visible edge of a cutaneous MCT, and a deep tissue margin beneath one fascial plane is described as adequate to completely resect the tumour and reactive zone. When cutaneous MCTs arise on the body the panniculus muscle (platysma, cutaneous trunci, preputialis, supramammarius) is a suitable deep fascial plane, and the thick fibrous fascia is suitable on the limbs. The origin of 3cm is unknown and it seems overly simplistic to suggest this rule equally applies, for example, to both a biologically inactive 5mm MCT present for two years, and also to a rapidly growing 5cm MCT which appeared a month previously.

The first reference to the 3cm margin in the literature is unclear and the 3cm preconception was challenged by Simpson and others (2004) who studied 23 grade 1 or grade 2 MCTs excised with 3cm margins. They successfully demonstrated that all grade 1 tumours were completely excised at the 1cm skin margin and that all grade 2 tumours were completely excised at the 2cm skin margin, although 2/20 dogs had tumour cells within 1mm of the deep margin (fascial tissue). This was true independent of tumour diameter (range 0.35‐5cm). A subsequent prospective study excising all grade 1 or 2 tumours (diameter range 0.4‐3.1cm) with a 2cm skin/one fascial plane deep margin achieved complete excision in 91% cases, and avoided the much larger tissue defects seen with 3cm margins (Fulcher, 2006).

Henderson has recently proposed an intriguing ’proportional margins’ theory for grade 1 or 2 MCTs less than 2 cm in diameter. In brief the width of the lateral skin margin and depth of the deep margin is equal to the tumour diameter. So a 0.5cm diameter MCT is excised with a 0.5cm skin margin and 0.5cm deep margin, whereas a 2cm MCT is resected with a 2cm skin margin and 2cm margin deep (or fascial plane). It would then follow from Fulcher’s work that a constant 2cm would be used in grade 1 or 2 tumours over 2cm in diameter. Henderson’s pilot work was presented at the Veterinary Cancer Society; results of a larger prospective study of grade 1 and grade 2 tumours excised using the proportional margin theory is currently under peer‐review. It has not been studied whether a 2cm margin is sufficient to obtain clean margins in MCTs over 5cm in diameter (the largest tumour in either Simpson or Fulcher’s papers). It is also not known whether any of these rules described above hold true for grade 3 MCTs.

The phenomenon in dogs of presenting with multiple simultaneous cutaneous MCTs was investigated by Mullins in 2006. 54 dogs with 153 tumours were retrospectively reviewed and it was found that the 1 year and 2‐5 year survival rates were 87% and 85% respectively. Rate of metastasis was 15%. The only negative prognostic factor in multivariate analysis was the presence of clinical signs at time of presentation (cutaneous ulcers, signs of pain, swelling, or gastro‐ intestinal signs such as vomiting, melaena or diarrhoea). This study prompted the authors to conclude that multiple cutaneous MCTs in dogs are associated with a low rate of metastasis and a good prognosis for long‐term survival with adequate excision. This is useful information for the common clinical scenario of finding 2‐10 MCTs on a dog, and finding 5 more whilst clipping. The overall outcome for the patient is favourable, and if the margins on these smaller masses can now be ‘down‐staged’ too, then many of these can be removed with conservative surgery or large skin punches.

Mast cell tumours can be successfully removed with the instrumentation available in most general practices. Cautery is recommended to minimize local bleeding and risk of post‐operative haematoma. If reconstructive procedures are necessary then skin hooks or using stay sutures to mobilise wound edges are advisable in order to avoid traumatizing the leading wound edges. If a large dead space exists after tumour excision, this should be closed by subcutaneous sutures rather than drains as placing the latter in oncological surgeries widens the potentially contaminated field if the excised tissue is reported as

histologically incomplete. Any subsequent adjunctive therapies (e.g. re‐excision, radiation) must then also include the path of the drain.

If multiple masses are to be removed under the same anaesthetic, these can all be draped out separately. A new set of instruments should be used for each mass to prevent tumour cells being transferred from one surgical wound to another. Surgery should also include excising any regional lymph nodes found to be cytologically positive for tumour metastases. Removal of the draining lymph node regardless of physical characteristics might be wise as a staging tool.

Following excision, all cut edges of the tissue (skin margin and deep margin) should be stained with India Ink, left to dry, and then placed into 10% formalin to fix, at a ratio at least 1 part mass: 10 parts formalin. On processing and interpretation, identifying tumour cells in ‘inked’ tissue represents tumour at the wound edge, and so an incomplete excision. Larger tissue masses may need to be ‘bread‐loafed’ prior to fixing, as the formalin cannot diffuse to a depth of greater than 1cm tissue. Alternatively some surgeons prefer to tie small suture knots at locations where concern exists regarding the width of the normal tissue margin, to encourage the pathologist to examine these areas more critically.

Surgical incisions typically heal without complication, although there is an increased risk of wound dehiscence with MCTs due to the local concentration of heparins, histamine and proteases which interfere with the acute inflammatory and proliferative phases of wound healing. Wound breakdown, along with intra‐ operative hypotension and excessive haemorrhage, is most often seen with large, poorly differentiated MCTs. In one study of grade II MCTs, wound dehiscence was seen in 10% after surgery (Séguin, 2001).

The aim of surgery must be clear prior to starting the resection. A wide excision and reconstruction is the simplest and fastest choice for a cure, but if the mass is not amenable to a wide excision then local excision within the reactive zone to remove gross disease may be performed, with curative‐intent adjunctive radiation to follow. Irradiating a flap, or placing a flap into a previously irradiated field, however, does carry a 77% risk of complications due to poorer wound healing and fibrosis (Séguin 2005).

The reported recurrence rate following complete surgical excision of grade 2 tumours is 11% with time to recurrence of 2‐24 months (Weisse, 2002), and the rate following incomplete excision is 23.3% (Séguin, 2006). Although many book chapters, review articles and papers state further surgery, radiation and/or chemotherapy are indicated in the face of incomplete excision, published work shows that most incompletely excised MCTs do NOT recur (18‐35% recurrence rate reported). In Séguin's study, a local recurrence rate of 23.2% (7/28) of incompletely excised cutaneous grade 2 MCTs was identified, with a median follow up of over 800 days. These authors identified that a combination of Ki‐67 and PCNA (proliferating cell nuclear antigen) scores was prognostic for local recurrence.

These studies contradict the assumption (just as we are now beginning to

appreciate with STS) that most incompletely excised MCTs locally recur in the dog’s lifetime, raising the question of value and efficacy of adjuvant therapies such as radiation or chemotherapy, plus raising the concern that we are over‐ treating a number of patients. However, since developing local recurrence is typically prognostic for decreased overall survival, until data becomes available regarding the impact of omitting additional treatment following an incomplete excision, adjuvant therapy remains the standard of care.

Radiotherapy is not always available, and is associated with increased expense and mild to moderate side effects. Alternatives to radiotherapy, such as hypotonic water, have been investigated. The literature is contradictory on the efficacy of hypotonic water in controlling residual mast cell disease. Surrounding mast cells with deionised or distilled water will cause them to lyse due to the salt content within the cell, but this is also true for most cells (in fact, hypotonic water is used as a RBC lysate in laboratories for exactly that reason). This helps explain the discomfort associated with hypotonic water injections. Early work (Grier, 1995) demonstrated a significant reduction in local recurrence following incomplete surgical excision when distilled water was injected into the wound (31.6% recurrence) versus surgery alone (63.6% recurrence). In another study using historical controls, 17 incompletely excised cutaneous MCTs were treated with adjunctive deionised water and only 2 (12%) recurred (Neyens, 2004). Jaffe (2000) however published conflicting work showing dogs treated by incomplete excision and deionised water had a worse prognosis with a shorter recurrence free period than those without adjuvant water therapy. Brocks (2008) in an attempt to settle this controversy published results of a randomised, double‐ blinded, placebo‐controlled study comparing distilled water to lactated Ringers in dogs with incompletely excised MCTs. They found no significant difference in either local recurrence or survival time between the 2 groups, hopefully laying this subject to rest.

Stanclift and Gilson (2008) published work on neo‐adjuvant prednisone prior to surgery to treat MCTs in anatomical locations where achieving wide surgical margins might be problematic. The aim was to use prednisone to consolidate the mass, reduce the volume and so hopefully allow for a smaller surgery to be performed. They studied 49 dogs and had 2 study arms, using doses of either 1mg/kg SID or 2.2mg/kg SID p.o. Dogs received steroids for a median of 10 days (range 3‐60 days) and mean reduction in size was 52‐78% with either the lower or higher dose of steroids respectively. 87% of masses underwent at least a 25% reduction in size. What was interesting in this study was that 89% of resections were described as clean (classified in this study as at least 1mm of normal tissue histologically between tumour and cut surface), and the tumour excision was taken from the ‘new’ boundary of the mass, i.e. the shrunken mass following steroid therapy. This paper prospectively investigated what was often already done by many clinicians, i.e. shrink the MCT with steroids and then excise, but previous discussion centred on whether the pre‐treatment or post‐treatment mass should be used to plan surgery. One criticism of this paper is that the authors state achieving clean margins with this protocol is very achievable (89%), and yet local recurrence rate following clean excision (17.6%) is higher than most would predict, and may be a consequence of generously describing 64

a 1mm margin as clean without acknowledging that satellite clusters may have been missed during tissue processing.

The ‘rules’ governing surgery of subcutaneous MCTs are poorly defined, and the 3cm margin is still widely advocated. Papers similar to the 1 cm and 2 cm margin studies in cutaneous MCTs have not yet been published with respect to subcutaneous or intestinal MCTs and at this time it seems prudent to maintain the status quo. Inevitably the deep margin for subcutaneous tumours is deeper than for cutaneous MCTs and so volumes of excised tissue are larger, and local reconstruction techniques are more commonplace. When the mass is a discrete gastro‐intestinal tumour, intestinal resection‐anastomosis is indicted to alleviate intestinal chronic obstructive symptoms. Incisional or excisional biopsies of the draining mesenteric lymph nodes should be acquired for disease staging. Wide (3‐5cm) margins of normal intestine should be resected with the mass but the overall prognosis for this tumour location remains poor.

Chemotherapy: New Molecular Targets for Diagnosis and Therapy in Mast Cell Tumours

Richard Elders MVB, CertSAM, MRCVS

1) KIT

KIT is a type III transmembrane receptor with a tyrosine kinase domain. It is the receptor for stem cell factor, which on ligation induces KIT dimerisation, phosphorylation and a downstream signalling cascade. This signalling is essential for the proliferation and long term survival of normal mast cells, and for the optimisation of several of their functions. Internal tandem duplication mutations in the KIT gene clustered around exon 11 are correlated with increased aggressiveness in canine mast cell tumours. This is consistent with the mutations interfering with the negative regulation of the kinase domain and inducing constitutive phosphorylation of KIT in the absence of stem cell factor promoting survival and proliferation. Up to half of mast cell tumours demonstrate a KIT duplication mutation, although the frequency varies, depending on the population surveyed (first opinion clinic vs. referral hospital patients). Despite the characterisation of many different mutations since 1999, few investigators have additionally characterised the mutations' effects on KIT phosphorylation status, however, all of those tested to date have induced constitutive phosphorylation. A few mutations have proven to be wholly within intron 11, which suggests they are unlikely to have any biological activity. Most studies of KIT mutation status have been based on genomic DNA extracted from tumour tissue (rather than mRNA, which would indicate whether the mutated gene was being transcribed). Despite this, in several survey populations, detection of an abnormally large KIT product by PCR using genomic DNA as template is correlated with a worse histopathological grade, more frequent local recurrence, more frequent tumour development at distant sites, shorter disease‐ free interval and shorter overall survival time. A recent paper reported a trend towards better outcomes for patients bearing a tumour with a KIT duplication mutation receiving adjunctive chemotherapy, compared to those receiving surgery alone.

Although not all mast cell tumours have mutated KIT, normal canine cells have not been shown to have such mutations. Detection of a KIT mutation can be considered confirmatory of the neoplastic transformation of mast cells. Demonstration of a KIT mutation in material from a draining lymph node biopsy/aspirate could be used as evidence of metastasis (although not finding such a mutation would not rule out metastasis). This could be useful in cases in which draining lymph node cytology results are ambiguous, and may assist in surgical planning, as may pre‐operative determination of the KIT mutation status of the primary mass. Several other proteins expressed by mast cells, such as tryptase and chymase, could also be used as molecular markers of mast cells'

presence in biopsy material, but they are not neoplastic mast cell‐specific, offering perhaps the same information as immunohisto/cytochemistry although with very high sensitivity.

KIT staining patterns have also been a major focus of research. Several authors have correlated cytoplasmic (diffuse/stippled) or perinuclear KIT protein localisation with mutated KIT, more frequent local recurrence, poorer differentiation, higher proliferation indices and shorter disease‐free interval and overall survival. However, there is some evidence that KIT protein is not always membrane‐associated in normal mast cells and some apparently well‐ differentiated tumours have also demonstrated intracellular KIT protein localisation. Revealing the KIT protein localisation pattern in mast cells within draining lymph node/visceral organ aspirates during staging by immunocytochemistry might also be useful in considering whether any mast cells found are normal or neoplastic, but compared to KIT mutational status this finding seems likely to be less robust.

Following several years of clinical trials and case reports/series on the use of drugs marketed for humans, several veterinary receptor tyrosine kinase inhibitors have been launched to combat mast cell diseases. Most of these drugs tend to compete with ATP for its binding site within the kinase domain of KIT and other related receptor tyrosine kinases. In previous surveys, not every tumour with a KIT duplication mutation responded, and proportionately fewer tumours lacking mutations tended to respond, possibly consistent with the lack of absolute specificity of these drugs in targeting mutated KIT. Acquired resistance to receptor tyrosine kinase inhibitors has been documented in human oncology, mainly through mutations in KIT, which interfere with the drugs' binding sites, although these drugs can also be substrates for efflux pumps. Some veterinary patients have been treated for prolonged periods without severe side effects, although at high doses, fatal complications have been reported.

2) Mast Cell Proliferation Markers

Many authors consider mast cells to be long‐lived, tissue‐dwelling cells, without the ability to proliferate. Mast cell accumulation in tissues is thought to result from prolonged survival of existing resident mast cells and a higher rate of migration of immature mast cells from the bone marrow via the bloodstream, rather than local proliferation in the tissues. Various proliferation markers have been used to predict aggressive biologic behaviour in canine mast cell tumours, each with their advantages and disadvantages, such as Ki67, AgNOR, PCNA and mitotic index. In several studies, the most useful marker seems to be Ki67 immunohistochemistry, although the markers offer complementary information. However, methods and prognostically significant cut‐off values vary between laboratories and papers. Some authors have advocated the use of the mitotic index as it is easily standardisable and can be performed on routinely stained histopathological sections, without necessitating additional processing.

3) TRAIL

Tumour Necrosis Factor alpha‐Related Apoptosis Inducing Ligand (TRAIL, CD253, Apo2‐L, TNFSF10) is a recently discovered member of the TNF superfamily of cytokines. Early experiments using this cytokine showed it preferentially killed neoplastic cells, sparing untransformed cells. TRAIL's receptor repertoire is unusually complex in the human. There are 2 death‐ inducing receptors (TNFRSF10A & TNFRSF10B), which ligate TRAIL extracellularly to trigger an intracellular pro‐apoptotic signalling cascade. There are also 2 membrane‐bound decoy receptors (TNFRSF10C & TNFRSF10D) which are smaller proteins and do not signal for apoptosis. Finally, there is a soluble decoy receptor (TNFRSF11B), which binds TRAIL off the cell membrane. Tumour cells tend to upregulate their expression of death inducing receptors while downregulating their decoy receptors. Initially, this differential receptor expression profile provided a convenient explanation for the greater sensitivity of neoplastic cells to the apoptosis inducing effects of TRAIL. However, many tumours have demonstrated both inherent and acquired resistance to TRAIL‐ based therapies. It seems that while the membrane expression of death‐inducing receptors is necessary for TRAIL to signal for apoptosis, this signal can be modulated and opposed by numerous intracellular proteins. Furthermore, TNFRSF10D and the death inducing receptors can activate NF?B, promoting cell survival; under experimental conditions TRAIL‐based therapy has been shown to enhance the growth and invasiveness of occasional resistant tumours. In general, however, TRAIL is thought to contribute to anti‐tumour immunosurveillance including metastasis suppression, with its elimination facilitating tumour initiation and metastasis in mice.

In contrast to early generalisations regarding TRAIL's effects, accumulating evidence supports specific effects of TRAIL in specific cellular systems. TRAIL expression is upregulated on several cells during allergic responses in humans and is involved in activation‐induced mast cell death, presumably to prevent uncontrolled mast cell accumulations. Additionally, mast cells derived from artificially stimulated human umbilical cord blood and human neoplastic mast cell lines demonstrate susceptibility to TRAIL‐induced apoptosis. The canine mastocytoma cell line C2, established from an aggressive mast cell tumour, undergoes apoptosis following exposure to recombinant human TRAIL. A more recent paper published by a pharmaceutical company demonstrated that several canine cancer cell lines were sensitive to the apoptosis‐inducing effects of recombinant canine TRAIL. These preliminary results suggest that species‐ specific TRAIL‐based therapies might become available in the near future for veterinary cancer patients. When combined with TRAIL‐based therapies, several other therapies demonstrate synergistic anti‐tumour effects, including receptor tyrosine kinase inhibitors, proteosomal inhibitors, COX‐2 inhibitors, several chemotherapeutics and radiation. As many soluble cytokines, including TRAIL, have short plasma half‐lives, agonistic anti‐death inducing receptor monoclonal antibodies have also been developed for use in humans. Several Phase I and II clinical trials of TRAIL‐based therapies are underway and some results are beginning to be reported. Although the primary effect of TRAIL‐based

therapies is apoptosis, specific anti‐tumour T‐cell responses have also been noted, resulting in enhanced long term and systemic anti‐tumour immunity.

References:

Ashkenazi A & Herbst RS (2008) To kill a tumour cell: the potential of pro‐apoptotic receptor agonists Journal of Clinical Investigation 118:6 pp 1979‐1990

Finnberg N & El‐Deiry WS (2008) TRAIL death receptors as tumour suppressors and drug targets Cell Cycle 7:11 pp 1525‐1528

London, CA (2007) The role of small molecule inhibitors for veterinary patient Veterinary Clinics of North America (Small Animal Practice) 37:6 pp 1121‐1136

London CA & Sequin B (2003) Mast cell tumours in the dog Veterinary Clinics of North America (Small Animal Practice) 33:3 pp 473‐489

Withrow & MacEwen’s Small Animal Oncology (4th Edition) WB Saunder, 2006

Use of Intra‐Pleural Chemotherapy for Management of Malignant Pleural Effusion

Ana Lara DVM MSc PhD DACVIM (Oncology) MRCVS Royal Veterinary College Lecturer in Veterinary Oncology

Malignant pleural or pericardial effusion (MPE) in the dog carries a poor prognosis. Pleural effusions in patients with cancer may be due to direct invasion of the pleural space by a neoplastic cells, or secondary compression of intrathoracic vessels by a space occupying mass which causes elevation of pulmonary or parietal pleural venous or lymphatic pressure. Malignant effusion can be debilitating and may lead directly to the patient’s death from respiratory failure; or it can lead to such a serious decline in patient’s clinical status that the owner seeks euthanasia. The survival time varies depending on the successful control of the clinical signs with periodic thoracocentesis or pericardiocentesis. For management of malignant pericardial effusion, pericardectomy or a pericardial window is recommended allowing subsequent fluid drainage by thoracocentesis.

Control of malignant effusion can be achieved by chemical pleurodesis and/or by direct cytotoxic activity against neoplastic cells. Several highly irritating substances like tetracycline, quinacrine or talc have been used to cause chemical pleurodesis in dogs. Treatment with these drugs can cause significant pain in the patient. Cytotoxic drugs help to control effusion because they either cause irritation and adhesion (pleurodesis) and they have direct cytotoxic activity. Drugs used in human medicine for this purpose include cisplatin, doxorubicin, carboplatin, mitoxantrone, 5‐fluorouracil and bleomycin. Any strategy directed at pleurodesis requires that the visceral and parietal pleura be effectively brought into contact after instillation of the irritant. Complete drainage of the pleural cavity before treatment is essential to ensure this contact between both parietal and visceral pleura.

There are few studies about the use and success of intracavitary chemotherapy (IC) including intrapleural administration in dogs with MPE. The use of IC for the treatment of MPE is based on the potential for increased exposure of the tumour to antineoplastic agents leading to increased cytotoxicity. Pharmacological studies have shown that when treated with IC the tumour is exposed to capillary blood supply to a concentration equivalent to that achieved by IV administration, and the surface cell layers to a concentration 1‐3 logs higher. However the tissue penetration of intracavitary chemotherapeutic agents is usually limited to 1‐ 3mm. Thus, IC may not be efficacious in bulky disease.

Cisplatin Moore et al. in 1991 reported the use of IC with cisplatin for treatment of thoracic or abdominal malignancies associated with effusion in 6 dogs. Cisplatin was administered every four weeks at a dosage of 50mg/m2 diluted in a

volume of 250ml/m2. Due to the nephrotoxicity associated with this drug, diuresis was performed for 14 hours prior and 6 hours post administration and antiemetic therapy with butorphanol. Haematological toxicity was not reported and vomiting was observed in five dogs after IC treatment. Four dogs with MPE (3 mesotheliomas and 1 dog with carcinomatosis) had stable disease and resolution of effusion for periods ranging from 129 to greater than 306 days and survival times from 129 to 325 days. A few additional case reports of pericardial mesotheliomas undergoing pericardiectomy and IC with cisplatin have been published with variable results.

Carboplatin and Mitoxantrone Charney et al. in 2005 reported the outcome of dogs with carcinomatosis, sarcomatosis and mesothelioma with or without malignant effusion using IC consisting of carboplatin and/or mitoxantrone. Mitoxantrone was used at a dosage of 5‐5.5mg/m2 diluted in 0.9% saline 1:1 and then dosed at 1 ml/5 kg. Carboplatin was used at a dosage of 300mg/m2 diluted in D5W to 10mg/ml and then dosed at 1 ml/5 kg. IC was administered every three weeks. Grade II gastrointestinal toxicity was observed in four dogs and grade I haematological toxicity in 3 dogs. Four dogs with MPE included in the study had resolution of the effusion but duration of resolution was not reported. A median overall survival time of 332 days was reported for these dogs, with time from first IC treatment to death or last evaluation (for dogs lost to follow up or alive at the end of the study) of 18, 43, 117 and 299 days.

5Fluorouracil (5FU) 5‐FU has been used for treatment of MPE in dogs, but there are no well‐ documented publications of this therapy. At Ohio State University (Columbus, OH, USA), 5‐FU is administered at a dosage of 150mg/m2 diluted in saline to a total volume of 60ml. Treatments are performed weekly for four treatments, then if the effusion decreases in volume measured per week, the 5FU administrations are performed biweekly, and thereafter every third week or monthly. The clinical decision of increasing time between treatments is based on response to therapy. Dogs treated with this protocol did not show haematologic or gastrointestinal toxicity. 5FU is inexpensive in the USA and UK.

Bleomycin It has been used anecdotally in dogs with MPE at weekly dosages of 10IU/m2. This drug has very little myelosuppressive effect. Intravenous administration can cause pulmonary fibrosis and hypersensitivity reactions with cumulative doses. I have used it in one dog with MPE that failed to all the IC described above; a decrease in the effusion drained per week was achieved, with no adverse effects observed.

Subcutaneous Access Ports for the Management of Malignant Pleural Effusions

The management of malignant effusions requires chronic thoracic drainage. Several techniques have been reported for dogs and cats. Percutaneous thoracocentesis is the method most commonly used, but drainage of persistent effusion requires repeated penetration of the thoracic cavity and potential complications associated to this procedure include pneumothorax, infection, and patient discomfort. Placement of thoracostomy tubes provides continuous or intermittent pleural drainage over an extended period; however ascending infections, tube failure, and tube removal or damage by the patient can limit the duration of their use. Pleuroperitoneal or pleurovenous shunting,which has been used to transfer pleural fluid from the pleural space to the peritoneal cavity, is contraindicated for neoplastic effusions, and reports of clinical effectiveness in veterinary patients are equivocal.

Vascular access ports were initially designed for repeated blood sampling and intravenous administration of chemotherapeutics in humans. Over the last few years, vascular access ports with fenestrated catheters have been successfully used for management malignant effusion in humans. These devices have two parts: a fenestrated silicone catheter; and a subcutaneously placed stainless steel or titanium portal chamber. A cath‐shield connector firmly connects the catheter to the chamber, which is accessed through the skin by use of a specially designed Huber‐point needle, allowing repeated puncture of the port without compromise of the integrity of the rubber diaphragm. These devices are used for thoracic drainage and administration of intracavitary chemotherapy, decreasing the prevalence of complications derived from repeated percutaneous thoracocentesis or chest tube. They can remain in the patient for extended periods of time, can be used in the home setting or by paramedic personnel, and are cost‐effective.

In veterinary medicine there is one study (Cahalane et al. 2007) describing the use of vascular access ports attached to Jackson‐Pratt drains for chronic thoracic drainage in three dogs, two of them with malignant effusion due to lymphoma and lymphangiosarcoma. A specific pleural access port (PAP) with a fenestrated silicone catheter for veterinary use, called PleuralPort® (Norfolk Veterinary products), has been commercially available since 2005 in two different sizes, one for dogs and a smaller one for cats.

Between 2005 and 2007, between Ohio State University, Florida Veterinary specialists and VCA Albuquerque, we managed nine dogs with malignant effusions by placement of PAPs, which were used for chronic thoracic drainage and intrapleural chemotherapy administration. Seven dogs had pleural carcinomatosis and two dogs mesothelioma. Six dogs had severe pleural effusion on presentation; one dog had a mild effusion and a mediastinal mass; two dogs had cardiac tamponade due to pericardial effusion. Thoracotomy and pericardiectomy were performed in both dogs with pericardial effusion.

PAP’s were surgically placed under general anesthesia on the hemithorax in which the effusion was considered more productive. Dogs were placed in

lateral recumbency, with the thoracic wall aseptically prepared. A 5 cm vertical skin incision was created at the level of the dorsal third of the 7th or 9th rib. A 2‐3 cm subcutaneous pocket just superficial to the muscle fascia was bluntly made cranial to the skin incision to house the port and help offset the skin closure and port site. Underlying musculature was incised to expose the intercostal space and a purse‐string suture was placed in the underlying intercostal muscle. The catheter was bluntly introduced through the purse‐string, and fed into the thoracic cavity in a dorsoventral direction. The purse‐string suture was tied firmly to create a watertight seal. The port was punctured with a Huber needle, and saline was infused and air aspirated to ensure patency of the catheter. Thoracic wall muscles were re‐apposed with absorbable sutures and the port was fixed to the muscle fascia cranial to the skin incision with monofilament non‐ absorbable sutures. Thoracic radiographs were taken after surgery in all patients to assess catheter placement. No complications were associated with port placement; only one dog developed seroma around the incision site.

All dogs received IC, eight with 5‐FU and one with carboplatin. Three dogs also received intravenous chemotherapy alternated with IC. There were no reported gastrointestinal adverse effects or haematologic toxicity related to the IC administration. Thoracic fluid drainage through PAP started the day of placement with initial volumes prior to IC ranging from 435 to 1700 ml (54‐78 ml/kg; mean 1245ml, 50 ml/kg). MAP did not resolve completely in any dog but drained volume decreased to 10‐625 ml (0.4‐27 ml/kg; mean 244ml, 10ml/kg). The interval between port placement and death or last evaluation ranged from 35 to 470 days (median 84 days). Two dogs were euthanised due to lethargy and inappetence, despite controlled MPE; four dogs had progressive disease; two dogs died of unrelated causes and one dog was alive at the end of the study.

Complications Observed after LongTerm use of Pleural Access Ports

• Rupture of silicone catheter in one dog 40 days after placement. Removal of the port and replacement was performed, remaining viable until the death; • Occlusion of PAP 44 days after placement in one dog. At the same time recurrence of thymic carcinoma was diagnosed, which could have been the cause of occlusion. Owners declined further work up at that point; • Inability to provide complete thoracic drainage due to imperforate mediastinum in one dog. A second PAP was placed in opposite hemithorax with thoracic drainage and IC continued using both PAP´s.

In conclusion, the prognosis of dogs with MPE can be improved if chronic thoracic drainage and IC are performed. There are several IC options that can be successful in controlling MPE for limited periods of time and PAPs can be useful tools in the long‐term management of dogs with MPE allowing immediate and frequent therapeutic thoracic drainage without the discomfort of percutaneous thoracocentesis and IC administration in a given schedule with or without existence of pleural effusion. PAP´s are associated with minimal complications and may improve the patient’s quality of life.

References and Recommended Reading:

1. Balli A, Lachat M, Gerber B, Baumgartner C, Glaus T. Cardiac tamponade due to pericardial mesothelioma in an 11‐year‐old dog: diagnosis, medical and interventional treatments. Schweiz Arch Tierheilkd. 2003 Feb;145(2):82‐7. 2. Calahane AK, Flanders JA, Steffey MA, Rassnic KM. Use of vascular access ports with intrathoracic drains for treatment of pleural effusion in three dogs. J Am Vet Med Assoc 2007; 230:527‐531. 3. Charney SC, Bergman PJ, McKnight JA, Farelly J, Novosad CA, Leibman NF, Camps‐Palau MA. Evaluation of intracavitary mitoxantrone and carboplatin for treatment of carcinomatosis, sarcomatosis and mesothelioma, with or without malignant effusions : A retrospective analysis of 12 cases (1997‐2002). Vet Comp Oncol 2005; 3 :171‐181.. 4. Closa JM, Font A, Mascort J. Pericardial mesothelioma in a dog: long‐term survival after pericardiectomy in combination with chemotherapy. J Small Anim Pract. 1999 Aug;40(8):383‐6. 5. Laing EJ, Norris AM. Pleurodesis as a treatment for pleural effusion in the dog. J Am Anim Hosp Assoc 1986; 22:193‐196. 6. Lara A, Kosarek C, Hosoya K, Alvarez F, Smeak D, Pozzi A, Kisseberth W, Ries P, Couto CG. Palliative management of malignant effusion with implantable vascular access ports in dogs. Proceedings Annual Meeting of the European Society of Veterinary Oncology, Cambridge, England, March 2007. 7. Madewell BR and Theilen GH. Tumors of the respiratory tract and thorax. In: Veterinary Cancer Medicine, GH Theilen and BR MAdewell eds., Philadelphia, Lea & Febiger, 1987:553‐554. 8. Moore AS, Kirk C, Cardona A. Intracavitary cisplatin chemotherapy experience with six dogs. J Vet Int Med 1991; 5:227‐231. 9. Morrison WB. Non pulmonary intrathoracic cancer. In: Morrison WB ed. Cancer in dogs and cats. Medical and surgical management. 2nd ed. Jackson: Teton New Media, 2002: 513‐527. 10. Reed DN, Vyskocil JJ, Rao V. Subcutaneous access ports with fenestrated catheters for improved management of recurrent pleural effusions. Am J Surg 1999; 177:145‐146. 11. SB. Regional chemotherapy. In: Holland JF, Frei E, Bast RC, Kufe DW, Morton DL, Weischelbaum RR, eds. Cancer Medicine. 3rd ed. Philadelphia: Lea &FEbiger, 1993:640‐652. 12. Shapiro W, Turrel J. Management of pleural effusion secondary to metastatic adenocarcinoma in a dog. J Am Vet Med Assoc 1988: 192:530‐532. 13. Shoji T, Tanaka F, Yanagihara K, Inui K, Wada H. Phase II study of repeated intrapleural chemotherapy using implantable access system for management of malignant pleural effusion. Chest 2002; 121:821‐824. 14. Verfaillie G, Van Herrweghe R, Noppen M, Sacre R, Lamotte J. Us of a Port‐a‐Cath system in the home setting for the treatment of symptomatic recurrent malignant pleural effusion. Eur J Cancer Care 2005; 14:182‐184.

Urinary Bladder and Urethral Tumours

Currently, bladder and urethral tumours remain very frustrating diseases to treat, typically presenting with stranguria or urinary obstruction due to advanced tumor stage (80% stage T2 and 20% stage T3), late in the course of the disease (Knapp et al. Urol Oncol, 2001). In the bladder, surgery is the only treatment that offers a potential cure, provided the neoplasm can be completely excised and metastasis has not occurred. In dogs trigonal disease is more common; in cats the distribution between cranial and caudal is more even.

A large volume of bladder (commonly reported to be 70%) can be resected with no lasting effect on urinary storage and voiding. It is common, however, for the animal to have transient urinary incontinence following large partial cystectomies. This should not need treating and tends to improve over a few weeks. Appositional repair is appropriate for closing bladder wounds. A single or double layer using a monofilament absorbable material is indicated. In select cases of extensive disease, or peri‐ureteral and obstructive trigonal disease, submucosal resection of superficial tumour can be performed to reduce tumour load. An incision is made in the mucosa and then this is undermined superficial to the submucosa allowing the mucosa to be peeled away. As far as is possible the remaining mucosa should be apposed over the exposed submucosa but if difficult, the tissue can be left to heal by second intention granulation, epithelialisation and contraction. Urothelium will cover the defect within 7‐10 days. This palliative measure may help with clinical signs and increases the probability of chemotherapy being effective by reducing tumour target volume.

When the tumour is surrounding one or both ureteric openings and the rest of the bladder mucosa appears unaffected, then vesico‐ureteric resection and neoureterostomy can be performed. The terminal ureter(s) is/are resected with at least 1‐2cm of palpably normal margin, and re‐implanted in a grossly normal area, usually the bladder apex. Transient hydronephrosis lasting 2‐3 weeks is common. The dogs are monitoring by ultrasound initially every week for signs of hydronephrosis and tumour recurrence, and after 4 rechecks, this can become every 3 months.

A combination of total cystectomy, partial urethrectomy, and ureter re‐ implantation into the vaginal remnant is possible for advanced diffuse cases of bladder and urethral neoplasia. It is technically very demanding to achieve clean tumour free margins and ensure ureteric patency; furthermore urinary incontinence is an unavoidable complication. 75

A technique for resection of the trigone and cranial urethra has recently been described in two dogs (Saulnier‐Troff et al, 2008), one with TCC, one with rhabdomyosarcoma. The technique involved elevating the serosa off the bladder neck and with it preserving the vascular supply of the caudal vesical artery and the nervous plexus to the bladder and urethral remnant, and then circumferentially resecting en bloc the bladder neck and proximal urethra. This is clearly a technically advanced procedure and cannot be recommended as a treatment option until more than 2 cases are described unless no other options exist to the owner. It carries high risks of local recurrence, bladder necrosis and urinary incontinence.

Surgical options for urethral neoplasia depend entirely on case selection and presentation. Diffuse disease extending from the urethral papilla through to the trigone are not candidates for surgical excision. Options for localized small volume disease include:

Caudal: Vaginourethroplasty: Following a pubic symphysiotomy, the caudal urethra is resected and re‐ anastomosed (end‐to‐side) to the vagina.

Urethral pullthrough: Similar in concept to a rectal pull‐through. Following an episiotomy, the vaginal wall is incised circumferentially around the papilla and the terminal urethra tractioned caudally, and repaired mucosa to mucosa.

Urethral resection; Prepubic Urethrostomy.

Central: Partial urethrectomy, endtoend anastomosis. Via prepubic urethrostomy.

Urethral Repair: Autogenous sublingual mucosal free graft recently described in experimental dogs as pre‐clinical study for urethral defects in humans. Free mucosal graft inverted and sutured into a tube, and placed across the urethral defect. Via prepubic urethrostomy.

Diffuse: TransUrethral Resection (TUR): Adapted from the treatment for BPH in men. A TUR device is inserted into the urethra and diffuse proliferative neoplastic tissue is shaved off the walls. Risks include full thickness urethral tears/perforations. Only been described in one small case series.

Tube Cystostomy: Foley or low‐profile catheter inserted at laparotomy for bladder

decompression. Is a treatment option for any level of urethral obstruction. Fast and easy to insert, suitable for both cats and dogs, and remarkably well tolerated.

Radiation

Reports dating back to 1987 show that XRT has been used successfully to control canine bladder TCC and prostatic carcinoma (Withrow SJ et al., Vet Surg 18, 1989; Anderson CR, et al., Vet Radiol Ultrasound, 2002.). However, irradiation of lower urinary tract tumours poses several unique challenges. These include the difficulty in tumour localisation, and proximity of the tumour to dose‐limiting structures such as the ureters and colon. Complications of therapy include cystitis and fibrosis as well as irritation to surrounding organs (McCaw DL, et al., Vet Radiol Ultrasound, 1988; Poirier VJ, et al, JAAHA, 2004). These sequelae can significantly detract from the patients’ quality of life and successful use of XRT for the treatment of canine carcinoma requires careful delivery of the radiation dose. Factors to consider when designing radiation protocols include the volume of bladder/urethra/prostate to be irradiated, the total dose to be delivered, the dose per fraction to be used and the dose to surrounding tissues.

To date, two different methods of delivering the prescribed radiation dose have been described: intraoperative radiotherapy (IORT), in which the bladder and/or prostate is exteriorized and single large doses (10‐40 Gy) are given (Withrow SJ et al., Vet Surg 18, 1989; Anderson CR, et al., Vet Radiol Ultrasound, 2002.; DL, et al., Vet Radiol Ultrasound, 1988; Kinsella TJ, et al. Int J Radiat Oncol Biol Phys, 1988); or fractionated treatments where multiple smaller doses (3‐ 5.75 Gy) in 6 to 12 fractions are given to total doses of 30‐48 Gy. These studies have shown tumour responses, but treatment‐related toxicity including colon perforation and stenosis of the ureters has been reported with IORT. Other major deterents to the use of radiation in treating lower urinary tract carcinomas are the need to perform surgery in the radiation facility in order to exteriorize the bladder or prostate as in IORT, and the need for laborious contrast studies and urinary catheterisation to ensure uniform bladder size as well as lesion localisation with fractionated treatments.

New applications of stereotactic radiosurgery (SRS) currently under investigation at the University of Florida include the treatment of lower urinary tract tumours, especially transitional cell carcinoma (TCC) of the urethra and/or bladder, and TCC and adenocarcinoma of the prostate. CT in conjunction with SRS allows precise lesion localisation followed by a highly conformal dose delivery to the tumour with minimal exposure to surrounding tissues. In a modification of the technique previously described for appendicular osteosarcoma (Farese et al, JAVMA), a modified bite‐plate fiducial array is temporarily attached to the pelvis, allowing precise anatomical referencing of the pelvic canal. As an added advantage, the CT, fiducial array attachment and SRS are all performed under a single anesthetic procedure. Our group is currently targeting urethral tumours and, although still early in the course of treatment, 6 female dogs with urethral TCC have been treated, and results have been

promising with regard to local control. Obstruction was relieved in all dogs. 2 dogs have died of metastatic disease and the other 4 are currently alive. Urethral stricturing from the radiation or disease progression outside the radiation field has been seen in 2 dogs and these have required urethral stenting. One dog which presented with urinary obstruction underwent urethral stenting prior to SRS.

It is important to note that SRS, like any radiation, is a local treatment, and so patients will require systemic chemotherapy, for example mitoxantrone q3 weeks and a daily COX‐2 inhibitor (piroxicam, meloxicam). The goal of SRS, fundamentally, is to offer local disease control in the urethra and move the emphasis of survival onto the chemotherapy.

Urethral Stenting

Palliative urethral stenting offers relief to dogs with urinary obstruction due to neoplastic infiltrative disease. It does not address local disease as such, only the symptoms, and the gold‐standard therapies for urethral neoplasia should still be discussed, including systemic chemotherapy. Prior to stenting, most patients undergo urethroscopy and cystoscopy to ensure disease is predominantly confined to the urethra. Cases with a large bladder/trigonal component and cases that extend out of the urethral papilla present additional problems. In the former there is an increased risk of incontinence post‐stenting; in the latter, an increased risk exists of persistent stranguria and discomfort from the stent pressing into the submucosa of the vaginal vault.

Stent sizing is the first step after endoscopy. This typically involves inserting a Foley catheter under anaesthesia into the distal urethra, increasing pressure in the bladder by applying caudal abdominal compression, and performing a fluoroscopic retrograde urethrogram to maximally distend the urethra. This not only identifies the length of the narrowing and its position within the urethra, but also helps identify the maximal achievable width under pressure. The stent chosen is typically 10% larger than the maximal distension, and then rounded up to an available diameter.

The stents used commonly are self‐expanding and made of nitinol (VetStent – Urethra, Infiniti Medical) and are laser cut from a single piece of material. Deployment requires fluoroscopic imaging to ensure accurate positioning, as in contrast to the tracheal stents, which can be re‐constrained mid deployment if the position is sub‐optimal and the stent re‐positioned; once deployment has started with a urethral stent it cannot be reversed.

Available stents are:

Expanded stent Expanded stent length Stent introducer diameter (mm) (mm) delivery system 6 40 6F – 75cm 6 60 6F – 75cm 8 40 6F – 75cm 8 60 6F – 75cm 10 40 6F – 75cm 10 60 6F – 75cm 10 80 6F – 75cm 12 60 7F – 85cm 12 80 7F – 85cm

Ideally deployment should immediately follow sizing, but if ordering and delivery of a stent is required then an indwelling Foley can be placed and the patient recovered. The principles of stenting are easily learnt; but just as with tracheal stenting, the consequences of misplacing a stent are disastrous. With long urethral constrictions, especially in males along the prostatic urethra, 2 overlapping stents might be required.

Occasionally phenoxybenzamine might be required to reduce urethral spasm and the discomfort associated with a stent, or phenylpropanolamine if the stent over‐dilates the bladder neck and causes a low‐grade urinary incontinence.

Stenting only offers temporary relief. Ultimately the TCC will continue to grow, will infiltrate between the holes in the stent filling the lumen once more, and result in urinary obstruction.

Interventional Oncology

Recently intra‐arterial chemotherapy has been described for management of bladder and prostate tumours in dogs (Weisse et al, 2008). The advantage of intra‐arterial delivery of chemotherapy is that by carefully selecting a major contributing artery to the tumour, the drug is delivered at a high dose to the target tissue, without the dilutional effect of intravenous administration, which is followed by cardiac, then pulmonary bed, then cardiac, then arterial dispersion before reaching the tumour.

In laboratory rabbits with bladder tumours, concentrations of carboplatin chemotherapy achieved in bladder tissue, when administered by the arterial route (internal iliac), was significantly higher than when given by intravenous administration. All tumours shrank in size and 37.5% disappeared; this compared to all tumours growing with chemotherapy administered by the intravenous route. Further work in dogs using the drug pirarubicin demonstrated that intra‐arterial drug achieved a concentration 8 times that of intravenous delivery in the bladder mucosa and muscle layers.

For treatment of spontaneous canine bladder tumours that have failed conventional local and/or systemic therapies, a pneumocystogram is first performed under anaesthesia to help delineate the bladder outline more clearly and increase the local contrast of the caudal abdomen. Then a cut‐down to either the carotid or femoral artery is made, and micro‐catheters advanced to the level of the branching of the aorta into the external iliacs. By serial angiograms, the catheter is advanced into as selective a vessel as possible, first the internal iliac, then internal pudendal, and finally prostatic artery if possible. The chemotherapy agent described for use is carboplatin at 250mg/m2, a slight reduction from the conventional intravenous dose of 300mg/m2. This is typically injected in 2 aliquots to target both sides of the mass. A predominantly unilateral lesion could theoretically receive the entire dose on one side. The artery is either repaired or ligated.

A recheck at 10‐14 days for a CBC to rule out myelosuppression is recommended, then at 4 weeks with repeat treatment performed if stable disease or measurable response based on ultrasound or pneumocystogram. Repeating intra‐arterial therapy if disease has progressed is probably pointless.

All data for this procedure is immature. Numbers of dogs treated are small, and through case‐selection, many had failed conventional intravenous chemotherapy, further worsening likely outcome and response. The potential, however, is high; it offers a minimally invasive therapy for unresectable bladder, urethral or prostate cancer.

Lymph Node Staging / Sentinel Lymph Node Studies

Sentinel node studies may hold the key to understanding the pattern and significance of lymph node metastasis in urothelial cancers. More and more cancer cells are being found in lymph nodes via multiple histologic sections with traditional or immunohistochemical staining, and the issue of whether these are cancer cells in transit through lymph nodes or are micrometastasis sites that may progressively develop into a mass in unresolved. One lesson learnt from sentinel node staging in dogs is that relying simply on ultrasound examination is unreliable, and will underestimate incidence of nodal disease. Tumour cells in the bloodstream, when encountering a regional lymph node, can have a variety of interactions. They may: 1. bypass the node through lymphaticovenous or lymphatic shunts; 2. pass through the node in peripheral sinuses; 3. lodge in the node, but not grow; 4. lodge in the node, and be destroyed; 5. lodge in the node, and progressively grow. Studies have provided evidence that detecting lymph node metastasis is most likely an indicator of biologic behaviour (malignant), rather than a governor of behaviour. This translates into shedding cells from lymph nodes seldomly if ever being the cause of new metastasis to distant parts, rather they are simply telling you that this is an aggressive neoplastic condition.

Treatment Options for Prostatic Carcinoma (PCA)

Henry L’Eplattenier Dipl. ECVS, MRCVS RCVS recognised specialist in small animal surgery VRCC Veterinary Referrals, Laindon, Essex Tel. 01268 564664 [email protected]

Canine prostate carcinoma (PCA) is uncommon, with an estimated prevalence of 0.2 to 0.6% (Bell et al. 1991). True prevalence is unknown as population‐based data is not available. Canine PCA has an invasive growth pattern and commonly metastasizes to the sublumbar lymph nodes; occasionally, metastases to the lungs and lumbar vertebrae are observed. Castration has no effect on disease progression, nor does it prevent occurrence of PCA; in fact, it appears that castrated males are at an increased risk of developing PCA compared with intact males (Teske et al. 2002). Clinically, canine PCA therefore resembles late stage, hormone‐independent human PCA and the dog is an appropriate model for understanding the pathogenesis of PCA in humans.

Most dogs diagnosed with PCA are not treated, either because of the presence of metastases, or because of the poor prognosis. Little data is available on life expectancy after diagnosis if no treatment is attempted. One study reports that 58 of 72 dogs were euthanised at the time of diagnosis and that the median survival time for the remaining 12 dogs was 30 days (Cornell et al. 2000). Whether treatment was attempted in these 12 dogs is not mentioned.

This abstract presents treatment options based on a review of the literature and on the author’s own research.

A. Literature Review

Endocrine therapy The principle of endocrine therapy is the elimination of prostate stimulation by androgens. Castration in dogs does not have a sparing affect on the development of PCA (Obradovich et al. 1987) and in fact PCA occurs more often in castrated dogs than intact males (Teske et al. 2002). In intact male dogs with PCA, castration does not have a favourable influence on outcome either. Therefore endocrine therapy plays no role in the treatment of PCA in dogs.

In humans, endocrine therapy is the mainstay for the management of advanced PCA that cannot be cured with radical local therapy, either surgery or radiation therapy.

Surgery

1. Total Prostatectomy Treatment with curative intent is rendered difficult by the fact that radical prostatectomy in dogs with prostatic disease is associated with a very high incidence of postoperative urinary incontinence (Basinger & Rawlings 1987; Basinger et al. 1987; Goldsmid & Bellenger 1991). Interestingly, radical prostatectomy can be performed in normal dogs without causing incontinence (Basinger et al. 1989; Price et al. 1996). It therefore seems that urethral sphincter malfunction is more likely due to primary prostatic disease than to the surgical procedure.

2. Partial Prostatectomy Because of the complications associated with radical prostatectomy, techniques for partial removal of the prostate have been described. These include partial prostatectomy using a Nd:YAG laser (Hardie et al. 1990) and intracapsular subtotal prostatectomy using electrocoagulation (Harari & Dupuis 1995) or an ultrasonic aspirator (Rawlings et al. 1994; Rawlings et al. 1997). However these techniques have been used for the management of non‐neoplastic prostate diseases in dogs (e.g. abcesses and cysts) and neither of them have been used to treat PCA. One prospective study (Vlasin et al. 2006) compared total and partial intra‐capsular prostatectomy in 10 and 11 dogs, respectively. Survival times were significantly better with the subtotal prostatectomy (SP) than with the total prostatectomy (TP), but were short for both groups (mean of 112 and 20 days).

3. Other Surgical Treatments Reports of other surgical treatments for PCA in dogs are limited to a few cases. Surgical placement of a retained urethral catheter in three dogs with PCA and stranguria enabled the dogs to survive 3 to 5 months after surgery (Mann et al. 1992). In a recent study, 3 male dogs with prostatic neoplasia (prostatic transitional cell carcinoma in 2 cases and undifferentiated carcinoma in one case) were treated with transurethral resection using an electrosurgical loop, combined with intraoperative radiation therapy in 2 of those 3 dogs (Liptak et al. 2004). Survival times were 32, 74 and 264 days.

Radiotherapy

There are no reports on radiotherapy alone in dogs. Radiotherapy has been attempted in combination with surgery. In one study 10 dogs underwent intraoperative orthovoltage radiation (Turrel 1987). The mean survival time was 114 days. In another study external beam radiation therapy was combined with administration of ketoconazole. Two dogs treated with this technique were reported to survive for 12 weeks and 4 months, respectively (Bell et al. 1991). Complications of radiation in the caudal abdomen are common and include colitis, thickening of the bladder wall with possible incontinence and stricture of the ureters with subsequent hydronephrosis.

In humans, radiotherapy is most often performed as brachytherapy, a technique involving the placement into the prostate of multiple small seeds of radioactive isotopes. Expectations for cure are between 80% and 90%.

Chemotherapy

No effective chemotherapeutic protocol is available for the treatment of PCA in dogs. In humans, most chemotherapeutic agents have been shown to palliate symptoms, but will not prolong survival of patients. The only exception is docetaxel.

COX2 Inhibitors COX‐2 is expressed in various types of canine cancer including PCA, and is thought to play a role in carcinogenesis. This is supported by evidence that treatment with COX‐2 inhibitors may be beneficial in the management of certain tumours in dogs. In particular piroxicam has been shown to effectively inhibit transitional cell carcinoma (TCC) in dogs (Knapp et al. 1992; Knapp et al. 1994; Henry et al. 2003). Although it has been shown that COX‐2 is expressed in canine PCA, the effect of NSAIDs on the clinical outcome of dogs with PCA has not been extensively investigated. Preliminary findings in dogs with PCA (Sorenmo et al. 2003) and in a mouse prostate carcinoma model suggest that COX‐2 inhibitors may play a significant role in the management of PCA in dogs.

Photodynamic therapy

Photodynamic therapy involves the use of a photosensitizing drug and subsequent delivery of light in the presence of oxygen. The resulting photochemical reaction releases various oxygen radicals that are highly toxic. Treatment of PCA with transurethral photodynamic therapy allowed the only dog reported to survive nearly 9 months after treatment (Lucroy et al. 2003). A number of different photosensitizers have been tested in healthy dogs. These include dihematoporphyrin ester/ether (Photofrin), meso‐tetra‐(m‐ hydroxyphenyl)‐chlorin (mTHPC), aluminium disulfonated phthalocyanine (AlS2Pc), 5‐aminolevulinic acid (ALA), tin ethyl etiopurpurin (SnET2) and Motexafin Lutetium (Lu‐Tex). Photofrin is a so‐called first generation photosensitizer and has been associated with significant skin photosensitization for several weeks after administration. The other photosensitizers tested in the dog are second‐generation photosensitizers with reduced skin photosensitization because of a faster elimination and are therefore more appropriate to use for PDT of prostate carcinoma in dogs. Dogs receiving AlS2Pc and ALA were kept in a dimly lit area and for the other studies the dogs were kept in normally illuminated cages but not exposed to direct sun light. The administration of light to the prostate after intravenous injection of the photosensitizer can be performed in a number of ways: interstitial PDT is the administration of light to the parenchyma of the prostate through several light fibres inserted into the prostate from the perineum. Transurethral PDT is the administration of light through a light cable inserted into the urethra like a

urinary catheter. This technique is technically easy but its efficiency is limited by the depth of light penetration into prostate parenchyma (usually less than 20 mm). The largest area of necrosis was obtained using SnET2 with lesions of up to 26 mm diameter reported after interstitial PDT and 10 mm necrosis around the urethral wall reported after transurethral PDT. With the other photosensitizers, the lesions seen were between 10 and 20 mm in diameter. After transurethral PDT the urethral mucosa is destroyed but regenerates within 3 weeks of treatment.

B. Own research

Partial Prostatectomy using a Nd:YAG laser for Management of Canine Prostate Carcinoma Henry F. L’Eplattenier, Sebastiaan A. van Nimwegen, Frederik J. van Sluijs, and Jolle Kirpensteijn Vet Surg 35:406411, 2006

Methods – Subcapsular partial prostatectomy, sparing the urethra and the dorsal aspect of the prostatic capsule, using Nd‐YAG laser dissection to remove the prostatic parenchyma and control hemorrhage was performed in 4 normal dogs and subsequently in 8 dogs with histologically confirmed PCA. Additional treatment of PCA dogs included local application of interleukin‐2 and systemic administration of meloxicam. Prostate size, complications, and survival time were recorded. Laser associated thermal damage to surrounding tissue was evaluated by histology. Results – In normal dogs, no damage to the dorsal prostatic capsule or urethra was detected. In PCA dogs, median survival was 103 days (range, 5‐239 days). Three dogs died from complications within 16 days, whereas 5 (median survival, 183 days; range, 91‐239 days) had improvement or resolution of clinical signs. Urinary incontinence did not occur. Conclusion – Laser assisted subcapsular partial prostatectomy can be performed in dogs with PCA without development of postoperative incontinence.

Preliminary results of Intraoperative Photodynamic Therapy with 5 Aminolevulinic acid in Dogs with Prostate Carcinoma H.F. L’Eplattenier, B. Klem, E. Teske, F.J. van Sluijs, S. A. van Nimwegen, J. Kirpensteijn In Press: The Veterinary Journal (2007), doi:10.1016/j.tvjl.2007.08.001(available online)

Six client‐owned dogs with prostate carcinoma were treated with a combination of partial subcapsular prostatectomy using an Nd:YAG laser, intraoperative photodynamic therapy using a halogen broad band lamp after local administration of a photosensitiser, and systemic treatment with meloxicam. Median survival time was 41 days (range 10‐68 days), which compared negatively with previous reports of both (1) subtotal laser prostatectomy

combined with topical interleukin‐2 administration and (2) photodynamic therapy alone. Despite treatment, the disease progressed locally, causing signs of stranguria to recur, and in the form of distant metastases. The recurrence of clinical signs due to the primary tumour despite photodynamic therapy is probably largely explained by insufficient penetration of light into the tissue. Better results may be obtained using other light sources (e.g. laser) and alternative techniques of light delivery, such as fibres or catheters allowing interstitial diffusion of light.

References:

Basinger RR and Rawlings CA. (1987). Compendium of Continuing Education, 9, 993‐999. Basinger RR, Rawlings CA, Barsanti JA and Oliver JE. (1989). Journal of the American Animal Hospital Association, 25, 385‐392. Basinger RR, Rawlings CA, Barsanti JA, Oliver JE, Jr. and Crowell WA. (1987). Vet Surg, 16, 405‐10. Basinger RR, Robinette, C.L., Hardie, E.M., Spaulding, K.A. (2003). Textbook of Small Animal Surgery. Slatter D (ed.). Saunders: Philadephia, pp 1542‐1557. Bell FW, Klausner JS, Hayden DW, Feeney DA and Johnston SD. (1991). J Am Vet Med Assoc, 199, 1623‐30. Cooley DM and Waters DJ. (2001). Small Animal Clinical Oncology. Withrow Sj and Macewen Eg (eds). Saunders: Philadelphia, pp 478‐489. Cornell KK, Bostwick DG, Cooley DM, Hall G, Harvey HJ, Hendrick MJ, Pauli BU, Render JA, Stoica G, Sweet DC and Waters DJ. (2000). Prostate, 45, 173‐83. Harari J and Dupuis J. (1995). Semin Vet Med Surg (Small Anim), 10, 43‐7. Hardie EM, Stone EA, Spaulding KA and Cullen JM. (1990). Vet Surg, 19, 348‐55. Henry CJ. (2003). Vet Clin North Am Small Anim Pract, 33(3), 597‐613. Knapp DW, Richardson RC, Bottoms GD, Teclaw R and Chan TC. (1992). Cancer Chemother Pharmacol, 29(3), 214‐8. Knapp DW, Richardson RC, Chan TC, Bottoms GD, Widmer WR, DeNicola DB, Teclaw R, Bonney PL and Kuczek T. (1994). J Vet Inern Med, 8(4), 273‐8. Liptak JM, Brutscher SP, Monnet E, Dernell WS, Twedt DC, Kazmierski KJ, Walter CU, Mullins MN, Larue SM and Withrow SJ. (2004). Vet Surg, 33, 505‐16. Lucroy MD, Bowles MH, Higbee RG, Blaik MA, Ritchey JW and Ridgway TD. (2003). J Vet Intern Med, 17, 235‐7. Mann FA, Barrett RJ and Henderson RA. (1992). Vet Surg, 21, 342‐7. Obradovich J, Walshaw R and Goullaud E. (1987). J Vet Intern Med, 1, 183‐7. Price DT, Chari RS, Neighbors JD, Jr., Eubanks S, Schuessler WW and Preminger GM. (1996). J Laparoendosc Surg, 6, 405‐12. Rawlings CA, Crowell WA, Barsanti JA and Oliver JE, Jr. (1994). Vet Surg, 23, 182‐9. Rawlings CA, Mahaffey MB, Barsanti JA, Quandt JE, Oliver JE, Jr., Crowell WA, Downs MO, Stampley AR and Allen SW. (1997). J Am Vet Med Assoc, 211, 868‐71. Sorenmo KU, Goldschmidt M, Schofer F, Goldkamp C and Ferracone J. (2003). Vet Comp Oncology, 1, 48‐56. Teske E, Naan EC, van Dijk EM, Van Garderen E and Schalken JA. (2002). Mol Cell Endocrinol, 197, 251‐5. Turrel JM (1987). J Am Vet Med Assoc, 190, 48‐52.