J 201Thallium SPECT; Accuracy in Astrocytoma Diagnosis and Treatment Evaluation

SE9900361

J 201Thallium SPECT; accuracy in astrocytoma diagnosis and treatment evaluation

Kristina Kallen, M.D.

LUND UNIVERSITY

3 1-04 201Thallium SPECT; accuracy in astrocytoma diagnosis and treatment evaluation

Akademisk avhandling

Som med vederborligt tillstand av medicinska fakulteten vid Lunds Universitet for avlaggande av medicine doktorsexamen kommer att forsvaras i Segerfalksalen Wallenberg Neurocentrum, centralblocket, Universitetssjukhuset i Lund, lordagen den 16 oktober 1999 Id. 9.00

Kristina Kallen leg. lakare

Institutionen for Klinisk Neurovetenskap Avdelningen for Neurologi Lunds Universitet

Fakultetsopponent: Docent Carsten Wikkelso Institutionen for Klinisk Neurovetenskap Universitetssjukhuset Sahlgrenska GSteborgs Universitet Organization Document name LUND UNIVERSITY DOCTORAL DISSERTATION Dept. of Clinical Neuroscience Dateofissue ^^ ^ ^ University Hospital S-221 B5 LUND, Sweden CODEN: ISRN LUMEDW/MENL_1016_SE Autlior(s) Sponsoring organization KRISHNA KAlLEN

Title and subtitle ^Thallium SPECT, accuracy in astrocytoma diagnosis and treatment evaluation

Abstract The aims of the studies included in this thesis were: • to investigate the reliability of M'thallium single photon emission computed tomography (2O1T1 SPECT) for preoperative diagnosis and histological staging of malignant astrocytomas in comparison with CT; • to develop a method for quantification of cerebral thallium uptake, and to evaluate the quantitative measurement in comparison with CT, for astrocytoma treatment follow-up purposes; • to compare quantitative MITI SPECT and proton magnetic resonance spectroscopy ('H-MRS) with conventional MR imaging for astrocytoma monitoring, and to evaluate associations between change of morphological tumour characteristics during treatment and changes of cerebral thallium uptake and metabolic ratios. Results and conclusions: • High wlTl-index, calculated as a ratio comparing tumour uptake to uptake in the contralateral hemisphere, is an indicator of highly malignant astrocytoma. Differentiation between the high-grade astrocytomas, the low-grade astrocytomas, and infectious lesions is only partial, with an overlap of TI- indexes between these groups. High-grade astrocytomas that do not show contrast enhancement on CT, and astrocytomas with central necrosis and moderate ring-enhancement, tend to be underestimated when evaluated by M1Tl-index calculation. MITI SPECT is not a reliable method for non-invasive tumour staging among the group of highly malignant astrocytomas. • Quantification of cerebral 2olTl-uptake, defining the volume of viable tumour tissue, is a new method for astrocytoma chemotherapy monitoring. Results suggest that the method provides prognostic information, and information of treatment efficacy, at an earlier stage than CT. • We did not find a higher accuracy of quantitative "'Tl SPECT than of MR for h monitoring purposes and our results indicated that treatment induced MR changes were interrelated with UJ ^"Tl-uptake variations. • Multi-voxel 'H-MRS was difficult to apply for astrocytoma treatment monitoring, due to the anatomical and histological heterogeneity of astrocytomas. gg

2u1 cywor s. astroCytoma, abscess, thallium, single-photon emission computed tomography, computed tomography, magnetic resonance imaging, protonmagnetic resonance SDectroscopy, chemntrmrany Classification system and/or index termes (if any):

Supplementary bibliographical information: Language Eno]i sh

ISSN and key title: ISBN 91-628-3684-6

Recipient's notes Number of pages Price Security classification

Distribution by (name and address) Kristina Kalian, Dept. of Neurology, University Hospital, 221 85 LUND, Sweden I. the undersigned, being the copyright owner or the abstract of the above-mentioned dissertation, hereby grant to all reference sources permission to publish and disseminate the abstract of the above-mentioned dissertation.

Signature Date ^ OJ 201Thallium SPECT; accuracy in astrocytoma diagnosis and treatment evaluation

Kristina Kdllen, M.D.

Department of Clinical Neuroscience Division of Neurology Lund University Hospital S-221 85 Sweden

1999

Academic dissertation ISBN 91-628-3684-6 © Kristina Kallen 1999 (pages 1-74, and paper V) Printed by Studentlitteratur Lund, Sweden Contents

List of papers 1 Abbreviations 2 Introduction 3 Clinical background 4 Diagnostic procedures 4 Treatment 5 Low grade astrocytomas 5 High grade astrocytomas 6 Treatment response assessment 11 Clinical parameters 11 Neuro-imaging methods for diagnosis and treatment follow-up 13 CT and MR imaging 13 20lThallium SPECT 15 Historical background 15 Diagnosis and staging 17 Comparison with anatomic imaging 18 Treatment evaluation 19 'H-MR spectroscopy 21 Metabolite characterisation 21 Diagnosis and staging 22 Treatment evaluation 23 Aims of the study 24 Materials and methods 25 Patient selection and follow-up 25 Techniques 27 CT 27 2OIThallium SPECT 27 MR imaging and 'H-MR spectroscopy 28 Data acquisition for therapy monitoring 29 Data analysis 29 Results 35 Pre-operative studies; Papers I and II 35 Quantification of cerebral 2o'thallium-uptake; Paper III 38 Post-operative studies; Papers IV and V 40 General Discussion 43 Diagnosis and staging 43 Treatment monitoring 45 Conclusions 51 Figures in colour 53 Acknowledgements 58 Summary in Swedish 60 References 64 List of papers

I. K. Kallen, A-M. Andersson, M. Heiling, S. Holtas, A. Brun, E. Ryding. Preoperative grading of glioma malignancy with thallium-201 single-photon emission CT: Comparison with conventional CT. AJNR Am JNeuroradiol 1996;17:925-932

II. K. Kallen, A-M. Andersson, M. Heiling, S. Holtas, A. Brun, E. Ryding, I. Rosen. Evaluation of malignancy in ring enhancing brain lesions on CT by thallium-201 SPECT. J Neurol Neurosurg Psychiatry 1997;63:569-574

III. K. Kallen, B. Geijer, A-M. Andersson, S. Holtas, E. Ryding, I. Rosen. Assessment of glioma viability by estimating 20;Tl SPET •tumour uptake volume. Nucl Med Comm 1999; 20:837-844

IV. K. Kallen, B. Geijer, P. Malmstrom, A-M. Andersson, S. Holtas, E. Ryding, I. Rosen. Quantitative Tl-201 SPECT imaging in brain tumour treatment follow-up. A sensitive tool for early identification of chemotherapy response? Accepted for publication in Nucl Med Comm

V. K. Kallen, I.M. Burtscher, S. Holtas, E. Ryding, I. Rosen. 201 Thallium SPECT and 'H-MRS compared to MRI in chemotherapy monitoring of high-grade malignant astrocytomas. Submitted

Papers I-III are reprinted by permission of the copyright owners. Abbreviations

AA anaplastic astrocytoma ATPase adenosine triphosphatase BBB blood-brain barrier Cho choline Cr creatine CT computed tomography 2DCSI two dimensional chemical shift imaging GBM glioblastoma multiforme 'H-MRS proton magnetic resonance spectroscopy Lac lactate Lip lipids MRI magnetic resonance imaging NAA N-acetyl aspartate PCV procarbazine, CCNU, vincristine PET positron emission tomography PRESS point-resolved spectroscopy ROI region of interest SE spin echo ST survival time TI 20lthallium index 201Tl SPECT 2O'thallium single-photon emission comp TTD time to treatment discontinuation TTP time to tumour progression TUV tumour uptake volume Introduction

Primary highly malignant brain tumours in adults comprise of astroglial cell neoplasms or astrocytomas. The astrocytomas represent a challenge to neurologists, oncologist and neurosurgeons responsible for patient management and tumour treatment, as patients often present with disabling neurological symptoms, respond poorly to therapy and undergo rapid clinical deterioration. Highly malignant astrocytomas are the fourth leading cause of cancer mortality in men and women between 15-34 years of age, and the fifth leading cause in men between 35-54 years of age ]'2. The patients are often in their most productive age; in the middle of their career, they are parents of young children and have extensive economic responsibilities. It is a demanding responsibility for all physicians engaged in astrocytoma management to offer these patients the best possible treatment and care. The effort to enhance the accuracy of preoperative diagnosis work up, and the reliability of treatment efficacy monitoring, is part of the challenge to improve the outcome of astrocytoma patients. This thesis addresses the question of diagnostic precision and treatment evaluation by functional brain investigation tools, focusing on 2O'thallium single- photon computed tomography (201Tl SPECT) assessment for astrocytoma diagnosis, tumour staging and treatment monitoring. Additionally, chemical brain analysis by proton magnetic resonance spectroscopy ('H-MRS) is evaluated during treatment follow-up. Clinical background

Diagnostic procedures

Neurological signs suggestive of brain tumour disease are; adult onset of seizures, development of increasing focal deficits, and general symptoms of elevated intracranial pressure such as headache or decline of cognitive functions. Characteristic features for astrocytomas are seen with computed tomography (CT) and magnetic resonance imaging (MRI), and a probable diagnosis can be provided by structural imaging in most cases. An experienced neuroradiologist will correctly diagnose 95% of malignant brain tumours 3. However, a proportion of the patients diagnosed to have a low-grade astrocytoma by CT or MRI, will harbour a highly malignant tumour within the lesion when histological diagnosis is obtained

Tissue specimen for pathological analysis is provided by either open surgery with tumour resection or by stereotactic biopsy. If the tumour is located in an eloquent area, or in central brain regions not accessible for cyto-reductive surgery, stereotactic biopsy for diagnosis always precedes treatment initiation. The biopsy needle is targeted towards the region appearing most malignant on structural imaging. However, as only a fraction of a heterogeneous tumour is analysed, there is always a risk for sampling error, and erroneous histo- pathological diagnosis 6.

The histological classification method of Kernohan et al. is used for astrocytoma staging by the neuro-pathologists in our centre 7. This system includes four grades of primary astrocytoma, oligodendroglioma or mixed oligo- astrocytoma (I, II, III, and IV) but astrocytomas with heterogeneous growth and areas of varying malignancy grades are common. Tumours graded I and/or II are grouped together as low-grade astrocytomas, tumours graded II-III, III and/or grade IV are grouped together as high-grade astrocytomas. With reference to the WHO classification this means that high-grade astrocytomas comprise anaplastic astrocytomas (AA) and oligodendro-astrocytomas and glioblastomas multiforme (GBM), the others belonging to the low-grade astrocytomas.

Treatment

Principles of treatment and management differ between low-grade and high- grade astrocytomas.

Low-grade astrocytomas The primary treatment for low-grade tumours is cyto-reductive surgery if feasible without creating additional neurological deficits. Data regarding treatment of these tumours strongly supports surgical removal, as extensive as possible, to prolong survivaly"". Radiation therapy is given to patients with low- grade astrocytomas in cases where surgical resection is limited by the astrocytoma localisation, i.e. when tumour progression directly violates eloquent brain areas and the patient is at risk for functional decline, if no treatment is initiated. Prospective randomised studies on low-grade astrocytoma management have not proven a survival benefit from conventional dose (58-60 Gy) radiotherapy as an adjuvant primary to surgery . High-grade astrocytomas Primary high-grade astrocytomas as well as low-grade tumours, which have undergone malignant transformation, are conventionally treated with extensive cyto-reductive surgery, followed by external beam radiation, and in selected cases adjuvant chemotherapy. Depending on patient age and clinical state decisions on further treatment options are made in each individual case. Even with aggressive therapy the median survival time is not longer than 18 months for patients with AA, or 12 months for patients with GBM n. Long-time survivors, with more than a 5 years survival, are less than five percent despite of combined treatment14'15. It is convincingly shown that prognosis is highly dependent on multiple prognostic variables, and that these variables have a greater impact on survival than most treatmentsl0'16'17. The elderly patients (> 60 years) have a very poor prognosis despite of multi-modality treatment, a low chance of treatment benefit, and an increased risk for treatment complications l8" 20. A survival analysis of more than 1500 patients with high-grade astrocytomas who entered clinical trials showed that the median survival of those patients with the best prognostic factors (AA, < 50 years old, good performance status and normal mental status) was 58 months, while those with the worst (GBM, > 50 years, performance status Karnofsky < 70 and abnormal mental status) survived less than 5 months '6. As a consequence, patients who are offered aggressive treatment must be carefully followed for tumour progression. They have an already very limited life expectance and good quality of life is not consistent with unjustified continuance of ineffective treatment.

Surgery Surgical treatment is never curative as highly malignant astrocytomas grow with diffuse infiltration and microscopic tumour cells spread to distant brain regions, and consequently, radical surgery is not feasible. Extensive surgical neoplasm cyto-reduction has been shown to be of benefit for prolonged survival as an independent prognostic variable 2"21'2.

Radiotherapy Radiation therapy plays a central role in the treatment of highly malignant astrocytomas, as it is the most effective adjunctive therapy following surgery. There is convincing evidence that irradiation improves survival, but it does not eradicate the tumour 23. Radiation treatment is initiated within three to five weeks of surgery. Standard radiotherapy comprises six weeks of fractionated external beam radiation administered by 6-8 MeV photons from linear accelerators. The target volume is defined as the contrast enhanced volume seen on CT with a 3-centimetre margin. Individual computer-based dose planning is carried out. Patients are given 58-60 Gy in two fractions, five days per week, in one series. This commonly used radiation protocol provides maximum treatment efficacy with acceptable risks for adverse reactions from irradiation. A total dose of > 60 Gy is associated with a markedly increased risk for cellular damage in normal brain tissue, and is also associated with a higher frequency of complications 24, but irradiation adverse events are still significant (up to 18%) when radio-therapeutic dosages < 60 Gy are given25. Brain injury by external beam radiation is classified according to time of onset into acute, early delayed and late reactions 26. 1. Acute reactions appear during the first weeks of radiotherapy and consist of headache, nausea, fever, somnolence and worsening of focal neurological deficits. Increased cerebral oedema is the likely cause of the symptoms 27'28.

2. Early delayed encephalopathy occurs one to four months after the completion of the irradiation. Common symptoms are headache and somnolence with or without deterioration of pre-existing focal deficits. Neuro-pathological autopsy reports of patients with fatal deterioration eight to twelve weeks after radiotherapy describe foci of demyelination, cell loss and mononuclear cell infiltrates. The changes reflect possible inflammatory or autoimmune reactions 29;3().

3. Development of focal radiation necrosis occurs with a latency of four months to several years after treatment. The clinical picture is equivalent with that of other space occupying lesions. The histo-pathological picture shows features consistent with extensive vascular injury. Commonly seen is fibrinoid necrosis and hyaline thickening of vessel walls, thrombotic occlusion of small vessels, proliferation of adventitial fibroblasts, and hyaline transudates surrounding the affected vessels 31 32.

4. For the few long-time survivors, irradiation of the central nervous system, with conventional doses of 60 Gy, almost invariably produces negative long- term effects of neuro-cognitive impairment due to diffuse cerebral injury26. The neuro-toxic effect of cranial irradiation results in variable severe neuro- psychological deficits ".

If chemo-sensitive subgroups among the patients with highly malignant astrocytomas can be identified by new functional methods, these patients would not need to be subjected to external beam radiation, and thus, the cognitive long term adverse events associated with this treatment would be avoided.

Chemotherapy The efficacy of post-operative chemotherapy has been difficult to demonstrate and conflicting data have been reported on this issue. Post- operative chemotherapy alone has little effect on survival, but there are data published showing an additional effect of combined chemotherapy plus radiation. A meta-analysis of 16 performed small trials suggested that a small survival advantage could be obtained by adjuvant chemotherapy compared with patients treated with radiation alone. An increase in survival of 10.1% was seen at 1 year and 8.6% at 2 years 34. However, a more recent large randomised trial was not able to confirm this benefit from conventionally used combined chemotherapy 35. Tumours with an oligodendral component are reported to have enhanced chemo-therapeutic sensitivity. Several small studies of malignant oligodendrogliomas and mixed oligodendro-astrocytomas, treated with chemotherapy administered before irradiation, have shown a 79-100% temporary, complete or partial, response rate 36"39.

In our centre, combined chemotherapy is offered to a subgroup of patients who are likely to benefit from this treatment. The therapeutic approach is based on results from a prospective, randomised study of patients with highly malignant astrocytomas, undertaken in the University hospital in Lund 40. The study demonstrated that, patients less than 50 years of age, treated with a combined chemotherapy regime and concomitant radiation therapy, had significantly longer survival than patients older than 50 years of age. Results were corrected for prognostic variables associated with better survival outcome over and above age effect alone (Karnofsky index 4I, areas of grade II, absence of extensive necrosis)40.

Chemotherapy is started with PCV (procarbazine, lomustine (CCNU), vincristine) four to five weeks post surgery in accordance with this local treatment regime 40. Procarbazine 75 mg/m is given oraly day 1 to day 28, CCNU 50 mg/m is given oraly and vincristine 1 mg/m is injected intravenously on day 1 and day 15. Cycle length is 56 days. Patients with stable disease receive maximum treatment comprising 10 courses of chemotherapy, i.e. a cumulative maximum treatment time of 18 months. Myelosuppression with trombocytopenia and peripheral neuropathy are dose limiting side effects. A fraction of PCV treated patients suffer from severe adverse effects. Liver toxicity reactions and nephrotoxicity are seen in approximately two percent of cases. As chemotherapy is associated with toxic side effects, it is important to discontinue treatment at an early stage if there is no evidence of treatment benefit. Consequently, patients benefit from sensitive and specific methods for treatment efficacy monitoring, even if today no alternative treatment regime can be offerd when chemotherapy failure is evident. Furthermore, it is clear that we must develope more selective and specific therapies to reduce neurotoxicity.

New treatment approaches Multiple novel chemotherapeutic agents and regimes have been given in phase 2 and phase 3 studies to patients with relapse of malignant astrocytoma. Despite some initiall promising results for some of these drugs, conventional regimes such as PCV remain as effective as the novel agents. New therapeutic ways of administration have been tested, including intra-thekal or intra-arterial administration, osmotic opening of the blood brain barrier (BBB), and the use of implants impregnated with chemotherapeutics. Preliminary results have been disappointing. Researchers dealing with chemotherapy in CNS tumours generally agree, that the benefit of cytotoxic drugs has reached a plateau; the future therapy probably lies in entirely new concepts of treatments 42"43. Ongoing studies are evaluating new treatment approaches such as interstitial brachytherapy and hyperfractionated external beam radiation, suicide gene therapy, such as herpes simplex thymidine kinase transfection, immunotherapy, antiangiogenesis and activation of tumour cell apoptosis by DNA damaging cytotoxic drugs l3;43'45.

10 Treatment response assessment

In the clinical setting, treatment response is evaluated by performance and imaging assessment. With disease progression tumour increase seen by neuro- imaging often preceedes clinical deterioration, but in some cases, functional decline is seen prior to radiological changes. Both clinical and imaging criteria have an element of subjectivity and both fluctuate with changes in steroid dose.

Clinical parameters

In clinical trials, assessment of response and outcome measures is methodologically complex. There are some special difficulties with regard to therapy evaluation in malignant astrocytoma trials. Pre-treatment variables influence conventionally used outcome measures and any benefit from new therapies must be separated from consequences of patient selection 17'46. Variables which strongly influence survival are; duration of symptoms, age at diagnosis, tumour grade, and postoperative performance status 18-19'4M9,

Response criteria usually include an evaluation of clinical state by standardised simple measures of neurological impairment, defined in performance scale instruments. Most commonly used are the Karnofsky scale and the Barthel disability index41;49. Many investigators tend to report clinically stable disease along with objective imaging response as possible evidence of efficacy of a treatment regime46.

Outcome measures commonly used in brain tumour trials are progression free survival, time to progression, and survival l7. Progression free survival or time to progression are measured either by time to neuro-imaging progression or by time to clinical progression. There are many difficulties and concerns with using these endpoints. They are time consuming, perhaps impossible to verify in a truly blind fashion and vulnerable to bias by investigator and subject expectations. Survival is considered a more reliable outcome measure and the "gold standard" as primary endpoint for randomised controlled trials 5°.

Neuro-imaging methods for diagnosis and treatment follow-up

CT and MR imaging

Anatomical imaging is an important part of the diagnosis, staging, and follow-up of patients harbouring brain tumours. MRJ is more sensitive than CT for detection of intra-cerebral astrocytomas and for demonstrating their extent and spread. Advantages of MRJ over CT include increased ability to detect contrast enhancement and to image in multiple planes 2:>.

Diagnosis Low-grade astrocytomas generally present as a diffusely defined region of low attenuation on CT. On MRI, lesions are usually more apparent with more definable margins characterised by homogenous iso-intensity or mild hypo- intensity appearance on Tl-weighted images and uniform hyper-intensity on T2-

12 weighted images. The amount of mass effect and oedema associated with low- grade lesions is, as a rule, minimal. Most of the tumours fail to demonstrate contrast enhancement, and if seen, it is very limited 25.

High-grade astrocytomas frequently demonstrate marked contrast enhancement on CT and MRI, but even though damage to the BBB is considered a reliable marker of high-grade malignancy 51 non-enhancing tumours are rather frequent among the moderately malignant AAs (Kernohan grade II-III and III)52;53. If contrast enhancement is seen in the AA, it is often a homogeneous enhancement as identified on CT. The amount of mass effect is commonly significant and extensive peri-tumoural oedema is seen within the adjacent white matter. The GBMs (Kernohan grade III-IV and grade IV) are usually seen as complex, multi-cystic and heterogeneous lesions. A thick, irregular rim of contrast enhancement, sometimes with a nodule, surrounds central non-enhancing regions containing necrotic tissue. MRI reflects the tumour heterogeneity both on Tl- and T2-weighted images correlating with areas of haemorrhage, necrosis, varying degrees of cellularity and peri-tumoural oedema. Neither CT nor MRI demonstrate the true extent of microscopic tumour cell spread within the brain and there is always a degree of uncertainty in defining the border of the tumour process. Not only AA, but also GBM can present without evidence of BBB disruption 52;53. In these cases microscopic dissemination of malignantly transformed cells within extensive brain regions is characterised as gliomatosis cerebri.

Infectious lesions can present with the same or similar CT characteristics as high-grade astrocytomas. Lesions identified on CT with ring enhancement are diagnosed either as malignant tumours or cerebral abscesses. A problem with differential diagnosis is common in several of these cases 3;54.

13 Follow-up Treatment efficacy in brain tumour studies evaluated by neuro-imaging is complex. CT and MRI assess morphological parameters such as change in tumour size and change in contrast enhancement. The contrast enhanced area is only an indirect marker of viable tumour tissue. Due to the often irregular and complex tumour shape, exact tumour size is difficult to measure by anatomical imaging. When the accuracy of MR volumetric measurement in high-grade astrocytomas before and after chemotherapy was studied 55, a significant intra- observer variation was seen, as well as a systematic variation between the observers. A necrotic core in the centre of the lesion, or multiple necrotic cysts in a heterogeneous tumour, act as confounding factors when trying to evaluate reduction of viable tumour cell burden. Changes of size and alterations in BBB properties do not only relate to true therapy efficacy but can also represent a steroid treatment effect or postoperative contrast enhancement along the resection margin 17;46;51i:>6_ A recent study of MR chemotherapy response criteria in malignant astrocytoma could not support the validity of current response grading (partial response > 50% reduction, minor response 25-50% reduction, progressive disease > 25% increase) and found no significant relationship between imaging response and TTP 48. Imaging criteria for progressive disease can also be erroneous as adverse effects of treatment, early and delayed, non- specific inflammatory reactions and focal radiation necrosis, can mimic tumour progression on CT and MRI5:51;"-59. 201Thallium SPECT

Historical background and mechanism of intracellular uptake The first study of biologic distribution of 20lTl in man was published in 1977 60. Thallium disappearance from the blood was seen extremely rapidly and intracellular deposition in the heart, kidneys, large bowel and thyroid was nearly immediate. The whole body half-time of 201Tl was demonstrated to be about 57 hours.

201Tl SPECT was initially used in the clinical setting for myocardial imaging. 20'Thallium was then noted to accumulate in malignant tumours of the lung, thyroid, breast, and in soft tissue sarcomas 61. 201Tl SPECT has relatively newly been introduced as a diagnostic tool for guiding the management of astrocytomas. The method has been suggested both for pre-operative diagnostic and for treatment efficacy measurement. It has been suggested as a method that can help to distinguish between astrocytomas and inflammatory lesions, a reliable method for tumour malignancy grading, and a method for evaluating the response to treatment, in particular by its ability to distinguish tumour recurrence from radiation induced changes6I"83.

201Tl SPECT is a functional imaging tool, which depends on physiological mechanisms and cell function for visualisation of uptake. It relies partly on an active uptake in viable tumour cells and thus provides a measurement of tumour viability. Several authors have discussed the mechanism of intracellular 201Tl- uptake, and there are different theories for the intracellular accumulation. The action of the adenosine triphosphatase (ATPase) system on the cell membrane that extrudes sodium from the cell and allows potassium ions to enter has been shown to be an important mechanism for Tl-uptake ' ' . Tl is a lipophilic

15 cation, and is considered to be a potassium-analogue, althought the uptake of Tl and potassium are not identical.

Experimental work on rabbit kidneys has shown 20ITl to compete with potassium for the K+ activating site in the activation of the Na-K-sensitive ATPase, having an affinity approximately 10 times greater than K+ 84. The cellular uptake of 20lTl is blocked by blockage of the sodium-potassium pump by ouabain, furosemide and digitalis 87. A study of K+and Tl+ uptake in cultured human glioma cell lines has shown that: 1) The specific steady state accumulation of Tl was significantly higher in glioma cells than in cultured cells from other malignant tumours; 2) Potassium channel blockers such as ouabain, furosemide, bumetanide, or a Na+- or Cl'-free medium inhibited uptake of both cations; 3) About 10% of the uptake was neither ouabain nor bumetanide sensitive, and other K+ channel blockers had no effect on Tl+ uptake. It was concluded that the uptake routes for both cations were similar and due to Na-K- ATPase-dependent transport and to Na-K-Cl cotransport. A recently published study has demonstrated the same mechanism for intracellular Tl+ uptake 88, but the uptake mechanism for a fraction comprising about 10% was not identified. Confirmation of the partial independence of the Na-K-pump for thallium in tumour cell lines was also published by Arbab et al. 1996 . The mechanism of uptake is still partly unknown and remains to be elucidated.

Based on the hypothesis that malignantly transformed cells loose their ability to differ between thallium and potassium °, intracellular 2 'Tl-uptake is presumed to relate to cell growth rate. Enhanced intracellular incorporation is expected in poorly differentiated cell populations with high cell density and high proliferative activity 65;91. This is the rationale for tumour grading by 201Tl SPECT examination. In vivo characterisation has reported a correlation between high cerebral 20lTl-uptake and bromodeoxyuridine activity, a thymidine

16 analogue, which is specifically incorporated into synthetic phase of the cell cycle, and thus represents high proliferative activity 78.

Diagnosis and staging 201Tl accumulation in primary and metastatic brain tumours was initially described in 1978 and 1980 92;93. Kaplan et a!50 graded 201Tl-uptake by lesion size, configuration and intensity and found an accurate correlation with tumour findings from autopsy data. Kim et a/.66 published a paper in 1990 describing 201 JJ gpgcj examination of 40 patients with low-grade tumours and 11 with high-grade lesions. The authors introduced a method for tissue attenuation correction by uptake index calculation. A region of interest (ROI) was created for the tumour region and a mirror ROI for the non-affected tissue in the contra- lateral hemisphere. They found a strong statistical difference between the 2O1T1- uptake indexes in high-grade versus low-grade lesions at a threshold of 1.5. Thereafter several authors have confirmed the ability of 201Tl SPECT to identify highly malignant astrocytomas and to distinguish between high-grade and low- grade lesions 5;6I;62;67"72;94;95.

Different methods are defined for 20lTl-index estimation. Comparing tumour 2()1Tl-uptake in the lesion, either to uptake in the contra-lateral brain 6l*6-69-14-9\ in the scalp 75"77, or in the heart %, has been suggested as the method of choice. 20lTl-uptake has been determined either by calculating a value from a few 201 63;66 67 adjacent pixels with the highest Tl-uptake within the lesion - ) or as an 66 68 9 73 74 77 96 average count per pixel within a ROI - * ; ; ; ; , Most authors apply the average method, as it is considered the most reproducible method 66. Ratios below a threshold of between 1.4-2.0 are reported to identify low-grade lesions, whereas ratios above 1.5-2.0 indicate highly malignant astrocytoma.

17 2<)IT1 SPECT has also been evaluated for tumour grading among the population of high-grade lesions, i.e. its ability to distinguish between tumours Kernohan grade II-III, III, III-IV and IV. Conflicting data are published upon this issue. Studies in favour of staging by 20lTl SPECT are pointing out; 1) the advantage of a non-traumatic technique and 2) the risk for sampling error when histo-pathologic diagnosis is obtained by stereotactic biopsy 65;78;97. However, recently published data do not support the staging ability of 201Tl SPECT94;98;".

201Tl SPECT has been suggested as a method that can help to distinguish between malignant astrocytomas and infectious abscesses as the two entities present with the same characteristics on MR or CT images. Distribution of thallium in tumour tissue and inflammatory lesions has been studied in rats by Ando et al. H)l) Their study demonstrated that 20lTl accumulated mainly in tumour tissue, but also, to a lesser extent in subcutaneous inflammatory tissue, specifically at sites with a high density of leukocytes and macrophages. 201Tl was barely detectable in necrotic tissue. Ruiz et a/."" suggested 2mTl SPECT to be of value in discriminating lymphomas from infectious processes in AIDS patients with intracranial mass lesions, as none of their patients with infectious lesions had a 201Tl- uptake. However, other authors have reported 2mTl-uptake both in extra-cranial and intra-cranial infectious lesions 102-103. Abnormal intracerebral 201Tl localisation was found in a bacterial brain abscess by Krishna et al.l04, and uptake in a cytomegalovirus encephalitis by Gorniak et al.105.

Comparison with anatomical neuro-imaging 20lTl-index calculation for astrocytoma diagnosis has mostly been validated by comparison with histo-pathological findings. A few papers have compared 20lTl SPECT findings with radiological characteristics. Ricci et al.m published a systematic evaluation of the influence of morphological characteristics of GBMs on cerebral thallium uptake in 1996. MRI parameters such as tumour dimension,

18 peri-lesional oedema, intra-tumoural necrosis and contrast enhancement were compared to degree of thallium uptake. Only necrosis was shown to significantly influence the degree of 201Tl-uptake, correlating with a decrease of 201Tl-uptake index. The use of small regular ROIs was found to diminish the influence of necrosis. Another indication of the influence of morphological features on 20lTl-uptake was reported by Kimura et a/.107. Indexes in separate lesions within the same brain varied between 2.5 and 1.5 due to cystic or solid tumour growth, despite of the identical histological diagnosis of GBM. Maria et a/.108 evaluated the correlation between gadolinium enhancement on MRI and the degree of thallium uptake on SPECT in a study published 1997. Thirteen children with brain-stem astrocytoma were evaluated, 74% showed concordance between the presence of contrast enhancement and 201Tl-uptake. There were no cases that demonstrated 20lTl-uptake but lacked gadolinium enhancement. Rollins et al. (1995)98 reported a tendency to underestimate tumour burden by 2(1 *T1 SPECT examination of non-enhancing astrocytomas. However, Nadvi et al. (1998)94 found no correlation between 20lTl-uptake and the following MR features: gadolinium enhancement, necrosis or location of the lesion within the brainstem. Published data are partly conflicting and the true influence of morphological characteristics on cerebral 20lTl-uptake remains to be determined.

Treatment evaluation Steroids, used as symptomatic therapy in astrocytoma treatment, are well known to influence CT and MRI findings by reduction of peri-tumoural oedema. Conflicting data have been published concerning the effect of steroids on cerebral 201Tl-uptake 7i;109. Ricci et al.m did not find any influence of peri- lesional oedema on degree of 201Tl-uptake when MRI characteristics were correlated to SPECT findings. Black et al.lx found that uptake was minimally affected by steroids, but Namba et al.m found uptake indexes decreased by 33% compared to the values before steroid administration. If 20lTl-uptake is

19 unaffected by concomitant steroid administration, in contrast to clinical and structural imaging evaluation, this carries important implications for therapy monitoring.

Post-operative 201Tl-indexes have shown an ability to differentiate between 69;73 75 96 110 residual tumour and focal radiation necrosis - ; ; ; an^ an ability to correctly detect post treatment tumour recurrence 63;69;73-74'76-77'80-82_ However, the accuracy varies between 57% and 100% in different studies. Sensitivity is decreased due to resolution problems, when small lesions are evaluated 69, and partial volume effects can lead to underestimation of 20lTl-uptake. Cerebral 20lTl-uptake post astrocytoma surgery indicates a residual viable tumour mass 65, and the presence of 201Tl in the brain may have important prognostic implications. 20lTl-uptake progression has been shown to correlate with negative patient outcome (higher mortality rates) 64;83;108:l" A few published studies have evaluated astrocytoma treatment by repeated scanning during chemotherapy comparing changes of 2lHTl-index level 99:ll()-ll2;l13, 201xi- index calculation for therapy monitoring purposes has been reported to be a less sensitive method for astrocytoma evaluation than other functional methods ("C- MET PET "0;"4, l8F-fluorodeoxygIucose PET w). Diagnostic precision in comparison with structural imaging is scarcely investigated. The only published prospective study evaluating Tl SPECT and CT for brain tumour burden found, that CT was unable to diagnose increase of tumour size in six out of nine patients correctly diagnosed by 20lTl SPECT "2. However, MRI is a more sensitive method than CT for astrocytoma treatment evaluation 25, and comparison between MR and functional imaging reliability remains to be investigated.

20 'H-MR spectroscopy

Metabolite characterisation Magnetic resonance spectroscopy characterises the metabolic profile in a defined brain region. Metabolites can be identified in the spectra due to their differences in resonance frequency. The metabolites are presented in a spectrum in a plotted graph as a series of peaks along the x-axis. The area under the curve, rather than amplitude, represents the metabolite concentration. Spectral patterns from astrocytomas differ from those in normal brain parenchyma. The following metabolites are of significance for brain tumour evaluation; 1. N-acetyl aspartat (NAA), identified at 2.0 ppm on the x-axis of the spectrum, is a neuronal marker. Decrease of NAA concentration represents neuronal death or dysfunction or replacement.

2. Creatine (Cr), identified at 3.0 ppm, represents the intracellular pool of phosphorylated and unphosphorylated creatine, part of the ATP synthesising process. Depletion is seen when substantial tissue loss or replacement occurs.

3. Choline (Cho), identified at 3.2 ppm, reflects cell membrane turnover. Elevation is seen with an increase of cell membrane turnover.

4. Lactate (Lac), identified at 1.3 ppm, is the end product from anaerobic glycolysis. It is therefore a marker of cerebral ischemia and seen in necrotic tissue.

5. Mobile lipids (Lip), identified at 0.9-1.3 ppm reflect mostly cellmembrane breakdown-products.

21 In brain areas with malignantly transformed cell populations 'H-MRS spectra show an increase in Cho combined with a decrease in NAA (Figure 1), which makes the NAA/Cho ratio a suitable marker for astrocytomas. NAA/Cho ratios > 2 are seen in the normal brain, ratios < 1 are definitely pathological. In necrotic tumour cores an extensive loss of NAA, Cho, and Cr is expected often in combination with detectable Lac.

NAA Cho

NAA Cho Cr Vu wUwW

Normal 'H-MRS spectrum A.strocyloma 'H-MRS spectrum NAA/Cho ratio = 1.6 NAA/Cho ratio = 0.9 Figure 1. Examples of normal bruin and astrocytoma spectra

Diagnosis and staging Contradictory results are published on the reliability of 'H-MRS characterisation for malignant astrocytoma diagnosis and staging ' " . When the most common technique is implemented, single-volume spectroscopy, tumour heterogeneity complicates interpretation of spectral analysis. This technique takes measurements from a restricted area (minimum 1,5 cm for each dimension), which is too large to avoid partial volume effects, but still too small to cover a complex and irregular GBM l26. Ott et alm evaluated 'H-MRS for

22 differential diagnosis. They found limitations for 'H-MRS use, as intra- individual differences between spectra of one tumour at different locations were often larger than differences between spectra of tumours with different histo- pathologic characteristics. Lactate is likely to be found in high-grade astrocytomas but its presence is not a reliable indicator of malignancy, it can also be found in cerebral infectious lesions, benign tumours, and in focal radiation necrosis 120;l2I127;l28. However, a recent study showed promising results from 'H-MRS analysis of tissue adjacent to the lesion as identified on MR images l29. Differences in NAA/Cho ratio in tissue, adjacent to and at one voxel distance from the lesion, could distinguish lesions with infiltrating growth from well-circumscribed lesions, i.e. high-grade astrocytomas from metastases.

Treatment evaluation Preliminary studies and case reports have evaluated !H-MRS for astrocytoma therapy monitoring I24-I3CM32_ Houkin et al,m investigated metabolic changes in brain tumours after treatment and no significant changes were observed in five cases with GBM after chemotherapy and radiotherapy. Speck et al.131 observed spectral changes after interstitial radiosurgery, which demonstrated the possibility of therapy monitoring. Waldrop et a/.130 investigated children after chemotherapy and whole brain radiation treatment, and showed that MR spectroscopy of brain tissue remote from the tumour reveals treatment-related biochemical changes.

23 Aims of the study

• To compare CT findings, 20lTl-uptake indexes and histological tumour staging in astrocytomas, and to evaluate the reliability of the two preoperative assessment methods.

• To investigate patients with cystic contrast enhancing lesions on CT, and to evaluate, if 20iTl SPECT indexes add to increased precision in preoperatiye differential diagnosis between astrocytomas and abscesses.

• To describe a quantitative 20lTl SPECT measurement, evaluating total 2O1T1- uptake within the tumour region, aimed for treatment efficacy monitoring.

To evaluate the quantitative 2(IIT1 SPECT measurement in comparison with CT for astrocytoma treatment follow-up purposes.

To compare quantitative 20ITl SPECT and 'H-MRS with conventional MRI for astrocytoma chemotherapy monitoring, and to evaluate associations between change of morphological tumour characteristics during treatment, identified by MRI, and changes of cerebral thallium uptake and metabolic ratios, characterised by 'H-MRS.

24 Materials and methods

Patient selection and follow-up

Paper I; Over a 3-year interval (1992 - 1994), 67 patients with suspected malignant astrocytoma, as judged from CT, underwent 201Tl SPECT brain scans. Thirty-seven patients with 201Tl SPECT performed preoperatively and histo- pathologic diagnosis of astrocytoma were included in study I. Patient age range was 10 to 75 years, seven patients had diagnostic biopsy and 30 had undergone tumour resection. Tumours were classified in accordance to the histological classification of Kemohan et al.7. The time interval between the SPECT examination and the surgery never exceeded 35 days (median 12 days).

Paper II; From the same patient sample as in study I, twenty-one patients with cystic, ring-enhancing lesions on CT were included in study II. Patient age ranged from 11 to 74 years, two patients underwent diagnostic biopsy, 15 patients underwent tumour resection. In four cases an infectious abscess was suspected. The cystic cavity was punctured and extirpated in two cases, the infectious diagnosis was verified by positive leukocyte SPECT in the other two cases. The clinical course of the patients was followed for a minimum of three years and a maximum of six years. All patients with highly malignant tumours, except for one, were followed until death.

25 Papers III, IV and V; From October 1, 1995 through May 31, 1997, 27 consecutive patients admitted to the University hospital of Lund, southern Sweden, were enrolled. All patients had histologically defined highly malignant astrocytoma, classified according to the method of Kernohan et al?. Patients were accepted for adjuvant three-agent-combined chemotherapy (PCV) in accordance with the adapted local treatment regime 40 with or without concomitant external radiotherapy (58-60 Gy). All patients, that were offered chemotherapy, were below 50 years of age, and with good functional status at treatment initiation. Five patients were lost from follow-up, as only base-line imaging was obtained. Treatment discontinuation was due to intolerability, rapid progression of disease, and, in one patient, death due to a non-treatment related accident.

The 16 primarily enrolled patients (inclusion period = 18 months) were followed during treatment with repeated 20lTl SPECT and CT (Paper IV), the six last enrolled patients (inclusion period = five months) were additionally followed by repeated MRI and 'H-MRS (Paper V). Sixteen patients received PCV chemotherapy in combination with external radiation treatment for newly diagnosed high-grade astrocytoma. Six patients received PCV chemotherapy for recurrent low-grade astrocytoma after histo-pathologic diagnosis of malignant transformation. Only one patient underwent re-operation after chemotherapy discontinuation due to progressive disease. No patient was treated with second line chemotherapy at tumour progression, which means that all but one patient received only symptomatic treatment and best palliative care after chemotherapy failure. The clinical course of the patients was followed until death, or for a minimum of 101 weeks and a maximum of 135 weeks.

26 Techniques and data acquisition

All patients (papers I-V) had a preoperative CT examination before and after administration of contrast. Different CT scanners were used. The slices were parallel to the orbito-meatal line and the thickness varied between 4-10 mm in the posterior fossa and 8-10 mm supra-tentorially. Contrast enhancement was achieved by a bolus injection of 100 ml iohexol 300 mg/ml or 1.5 ml/kg body weight to patients with a body weight below 60 kg.

201Thallium SPECT

The 20lTl SPECT measurements (papers I-V) were made in a SPECT camera (Ceraspect) dedicated to brain imaging. Each patient was given an intravenous injection of 100 MBq 20lTl in isotonic sodium chloride. The patients were positioned in the SPECT camera five minutes after thallium administration. The 2

27 CT images were used as anatomic guide. The 20lTl-uptake was measured in the slices corresponding to the location of the lesion identified on CT.

MR imaging and 'H-MR spectroscopy

All MRI and 'H-MRS measurements were performed on a 1.5 T unit (Siemens Magnetom VISION, Siemens, Erlangen, Germany) using a circularly polarized headcoil and a standard spectroscopy package provided by the manufacturer. MR included T1-, T2-, and PD-weighted sagittal and axial pre- contrast series and axial and coronal Tl-weighted post-contrast series performed after the 'H-MRS measurements. MR sequences used were Tl-weighted (TR/TE 450-630/14-17) spin echo (SE) sequences and a PD-, T2-weighted (TR/TE 3710/22,90) fast SE sequence. Three series of orthogonal Tl-weighted gradient echo scout images (TR/TE/flip angle 100/4.8/80°) covering the lesion, were used to position the measurement volume in 'H-MRS. For 'H-MRS a two- dimensional chemical shift imaging (2D CSI) point-resolved spectroscopy (PRESS) sequence with a TR of 1500 ms and a TE of 270 ms was used.

'H-MR spectroscopy data were post-processed and MR spectra were evaluated for the following substances; choline containing compounds (Cho); creatine (Cr); N-acetylaspartate (NAA) and lactate or mobile lipids, summarized under lactate-lipids (LL).

28 Data acquisition for therapy monitoring

Repeated 201Tl SPECT, CT (paper IV), MRI and 'H-MRS (paper V) examinations assessed treatment response. Evaluations were performed prior to the first course of chemotherapy and then prior to the start of every second PCV course until treatment discontinuation. Maximum imaging follow-up time for progression was 74 weeks (last scan = follow-up 5, before start of PCV cycle 10) equivalent to a total of six evaluations. The pre-chemotherapy measurement was used as the baseline, a second measurement was performed eight weeks after treatment initiation and thereafter with a 16-week interval as illustrated in Figure 2.

PCV courses i=> •=> £=> •=> <=> !==> <==> •=> <==> c=C>

0 8 16 24 32 40 48 56 64 74 78 Weeks from chemotherapy initiation

baseline follow-up 1 follow-up 2 follow-up 3 follow-up 4 follow-up 5 Imaging

Figure 2. Examinations in relation to treatment

29 Data analysis

Paper I; The reliability of 201Tl SPECT was compared to CT in grading the malignancy of astrocytomas. Interpretators were blinded to results of each others. From CT images, total volume of the tumour and the volume of the tissue with BBB damage were measured. The presence of BBB damage was defined by a visual comparison of pre- and post contrast images. To measure the volume of the tumours and their enhancement, a 5 x 10 cm grid was drawn on a disposed transparent x-ray film using cm-scale, indicated on the image. The grid was placed on the tumor and areas of enhancement in every slice of the tumour and the area of enhancement or tumour volume was measured. The total volume was then calculated by multiplying the area of each slice with the slice thickness, and the volumes on each slice showing pathology were added to each other. In each case the tumor was further categorized in terms of solidity or diffuse tumor growth, and the presence of necrotic or cystic tumour components.

Cerebral 201Tl -uptake in the astrocytomas was estimated by index calculation. A cross section profile was drawn through the part of the tumour with the highest 20iTI-uptake. The average value of the three adjacent pixels with the highest thallium uptake was used. For each patient we determined a 201Tl- index of the uptake within the tumour, relative to the mean of three homologous pixels in the contra-lateral side. 201Tl-index = tumour uptake/contra-lateral uptake. In cases where the lesion was in the midline, and consequently the 2 Tl - index would have been 1.0, the tumour uptake was compared to the uptake in a low uptake region in the lateral part of one hemisphere.

Paper II; Cerebral 20lTl-uptake in cystic contrast enhancing lesions on CT was measured by 20lTl -index calculation and evaluated for differential diagnosis

30 between astrocytomas and abscesses. Uptake indexes were compared with CT enhancement volumes, histo-pathology, and survival times. Methods for tumour volume and contrast enhancement volume estimation, and for 201Tl-index calculation were the same as described for paper I.

Paper III; Three methods for quantification of 20lTl-uptake in the region with post-operative residual viable tumour were compared with early post-operative tumour volumes calculated from CT. 201Tl-indexes calculated in two different ways were compared with histo-pathologic findings.

Calculation of tumour volume was made directly from the CT films. The total volume measured included the area with BBB damage plus areas with necrosis or cysts. Volumes were measured using three orthogonal diameters (two of them in the plane of the film). The half value of the arithmetic average of the three diameters was taken as the radius of a sphere.

For each 201Tl SPECT measurement we determined a tumour uptake volume (TUV) within a defined 3-dimensional tumour region of interest (ROI). The ROI was drawn around the lesion with a border localised well outside the lesion. An example of a SPECT image with a marked ROI surrounding the lesion is shown in Figure 3 (page 53). The TUV was calculated from the number of voxels that had a 201Tl-uptake above a threshold calculated from the uptake level on the contra-lateral side. If the tumour extended over the midline, the reference ROI was obtained in an anterior or posterior direction. The voxel volume was 0.013 cm3. The threshold level was selected with the aim to exclude any tissue that did not have a 20ITl-uptake corresponding to a highly malignant astrocytoma. Previous studies have shown that a thallium index > 1.4 in an astrocytoma is a 66 68;71;74 predictor of high grade malignancy " _ in consequence, 1.4 times the mean 201Tl-uptake in the reference region was used as threshold. The scatter of 201Tl-

31 uptake in individual voxels may give an overestimation of malignancy grading in individual pixels. To compensate for the corresponding overestimate of the malignant tumour volume, three different cut-off levels were used: Method a) the threshold was increased to a 96.7 % confidence level estimated from the standard deviation (SD) of the 20lTl-uptake in the pixels of the reference region [mean ref*1.4+2SDJ, the volume of voxels in the reference ROI with a 201Tl-uptake above the threshold level was subtracted from the number in the tumour ROI, with compensation for unequal ROI sizes; Method b) the threshold was increased to a 99.9 % confidence level

[mean ref*1.4+3SD]\ Method c) the volume of the pixels in the tumour ROI with a 01Tl-uptake that exceeded the highest uptake in the reference ROI [max refj.

201Tl-uptake indexes were calculated in two different ways:

1. the pixel with maximum counts in the tumour ROI = [max tumour/mean ref] average counts per pixel in the reference ROI

2. average counts per pixel in the tumour ROI . = [mean tumour/mean ref] average counts per pixel in the reference ROI

Paper IV; Reliability of quantitative 2 25%. Tumour volumes were calculated from CT films, and

32 TUV was calculated from SPECT scans by the method b) previously described (data analysis; paper III; [mean ref*1.4+3SD]).

Outcome parameters were defined as: 1. time to treatment discontinuance (TTD) = number of weeks from day 1 of chemotherapy until treatment was discontinued due to definite tumour progression (defined by CT criteria and clinical assessment, evaluating performance and steroid dose);

2. survival time (ST) = number of weeks from initiation of chemotherapy until death.

Paper V; Results from repeated examinations with 2O1T1-SPECT, MRI and 'H-MRS during chemotherapy, were compared with the clinical course of the patients. Examination interval is illustrated in Figure 2 (page 29). Changes of MRI parameters were compared with changes of 20lTl-uptake volumes and 'H- MRS metabolic ratios. Associations between morphological changes and 201Tl SPECT and ' H-MRS results were assessed.

MRI follow-up examinations were evaluated regarding tumour size, size of contrast enhancement, oedema, necrosis, haemorrhage, and mass effect. Recurrence, local or distant, of tumour after resection was noted. The disease was defined as being in a state of overall progression or regression when it changed category (four defined levels) of tumour size and/or size of contrast enhancement. Increase of necrosis or haemorrhage was defined as progression. Exclusive changes in oedema were not defined as overall progression or regression of disease.

33 20iTl-uptake for baseline and follow-up examinations were calculated by quantitative 20lTl SPECT method b) [mean ref*1.4+3SDJ. Examinations were evaluated for: 1) regression = TUV reduction > 25%; 2) status quo = TUV reduction or increase < 25%; and 3) progression = TUV increase > 25%.

Metabolite ratios calculated from MR spectroscopy were: NAA/Cho, NAA/Cr and Cr/Cho. The presence or absence of LL was noted. Criteria for voxel-selection in base-line and follow-up examinations were: 1) the tissue covered by the voxel could be identified in all evaluated measurements according to anatomic structures on MRS-scout images and MR images (excluding voxels covering tissue displaced by extensive mass effect or exclusively covering necrotic cyst material); 2) the voxel evaluated should, if possible, cover the lesion, or otherwise tissue adjacent to the lesion; 3) among the eligible voxels those with lowest NAA/Cho ratios were chosen.

Only NAA/Cho ratios < 1.0 were defined as pathologic. Regression or progression of disease was defined as changes in NAA/Cho ratios from < 1.0 to > 1.0 or from > 1.0 to < 1.0 respectively. Loss of detectable NAA or Cr in an evaluated voxel was defined as progression of disease.

34 Results

Table 1. Overview of number of patients in studies I-V

201TISPECT studies Astrocytoma Cerebral abscess high-grade low-grade (n) (n) (n) Preoperative studies I Reliability in astrocytoma malignancy grading 31 6 II Evaluation of malignancy in cystic lesions 14 3 4 Post-operative studies III Validation of quantitative method 15 IV Treatment monitoring; comparison to CT 16 V Treatment monitoring; comparison to MRI 6

Pre-operative studies; Papers I and II

201Tl-indexes for different lesions grouped together according to histo- pathologic diagnosis are given in Table 2. We found a significant difference between low-grade astrocytomas and astrocytomas grade III and/or grade IV, but not between low-grade gliomas and gliomas of grade II-III.

35 Table 2. 2nl Tl-indexes for different lesions in studies I and II

Diagnoses and morphology (n) 201Tl-indexes *

1. Highly malignant astrocytomas 31 2.1(1.0-5.1; - solid or mixed 11 2.8(1.8-5.1) - cystic tumours 15 1.7(1.0-4.2) - diffuse or non-enhancing 5 1.6(1.2-1.8) 2. Astrocytomas grade I-II 6 1.2(1.0-1.4) 3. Pleomorphic xanthoastrocytoma 1 6.2 4. Abscesses 4 1.6(1.0-1.9)

*median and range

The pre-operative studies clearly demonstrated that 20lTl-uptake was dependent on tumour morphology, in particular extent of blood-brain barrier breakdown. Both studies found a significant relationship between uptake indexes and contrast enhancement volume as illustrated by the regression plots in Figure 4.

-10 0 10 20 30 40 50 60 0 10 2030405060708090 Contrast enhancement volume high-grade gliomas (ml) Contrast enhancoral wlureQStic lesions (ni)

r=0.4l,p=0.02l4 r=0.75, p=0.005

Figure 4. Correlation between thallium indexes and CT enhancement volumes in high-grade gliomas and in cystic lesions

36 Tumours that did not show contrast enhancement on CT, were not accessable for thallium. Non-enhancing high-grade astrocytomas showed 201Tl-uptake and 20lTl-index in the same range as low-grade tumours and infectious lesions.

Cystic tumours with extensive areas of necrosis were also noted to accumulate thallium to a lesser degree and to show uptake indexes in the lower range (Paper I). 20lTl-index > 2 was consistent with tumour diagnosis, but overlapping indexes were seen between high-grade and low-grade lesions (Papers I and II).

Accuracy of 20ITl SPECT and CT in identifying high-grade lesions (paper I) is given in Table 3.

Table 3. Reliability of three different imaging methods in astrocytoma diagnosis and treatment monitoring

Number and aim of study Accuracy (%) 201 Tl SPECT CT MRI

I To differentiate between high-grade and low-grade astrocytomas 78 84 (evaluated lesions = 37)

IV Early identification of treatment failure 75 56 (SPECT/CT measurements = 59/59)

V Diagnostic precision in chemotherapy efficacy monitoring 78 - 83 (SPECT/MR measurements = 18/18)

37 Overlapping indexes were also seen between the different groups of highly malignant astrocytomas (grade II-III, grade III, and grade III-IV/IV) as shown in Figure 5.

grl-II grll-III grill gr III-IV/IV Kernohan histological grading

Figure 5. Differences between thallium indexes in Kernohan histologic groups (mean and SEM) for tumours of grade I-II, n=6; grade II-III, n=6; grade III, n=lI; grade III-IV/or grade IV n=14.

O C\ 1 Quantification of cerebral Tl-uptake; Paper III

The quantitative method for 20lTl-SPECT evaluation was validated by comparison with early postoperative CT images. Simple linear regression analysis gave a significant correlation between all three TUV methods and CT measured tumour volumes (p=0.009-0.017). Method b) [mean ref*1.4+3SD] showed the highest correlation with CT. 2olTl-indexes, calculated by either [max tumour/mean ref ] or [mean tumour/mean ref J, did not show a significant correlation with histological grading within the group of highly malignant

38 astrocytomas. The study confirmed the difficulties in tumour staging by 201Tl- indexes calculation.

An example is given in Figure 6 (page 54) to illustrate the impact of tumour morphological characteristics on 201Tl-indexes. CT and SPECT images from our follow-up series of two patients with highly malignant astrocytomas during chemotherapy are shown. One patient had a homogeneous contrast-enhancing tumour (patient A) and the other patient harboured a cystic tumour with ring- enhancement (patient B). For both patients, chemotherapy was discontinued due to progression of disease. Indexes calculated from [max turn/mean ref ] ~ max 201TI were compared to indexes calculated from [mean turn/mean ref ] — average 21>ITI within the defined ROIs, and also compared to TUV measurements.

In patient A; average 20!TI, as well as TUV calculation, monitors disease progression by showing increase of tumour burden at repeated scanning, while the max 2OITI method gave unstable results with low precision.

In patient B; tumour progression is neither monitored by average 20lTI calculation, nor by max ' TI. Results from index calculation in patient B, are not concordant with results from visual analysis showing initial decrease of 2t)1Tl-uptake at follow-up 1. In contrast to indexes, TUV calculation is concordant with visual analysis, as well as concordant with the clinical course of the patient. The TUV method provides a more robust measure than 201Tl-index calculation for astrocytoma longitudinal follow-up purposes.

39 Post-operative studies; Papers IV and V

Paper IV; Sixteen patients with histologically verified highly malignant astrocytomas were followed with repeated scanning during chemotherapy, median follow-up before treatment discontinuation was 27 weeks (range 16-78). A total of 59 SPECT and 59 CT examinations were performed.

The comparative reliability of the SPECT and CT methods was evaluated. 20lTl SPECT showed the highest accuracy in early identification of astrocytoma treatment failure at week-24 assessment. The accuracies of the two methods are given in Table 3 (page 37). Patients with > 25% tumour increase (n=7), as measured by quantitative 201Tl SPECT, after three courses of chemotherapy, had a significantly reduced TTD (p=0.040).

Uptake volumes among the patients who received simultaneous radiation and chemotherapy postoperatively (n = 10) fell into two categories at assessment after three courses of chemotherapy: a) TUV < 10ml (n=5); and b) TUV > 10ml (n=5). Patients with TUV > 10ml at this assessment point had a significantly shorter TTD (p=0.016) and a significantly reduced ST (p=0.024).

Figure 7 (page 55) shows 201Tl SPECT images, TUV measurements and CT images from a patient with recurrent low-grade astrocytoma receiving chemotherapy. Arrows indicate when the criterion for progressive disease was met by the two different methods for treatment evaluation. Treatment was discontinued after seven courses of chemotherapy due to disease progression.

40 Paper V; Six patients with histologically defined highly malignant astrocytomas were followed with repeated scanning during chemotherapy, median follow-up before treatment discontinuation was 45 weeks (range 16-78). A total of 24 examinations including SPECT, MRI and 'H-MRS were performed.

Functional and anatomical imaging and ratios from chemical analysis were compared with the clinical course of the patients and gave the following results; 1. Five patients developed clinical progression of disease, 4 out of 5 20!Tl SPECT assessments, 4 out of 5 MRI and 1 out of the 2 interpretable 'H-MRS examinations met the criterion for progression at final measurement before clinical progression of disease.

2. During the phase of clinically stable disease: a) the criterion for regression or status quo was met in 10 out of 13 2OIT1 SPECT assessments, 11 out of 13 MRI assessments, and 8 out of the 9 interpretable 'H-MRS examinations; b) the criterion for progression was met in 3 out of 13 2O1X1 SPECT assessments, 2 out of 13 MRI, and 1 out of 9 'H-MRS examinations.

The accuracy of 201Tl SPECT was slightly lower than that of MRI for astrocytoma chemotherapy monitoring as summarised in Table 3 (page 37).

Evaluation of associations between morphological changes and 201Tl SPECT and 'H-MRS examination gave the following results: a) Regression of disease as defined by 201Tl SPECT was associated with decrease in tumour size, contrast enhancement, oedema and haemorrhage on MRI; b) Progression of disease as defined by 20ITl SPECT was associated with increases in the same four MRI parameters, and also a increased degree of necrosis; c) Regression of disease as defined by ' H-MRS was associated with less oedema;

41 d) Progression of disease as defined by H-MRS was associated with an increased tumour size and haemorrhage, and an increase or decrease of contrast enhancement on MRI.

Patients without any or with limited (< 5ml) contrast enhancement on MRI had a small volume (< 10 ml) of detectable cerebral 201Tl-uptake. This is illustrated in Figure 8 (page 56), which shows, that a distinct uptake of 201Tl was seen in the tumour region simultaneously as gadolinium enhancement became evident on MRI at follow-up two. In this patient ' H-MRS showed NAA/Cho ratio decrease concordant with progression of disease at follow-up 1.

42 General discussion

Diagnosis and staging

Our results are in agreement with those of several earlier reports in showing a higher uptake of 201Tl in highly malignant astrocytomas than in low-grade and infectious lesions, when lesions were grouped together 61;62;65;97. However, differences between these three groups were not statistically significant as indexes were partly overlapping. We compared 201Tl SPECT findings with results from CT, and there was no evidence of superiority for the SPECT method in preoperative diagnostic precision or tumour staging.

Certain tumour characteristics seem to influence 201Tl-uptake. Findings were consistent throughout this study, in pre-operative, as well as in post-operative evaluation of 201Tl SPECT examinations. One of the findings was the fact that 20lTl did not accumulate in tumours that did not show contrast enhancement on CT or MR images. This is in accordance with results from recently published data (Maria et al. 1997) 1OS, blood-brain barrier breakdown seems to be a prerequisite for 201Tl penetration into the brain. This carries important implications for the use of 201Tl SPECT in tumour diagnosis. Low-grade astrocytomas are as a rule non-enhancing and 201Tl SPECT does not add information to structural imaging in these cases. High-grade lesions without contrast enhancement on CT or MRI are falsely judged as low-grade by both methods.

43 One benign tumour of uncommon cell type with marked contrast enhancement on CT was found to accumulate 20lTl to an extremely high extent (Paper II). Ueda et al.68 have shown that time sequential scanning is of importance for differentiation between hypervascular benign and malignant tumours. A high retention index with early uptake but slow washout of 2OIT1 indicates a highly malignant lesion. In our study we have only performed single scanning, and therefore, retention indexes have not been calculated. It is possible that early and delayed scanning would have added important information in this special case.

There is need for new diagnostic methods with enhanced specificity for noninvasive tumour staging, or with improved abilities to guide stereotactic biopsy towards the region with the highest grade of malignancy within the lesion. In accordance with data from recently published studies (Nadvi et al. 1998) 94, we were not able to confirm the staging ability of 20lTl SPECT. 201Tl-indexes did not correlate with histological grades of malignancy and biologic aggressiveness. Statistically, there was no evidence of less aggressive behaviour among the tumours graded as highly malignant but showing a low 'Tl-uptake. Tumour heterogeneity and morphological characteristics, as damage to the blood-brain barrier and cystic or solid growth, seem to be of critical importance for degree of 20lTl-uptake and index level also among the group of highly malignant astrocytomas.

Many authors have pointed out that 2UI Tl-uptake is not solely dependent on BBB injury, but also on cell turn over rate and cell differentiation. In vivo studies have reported enhanced intracellular Tl incorporation in poorly differentiated cell populations with high proliferative activity 65;91. One mechanism of uptake is an active transport by less selective ion channels in malignantly transformed cells by affinity to cell membrane potassium activating

44 sites 84. However, the actual affinity for Tl+ in comparison with K+ for the Na-K- ATPas site in normal astroglial cells is not sufficiently established. To the best of our knowledge, there has been no study undertaken to elucidate if primary cultures from astroglial cells differ from human glioma cells cultures by higher 20lTl-uptake. Additional to an active uptake mechanism, 201Tl-uptake by astrocytes may be largely passive and a function of the membrane potential gradient. The brain is shielded from 201Tl influx by the BBB, and actual accumulation of 201Tl in an unshielded brain with normal astroglial cells is unknown.

Treatment monitoring

Evaluation by conventional imaging during treatment is essential; clinical deterioration due to tumour progression can be separated from development of hydrocephalus and other treatable mass effects. Structural imaging is therefore necessary for guidance of surgical intervention and steroid therapy and provides reliable information relevant for clinical decisions. However, both clinical evaluation and neuro-imaging for astrocytoma treatment follow-up has its shortcomings 5I;56. There is a need for additional methods for treatment evaluation, both in the clinical setting, and for future trials of new therapy strategies. New methods for treatment evaluation are needed, but they must give additional information of clinical relevance, which will influence management decisions l33. Reliable methods are needed for differentiation of tumour progression and treatment adverse effects. Methods should be able to generate specific information regarding therapy benefit, and justification of therapy.

45 We suggest that the 20lTl SPECT tumour-uptake-volume is a quantitative measurement that reflects metabolically highly active tissue and residual tumour viability after surgery. TUV identifies high uptake regions within the tumour ROI. By setting a threshold the TUV method excludes low uptake regions as necrotic tumour cysts. The chosen threshold evaluates the total volume of highly malignant cells. The tumour viability assessment by TUV estimation is not influenced by local variability such as focal regions of extremely high uptake. Setting a cut off level compensates for overestimation due to uptake scatter. A semi-quantitative method has previously been described by Mountz et al.' 4, yielding a value defined as "tumour bulk" expressed in units of background pixel equivalents. Methods quantifying cerebral 201Tl-uptake are, to a lesser extent than 201Tl-indexes, influenced by changes in morphological parameters and tumour heterogeneity during treatment.

Many factors reduce the reliability of 20lTl SPECT for treatment monitoring purposes. If cerebral 201Tl-uptake is influenced by certain morphological parameters, also uptake volume can be underestimated due to tumour necrosis 106;107, characteristics concordant with CT and MRI features that can be seen in tumour progression. Underestimation of TUV can also be a consequence of an intact BBB. However, our results in favour of quantitative 2OIT1 SPECT for chemotherapy monitoring are supported by a study recently published (Roesdi et al. December 1998), evaluating 20lTl SPECT as a response parameter in 10 patients treated with PCV. In 9 out of 13 patients early 20lTl SPECT changes (decrease in volume and intensity) were more pronounced than radiological findings. 20lTl SPECT results after completion of chemotherapy correlated with clinical findings during follow-up 13\ Volume quantification of 201Tl-uptake seems to be more robust than 20lTl-indexes for treatment monitoring purposes.

46 Our results indicate a higher reliability of quantitative 20!Tl -SPECT than of CT for early signs of astrocytoma progression (Paper IV). As the number of evaluated patients in this study is small, the results may be biased by patient selection. However, median time to progressive disease and treatment discontinuation in relation to patient median age and tumour grade are in concordance with several other brain tumour studies 1'14'1 '34'4 '13 "13 . Even if the selected patients belong to a group with a likelihood of chemotherapy response, published controlled trials of nitrosourea chemotherapy have shown only slight improvement in median time to progression 34. In our material 15% (n=3) received PCV treatment for ten cycles with base-line and five follow-up imaging series. The number of patients with stable disease during therapy is in concordance with other studies l4'4.

We found a significant correlation between 201Tl SPECT assessed progressive disease and reduced time to treatment discontinuation, and a higher accuracy of early identification of treatment failure by SPECT than by CT (Paper IV). In 13 cases treatment was discontinued before completion of chemotherapy course 10, in 5 out of 13 cases SPECT visualised tumour progression preceeded CT visible progressive disease. Functional assessment of tumour viability by 201Tl SPECT seems to add information to morphological screening by CT. Increase of viable tumour volume with high 201Tl-uptake can probably at an early stage identify patients for whom treatment is not beneficial. This is valuable information as treatment toxicity can be avoided for these patients. However, in several cases SPECT evaluation did not add any new information. In six cases 201Tl SPECT and CT showed progressive disease simultaneously, and in two cases 20lTl SPECT as well as CT was negative in spite of obvious clinical deterioration leading to treatment discontinuation.

47 Five patients with TUV < 10 ml at week-24 assessment had a significantly longer TTD and ST than patients with TUV > 10 ml (Paper IV). Degree of I39 I41 contrast enhancement on CT is shown to be a predictive factor ~ 5 and similarly, low 201Tl-uptake volume reflecting a small viable tumour burden is probably a positive prognostic factor of equal significance. In Paper V, we present a small number of astrocytoma patients that were followed during chemotherapy. Due to early clinical progression of disease in the majority of cases, a maximum follow-up, comprising 6 repeated imaging series during 18 months, was obtained only in one case. The strength of our study is the prospective follow-up design with repeated scanning at fixed intervals. To our knowledge, data have not been published previously from a comparative longitudinal study with repeated 201Tl SPECT, MRI and 'H-MRS during astrocytoma chemotherapy.

201Tl SPECT has been reported to differentiate focal radiation necrosis from 75;ll0 201 tumour relapse ; and Tl SPECT evaluation has been proposed for identification of inflammatory reactions imitating tumour progression. Our results do not support the hypothesis of SPECT superiority over MRI in these aspects (Paper V). During the period of active chemotherapy and clinically stable disease, concordance was seen between MRI tumour characteristics indicating disease progression and quantitative SPECT measurements. Increase in tumour size and enhancement was in every case associated with increase of 201Tl-uptake. If the progression of disease as identified by MRI was a false finding, then both SPECT and MRI were as misleading in this situation. On the other hand, SPECT results were not always in agreement with MRI findings. At the last scanning before treatment discontinuation and clinical progression, MRI failed to identify progression of disease in one case, 201Tl SPECT in another. Complementary information is provided by the different methods in selected cases.

48 Isolated oedema regression on MRJ was not interpreted as regression of disease by the radiologist, which is in accordance with clinical practice. Steroids, used as anti-oedema therapy in astrocytoma treatment, are well known to influence MRI findings and thus be responsible for apparent response S1. Conflicting data have been published concerning the effect of steroids on cerebral 201Tl-uptake l09;I12. Our study showed that isolated oedema reduction seen on MRI was associated with decrease of 20lTl-uptake and increase of NAA/Cho ratio (Paper V). These findings are interesting, they indicate that the effects of steroids have to be taken into account when 201Tl SPECT analysis or 'H-MRS is used for treatment evaluation.

Metabolic characterization by 'H-MR spectroscopy was only accomplished in 7 out of 18 assessments (Paper V). Difficulties were encountered when assessing metabolic changes close to post-operative cavities and in the frontal region due to susceptibility artifacts. Further, changes in tumour size, mass effect and oedema in different follow-up series often made it impossible to define one or more voxels covering the same portion of the lesion in all treatment assessment measurements, and caused therefore difficulties in the usage of MR spectroscopy as a tool in therapy control.

The single volume spectroscopy technique was used in previous astrocytoma treatment studies 130'132. According to our experience, using a multivoxel technigue, the main limitation of MR spectroscopy in treatment monitoring was, that it only allowed measurements of a restricted area (maximum of 10x10 cm in the axial plane). The anatomic and histologic heterogeneity of astrocytomas i24;i42;i43^ an(j tke fac^ mat the location of tumour recurrence or progression can not be predicted, add to the difficulties. A multi-slice technique, and fat saturation pulses, would permit evaluation of larger parts of the tumour, covering also cortically situated lesions.

49 Two indications, where 'H-MRS showed potential in answering specific questions regarding treatment control, were seen. In one case 201Tl SPECT and MRI interpreted a new hemorrhage during the active treatment phase as disease progression (follow-up 3 after 5 courses of PCV). The patient was stable in performance and received maximum treatment with a total of 10 cycles of PCV without signs of neurological decline during long term follow-up (Figure 9). In contrast to MRI and SPECT, ' H-MRS showed status quo as the metabolic spectroscopic parameters in viable tissue adjacent to the hemorrhage were unchanged. In the other case early signs of tumour progression were seen with 'H-MRS 16 weeks before progression of disease was evident with 20lTl SPECT and MRI (Figure 8). 'H-MRS voxels were in both cases placed close to the lesion border. Spectral analyses from voxels adjacent to the border are likely to show disease progression as astrocytomas grow with diffuse infiltration of surrounding tissue l29. However, according to Waldrop et a/.130 similar remote spectral pattern changes can also reflect treatment related biological changes. Our observations must be interpreted with caution and further studies are needed to explore the usefulness of' H-MRS for astrocytoma treatment follow-up.

50 Conclusions

From the findings in the studies on which this thesis is based, the following conclusions can be drawn:

• Damage to the blood-brain barrier is a precondition for 'Tl-uptake in astrocytomas.

• High 20lTl-uptake in enhancing lesions is an indicator of highly malignant astrocytoma. However, the differentiation between the high-grade astrocytomas and low-grade astrocytomas or abscesses by ZOIT1 SPECT is only partial with an overlap of 20iTl-indexes between these groups.

• Astrocytomas with central or multiple necrotic areas and moderate ring enhancement on CT tend to be underestimated when evaluated by means of 20ITl SPECT by 201Tl-index calculation.

• Overlapping 20lTl-indexes were found between the histologic groups among the highly malignant astrocytomas. 20ITl SPECT is not a reliable method for tumour staging.

• Quantification of 20ITl-uptake by 201Tl SPECT tumour volume estimation is a new method for astrocytoma chemotherapy monitoring. Results suggest, that the method provides prognostic information and information of treatment

51 efficacy at an earlier stage than CT. However, we did not find a higher accuracy of this method than of MRI for astrocytoma chemotherapy efficacy monitoring. Future studies are needed to define the potential of quantitative 201TI SPECT for astrocytoma treatment evaluation purposes.

'H-MRS, with measurement from a restricted area, was difficult to apply for astrocytoma treatment monitoring due to the anatomical and histo- pathological heterogeneity of astrocytomas. Improvements regarding size of measurement area and fat suppression pulses are needed.

52 g 3. CTand 201Thallium SPECT scans in a 41-year old man (study HI) with a mixed solid and cystic astrocytoma grade IV in the right temporo-parietai region. Tumour ROls and reference ROls are indicated on SPECT slices by white border squares. Figure 6. Baseline CT and repeated SPECT examinations during chemotherapy: patient A had a homogeneous contrast enhancing astwcytoma grade IV; patient B a cystic ring-enhancing astrocytoma grade III-IV. Comparison between monitoring of treatment efficacy by 1) thallium index from average uptake calculation (average -0ITI); 2) thallium index from maximum uptake calculation (max ~01TI); and by 3) tumour uptake volume calculation (TUV). In contrast to thallium indexes, TUV results were in both patients concordant with visual analysis of thallium uptake, as well as with the clinical course of the patients. >d

1«" s '- s-

' *:«

^ * **<• Xr L \ -

Figure 7. Repeated CT and20 hi luring chemotherapy (study IV) in a patient with a recurrent low-grade astrocytoma, malignantly transformed to an astrocytoma grade IV Arrows indicate when disease progression (> 25% increase) was seen by: 1) tumour uptake volume calculation (TUV); 2) CT volume calculation. Treatment was discontinued after follow-up 4 due to clinical progression of disease. Figure 8. Repeated SPECT, MRI and MR spectroscopy examinations during chemotherapy (study V), in a 33 year-old man with an astrocytoma grade IV, located in the right occipital lobe. iH-MRS showed progression at first follow-up (NAA/Cho ratio decrease from > 1 to < I), the voxels were placed in the posterior part of the lesion at 2-3 cm distance from the resection cave. MRI and SPECT showed progression at final evaluation before clinical deterioration (increase of MRI contrast enhancement and 201thallium TUV > 25%). CO

Figure 9. Repeated SPECT and MRI examinations monitoring chemotherapy (study V) in a 36 year-old woman with an anaplastic astrocytoma in the right frontal lobe. SPECT showed a decrease of201 thallium uptake at follow-up I, 2, 4 and 5, and the patient was stable in performance evaluation during the 18 months of treatment. SPECT and MRI showed progression at follow-up three due to a new haemorrhage. 'H-MRS results showed stable disease, the voxel covered tissue adjacent to the lesion. Acknowledgements

I wish to thank everyone who has helped and encouraged me in completing my thesis, and especially:

• Professor Ingmar Rosen, Div. of Clinical Neurophysiology, Dept. of Clinical Neuroscience, Lund, my supervisor, for friendly and constructive support and scientific advise, for fund raising and providing time for research, for excellent and prompt feed-back during the process of article writing; • Dr. Hans Sjoholm, Dept. of Neurophysiology in Ulleval Hospital, Oslo, and my co-supervisor Professor Leif G Salford, Div. of Neurosurgery, Dept. of Clinical Neuroscience, Lund, for initiating this project during the early nineties, admitting patients to SPECT investigation, providing the material for the pre-operative studies in papers I and II; • Associate Professor Bengt Nilsson, my clinical supervisor and colleague in the Dept. of Neurology, for his encouragement and constructive support during the initiation phase of this project, for his enthusiasm, and for always being available for discussions when difficulties of practical or scientific nature were encountered; • Associate Professor Erik Ryding, head of the SPECT department, and Dr. Isabella Burtscher, expert on the MRS method, both co-authors, for their kind helpfulness, and for their technical and scientific advice; • Technician and co-author Ann-Margret Andersson, Dept. of Neurophysiology, for sharing her knowledge so unselfishly, and for dedicating her time to make such a significant contribution to this thesis; • Colleagues and staff members of the Dept. of Neurology, Dept. of Neurosurgery, Dept. of Neurophysiology, and Dept. of Neuroradiology, for their care and concern for the patients, with special thanks to the nurses

58 Susanne Fridolfsson, Ingrid Lindqvist, and Kerstin Sanfridsson, who helped me out with practical issues during the prolonged phase of data collection in the post-operative treatment studies; • Professor Arne Brun, Professor Stig Holtås, Associate Professor Per Malmquist, Dr. Bo Geijer, and Dr. Marianne Heiling, co-authors and contributors to this thesis; • Magnus Lindgren, and Roland Perfekt, for statistical advice; • My family Torben, Anna, Josef, and Niels Brattström, for being themselves whenever and whatever happens, for their love and understanding during the decades that we have now known one another.

The project was supported by grants from the Swedish Cancer Fund (#0233), the Swedish Medical Research Council (#084), the Medical Faculty of the Lund University, the Foundations and Donations for reasearch Lund Health Care District, the Rut and Erik Hardebo Donation Fund, the Elsa Schmitz Foundation, and the Segerfalk Foundation. The SPECT camera was financed by a grant from Knut and Alice Wallenberg Foundation.

59 Sammanfattning på svenska (Summary in Swedish)

Introduktion Malignt astrocytom är den vanligaste formen av primär hjärntumör. Vid klinisk misstanke om hjärntumör undersöks patienten med skiktröntgen av hjärnan, sk datortomografi (DT), eller med magnet resonans (MR), sk magnetkamera. Typiska DT och MR fynd vid hög-malignt astrocytom är en oregelbunden förändring, med en mörkare lågattenuerande kärna, samt ett omgivande kontrastuppladdande bräm. Kontrast-uppladdningen är ett resultat av en skadad blod-hjärn barriär till följd av utväxt av sjukligt förändrade tumörkärl. Just barriärskade markören är det radiologiska tecken som särskiljer misstänkt låg-malign från hög-malign tumör, men radiologiska kriterier för malignitetsgradering är inte 100 procentiga. Differentialdiagnostiskt föreligger också svårigheter att utifrån radiologiska fynd säkert skilja astrocytom från infektiösa förändringar, som till exempel abscesser.

För säkerställande av tumördiagnos krävs mikroskopisk eller histolo- patologisk undersökning av tumörvävnad. Lågmaligna astrocytom klassificeras som grad I-II, hög-maligna former som grad III-IV. Malignitetsgraden är avgörande för val av behandling. Hög-maligna astrocytom behandlas aggressivt med kirurgisk resektion, strålbehandling i maximalt tolererad dos, samt i vissa fall tilläggsbehandling med cytostatika. Vid centralt växande tumörer som ej är tillgängliga för resektion görs istället en provtagning, sk stereotaktisk biopsi. Endast ett fragment av tumören analyseras i dessa fall, och undantagsvis kan då även den histo-patologiska diagnosen vara missvisande.

På Universitetssjukhuset i Lund erbjuds patienter med hög-malignt astrocytom cytostatika behandling i enlighet med ett särskilt schema.

60 Behandlingscykler om fyra veckor återupprepas med två månaders intervall. Behandlingseffekten utvärderas med återkommande undersökningar, 6 gånger under det sammanlagda behandlings-förloppet på 18 månader. Det föreligger med i dag rutinmässigt använda metoder svårigheter att utvärdera behandlingseffekt. Post-operativa förändringar såväl som godartade strålrelaterade biverkningar kan imitera ett tumöråterfall. DT och MR visar dels aktiv tumörvävnad, dels döda eller nekrotiska områden och peritumoral svullnad. Det är inte möjligt att via DT/MR säkert avgränsa viabel tumörvävnad från omgivningen.

Experimentella arbeten har påvisat ett upptag av 20ithallium-joner (201Tl) i tumöromvandlade hjärn celler. Det selektiva upptaget i tumörvävnad har påvisats stå i proportion till grad av malignitet. Avhandlingens syfte är att utvärdera en klinisk metod som mäter hjärncellernas upptag och ansamling av 201Tl. Metoden mäter således cell-fonktion, till skillnad från konventionell DT och MR vilka avbildar hjärnans uppbyggnad eller struktur. I avhandlingen utvärderas 201Tl single photon computed tomography (2O1T1 SPECT) för pre- operativ diagnostik och malignitetsgradering, samt för särskiljande av tumörvävnad från hjärnvävnads reaktioner orsakade av behandling.

Pre-operativ diagnostik med hjälp av 2O1T1 SPECT I Delarbete I undersöks diagnostisk noggrannhet vid pre-operativ malignitetsgradering av astrocytom med hjälp av 20ITl SPECT. Trettiosju patienter med astrocytom är undersökta (31 hög-maligna, 6 låg-maligna). Pre- operativt 201Tl-upptags index i tumör relativt normal vävnad beräknas och korreleras till DT-fynd samt histo-patologisk diagnos. I Delarbete II studeras 201-pj spiers diagnostiska tillförlitlighet vid undersökning av 21 patienter med ringformigt kontrastuppladdande hjärn förändringar (17 astrocytom, 4 abscesser).

61 Slutsats pre-operativa studier; En skadad blod-hjärn-barriär är en förutsättning för intracerebralt inflöde av 20ITl-joner. Detta medför att 20!Tl SPECT är en metod lämpad för pre-operativ tumördiagnostik endast om lesionen på DT är kontrastuppladdande. Ett högt 20lTl-index i sådana fall talar för diagnosen hög-malignt astrocytom. Icke kontrast-uppladdande hög-maligna astrocytom, och tumörer med nekrotisk kärna, har ett lågt 201Tl-index i nivå motsvarande den som uppmäts hos låg- maligna tumörer. Även infektiösa förändringar har ett lågt 201Tl-index. Upptags index för de olika grupperna är delvis överlappande, den diagnostiska noggrannheten vid 20ITl SPECT motsvarar den vid DT undersökning. Metoden kan ej med tillfredsställande säkerhet malignitetsgradera tumörer inom den hög- maligna gruppen, dvs skilja mellan astrocytom grad III, III-IV och IV.

Post-operativa studier av behandlingseffekter med hjälp av 201Tl SPECT I Delarbete III beskrivs en ny metod för kvantifiering av post-operativt2 'Tl- upptag i tumörvävnad, samt validerar metoden i tidig post-operativ fas genom jämförelse med DT fynd. I Delarbete IV jämförs DT och kvantitativt 20lTl SPECT för tidig identifikation av tumörprogress (> 25% ökning av tumör volym) under cytostatika-behandling. Sexton patienter med histologiskt säkerställt hög-malignt astrocytom har följts sekventiellt med 201Tl SPECT och DT. Patienterna har följts tills cytostatika behandling avslutats, antingen p.g.a. fullföljd maximal behandling, eller avbrutits på grund av kliniska och radiologiska tecken talande för tumör progress. I Delarbete V har behandlingsstudien kompletterats med sekventiell MR för 6 patienter.

Slutsats post-operativa studier; Kvantitativt 201Tl SPECT ger ett mått på volymen metabolt aktiv vävnad i tumörområdet utifrån det antal pixels (en pixel=0,013 cm3) som uppvisar ett 201Tl-upptag överstigande ett definierat tröskelvärde. Detta mått på metabolt

62 aktiv vävnad är användbart för värdering av cytostatika behandlingseffekt. Med hjälp av kvantitativt 201Tl SPECT kan behandlingssvikt identifieras i tidigt skede, och med en bättre känslighet än med DT. Liten påvisbar volym metabolt aktiv vävnad är ett prognostiskt gynnsamt fynd. Vi har dock ej kunnat påvisa en bättre noggrannhet för 20lTl SPECT än för MR vid behandlingsuppföljning. De två metoderna visar likvärdig förmåga att särskilja behandlingsorsakade hjärnvävnads-reaktioner, som radiologiskt imiterar behandlingssvikt, från sann tumör-progress. I vissa fall kan kunskap från de båda undersökningsmetoderna komplettera varandra.

63 Reference List

1. Lesser GJ, Grossman SA. The chemotherapy of adult primary brain tumors. Cancer Treat.Rev. 1993;19:261-81.

2. Ammirati M, Vick N, Liao YL, Ciric I, Mikhael M. Effect of the extent of surgical resection on survival and quality of life in patients with supratentorial glioblastomas and anaplastic astrocytomas. Neurosurgery 1987;21:201-6.

3. Choksey MS, Valentine A, Shawdon H, Freer CE, Lindsay KW. Computed tomography in the diagnosis of malignant brain tumours: do all patients require biopsy? J.Neurol.Neurosurg.Psychiatry 1989;52:821-5.

4. Dean BL, Drayer BP, Bird CR, Florn RA, Hodak JA, Coons SW, Carey RG. Gliomas: classification with MR imaging. Radiology 1990;174:411-5.

5. Zimmerman RA. Imaging of adult central nervous system primary malignant gliomas. Staging and follow-up. Cancer 1991;67:1278-83.

6. Berger MS. Malignant astrocytomas: surgical aspects. Semin.Oncol. 1994;21:172-85.

7. Kernohan J.W., Mabon R.F., Svein H.J., Adson A.W. A simplified classification of gliomas. Proc Staff Meet Mayo Clinic 1949;24:71-5.

8. Kleihues P, Burger PC, Scheithauer BW. The new WHO classification of brain tumours. Brain Pathology 1993;3:255-68.

9. Scerrati M, Roselli R, Iacoangeli M, Pompucci A, Rossi GF. Prognostic factors in low grade (WHO grade II) gliomas of the cerebral hemispheres: the role of surgery. J.Neurol.Neurosurg.Psychiauy 1996;61:291-6.

10. Shinoda J, Sakai N, Nakatani K, Funakoshi T. Prognostic factors in supratentorial WHO grade II astrocytoma in adults. Br.J.Neurosurg. 1998;12:318-24.

11. Berger MS, Deliganis AV, Dobbins J, Keles GE. The effect of extent of resection on recurrence in patients with low grade cerebral hemisphere gliomas [see comments]. CaMce/-1994 ;74:1784-91.

12. Kiebert GM, Curran D, Aaronson NK, Bolla M, Menten J, Rutten EH, Nordman E, Silvestre ME, Pierart M, Karim AB. Quality of life after radiation therapy of cerebral low-grade gliomas of the adult: results of a randomised phase III trial on dose response (EORTC trial 22844). EORTC Radiotherapy Co-operative Group. Eur.J.Cancer 1998:34:1902-9.

13. Shapiro WR. Current therapy for brain tumors: back to the future. Arch.Neurol. 1999;56:429-32.

14. Chandler KL, Prados MD, Malec M, Wilson CB. Long-term survival in patients with glioblastoma multiforme. Neurosurgery 1993;32:716-20.

64 15. Grossman SA, Norris LK. Adjuvant and neoadjuvant treatment for primary brain tumors in adults. Semin.Oncol. 1995;22:530-9.

16. Curran WJ, Jr., Scott CB, Horton J, Nelson JS, Weinstein AS, Fischbach AJ, Chang CH, Rotman M, Asbell SO, Krisch RE. Recursive partitioning analysis of prognostic factors in three Radiation Therapy Oncology Group malignant glioma trials [see comments]. J.Natl.Cancer Inst. 1993;85:704-10.

17. Perry JR, DeAngelis LM, Schold SC, Jr., Burger PC, Brem H, Brown MT, Curran WJ, Scott CB, Prados MD, Kaplan R, Cairncross JG. Challenges in the design and conduct of phase III brain tumor therapy trials. Neurology 1997;49:912-7.

18. Grant R, Liang BC, Page MA, Crane DL, Greenberg HS, Junck L. Age influences chemotherapy response in astrocytomas. Neurology 1995;45:929-33.

19. Meckling S, Dold O, Forsyth PA, Brasher P, Hagen NA. Malignant supratentorial glioma in the elderly: is radiotherapy useful? Neurology 1996;47:901-5.

20. Whittle IR. Management of primary malignant brain tumours [editorial]. J.Neurol. Neurosurg. Psychiatry 1996;60:2-5.

21. Jeremic B, Grulicic D, Samardzic M, Antunovic V, Joksimovic M, Djuric L, Milicic B, Nikolic N. [The effect of extent of tumor resection on the outcome of combined therapy in patients with glioblastoma multiforme]. Srp.Arh.Celok.Lek. 1997;125:93-8.

22. Nitta T, Sato K. Prognostic implications of the extent of surgical resection in patients with intracranial malignant gliomas. Cancer 1995;75:2727-31.

23. DeAngelis LM. Brain tumor therapy: new horizons, new hope [editorial; comment]. Neurology 1998;50:1209-10.

24. Shrieve DC, Loeffler JS. Advances in radiation therapy for brain tumors. Neurol.Clin. 1995;13:773-93.

25. Madison MT, Hall WA, Latchaw RE, Loes DJ. Radiologic diagnosis, staging, and follow-up of adult central nervous system primary malignant glioma. Radio!. Clin.North Am. 1994;32:183-96.

26. Dropcho EJ. Central nervous system injury by therapeutic irradiation. Neurol.Clin. 1991;9:969-88.

27. Sheline GE, Wara WM, Smith V. Therapeutic irradiation and brain injury. Int.J. Radial. Oncol.Biol.Phys. 1980;6:1215-28.

28. Leibel SA, Sheline GE. Radiation therapy for neoplasms of the brain. J.Neurosurg. 1987;66:l-22.

29. Lampert P. Delayed effects of radiation on the human central nervous system: Early and late delayed reactions. Neurology 1964;14:912-4.

30. Lampert P. Dissiminated demyelination of the brain following cobolt gamma irradiation. Arch Pathol (AMA) 19590;68:322-32.

65 31. Martins AN, Johnston JS, Henry JM, Stoffel TJ, Di Chiro G. Delayed radiation necrosis of the brain. J.Neurosurg. 1977;47:336-45.

32. Rottenberg DA, Chernik NL, Deck MD, Ellis F, Posner JB. Cerebral necrosis following radiotherapy of extracranial neoplasms. Ann.Neurol. 1977;l:339-57.

33. Silber JH, Radcliffe J, Peckham V, Perilongo G, Kishnani P, Fridman M, Goldwein JW, Meadows AT. Whole-brain irradiation and decline in intelligence: the influence of dose and age on IQ score [see comments]. J.Clin.Oncoi. 1992;10:1390-6.

34. Fine HA, Dear KB, Loeffler JS, Black PM, Canellos GP. Meta-analysis of radiation therapy with and without adjuvant chemotherapy for malignant gliomas in adults. Cancer 1993;71:2585-97.

35. Thomas DGT, Bleehen NM, Roberts JT, Abram P, Brada M, Lantos PL, Moss TH, Ironside JW, Whaley JB, and Stenning SP. MRC randomised trial of adjuvant chemotherapy in high grade glioma - BR05. Journal of Neuro-oncology 39(2), 102- 102. 9-1-1998. Ref Type: Abstract

36. Glass J, Hochberg FH, Gruber ML, Louis DN, Smith D, Rattner B. The treatment of oligodendrogliomas and mixed oligodendroglioma- astrocytomas with PCV chemotherapy. J.Neurosurg. 1992;76:741-5.

37. Perry JR, Louis DN, Cairncross JG. Current treatment of oligodendrogliomas. Arch.Neurol. 1999;56:434-6.

38. Caimcross G, Macdonald D, Ludwin S, Lee D, Cascino T, Buckner J, Fulton D, Dropcho E, Stewart D, Schold C, Jr. Chemotherapy for anaplastic oligodendroglioma. National Cancer Institute of Canada Clinical Trials Group. J.Clin.Oncoi. 1994;12:2013-21.

39. Cairncross JG, Macdonald DR. Successful chemotherapy for recurrent malignant oligodendroglioma. Ann.Neurol. 1988;23:360-4.

40. Sandberg-Wollheim M, Malmstrom P, Stromblad LG, Anderson H, Borgstrom S, Brun A, Cronqvist S, Hougaard K, Salford LG. A randomized study of chemotherapy with procarbazine, vincristine, and lomustine with and without radiation therapy for astrocytoma grades 3 and/or 4. Cancer 1991;68:22-9.

41. Karnofsky DA. Problems and pitfalls in the evaluation of anticancer drugs. Cancer 1965;18:1517-28.

42. DeAngelis LM, Burger PC, Green SB, Cairncross JG. Malignant glioma: who benefits from adjuvant chemotherapy? Ann.Neurol. 1998;44:691-5.

43. Rosenberg RN. Neuro-oncology 1999 [editorial]. Arch.Neurol. 1999;56:397-8.

44. Fine HA. Novel biologic therapies for malignant gliomas. Antiangiogenesis, immunotherapy, and gene therapy. Neurol. Clin. 1995;13:827-46.

45. Levin VA. Neuro-oncology: an overview. Arch.Neurol. 1999;56:401-4.

66 46. Kaplan RS. Complexities, pitfalls, and strategies for evaluating brain tumor therapies. Curr.Opin.Oncol. 1998;10:175-8.

47. Perry JR, DeAngelis LM, Schold SC, Jr., Burger PC, Brem H, Brown MT, Curran WJ, Scott CB, Prados MD, Kaplan R, Cairncross JG. Challenges in the design and conduct of phase III brain tumor therapy trials. Neurology 1997;49:912-7.

48. Grant R, Liang BC, Slattery J, Greenberg HS, Junck L. Chemotherapy response criteria in malignant glioma. Neurology 1997;48:1336-40.

49. Grant R, Slattery J, Gregor A, Whittle IR. Recording neurological impairment in clinical trials of glioma. J.Neurooncol. 1994;19:37-49.

50. Kaplan WD, Takvorian T, Morris JH, Rumbaugh CL, Connolly BT, Atkins HL. Thallium-201 brain tumor imaging: a comparative study with pathologic correlation. J.Nucl.Med. 1987;28:47-52.

51. Byrne TN. Imaging of gliomas. Sewn. Oncol. 1994;21:162-71.

52. Gruber ML, Hochberg FH. Systematic evaluation of primary brain tumors. J.Nucl.Med 1990;31:969-71.

53. Chamberlain MC, Murovic JA, Levin VA. Absence of contrast enhancement on CT brain scans of patients with supratentorial malignant gliomas. Neurology 1988;38:1371-4.

54. Kondziolka D, Lunsford LD, Martinez AJ. Unreliability of contemporary neurodiagnostic imaging in evaluating suspected adult supratentorial (low-grade) astrocytoma [see comments]. J.Neurosurg. 1993;79:533-6.

55. Chisholm RA, Stenning S, Hawkins TD. The accuracy of volumetric measurement of high-grade gliomas. Clin.Radiol. 1989;40:17-21.

56. Cairncross JG, Pexman JH, Rathbone MP, DelMaestro RF. Postoperative contrast enhancement in patients with brain tumor. Ann.Neurol. 1985;17:570-2.

57. Curran WJ, Hecht-Leavitt C, Schut L, Zimmerman RA, Nelson DF. Magnetic resonance imaging of cranial radiation lesions. Int.J.Radiat.Oncol.Biol.Phys. 1987;13:1093-8.

58. Dooms GC, Hecht S, Brant-Zawadzki M, Berthiaume Y, Norman D, Newton TH. Brain radiation lesions: MR imaging. Radiology 1986;158:149-55.

59. Wallner KE, Galicich JH, Krol G, Arbit E, Malkin MG. Patterns of failure following treatment for glioblastoma multiforme and anaplastic astrocytoma. Int.J.Radiat.Oncol.Biol.Phys. 1989;16:1405-9.

60. Atkins HL, Budinger TF, Lebowitz E, Ansari AN, Greene MW, Fairchild RG, Ellis KJ. Thallium-201 for medical use. Part 3: Human distribution and physical imaging properties. J.Nucl.Med. 1977;18:133-40.

67 61. Abdel-Dayem H. 201T1-chloride uptake ratios in differentiating benign from malignant lesions: recommendations for ratio calculations and interpretation. Nucl.Med.Commun. 1995;16:145-9.

62. Staffen W, Hondl N, Trinka E, Iglseder B, Unterrainer J, Ladurner G. Clinical relevance of 201Tl-chloride SPET in the differential diagnosis of brain tumours. Nucl MedCommun. 1998;19:335-40.

63. Zhang JJ, Kim SM, Andrews D, Intenzo CM. Comparative assessment of four different Tl-201 tumor uptake indices in the evaluation of brain tumors. Clin.Nucl Med 1998;23:691-4.

64. Schwartz RB, Holman BL, Polak JF, Garada BM, Schwartz MS, Folkerth R, Carvalho PA, Loeffler JS, Shrieve DC, Black PM, Alexander E, III. Dual-isotope single-photon emission computerized tomography scanning in patients with glioblastoma multiforme: association with patient survival and histopathological characteristics of tumor after high-dose radiotherapy. J Neurosurg. 1998;89:60-8.

65. Tonami N. Thallium-201 SPECT in the evaluation of gliomas [editorial; comment]. J.Nucl.Med. 1993;34:2089-90.

66. Kim KT, Black KL, Marciano D, Mazziotta JC, Guze BH, Grafton S, Hawkins RA, Becker DP. Thallium-201 SPECT imaging of brain tumors: methods and results. J.Nucl.Med. 1990;31:965-9.

67. Sjoholm H, Elmqvist D, Rehncrona S, Rosen I, Salford LG. SPECT imaging of gliomas with Thallium-201 and Technetium-99m-HMPAO. Ada Neurol.Scand. 1995:91:66-70.

68. Ueda T, Kaji Y, Wakisaka S, Watanabe K, Hoshi H, Jinnouchi S, Futami S. Time sequential single photon emission computed tomography studies in brain tumour using thallium-201. Eur.J.Nud.Med. 1993:20:138-45.

69. Yoshii Y, Satou M, Yamamoto T, Yamada Y, Hyodo A, Nose T, Ishikawa H, Hatakeyama R. The role of thallium-201 single photon emission tomography in the investigation and characterisation of brain tumours in man and their response to treatment. Eur.J.Nucl.Med. 1993:20:39-45.

70. Biersack HJ, Grunwald F, Kropp J. Single photon emission computed tomography imaging of brain tumors. Semin.Nucl.Med. 1991 ;21:2-10.

71. Black KL, Hawkins RA, Kim KT, Becker DP, Lerner C, Marciano D. Use of thallium- 201 SPECT to quantitate malignancy grade of gliomas. J.Neurosurg. 1989;71:342-6.

72. Carvalho PA, Schwartz RB, Alexander E, III, Loeffler JS, Zimmerman RE, Nagel JS, Holman BL. Extracranial metastatic glioblastoma: appearance on thallium-201 - chloride/technetium-99rn-HMPAO SPECT images. J.Nucl.Med. 1991;32:322-4.

73. Kosuda S, Fujii H, Aoki S, Suzuki K, Tanaka Y, Nakamura O, Shidara N. Reassessment of quantitative thallium-201 brain SPECT for miscellaneous brain tumors. Ann.Nucl.Med. 1993;7:257-63.

68 74. Slizofski WJ, Krishna L, Katsetos CD, Black P, Miyamoto C, Brown SJ, Vender J, Chevres A, Khan AS, McEwan J. Thallium imaging for brain tumors with results measured by a semiquantitative index and correlated with histopathology. Cancer 1994;74:3190-7.

75. Schwartz RB, Carvalho PA, Alexander E, III, Loeffler JS, Folkerth R, Holman BL. Radiation necrosis vs high-grade recurrent glioma: differentiation by using dual- isotope SPECT with 201TI and 99mTc-HMPAO. AJNR Am.J.Neuroradiol. 1991;12:1187-92.

76. Rodrigues M, Fonseca AT, Salgado D, Vieira MR. 99Tcm-HMPAO brain SPECT in the evaluation of prognosis after surgical resection of astrocytoma. Comparison with other noninvasive imaging techniques (CT, MRI and 201T1 SPECT). Nucl.Med.Commun. 1993;14:1050-60.

77. Carvalho PA, Schwartz RB, Alexander E, III, Garada BM, Zimmerman RE, Loeffler JS, Holman BL. Detection of recurrent gliomas with quantitative thallium- 201/technetium-99m HMPAO single-photon emission computerized tomography. J.Neurosurg. 1992;77:565-70.

78. Oriuchi N, Tomiyoshi K, Inoue T, Ahmad K, Sarwar M, Tokunaga M, Suzuki H, Watanabe N, Hirano T, Horikoshi S, Shibasaki T, Tamura M, Endo K. Independent thallium-201 accumulation and fluorine-18- fluorodeoxyglucose metabolism in glioma. J.Nucl.Med. 1996;37:457-62.

79. Oriuchi N, Tamura M, Shibazaki T, Ohye C, Watanabe N, Tateno M, Tomiyoshi K, Hirano T, Inoue T, Endo K. Clinical evaluation of thallium-201 SPECT in supratentorial gliomas: relationship to histologic grade, prognosis and proliferative activities [see comments]. J.Nucl.Med. 1993;34:2085-9.

80. Zhang JJ, Park CH, Kim SM, Ayyangar KM, Haghbin M. Dual isotope SPECT in the evaluation of recurrent brain tumor. Clin.Nucl.Med. 1992;17:663-4.

81. Holman BL, Zimmerman RE, Johnson KA, Carvalho PA, Schwartz RB, Loeffler JS, Alexander E, Pelizzari CA, Chen GT. Computer-assisted superimposition of magnetic resonance and high- resolution technetium-99m-HMPAO and thallium-201 SPECT images of the brain. J.Nucl.Med. 1991;32:1478-84.

82. Moustafa HM, Omar WM, Ezzat I, Ziada GA, el Ghonimy EG. 201 Tl single photon emission tomography in the evaluation of residual and recurrent astrocytoma. Nucl.Med.Commun. 1994;15:140-3.

83. Black KL,-Emerick T, Hoh C, Hawkins RA, Mazziotta J, Becker DP. Thallium-201 SPECT and positron emission tomography equal predictors of glioma grade and recurrence. Neurol.Res. 1994;16:93-6.

84. Britten JS, Blank M. Thallium activation of the (Na+~K+)-activated ATPase of rabbit kidney. Biochim.Biophys.Acta 1968;159:160-6.

69 85. Ballinger JR, Sheldon KM, Boxen I, Erlichman C, Ling V. Differences between accumulation of 99mTc-MIBI and 201Tl-thallous chloride in tumour cells: role of P- glycoprotein. Q.J.Nucl.Med. 1995;39:122-8.

86. Brismar T, Anderson S, Collins VP. Mechanism of high K+ and T1+ uptake in cultured human glioma cells. Cell Mol.Neurobiol. 1995;15:351-60.

87. Sessler MJ, Geek P, Maul FD, Hor G, Munz DL. New aspects of cellular thallium uptake: Tl+-Na+-2Cl(-)-cotransport is the central mechanism of ion uptake. Nuklearmedizin 1986;25:24-7.

88. Mori K, Yamaguchi T, Maeda M. Mechanism of 201 thallium-chloride uptake in tumor cells and its relationship to potassium channels. Neurol.Res. 1998;20:19-22.

89. Arbab AS, Koizumi K, Toyama K, Araki T. Uptake of technetium-99m-tetrofosmin, technetium-99m-MIBI and thallium- 201 in tumor cell lines. J.Nucl.Med. 1996;37:1551-6.

90. Gehring PJ, Hammond PB. The interrelationship between thallium and potassium in animals. J.Pharmacol.Exp.Ther. 1967;155:187-201.

91. Hoshino T, Nagashima T, Cho KG, Davis RL, Donegan J, Slusarz M, Wilson CB. Variability in the proliferative potential of human gliomas. J.Neurooncol. 1989;7;137- 43.

92. Ancri D, Basset JY, Lonchampt MF, Etavard C. Diagnosis of cerebral lesions by Thallium 201. Radiology 1978;128:417-22.

93. Ancri D, Basset JY. Diagnosis of cerebral metastases by thallium 201. Br.J.Radiol. 1980;53:443-53.

94. Nadvi SS. Ebrahim FS, Corr P. The value of 201thallium-SPECT imaging in childhood brainstem gliomas. Pediulr.Radial. 1998;28:575-9.

95. Soricelli A, Cuocolo A, Varrone A, Discepolo A, Tedeschi E, Mainenti PP, Grivet- Fojaja MR, Covelli EM, Postiglione A, Salvatore M. Technetium-99m-tetrofosmin uptake in brain tumors by SPECT: comparison with thallium-201 imaging. J.Nucl.Med. 1998;39:802-6.

96. Mountz JM, Stafford-Schuck K, McKeever PE, Taren J, Beierwaltes WH. Thallium- 201 tumor/cardiac ratio estimation of residual astrocytoma. J.Neurosurg. 1988;68:705- 9.

97. Taguchi A. Clinical significance of thallium-201 single-photon emission computerized tomography (Tl-201 SPECT) in the evaluation of viability of gliomas. Kurume Med.J. 1992;39:267-78.

98. Rollins NK, Lowry PA, Shapiro KN. Comparison of gadolinium-enhanced MR and thallium-201 single photon emission computed tomography in pediatric brain tumors. Pediatr.Neurosurg. 1995;22:8-14.

70 99. Tamura M, Shibasaki T, Zama A, Kurihara H, Horikoshi S, Ono N, Oriuchi N, Hirano T. Assessment of malignancy of glioma by positron emission tomography with 18F- fluorodeoxyglucose and single photon emission computed tomography with thallium- 201 chloride. Neuroradiology 1998;40:210-5.

100. Ando A, Ando I, Katayama M, Sanada S, Hiraki T, Mori H, Tonami N, Hisada K. Biodistributions of 201T1 in tumor bearing animals and inflammatory lesion induced animals. Eur.J.Nucl.Med. 1987;12:567-72.

101. Ruiz A, Ganz WI, Post MJ, Camp A, Landy H, Mallin W, Sfakianakis GN. Use of thallium-201 brain SPECT to differentiate cerebral lymphoma from toxoplasma encephalitis in AIDS patients. AJNR Am.J.Neummdiol. 1994;15:1885-94.

102. Tonami N, Matsuda H, Ooba H, Yokoyama K, Hisada K, Ikeda K, Yamashita J. Thallium-201 accumulation in cerebral candidiasis: unexpected finding on SPECT. Clin.Nucl.Med 1990;15:397-400.

103. Aktolun C, Demirel D, Kir M, Bayhan H, Maden HA. Technetium-99m-MIBI and thallium-201 uptake in pulmonary actinomycosis [see comments]. J.Nucl.Med. 1991;32:1429-31.

104. Krishna L, Slizofski WJ, Katsetos CD, Nair S, Dadparvar S, Brown SJ, Chevres A, Roman R. Abnormal intracerebral thallium localization in a bacterial brain abscess. J.Nucl.Med. 1992;33:2017-9.

105. Gorniak RJ, Kramer EL, McMeeking A A, Zagzag D. Thallium-201 uptake in cytomegalovirus encephalitis. J.Nucl.Med. 1997;38:1386-8.

106. Ricci M, Pantano P, Pierallini A, Di Stefano D, Santoro A, Bozzao L, Lenzi GL. Relationship between thallium-201 uptake by supratentorial glioblastomas and their morphological characteristics on magnetic resonance imaging. Eur.J.Nucl.Med. 1996;23:524-9.

107. Kimura H, Takeno Y, Fukushima T. Variation in appearance of glioma on serial thallium-201 single photon emission computed tomography—case report. Neurol.Med.Chir (Tokyo) 1995;35:317-20.

108. Maria BL, Drane WB, Quisling RJ, Hoang KB. Correlation between gadolinium- diethylenetriaminepentaacetic acid contrast enhancement and thallium-201 chloride uptake in pediatric brainstem glioma. J.ChildNeurol. 1997;12:341-8.

109. Namba H, Togawa T, Yui N, Yanagisawa M, Kinoshita F, Iwadate Y, Ohsato K, Sueyoshi K. The effect of steroid on thallium-201 uptake by malignant gliomas. Eur.J.Nucl.Med. 1996;23:991-2. /

110. Sonoda Y, Kumabe T, Takahashi T, Shirane R, Yoshimoto T. Clinical usefulness of 11C-MET PET and 201T1 SPECT for differentiation of recurrent glioma from radiation necrosis. Neurol.MedChir (Tokyo) 1998;38:342-7.

71 111. Vertosick FT, Jr., Selker RG, Grossman SJ, Joyce JM. Correlation of thallium-201 single photon emission computed tomography and survival after treatment failure in patients with glioblastoma multiforme. Neurosurgery 1994;34:396-401.

112. Lorberboym M, Baram J, Feibel M, Hercbergs A, Lieberman L. A prospective evaluation of thallium-201 single photon emission computerized tomography for brain tumor burden. lnt.J.Radiat.Oncol.Biol.Phys. 1995;32:249-54.

113. Bokstein F, Cohen O, Gomori JM, Lossos A, and Siegal T. Comparison of conventional and functional neuroimaging features of anaplastic astrocytomas and anaplastic oligodendrogliomas. J.NeurooncoI. 39(2), 100-100. 1999. Ref Type: Abstract

114. Sasaki M, Kuwabara Y, Yoshida T, Nakagawa M, Fukumura T, Mihara F, Morioka T, Fukui M, Masuda K. A comparative study of thallium-201 SPET, carbon-11 methionine PET and fluorine-18 fluorodeoxyglucose PET for the differentiation of astrocytic tumours. Eur.JNud Med 1998;25:1261-9.

115. Bruhn H, Frahm J, Gyngell ML, Merboldt KD, Hanicke W, Sauter R, Hamburger C. Noninvasive differentiation of tumors with use of localized H-l MR spectroscopy in vivo: initial experience in patients with cerebral tumors [see comments]. Radiology 1989;172:541-8.

116. Bruhn H, Michaelis T, Merboldt KD, Hanicke W, Gyngell ML, Hamburger C, Frahm J. On the interpretation of proton NMR spectra from brain tumours in vivo and in vitro. NMR Biomed. 1992;5:253-8.

117. Castillo M, Kwock L. Proton MR spectroscopy of common brain tumors. Neuroimaging Clin.N.Am. 1998;8:733-52.

118. Go KG, Kamman RL, Mooyaart EL, Heesters MA, Pruim J, Vaalburg W, Paans AM. Localised proton spectroscopy and spectroscopic imaging in cerebral gliomas, with comparison to positron emission tomography. Neworadiology 1995;37:198-206.

119. Falini A, Calabrese G, Origgi D, Lipari S, Triulzi F, Losa M, Scotti G. Proton magnetic resonance spectroscopy and intracranial tumours: clinical perspectives, J.Neurol. 1996; 243:706-14.

120. Poptani H, Gupta RK, Jain VK, Roy R, Pandey R. Cystic intracranial mass lesions: possible role of in vivo MR spectroscopy in its differential diagnosis. Magn Reson. Imaging 1995;13:1019-29.

121. Poptani H, Gupta RK, Roy R, Pandey R, Jain VK, Chhabra DK. Characterization of intracranial mass lesions with in vivo proton MR spectroscopy. AJNR Am.J.Neuroradiol. 1995;16:1593-603.

122. Shimizu H, Kumabe T, Tominaga T, Kayama T, Hara K, Ono Y, Sato K, Arai N, Fujiwara S, Yoshimoto T. Noninvasive evaluation of malignancy of brain tumors with proton MR spectroscopy. AJNR Am.J.Neuroradiol. 1996;17:737-47.

72 123. Ott D, Hennig J, Ernst T. Human brain tumors: assessment with in vivo proton MR spectroscopy. Radiology 1993;186:745-52.

124. Preul MC, Leblanc R, Caramanos Z, Kasrai R, Narayanan S, Arnold DL. Magnetic resonance spectroscopy guided brain tumor resection: differentiation between recurrent glioma and radiation change in two diagnostically difficult cases. CanJ.Neurol.Sci. 1998;25:13-22.

125. Preul MC, Caramanos Z, Collins DL, Villemure JG, Leblanc R, Olivier A, Pokrupa R, Arnold DL. Accurate, noninvasive diagnosis of human brain tumors by using proton magnetic resonance spectroscopy. Nat.Med. 1996;2:323-5.

126. Cousins JP. Clinical MR spectroscopy: fundamentals, current applications, and future potential. AJR Am.J.Roentgenol. 1995;164:1337-47.

127. Demaerel P, Johannik K, Van Hecke P, Van Ongeval C, Verellen S, Marchal G, Wilms G, Plets C, Goffin J, Van Calenbergh F. Localized 1H NMR spectroscopy in fifty cases of newly diagnosed intracranial tumors. J.Comput.Assist.Tomogr. 1991;15:67-76.

128. Fulham MJ, Bizzi A, Dietz MJ, Shih HH, Raman R, Sobering GS, Frank JA, Dwyer AJ, Alger JR, Di Chiro G. Mapping of brain tumor metabolites with proton MR spectroscopic imaging: clinical relevance. Radiology 1992;185:675-86.

129. Burtscher I. Proton magnetic resonance spectroscopy in brain tumors. 1999. Lund university. 4-16-9990. Ref Type: Thesis/Dissertation

130. Waldrop SM, Davis PC, Padgett CA, Shapiro MB, Morris R. Treatment of brain tumors in children is associated with abnormal MR spectroscopic ratios in brain tissue remote from the tumor site [see comments]. AJNR Am.J.Neuroradiol. 1998;19:963-70.

131. Speck O, Thiel T, Hennig J. Grading and therapy monitoring of astrocytomas with lH-spectroscopy: preliminary study. Anticancer Res. 1996;16:1581-5.

132. Houkin K, Kamada K, Sawamura Y, Iwasaki Y, Abe H, Kashiwaba T. Proton magnetic resonance spectroscopy (1H-MRS) for the evaluation of treatment of brain tumours. Neuroradiology 1995;37:99-103.

133. Bloom M, Jacobs S, Pile-Spellman J, Pozniakoff A, Mabutas MI, Fawwaz RA, Van Heertum RL. Cerebral SPECT imaging: effect on clinical management. J.Nucl.Med. 1996;37:1070-3.

134. Mountz JM, Duetsch G, Kuzniecky R, Rosenfeld SS. Brain SPECT. In: Freeman LM, ed. Nuclear Medicine Annual. New York: Raven press, Ltd 1994: 48-9.

135. Roesdi MF, Postma TJ, Hoekstra OS, van Groeningen CJ, Wolbers JG, Heimans JJ. Thallium-201 SPECT as response parameter for PCV chemotherapy in recurrent glioma. J.Neurooncol. 1998;40:251-5.

73 136. Stenning SP, Freedman LS, Bleehen NM. An overview of published results from randomized studies of nitrosoureas in primary high grade malignant glioma. Br.J.Cancer 1987;56:89-90.

137. Kornblith PL, Walker M. Chemotherapy for malignant gliomas. J.Neurosurg. 1988;68:1-17.

138. Levin VA, Silver P, Hannigan J, Wara WM, Gutin PH, Davis RL, Wilson CB. Superiority of post-radiotherapy adjuvant chemotherapy with CCNU, procarbazine, and vincristine (PCV) over BCNU for anaplastic gliomas: NCOG 6G61 final report. Int.J.Radiat.Oncol.Biol.Phys. 1990;18:321-4.

139. Butler AR, Horii SC, Kricheff II, Shannon MB, Budzilovich GN. Computed tomography in astrocytomas. A statistical analysis of the parameters of malignancy and the positive contrast-enhanced CT scan. Radiology 1978;129:433-9.

140. Wood JR, Green SB, Shapiro WR. The prognostic importance of tumor size in malignant gliomas: a computed tomographic scan study by the Brain Tumor Cooperative Group. J.Clin.Oncol. 1988;6:338-43.

141. Murovic J, Turowski K, Wilson CB, Hoshino T, Levin V. Computerized tomography in the prognosis of malignant cerebral gliomas. J.Neurosurg. 1986;65:799-806.

142. Luyten PR, Marien AJ, Heindel W, van Gerwen PH, Herholz K, den Hollander JA, Friedmann G, Heiss WD. Metabolic imaging of patients with intracranial tumors: H-l MR spectroscopic imaging and PET. Radiology 1990;176:791-9.

143. Segebarth CM, Baleriaux DF, Luyten PR, den Hollander JA. Detection of metabolic heterogeneity of human intracranial tumors in vivo by 1H NMR spectroscopic imaging. Magn Reson.Med. 1990;13:62-76.