and Prostatic Diseases (1997) 1, 2±10 ß 1997 Stockton Press All rights reserved 1365±7852/97 $12.00 Review Pre-treatment staging of : recent advances and future prospects

RS Kirby Department of Urology, St George's Hospital, Blackshaw Road, London SW17 0QT

Keywords: prostate

Introduction in fact is present. Only locally extensive T3 tumours can be reliably staged by DRE. This ®nding prompts Accurate staging of prostate cancer is essential for the complementary tests such as ultrasound-guided tissue- correct selection of therapy, and, together with an assess- sampling to con®rm, for example, seminal vesicle in- ment of the Gleason score, also affords important prog- vasion. By contrast, DRE is of little value for patients nostic information. Unfortunately, in spite of many recent with stage T1 and particularly those patients whose advances, the precision with which an individual suffer- tumours were exclusively discovered by an elevated ing from prostate cancer can be staged pre-operatively is serum PSA (stage T1c), since, by de®nition, no DRE still sub-optimal. The underlying reason for this is that abnormality is present. extracapsular extension (ECE) of the tumour is generally microscopic rather than macroscopic. Recent progress has been made, however, and future technological advances, particularly in the ®eld of molecular staging, hold the Prostate speci®c antigen and staging promise of allowing clinicians to distinguish with greater accuracy those patients with truly localised disease, who It would be most convenient if the serum PSA value were are most likely to bene®t from radical therapy, from those capable of providing an accurate quantitative re¯ection of with more advanced disease, in whom systemic or endo- prostate tumour volume, and, by extension, an estimate crine therapy is more likely to be bene®cial. of tumour stage. But how well does serum PSA in fact correlate with the clinical stage? In one study of 553 patients presenting with various stages of prostate cancer, serum PSA was elevated in 447 cases (81%). The Digital rectal examination in the percentage of patients with elevated serum PSA increased staging of prostate cancer progressively by stage. Serum PSA exceeded 4.0 ng/ml in 63% of patients with stage A, 72% with stage B, 81% stage For many decades, digital rectal examination (DRE) has C; and 88% stage D. Hudson et al1 also found that serum been the cornerstone of the local staging of prostatic PSA increased progressively with clinical stage. However, cancer. DRE may estimate the local extension of disease, considerable overlap in clinical stages has been observed namely stage, but lacks sensitivity and speci®city to in these and other studies. Serum PSA is thus not suf®- determine either tumour volume or the presence or ciently reliable for precisely determining clinical stage on absence of extraprostatic extension of prostate cancer. In an individual patient basis. general, DRE underestimates both the extent of intracap- Data for pre-operative serum PSA and post-operative sular cancer and focal extracapsular . Therefore, pathological stage demonstrate similar imprecision.2 distinction between T2a, T2b and T2c based on DRE is Serum PSA certainly increases with advancing patholo- often unreliable. The same is true in separating T2 from gical stage.3 But, when correlating PSA and clinical stage, T3 tumours. In this regard, suspected in®ltration of the tremendous overlap is found in PSA values for all patho- prostate landmarks such as the lateral sulci and median logical stages. Higher pre-operative serum PSA is not furrow, as well as induration of the base of the seminal necessarily indicative of extracapsular extension, and vesicle as noted by a carefully performed examination lower values do not necessarily imply that a cancer is, may provide important staging information, but may in fact, con®ned to the prostate gland. Oesterling et al4 sometimes suggest extracapsular extension when none found false positive rates of PSA as a predictor of extra- capsular extension as 74% and 65% for a cut-off level of Received 9 June 1997; revised 16 July 1997; accepted 17 June 1997 4.0 ng/ml and 10.0 ng/ml, respectively. Pre-treatment staging of prostate cancer RS Kirby

Nevertheless, pre-operative serum PSA can provide To ensure that the cDNA fragment being ampli®ed was 3 some useful predictive information on pathological indeed a fragment of the PSA gene, the cDNAs recovered stage. In a review of more than 925 patients who under- from both human prostate cell line (LNCaP) preparations, went radical , Kleer et al5 observed 70% as well as from human prostate cancer specimens, were with PSA value < 2.0 ng/ml to have organ con®ned isolated and sequenced. In each instance, the cDNA was disease. At PSA values of 50 ng/ml, 82% had positive identical to the sequence of the PSA gene overlapping surgical margins or positive pelvic lymph nodes, or both. three of its six exons. The sensitivity of the assay devel- The predictive value of PSA has been shown to be oped was tested by examining serial dilutions of an enhanced by taking into account clinical stage, as well LNCaP tissue culture diluted with cultured B lympho- as the pathological features of a cancer in speci- cytes. The initial assay sensitivity allowed the detection of mens. Using a statistical model that combines clinical one prostate diluted in 106 white cells. One stage and tumour grade on biopsy with serum PSA, hundred and thirty-eight patients with clinically localised Partin et al6 plotted a probability curve for predicting prostate cancer, scheduled to undergo radical retropubic the degree of tumour extension at the time of . prostatectomy, were studied. A blood sample was taken Furthermore PSA values also give useful guidance as to from each patient pre-operatively. A comparison was the probability of the presence or absence of metas- made between the results of the RT-PCR assay and the tases. Oesterling et al4 investigated 2064 consecutive histology review. Of the 138 cases, 54 (39%) exhibited patients with prostate cancer, 853 of whom had newly extraprostatic extension of prostate cancer. Forty patients diagnosed, untreated and serum PSA < 20 ng/ml at pre- (29%) were judged to be potential surgical failures, as sentation. Out of 852 patients, 7 (0.8%) had a positive de®ned by a positive surgical margin or seminal vesicle bone scan and of these, 4 had a serum PSA > 10 ng/ml involvement. A total of 98 patients (71%) had specimen- and 5 had signi®cant skeletal symptoms. The 95% upper con®ned disease, including 84 (61%) who had organ- con®dence limit for the false-negative rate is thus approxi- con®ned cancer. The RT-PCR assay prior to radical pros- mately 1.5%. tatectomy was negative in 94 patients (68%) and positive The current position is then that the serum PSA value in 44 (32%). In patients with capsular perforation, the alone (and for that matter the ratio of free to total PSA) is assay was positive in 74% of cases, and in seminal vesicle inadequate for predicting clinical and pathological stages invasion, the assay was positive in 67% of cases. Overall, on an individual basis. However some prediction of of the 40 patients with pathological ®ndings suggestive of the ®nal pathological stage is possible by serum PSA, potential surgical failure, 70% had a positive pre-opera- especially taking also into consideration other factors, tive RT-PCR assay. such as clinical stage and the pathological features of The ability of the RT-PCR assay to predict potential cancer of biopsy specimens. surgical failures was compared with that of imaging, DRE, pre-operative serum PSA levels, as well as the Gleason score. The break points analysed for PSA and Molecular staging of prostate cancer Gleason score were standard: less than 10 ng/ml vs more than or equal to 7. In patients who experienced potential In 1993 Kary Mulis won the Nobel prize for the invention surgical failure (those with positive margins or seminal of the polymerase chain reaction (PCR) and it soon vesicle involvement), the RT-PCR assay had positive and became clear that there might be a role for this novel negative predictive values of 64 and 87%, respectively. technique in staging patients with clinically localised This compared favourably with similar values for ima- prostate cancer. The molecular technology of interest ging (23 and 72% respectively), DRE (9 and 73% respec- was the reverse transcriptase-polymerase chain reaction tively), and Gleason score (41 and 81% respectively). It (RT-PCR).7 RT-PCR can be used to identify cells expres- was concluded by this group that the blood assay using sing certain speci®c genes; in the case of prostate cancer RT-PCR for PSA was a potentially useful staging modality the most obvious gene of interest is PSA. The RT-PCR in this group of patients. technique involves preparing RNA sample and reverse Another possible target for the molecular staging of transcribing it into DNA. Following this, speci®c cDNAs prostate cancer is prostate speci®c membrane antigen can be ampli®ed by the use of primers that are speci®c for (PSMA). This is a glycoprotein of molecular weight the particular gene of interest. 100 000. Several groups have investigated RT-PCR detec- Deguchi et al8 detected mPSA encoding for PSA in the tion of PSMA in the bloodstream of prostate cancer lymph nodes from patients with prostate cancer. Their patients. Loric et al13 found that the detection of PSMA RT-PCR method could detect as little as 500±5000 copies was more often positive for PSMA than PSA in both of mRNA. Morento et al9 and Jaakkola et al10 suggested patients with localised and metastatic disease, as did that all patients with early prostate cancer were negative Isreali et al.9 By contrast, Cama12 achieved higher detec- for PSA mRNA by PCR, but 50±70% of patients with tion rates with PSA than PSMA. proven metastases were also negative. This they attribu- Several controversial issues remain unresolved. The ted to either methodological problems or loss of PSA ®rst involves the signi®cance of positive circulating RT- producing capacity of poorly differentiated prostate PCR signals and whether they truly represent malignant cancer cells. spread. Tumour biology predicts that most By contrast, in a study reported by Olsson,11,12 RT-PCR shed millions of cells into the circulation before the employed to evaluate the presence of messenger RNA primary tumour is clinically detectable. Furthermore, (mRNA) for PSA performed more favourably. The PSA the biological signi®cance of circulating tumour cells is primers were designed to isolate a cDNA fragment of the unclear. Intravasation of tumour cells into the circulation PSA gene containing 710 base pairs. is just one of a multistep process that ultimately results in Pre-treatment staging of prostate cancer RS Kirby

4 symptomatic metastatic disease. Such cells must not only bulging or irregularity of the boundary echo adjacent to a enter the circulation but also must survive long enough to peripheral zone hypoechoic lesion may be a feature of establish a metastatic colony and then to proliferate and ECE15,16 (Figure 1). Invasion of the may ¯ourish at that site. occasionally be apparent on TRUS.17 Extensive ECE is With the advent of RT-PCR technology, the detection of generally regarded as a contraindication for radical pros- circulating prostate-derived cells has new possibilities. tatectomy because of the very high incidence of positive However, whether such signals represent hitherto unde- margins and subsequent disease relapse. Sometimes, tected circulating benign cells or actively metastasizing however, neo-adjuvant hormonal downstaging or a more neoplastic elements remains to be elucidated. The pre- extensive procedure, including excision of the neuro- sence of circulating neoplastic cells does not necessarily vascular bundle adjacent to the ECE, might enable the equate with either clinically signi®cant metastatic disease surgeon to achieve complete tumour removal. More or a poor prognosis. Furthermore, patients undergoing usually radiotherapy is employed. Although TRUS in prostate needle biopsy, radical prostatectomy and cysto- isolation may sometimes determine the presence of scopy often have evidence of transient circulating signals ECE, it is now apparent that the careful consideration of that eventually clear from the circulation. DRE ®ndings and PSA values in conjunction with TRUS At present, no de®nitive conclusions on the utility of can achieve a higher speci®city in the identi®cation of RT-PCR molecular staging for prostate cancer can be extracapsular extension. The main contemporary value of reached. It is a powerful scienti®c technique that will, TRUS in this respect is to allow accurate guidance of the no doubt, help to resolve many currently unknown automated biopsy device to achieve pathological infor- aspects of prostate cancer growth and development. mation. In the future, however, the use of image enhan- Because positive RT-PCR results do often correlate with cers looks set to permit more accurate local staging of known metastatic disease, and since positive results are prostate cancer with this modality. rare in patients without prostate cancer, positive signals in a patient with pathologically localized prostate cancer could potentially indicate evidence of early metastatic TRUS-guided prostatic biopsy disease or portend early treatment failure. However, only with prolonged follow-up and larger study popula- Histological con®rmation of prostate cancer is an impor- tions will the true signi®cance of circulating prostate- tant prerequisite before initiating treatment for prostate tissue-derived cells be determined. Until that time, in cancer. A number of different biopsy techniques have the authors' opinion, RT-PCR molecular staging assays been used. Digital guidance can be used to direct should remain an investigational tool only and are cer- biopsy of palpable lesions via either the transrectal or tainly not suf®ciently re®ned to be employed in decision- transperineal routes. However, TRUS guidance has been making in the clinical management of men with prostate shown to be superior digital guidance in directing biopsy disease. of both palpable and impalpable prostate . Recently it has become clear that the information from TRUS guided are of considerable additional Transrectal ultrasound bene®t in staging prostate cancer. There has been some discussion regarding the optimum route for biopsy. The Transrectal ultrasound (TRUS) by itself is only of limited transperineal route carries the advantage that it is less value as a means of staging prostate cancer.14 However, likely to introduce infection, but the disadvantages of

(a) (b)

Figure 1 (a) and (b) Transrectal ultrasound images of prostate cancer seen in sagittal and longitudinal planes. Pre-treatment staging of prostate cancer RS Kirby requiring local anaesthesia and less easy guidance with tion by digesting the various protein barriers, for 5 the TRUS probe. In contrast, the transrectal route can be example connective tissue matrix and basement mem- utilised without the need for injection of local anaesthesia brane, that stand between them and the lumina of the and can be guided accurately through the biopsy port tumour vasculature. incorporated in the TRUS probe. It is now standard practice to administer a broad-spectrum antibiotic prior MVD satis®es many criteria for being a useful biomar- to biopsy and to give a subsequent course of oral anti- ker in prostate cancer. In combination with Gleason score biotics, usually of the quinolone variety, after the biopsy. and serum PSA, it signi®cantly increases the predictive This reduces infective complications to less than 2%. accuracy for pathological stage and potential outcome. Although the alternative technique of ®ne-needle aspira- Brawer et al19 found that the addition of MVD to a tion has been advocated by some, the cellular aspirate is Gleason score/PSA model increased its ability to identify sometimes dif®cult to interpret cytologically, resulting in extraprostatic extension correctly, from 53±85%. false negative ®ndings. A study comparing aspiration Most recently, optimised microvessel density (OMVD) cytology with core biopsy in 30 patients with prostate as a predictor of extraprostatic extension has been the cancer showed that while core biopsy made the correct subject of a multicentre study of 186 patients by Bostwick diagnosis in all cases there was a false positive rate of 30% et al,20 utilising computer-enhanced image analysis. The for aspiration cytology.18 At present therefore most pro- study concluded that the prediction of extraprostatic static biopsies are taken using an automatic biopsy extension was signi®cantly improved when OMVD was device. incorporated in a logistic regression model with the The number of positive biopsies, for example 4 out of 6, Gleason score from the diagnostic biopsy and the pre- their Gleason score and especially the presence or absence biopsy serum PSA value. For example, a patient with a of seminal vesicle invasion, as determined by biopsies Gleason score of 7 and a PSA value of 8 had a 53% guided to the base of these structures, are all useful probability of extraprostatic extension. Adding OMVD to predictors of the eventual pathological stage of prostate the model distinguished between probabilities of 37 and cancer. 89% for the same patient. With a given Gleason score and PSA value, the study indicated that the odds of extrapro- static extension increased 5.97 times with every unit change of log OMVD. Neovascularity as a staging tool A most promising new technique for predicting prostate cancer tumour extent relies on the quanti®cation of The value of CT scan and staging tumour angiogenesis or microvessel density (MVD). MVD is a quantitative measurement of the number of prostate cancer small blood vessels within a given area of tissue on a histological slide. It is well known that, in order to grow By contrast, the clinical utility of computed tomography larger than 1 mm2 in area, cancers must recruit new blood (CT) in the staging of prostate cancer has generally been vessels from the host, and this neovascularity can be disappointing.21 CT has proven to be of little or no value readily identi®ed by immunohistochemistry in formalin- in the assessment of the local disease. It is not capable of ®xed paraf®n-embedded tissue using an antibody to distinguishing between prostate cancer from benign con- human von Willebrand factor (factor VIII-related anti- ditions in the prostate and is unable to reliably identify gen). An association between a high degree of neovascu- extracapsular penetration or seminal vesicle invasion. It larization and lowered patient survival, has been would not appear to provide much additional informa- observed in studies of patients with cancers of the blad- tion to that obtained by digital rectal examination or der, brain, breast, cervix, endometrium, stomach, head transrectal ultrasound for the staging of local disease. and neck, lung, melanocytes, oral cavity, ovary, rectum, There has been interest in the use of CT to identify the testis and prostate. The following factors, probably acting presence of lymphadenopathy in the pelvis. However, in concert, are the likely reasons for this association: because of the current low incidence of positive lymph nodes, the utility of CT in this situation is almost negli- 1. High vascular density increases the area of vascular gible. Most nodal metastases are microscopic and do not surface, which may facilitate the escape of tumour cells enlarge or distort the lymph nodes. There probably are into the circulation. subsets of patients in whom risk for grossly positive 2. Newly formed blood vessels in tumours often are nodes might be higher. A PSA in excess of 25±35 ng/ml `leaky' which allows malignant cells to cross their or a poorly differentiated tumour (Gleason score > 7 on walls with relative ease. biopsy) may identify a risk for gross involvement and 3. An angiogenic cell leaving a primary tumour probably justify a CT scan. It is important to note that an enlarged is more likely than a non-angiogenic cell to develop is identi®ed on CT, a CT guided aspiration into a detectable secondary tumour because secondary biopsy is necessary to con®rm the presence of malig- tumour growth also depends on neovascularisation. nancy. 4. Solid tumours comprise both malignant cells and It has been estimated that as much as $75 million could stroma. Tumour MVD may be a measure of the success be saved annually in the United States alone by eliminat- of a tumour in forming essential stromal support. ing CT in patients with a PSA less than 10 ng/ml. An even 5. Enzymes secreted by cells at the tips of growing larger amount of money could be saved worldwide with capillaries may help tumour cells to enter the circula- an even more restrictive use. Pre-treatment staging of prostate cancer RS Kirby 6 Magnetic resonance imaging MRI scanning will prove to be more accurate than current techniques in the local staging of prostate malignancy, Magnetic resonance imaging (MRI) is a new and rapidly although cost constraints will always limit their avail- advancing, but expensive, technique that utilises a mag- ability for every patient. netic ®eld, rather than ionising radiation, to produce images of startling clarity. MRI scanners are capable of producing images in the transverse axial, coronal and sagittal planes, and endorectal probes are available for The role of radionuclide bone scans and enhanced de®nition. On T1-weighted images it is not possible to demon- correlative imaging studies in the strate the internal architecture of the prostate although a staging of prostate cancer clear demarcation is apparent between the prostate and seminal vesicles and the surrounding fat. By contrast T2- The axial skeleton is the favoured location for prostate weighted MRI images are able to demonstrate the internal cancer metastases (Figure 3). This `bone homing' mechan- architecture of the prostate. Typically the peripheral zone ism, a direct consequence of complex prostate epithelial- of the prostate returns high intensity signals whilst pros- bone stromal interactions, has made it routine clinical tate cancer produces relatively low intensity signals. practice to include a radionucleide bone scan and corre- However, most other abnormalities within the prostate lative imaging studies in routine prostate cancer staging. such as `', corpora amylacea and prostatitis also Recent literature, however, has questioned the utility of return low intensity signals so that unfortunately MRI is the clinical information gained from radionucleide ima- not especially accurate at identifying localised prostate ging and correlative radiographs for staging of prostate cancer. Disappointingly MRI has not proven very much cancer, especially when examined from the perspective of more accurate than TRUS in local staging of prostate cost-effectiveness. Typically, men with suspected prostate cancer.22 cancer undergo a digital rectal examination, serum pros- However, supplementation of transverse axial scans tate speci®c antigen (PSA) measurement, and transrectal with coronal and sagittal scanning has been shown to ultrasound with biopsy of the prostate. Routine skeletal enhance local staging accuracy from 61±83%. Extracap- imaging has been shown to be unnecessary in a signi®- sular extension is likely if there is apparent asymmetry, cant proportion of men with prostate cancer, with most irregularity or breaching of the peri-prostatic fat which, studies advocating the application of radionucleide unlike the prostate itself, sends low intensity signals. bone scan and correlative imaging to a more selective Seminal vesicle invasion is suggested if there are areas paradigm. of asymmetric low intensity within the normally high Radionucleide imaging was developed in 1963 and intensity vesicles. remains the gold standard for evaluation of patients A recent multicenter trial comparing MRI with TRUS with suspected skeletal metastases. This modality is in the staging of localised prostate cancer concluded that quite sensitive in terms of identifying metastases. Radio- there was no statistically signi®cant difference between nucleide imaging is far superior to clinical evaluation, the two. radiographs, or serum alkaline phosphatase, in the detec- There undoubtedly exists the prospect of more accurate tion of occult metastases. False negative scans are staging of localised prostate cancer. Ever improving reported in less than 1% of patients with con®rmed endorectal MRI surface coils are currently being evalu- ated.23,24 The images of the internal architecture of the prostate produced from these devices are of exceedingly high quality (Figure 2). It seems likely that endorectal

Figure 2 MRI of the prostate showing a peripherally located prostate cancer in®ltrating the neurovascular bundle. Figure 3 A positive bone scan due to metastatic prostate cancer. Pre-treatment staging of prostate cancer RS Kirby bone metastases, though false positive scans are more disease at presentation, as a means of staging and pre- 7 common. False positive studies, due to previous trauma, dicting response to therapy and overall survival. degenerative disease, or metabolic abnormalities, may at In summary, recent literature advocates a more selec- times be discerned on the basis of the appearance of the tive implementation of radionucleide bone scanning and lesion(s) on the scan. Radionucleide bone scans re¯ect correlative imaging in the diagnosis and treatment of osteoblastic activity in areas of cortical bone with prostate cancer. For patients with localised disease, the increased bone turnover. Similarly, plain radiographs serum PSA level should dictate whether a bone scan is re¯ect the calcium content of cortical bone which is performed in the preoperative evaluation unless there are altered at the site of clinical metastases. Neither radio- other risk factors such as bone pain. Patients with clinical nucleide imaging nor radiography is capable of imaging evidence of metastatic disease, such as an elevated alka- bone marrow, thought to be the initial site of prostate line or acid phosphatase level, or a signi®cantly elevated metastatic growth, which is then followed by cortical PSA level, should still undergo radionucleide bone scan- involvement later in the natural history of the lesion. ning as an initial imaging modality. If there are positive Plain radiographs are the least sensitive in evaluating ®ndings, or clinical suspicion in the face of negative the axial skeleton for metastases. Fifty percent of bone ®ndings, plain radiographs of the area of interest are mineral content must be altered by the metastatic process indicated. However, radiography may not offer much before a lesion can be detected radiographically. Radio- additional information with regard to the presence of nucleide bone scans can show lesions 3±6 months before metastatic disease until late in the natural history of the there is radiologic evidence of metastatic disease. Mag- lesion. If plain ®lms are equivocal or negative and clinical netic resonance imaging (MRI) has shown promise for suspicion is high, an MRI of the area may document the evaluation of the bone marrow, detecting occult skeletal presence or absence of a metastatic focus or foci. metastases missed by radionucleide imaging in certain areas of the axial skeleton. Still radionucleide imaging must be regarded as the ®rst-line modality when osseous Immunoscintigraphy in prostate cancer metastases are suspected, with plain radiographs or MRI serving a more corroboratory role. The possibility of demonstrating prostate cancer metas- Chybowski and associates25 were the ®rst to produce tases and the identi®cation of their anatomical distribu- evidence that radionucleide imaging of the skeleton tion and of suspected local recurrence after radical might be unnecessary in a signi®cant proportion of men prostatectomy by immunoscanning is of great potential presenting with clinically localised prostate cancer. In this importance, since this can not reliably be done by con- retrospective study, serum PSA level was a powerful ventional technetium scanning, CT, MR or US scanning. predictor for the presence of osseous . Fewer Immunoscintigraphy, however, has also its limitations. than 1% of patients with a PSA level less than 20 ng/ml Spatial details are limited, but may be improved by had a positive bone scan. No patient with a PSA level less single-photon emission CT (SPECT). Also the imaging than 15 ng/ml had a positive scan. This ®nding was of hepatic lesions with Indium isotopes is intrinsically con®rmed by Oesterling,26 who demonstrated a positive suboptimal because of normal intense liver uptake. A bone scan rate of 0.8% in patients with a PSA level of combination of functional and anatomical modalities may 20 ng/ml or less. No patient in the study with a PSA level be advantageous compared with either modality alone.28 of 10 ng/ml or less had a positive bone scan. These The advent of monoclonal antibody (MAB) technology studies looked at the contribution of tumour grade, has permitted the development of radiolabelled tumour clinical stage, and acid phosphatase level, to the positive reactive probes that can be used in conjunction with predictive value of the serum PSA level and demon- gamma camera imaging equipment. However, the use strated no signi®cant bene®t over PSA level alone in of MABs for tumour imaging is compromised by the fact predicting a positive radionucleide scan. The authors that tumour speci®c antigens do not currently exist and concluded that a radionucleide bone scan was unneces- all target antigens of practical relevance are also sary in men who present with prostate cancer and a PSA expressed to some degree on other malignant and non- level of 10 ng/ml or less, which nowadays represents malignant tissues. more than 40% of men with newly diagnosed prostate Radioisotopes may be complexed to MABs against cancer. Although it must be borne in mind that occasional either acid phosphatase (PAP) or prostatic speci®c antigen exceptions to this rule will be seen.27 (PSA). The use of MAB against PAP for the immunode- The role of radionucleide bone scans in patients with tection of metastases is hampered by its rapid removal known metastatic disease has also been studied. While from the blood and its subsequent sequestration by the not quantitative, the bone scan does assess the number liver. To circumvent this a number of studies have utilised and location of osseous metastases. Studies show that iodine 12329 or Indium labelled30 monoclonal anti-PAP radionucleide imaging provides prognostic information monoclonal antibody F(ab)2 fragment in studies using as well as predicts the duration of response to hormonal lymphatic administration (intraprostatic injection) and/or ablative therapy. Patients who have metastatic disease intravenous injection. limited to the pelvis and lumbar spine, with a limited Generally, radioisotopes complexed to MABs against number of metastatic foci, as determined by radionu- acid phosphatase and F(ab)2 fragment have been dis- cleide bone scan, demonstrate a signi®cantly better appointing due to lack of speci®city and sensitivity of response to hormonal therapy and have a longer mean these cytosol-directed MABs,31 even though preliminary survival than patients with more extensive disease at the studies originally looked promising.29,30 time of presentation. These studies support the need for Radiolabelled MABs raised against prostate cancer radionucleide bone scanning in patients with metastatic cells would have potential for immunoscanning using Pre-treatment staging of prostate cancer RS Kirby

8 gamma camera imaging. One such MAB is cytogen 356, a therefore lymph node staging has been performed by an murine immunoglobulin G monoclonal antibody to a open operation, either as a staging procedure prior to glycoprotein antigen, described by Horoszewicz32 that is radiotherapy or radical perineal prostatectomy, or as part localised on the membrane of benign and malignant of radical retropubic prostatectomy. There now exists, prostate epithelium that can be bound to an Indium however, the possibility of undertaking this procedure isotope. This antibody is commercially available. It has using minimally invasive techniques of laparoscopic been investigated in a number of preliminary studies, and technology.38,39 Laparoscopic pelvic lymphadenectomy so far it seems to be superior to conventional methods to (LPLND) is now a feasible option which allows accurate localise overt prostate cancer metastases and local extra- evaluation of the nodal status of patients with minimal capsular growth.33±35 The results of large phase III studies attendant morbidity. The post-operative hospital stay is that are underway are awaited with anticipation. only 24±48 h and most patients return to work within a Cytogen 351 MAB bound to technetium 99 m which week. has excellent imaging characteristics, also seems to have Clearly, pelvic lymph node dissection is not appropri- potential usefulness in demonstrating local primary, sec- ate for every patient, but should probably be con®ned to ondary or recurrent prostate cancer.36 individuals with either higher Gleason score (  7) pros- Another indirect immunoscintigraphy or `cold spot' tate cancer on biopsy or a PSA greater than 20±25 ng/ml immunoscintigraphy has also been used. It utilises 99 m who have no other evidence of metastatic disease and technetium labelled monoclonal murine anti-NCA-95 (a who are candidates, by virtue of both their tumour stage differentiation antigen of granulopoiesis, present on and their life expectancy, for curative therapy. Prior to mature granulocytes and their precursors in bone undertaking de®nitive local therapy, especially radical marrow) antibody (an IgG 1 antibody) that binds to prostatectomy, either by the perineal or retropubic NCA-95 (a non-speci®c cross reacting antigen of 95 kD). route, it could be argued that the pelvic lymph nodes The anti-NCA-95 MAB immunoscintigraphy provides a should be examined for the presence of metastatic cancer, homogeneous radionucleide image of the bone marrow, since if the lymph nodes are found to be positive, it is where metastases can be seen as cold spots.37 generally considered a sign of disseminated disease and therefore a contraindication to extirpative prostatic surgery. A further argument in favour of LPLND is that it Lymphangiography permits histological lymph node staging in individuals undergoing non-extirpative therapies such as de®nitive Pedal lymphangiography was once a popular technique external beam radiotherapy, especially those patients in for imaging pelvic lymph nodes. However, the technique studies where the results are being compared with those is invasive, uncomfortable for the patient, time-consum- of surgery. Lack of staging data in the past have often ing and carries the risk of morbidity from pulmonary skewed the results of trials and now embolism due to the lipid based contrast medium. With this new minimally invasive technique can facilitate more the advent of CT and MRI scanning the use of lymphan- complete pre-treatment staging, allowing more valid giography has virtually fallen into abeyance. Neverthe- comparisons between surgical and radiotherapeutic treat- less lymphangiography, by examining the internal ment options to be made. architecture of lymph nodes, does provide images that The advantages of LPLND lie in its accuracy in terms cannot be duplicated by other techniques, although of de®nitive con®rmation of the presence or absence of metastases must be at least 5 mm in diameter to be lymph node involvement and in the short hospital stay. detectable. Good images can be obtained of the external However, since only a small proportion of patients under- iliac, common iliac and para-aortic nodes, but doubt going radical prostatectomy are now found to have has been cast upon the ability of lymphangiography to positive lymph nodes, and this is nearly always in the adequately image the hypogastric and obturator nodes, patients with larger volume, higher stage tumours with which are precisely those which become involved earliest signi®cantly raised PSA values, it seems logical to restrict in the natural history of metastatic prostate cancer. The LPLND to that already dwindling category of case. This accuracy of lymphangiography has been tested in 40 will reduce the proportion of cases in whom the lymph patients who subsequently underwent open pelvic lym- nodes would subsequently be found, on histological phadenectomy as a staging manoeuvre for prostate. Lym- evaluation, to be negative. phangiography produced a false-positive rate of 59% and a false-negative rate of 36%. Most urologists accept that lymphangiography now has little role to play in contem- porary clinical practice in staging prostate cancer. Conclusions of the prostate is a highly prevalent and Laparoscopic pelvic lymphadenectomy insidious disease. Unless early detection strategies are employed most patients will present with cancer that As mentioned previously lymph nodes involved by pros- has already spread beyond the con®nes of the gland tate cancer are often indurated but not markedly and is therefore incurable. Early detection by a combina- enlarged, consequently non-invasive imaging has been tion of DRE and PSA testing has been shown to increase rather disappointing as a means of accurate evaluation of the yield of cancers con®ned to the prostate, but as yet the obturator and hypogastric lymph nodes. In the past there has been no con®rmation in randomised, controlled Pre-treatment staging of prostate cancer RS Kirby studies that this in itself reduces the disease-speci®c 3 Partin A, Carter H, Chan D. Prostate speci®c antigen in the 9 mortality. staging of localised prostate cancer: in¯uence of tumour differ- Accurate staging is crucial for the selection of the most entiation, tumour volume and benign . J Urol 1996; 143: 747±751. appropriate treatment modality. However, pre-operative 4 Oesterling JE, Martin SK, Bergstralh EJ, Lowe FC. The use of staging is, as yet, an inexact science. DRE and PSA Prostate-Speci®c Antigen in staging patients with newly diag- provide inexact indications of the local stage, but new nosed prostate cancer. JAMA 1993; 269: 57±60. techniques of molecular staging using PCR technology 5 Kleer E, Larson-Keller BS, Zincke H, Oesterling JE. Ability of hold the promise of a more precise method of detecting preoperative serum prostate-speci®c antigen value to predict extracapsular spread. TRUS may provide some indication pathologic stage and DNA ploidy. In¯uence of clinical stage of capsular or seminal vesicle involvement but TRUS- and tumour grade. Urology 1993; 41: 207±216. 6 Partin AW et al. The use of prostate speci®c antigen, clinical stage guided biopsies are necessary to provide histological and gleason score to predict pathological stage in men with con®rmation. The of the biopsies, the percentage localized prostate cancer. J Urol 1993; 150: 110±114. of the cores in®ltrated and the presence or absence 7 Diamandis EP, Yu H. Prostate cancer, prostate-speci®c antigen, and of microvascular or perineural invasion, as well as semi- the polymerase chain reaction. Clin Chem 1995; 41(2): 177±179. nal vesicle invasion, are all useful ancillary features, 8 Deguchi T et al. Detection of micrometastatic prostate cancer cells which when taken with the PSA value, provide an in lymph nodes by reverse transcriptase-polymerase chain reac- tion. Cancer Res 1993; 53: 5350±5354. estimate of the probability of extracapsular extension in 9 Moreno JG et al. Detection of hematogenous micrometastasis in the individual case. Quantitative measures of angiogen- patients with prostate cancer. Cancer Res 1992; 52: 6110±6112. esis by OMVD seem to add further valuable staging 10 Jaakkola S et al. Detection of prostatic cells in peripheral blood: information. correlation with serum concentrations of prostate-speci®c anti- CT scanning and MRI have been disappointing as gen. Clin Chem 1995; 41: 182±186. staging tools, although recent reports using endorectal 11 Katz AE, Olsson MD, Raffo A Je. Molecular staging of prostate cancer with the use of an enhanced reverse transcriptase-PCR MRI technology is more encouraging. Bone scanning is assay. Urology 1994; 43: 765±775. now only recommended in those with PSA values 12 Cama C et al. Molecular staging of prostate cancer. II. A compar- >10 ng/ml or those with high Gleason scores and/or ison of the application of an enhanced reverse transcriptase poly- skeletal symptoms since this investigation in patients merase chain reaction assay for prostate speci®c antigen and with lower values are almost always negative. prostate speci®c membrane antigen. J Urol 1995; 153: 1373±1378. Immunoscintigraphy is currently in its infancy but 13 Loric S et al. Enhanced detection of hematogenous circulating holds some promise. Laparoscopic lymph node sampling prostatic cells in patients with prostate adenocarcinoma by using nested reverse transcription polymerase chain reaction assay may be considered in those relatively few patients with based on prostate-speci®c membrane antigen. Clin Chem 1995; either poorly differentiated tumours and/or PSA values 41: 1698±1703. > 20 ng/ml who, by virtue of their life expectancy, are 14 Scardino PT, Shinohara K, Wheeler TM, Carter SSC. Staging of otherwise candidates for radical therapy. Prostatic CancerÐvalue of ultrasonography. Urol Clin NA 16 The challenge for the future is to further enhance 1989; 4: 713±734. the precision of prostate cancer staging. This will serve 15 Rifkin M, Sudakoff G, Alexander A. Prostate: techniques, results and potential applications of colour Doppler US scanning. Radi- two important purposes: ®rstly to reduce the number ology 1993; 186: 509±513. of patients undergoing procedures that in fact carry 16 Rifkin M, Choi H. Implications of small, peripheral hypoechoic little prospect of rendering them disease free, and lesions in endorectal US of the prostate. Radiology 1988; 166: 619± secondly to enhance the results of potentially curative 622. therapies by accurately targeting them at those most 17 Ohori M et al. Detection of microscopic extracapsular extension likely to bene®t. prior to radical prostatectomy for clinically localised prostate cancer. Br J Urol 1994; 74: 72±79. 18 Narayan P, Jajodia P, Stein R. Core biopsy instrument in the diagnosis of prostate cancer superior accuracy to ®ne needle aspiration. J Urol 1991; 145: 795±799. 19 Brawer MK, Bigler SA, Deering RE. Quantitative morphometric Acknowledgements analysis of the microcirculation in prostate carcinoma. J Cell Biochem 1992; 161: 62±64. The author would like to thank the members of commit- 20 Bostwick DG et al. Optimized microvessel density analysis tee number 6 of the International Consultation on prostate improves prediction of cancer stage from prostate needle biop- cancer for their assistance in clarifying the issues sur- sies. Urology 1996; 48: 47±57. rounding staging of prostate cancer. Committee members 21 Forman H et al. CT screening for comorbid disease in patients included: C Abbou, S Akdas, S Egawa, S Graham, with prostatic carcinoma. Is it cost-effective? AJR 1994; 162: 1125± 1128. G Grechi, R Hartung, V Marshall, J Montie, G Vallancien, 22 Hricak H et al. Prostatic carcinoma assessment by clinical assess- A Von Eschenbach, and H Wolf. ment, CT and MRI imaging. Radiology 1987; 162: 331±336. 23 Pollack HM, Schnall MD. Magnetic resonance imaging in carci- noma of the prostate. Prostate 1992; 4: 17±31. 24 Chelsky MJ, Schnall MD, Seidmon EJ, Pollack HM. Use of endorectal surface coil magnetic resonance imaging for local References staging of prostate cancer. J Urol 1993; 150: 391±395. 25 Chybowski FM et al. Predicting radionuclide bone scan ®ndings 1 Hudson MA, Bahnson RR, Catalona WJ. Clinical use of prostate in patients with newly diagnosed, untreated prostate cancer. speci®c antigen in patients with prostate cancer. J Urol 1989; 142: Prostate-speci®c antigen is superior to all other clinical para- 1011±1017. meters. J Urol 1991; 145: 313. 2 Lange PH et al. The value of serum prostate speci®c antigen 26 Oesterling J. Using prostate speci®c antigen to eliminate the determinations before and after radical prostatectomy. J Urol staging radionuclide bone scan: signi®cant economic implica- 1989; 141: 873±879. tions. Urol Clin North Am 1993; 20: 705±712. Pre-treatment staging of prostate cancer RS Kirby

10 27 Miller P, Eardley I, Kirby R. Prostate speci®c antigen and bone 34 Burgers JK, Hinkle GH, Haseman MK. Monoclonal antibody scan correlation in staging and monitoring of prostate cancer. Br J imaging of recurrent and metastatic prostate cancer. Semin Urol Urol 1992; 70: 295±298. 1995; 13: 103±112. 28 Neal CE, Meis LC. Correlative imaging with monoclonal anti- 35 Sanford E et al. Prostate cancer imaging with a new mono- bodies in colorectal, ovarian and prostate cancer. Sem Nucl Med clonal antibody: A preliminary report. Ann Surg Oncol 1994; 1: 1994; 24: 272±285. 400±404. 29 Leroy M, Teillac P, Rain J De. Radioimmunodetection of lymph 36 Feneley MR et al. Prostatic radio-immunoscintigraphy: Prelimin- node invasion in prostatic cancer. The use of iodine 123 (123I)- ary results using technetium-labelled monoclonal antibody, CYT- labelled monoclonal anti-prostatic acid phosphatase (PAP) 227 A 351. BJU 1996; 77: 373. F (ab0)2 antibody fragments in vivo. Cancer 1989; 64: 1±5. 37 Zoeller G, Sandrock, Munz DL, Ringert RH. Bone marrow 30 Vikko P, Heikkila I, Kontturi M. Radioimaging of the prostate and immunoscintigraphy versus conventional bone scintigraphy in metastases of prostatic carcinoma with 99m Tc labelled prostatic the diagnosis of skeletal metastases in urogenital malignancies. acid phosphatase speci®c antibodies and their FAB fragments. Eur Urol 1994; 26: 141±144. Ann Clin Res 1984; 16: 51±52. 38 Thomas R, Steele R, Smith R, Brannen W. One stage laparoscopic 31 Ahonen A, Kairemo K, Karnani Pe. Radioimmunodetection of pelvic lymphadenectomy and radical prostatectomy. J Urol 1994; prostatic cancer by 111 In-labelled monoclonal antibody against 152: 1174±1177. prostatic acid phosphatase. Acta Oncol 1993; 32: 723±727. 39 Mazeman E, Wurtz A, Gilliot P, Biserte J. Extraperitoneal pelvio- 32 Horoszewicz JS, Kawinski E, Murphy GP. Monoclonal antibodies scopy in lymph node staging of bladder and prostate cancer. to a new antigenic marker in epithelial prostatic cells and serum J Urol 1992; 147: 366±370. of prostatic cancer patients. Anticancer Res 1987; 7: 927±935. 33 Podoloff DA, Neal CE, Babaian R Je. Detection of lymph node metastases in prostatic carcinoma with In-111-labelled CYT 356. Radiology 1993; 189: 334±340.