Clinical Radiology (1992) 46, 225 231
Review Staging Prostatic Cancer R. C L E M E N T S , G. J. G R I F F I T H S and W. B. P E E L I N G * Departments of Clinical Radiology and * Urology, Royal Gwent Hospital, Newport, Gwent
Prostatic carcinoma is the second most c o m m o n neoplasm in men and currently the second most c o m m o n cause of cancer death. Inevitably an ageing population with an ever increasing number of men over the age of 50 years, will mean further increases in the prevalence of the disease. Early detection and treatment are considered important if the survival rate and morbidity of prostatic cancer are to be decreased Accurate pre-operative staging of the extent of the disease is important because the mode of therapy is determined by the clinical stage. In general, patients with localized disease are either treated with localized radiotherapy or radical prostatectomy or given deferred treatment after a period of observation; patients with advanced disease are treated by hormonal manipulation and symptomatic metastatic disease may be treated by radiotherapy. Systems for clinical staging of prostatic cancer categorize patients in terms of the physical extent of their disease, so that those with a potentially curable situation can be recognized and separated from those in whom palliation is all that can be offered. Accurate staging is therefore a guide to prognosis and forms the basis upon which initial management of patients with prostatic cancer is decided. Prostatic cancer is usually staged either by the tumournode-metastases ( T N M ) system which is the method most widely used in Europe, or by a modification of the Whitmore-Jewett classification which is widely employed in the USA [1-3]. It is important to be familiar with both methods so that comparison can be made between European and American studies. The fourth revision of the T N M classification was introduced in 1987. This introduced some changes in the staging of urological tumours from the 1978 third edition that have not been acceptable to many Urologists. It is the editorial policy of the British Journal o f Urology to publish articles only using the older third edition of the classification and hence it is this version that will be described in the present article. A simplified comparison of the European and American staging system for prostatic cancer together with the instance of pelvic lymph node metastases is shown in Table 1 [4,5]. TNM
STAGING
SYSTEM
The T N M staging system is the staging system most commonly used in the UK. In the third edition of the T N M classification, Stage To represents a tumour that is not clinically appreciable on physical examination, i.e. the tumour is not palpable, the tumour tissue being found incidentally in a prostatectomy specimen that has been obtained in an operation performed for outflow tract Correspondence to: Dr R. Clements, Consultant Radiologist, Department of Clinical Radiology, Royal Gwent Hospital, Cardiff Road, Newport, Gwent NP9 2UB.
obstruction. As with the American Stage A disease this category is sub-divided into two groups, T0a and T0b which are roughly equivalent to Stage A1 and A2 respectively. T0a represents a well differentiated carcin o m a involving less than one lobe whereas T0b is a multifocal or diffuse, moderately or poorly differentiated tumour. Stage T1 is an intracapsular tumour surrounded by palpably normal prostatic tissue whereas Stage T2 is an intracapsular tumour deforming the shape of the gland leaving the lateral sulci and seminal vesicles uninvolved. A Stage T3 tumour extends beyond the capsule with or without involvement of the lateral sulci and/or the seminal vesicles. Finally, Stage T4 tumours represent locally advanced lesions that are fixed to the pelvic wall or infiltrating neighbouring structures. The N category gives the lymph node status. No patients have no evidence of regional lymph node involvement, whereas N 1 patients have evidence of involvement of a single ipsilateral regional lymph node. In N2 disease there is evidence of involvement of contralateral or bilateral or multiregional lymph nodes. N3 disease shows evidence of involvement of fixed regional lymph nodes whereas an N4 tumour shows evidence of involvement ofjuxtaregional lymph nodes. The M category is used for assessing distant metastases: in M0 disease there is no evidence of metastases, but in M1 disease metastases are present. The problems associated with the fourth edition of the T N M classification as applied to the prostate gland are described by Schr6der et al. [6]. Much attention has been focused in recent years in North America on the accurate staging of prostatic cancer prior to radical prostatectomy. This major operation has become increasingly popular in North America and in some European countries and recently has started to be performed in a few specialist centres in the UK. Meticulous histological examination of the resected specimens obtained at radical prostatectomy has allowed extremely precise correlation between the abnormality detected by clinical, serological or imaging studies and the extent of the primary tumour. Unfortunately, at present the usual clinical presentation of prostatic cancer in the Table 1 - Simplified comparison of T N M and Whitmore-Jewett classifications for staging prostatic cancer
Digital palpation Impalpable Palpable confined Palpable unconfined Any primary
TNM
Whitmore-Jewett (modified) system
Approximate% lymphnode involvement
T0a T0b T1 T2 T3 T4 N+
A1 A2 B1 B2 C1 C2 D1 D2
0 34 18 34 50 80 100
M+
226
CLINICAL RADIOLOGY
U K is with advanced disease where hormonal manipulation is the only appropriate treatment modality.
A S S E S S M E N T OF THE PRIMARY T U M O U R (T C A T E G O R Y )
Digital examination is unlikely to be superceded as the main initial method of examining the prostate to assess primary prostatic cancer but obviously the accuracy of interpretation depends on experience and the care taken by the clinician. Even so there are limitations: even the most sensitive index finger may have difficulty with prostatic cancer in its early stages. Diagnosing small prostatic cancers by palpation is notoriously inaccurate; it is well known that 50% of palpably suspicious nodules of the prostate are found to be benign [2]. Pathological studies have also shown that clinicians are p o o r at estimating the location of a cancer within the prostate probably because these lesions feel very similar to chronic prostatitis, if this is accompanied by prostatic calcification. Chronic prostatitis arises in the peripheral part of the gland and can be present alongside malignant areas [7]. Perhaps the most serious drawback of digital evaluation of prostatic cancer is the high rate of inaccuracy when staging confined (T1, T2) tumours, as m a n y capsular breaches occur anteriorly away from the examining finger. This error rate has been shown to be of the order of 25% [7]. These are serious limitations not only for clinical practice but also attract attention to the grave errors that could be introduced into clinical trials which rely on digital examination of the primary tumour. Three main areas of potential error occur in relation to staging the earlier stages of prostatic cancer: 1 Detection of impalpable subclinical cancer. 2 Recognition of cancer in small palpable nodules or areas of induration. 3 Accurate identification of those cancers that are confined within the prostatic capsule. It is important to realize that the prostatic capsule is only one or two cells thick and is deficient near the apex of the gland. It is therefore not visualized by any current imaging method. In sonographic terms, we speak of the prostatic capsule as the boundary between prostatic parenchyma and the echogenic periprostatic fat. Further investigation to assess the primary tumour of a patient with prostatic cancer is currently performed by serum tests - such as prostate specific antigen (PSA) and imaging tests such as transrectal sonography (TRUS), computed tomography (CT) and magnetic resonance imaging (MR).
S E R U M TESTS FOR THE D I A G N O S I S A N D STAGING OF PROSTATIC C A N C E R
Prostate specific antigen (PSA) is a glycoprotein protease that is produced solely by the epithelial cells of the prostate. PSA was first isolated in 1979 [8] and has recently been increasingly used in the management of prostatic disease, particularly in the management of prostatic cancer. It is now accepted to be a more sensitive serum marker for prostatic cancer than prostatic acid phosphatase [9]. PSA measurements have a number of applications in the management of patients with prostatic
cancer, a n d can be used as an aid to the staging of established disease, for monitoring the response to treatment and potentially for screening and the early detection of prostatic cancer. PSA measurements have been found of value in the initial staging of patients with prostatic cancer and may be used in conjunction with imaging studies. There are however some difficulties with the interpretation of PSA measurements of which radiologists should be aware. Firstly, two main assays are currently commercially available, one produced by Hybritech Inc. the Tandem R-PSA assay - a two site immuno-radiometric test with monoclonal antibodies and a second produced by Yang Laboratories-the Pros check PSA assay which uses a conventional radioimmunoassay technique with a polyclonal serum. There is no clinical superiority of one over the other, but the Yang method consistently results in lower values and it is important that clinical laboratories should state the name of their assay along with their results. Secondly, PSA is produced by normal and hyperplastic prostatic tissue and by prostatic cancer. Since benign prostatic hyperplasia (BPH) and prostatic cancer are most prevalent during the same age range, the influence of BPH tissue on the serum PSA concentration must be known if PSA is to be a useful marker for prostatic cancer. Both BPH and carcinoma may cause elevation of the PSA, although the elevation caused by BPH is usually slight. PSA levels are elevated approximately 0.3 ng/ml/g in benign hyperplasia compared with 3.5 ng/ml/g of cancer tissue [10]. Thus a patient with a 60 ml prostate with benign hyperplasia would produce a PSA value indistinguishable from that produced by a 30 ml patient with a 3 ml cancer. Large rises in serum PSA levels indicate prostatic malignancy and cause few problems, but there may be difficulty with interpreting smaller elevations, when there is co-existing BPH. We have recently attempted to quantify the expected element of PSA elevation caused by BPH by relating it to the volume of the gland measured ultrasonically [11]. When volume-corrected serum PSA levels from patients with M1 cancers were compared with the volume-corrected PSA determination in patients with BPH, all patients with M1 disease were clearly discriminated from the BPH pattern. When PSA from patients with M0 cancers were compared with the BPH PSA pattern, 83% of patients were discriminated from the BPH pattern by the use of volume correction. It is likely that volume correction of PSA measurements will become increasingly important in the future. Various authors have demonstrated that serum PSA concentrations increase in proportion with the advancing clinical stage of prostatic cancer. Using the Yang assay in a study of 209 patients [ 10], it was found that serum PSA was directly proportional to clinical stage and was able to distinguish between all clinical stages except between Stages B2 and B3 and Stages C and BI disease. Using the Hybritech assay, different authors [12,13] have found serum PSA measurements increased proportionally with advancing clinical stage; there was however considerable overlap amongst all clinical stages. Other studies have investigated the relationship between the pathological stage and pre-operative serum PSA level in patients undergoing radical prostatectomy and have again demonstrated a rise in PSA concentration with advancing pathological stage [14,15]. There is however overlap between different pathological stages. Thus although
STAGING PROSTATIC CANCER
Fig. 1 TRUS; Axial scan: a T2 hypoechoic cancer is present in the right peripheral zone (arrow). The adjacent capsule is intact.
there is a direct correlation with serum PSA concentration and advancing clinical and pathological stage of prostatic cancer, PSA is not sufficiently reliable to determine the stage on an individual basis. PSA alone cannot be used for local tumour staging as it does not reliably distinguish patients with extra-capsular tumour spread. It is however a simple initial test that can give an indication of the likelihood of extra-prostatic metastatic disease; its role in relation to bone scintigraphy is discussed below. IMAGING TESTS FOR DIAGNOSING PROSTATIC CANCER Any diagnostic imaging technique that attempts to rectify the shortcomings of digital assessment o f the prostate will need to fulfil the following criteria: 1 The technique should be capable of demonstrating the zonal anatomy of the prostate to provide accurate localization within the prostate of suspect areas so that they can be biopsied accurately. 2 The prostatic 'capsule' should be shown throughout its periphery and at all levels of the gland. 3 The seminal vesicles, base of bladder and the rhomboid area distal to the apex of the gland should be demonstrated. Many earlier reports of the accuracy of imaging methods in staging prostatic cancer may now be considered invalid because of suboptimal imaging techniques. With sonography, early reports were based on lower frequency probes, and with MR, reports were based on poor quality scanners or suboptimal pulse parameters. There was usually a lack of direct pathological correlation with prostatectomy specimens. TRANSRECTAL ULTRASOUND The most widely used imaging technique for the diagnosis and staging of prostatic cancer is transrectal ultrasound (TRUS). Cancer usually appears sonographi-
227
Fig. 2 - TRUS; Axial scan: a transition zone hypoechoic area is present (arrow) adjacent to the defect of the previous transurethral resection (broad arrow).
cally as a hypoechoic area but both isoechoic and hyperechoic cancers can occasionally be found [16,17]. Prostatic cancer generally occurs in the peripheral zone (Fig.l), but can also arise in the central zone and transition zone (Fig. 2) [18]. It is now accepted that prostatic cancer may be a multifocal process with separate discrete foci in different areas of the gland. In these circumstances, a TO cancer may represent a focus of cancer within the transition zone but the peripheral zone also needs to be scrutinized carefully as a small peripheral zone lesion m a y co-exist and may well be the focus that subsequently spreads outside the gland and eventually metastasizes. The diagnosis o f a hypoechoic area must be established by biopsy as there are m a n y causes of hypoechoic areas within the prostate. Biopsy of the prostate under ultrasound guidance by the transrectal route is now an easy procedure [19]. T R U S can be used to stage the extent of prostatic cancer by assessing the integrity of the capsule and the likelihood of seminal vesicle invasion. A capsular breach indicates spread of cancer into the periprostatic fat (Figs 3, 4, 5). In our experience of 121 patients with confined prostatic cancer, the site of the capsular breach is more commonly in the anterior half of the gland than the posterior half [16]. There is however difficulty with any imaging technique in assessing the likelihood of microscopic invasion of the capsule in patients in whom the capsule appears macroscopically intact. In these circumstances a peripheral zone focus adjacent to the neurovascular bundle causing a bulge of the outline (Fig. 6) should be regarded as microscopic penetration of the capsule. Such microscopic disease can be important clinically in patients being considered for radical prostatectomy and is one of the reasons that imaging techniques have been found generally to be less accurate in assessing the precise state of capsular integrity when they are compared with microscopic histological appearances of specimens obtained at radical prostatectomy. Seminal vesicle involvement by prostatic cancer may also be assessed sonographically. In early reports, the criteria of seminal vesicle involvement used was loss of the
228
CLINICAL RADIOLOGY
Fig. 3 - TRUS; Axial scan: a large T3 cancer is present mainly in the left lobe involving over 50% of the prostate gland. There is a posterolateral capsular breach (arrow).
Fig. 5 TRUS; Axial scan: a large posterior capsular breach is present (arrow) in this T3 tumour.
Fig. 4 TRUS; Saginal scan: a small T3 hypoechoic cancer with adjacent capsular irregularity is present (arrow).
angle between the seminal vesicles and its adjacent prostate and bladder base. Other authors subsequently considered a 2 cm hypoechoic area at the prostate base to be synonymous with seminal vesicle invasion and Lee and colleagues [20] paid attention to involvement of the 'invaginated extra-prostatic space' around the ejaculatory ducts as indicative of seminal vesicle involvement. The sonographic features of seminal vesicle involvement have recently been assessed fully in 95 subjects without prostatic cancer and 205 patients with seminal vesicle invasion [21] (Fig. 7). The appearances were correlated with histological examination of the seminal vesicles either from staging core biopsies or prostatectomy speci-
Fig. 6 TRUS; Axial scan: a peripheral zone cancer (large arrow) adjacent to the neurovascular bundle (arrows) is likely to have microscopic extra-capsular extension.
mens. The ultrasonic findings correlating best with tumour invasion of seminal vesicles were hyperechogenicity and a combination of two or more of the following abnormalities cystic dilatation, asymmetry, enlargement, and anterior displacement. Of the 38 patients with histologically-proven seminal vesicle invasion by prostatic cancer, 35 (92%) had an abnormal appearance of the seminal vesicles on sonography. Of 167 patients with prostatic cancer without histological involvement of the seminal vesicles, sonograms showed abnormal seminal
STAGING PROSTATIC CANCER
229
Fig. 8 Endorectal MR scan (Courtesy Dr H. I~I. Pollack): T2 cancer in left peripheral zone with no extension beyond the prostatic capsule.
Fig. 7 - TRUS; Axial scan: there is right seminal vesicle invasion. The seminal vesicle is hyperechoic and enlarged.
vesicles in 42 (25%). In 95 patients with histologicallynormal prostates and seminal vesicles, the sonograms showed abnormal seminal vesicles in 11 (12%). The other area that must be carefully assessed sonographically in relation to the prostatic capsule in patients being considered for radical surgery is the apex of the gland. This is best demonstrated by T R U S on the sagittal view where involvement of the rhomboid area formed by the distal urethra, recto-urethralis muscle, rectal wall and apex of the gland can be assessed.
CT SCANNING The prostate appears homogeneous at CT, and CT has been found to be of no value in assessing prostatic cancer confined to the gland. CT may have a limited role in the staging of prostatic cancer, particularly if T R U S and M R is not available. It is not however the optimal investigation. In a comparative study of 27 patients with biopsy-proven carcinoma of the prostate examined with TRUS, CT and MR, T R U S was found to be superior for assessing capsular disruption and M R was found to be superior to both T R U S and CT for evaluating seminal vesicle invasion. MR was also slightly better than CT for detecting lymphadenopathy [22].
MR IMAGING The zonal anatomy and intraglandular texture of the prostate may be well demonstrated with MR imaging and MR has in recent years played an increasing role in staging prostatic cancer. This has been particularly true of patients prior to radical surgery. A comparative study of MR imaging and ultrasound in staging early prostatic cancer was reported by Rifkin et al. [23]. This multicentre study involved five prestigious institutions in North America. The patients were subsequently treated by radical prostatectomy and the surgical specimens were
serially sectioned at 5 mm intervals and examined in their entirety. Two-hundred and thirty patients with clinicallylocalized prostatic cancer were evaluated pre-operatively by both techniques. M R correctly staged 77% of cases of advanced disease and 57% of cases with localized disease; T R U S correctly staged 66% of cases of advanced disease and 46% of localized disease. The figures did not vary significantly between readers, and simultaneous interpretation of MR and T R U S scans did not improve the accuracy. Various reasons were advanced for the disappointing results of this survey in comparison with previous studies but the main reason appears to be that all the patients were thought on the basis of initial assessments to have surgically resectable disease thus minimizing the possibility of including patients with obvious advanced disease which is easier to stage with imaging studies. Additionally in previous studies the entire radical prostatectomy specimens were not always 'step sectioned' histologically to correlate exactly with the imaging. None of the imaging techniques available including TRUS, CT and M R are able to identify microscopic disease. They can only detect macroscopic disease, whether or not there is an attempt to evaluate local disease, periprostatic or seminal vesicle invasion, lymph node involvement or distant spread of the tumour. There have been further improvements in M R recently with the introduction of endorectal surface coils and this technique appears to offer improvements in relation to staging when compared with MR imaging with conventional body surface coils. The endorectal coil improves the signal-to-noise ratio dramatically because of the decreased field of view and the use of thinner (3 mm) slices. The improved spatial resolution gives better delineation of the internal anatomy of the gland and should improve its accuracy in relation to staging. Schnall et al. [24] quote an accuracy of 82% in differentiating Stage B from Stage C disease with endorectal surface coils (Figs 8-10). Two recent further developments, viz. the use of fast spin echo sequences and multi-coil array imaging with one endorectal coil, and a second body surface coil on the anterior or posterior surface of the pelvis should further improve the accuracy of MR. With these techniques a staging accuracy of approximately 90% may be
230
CLINICAL RADIOLOGY
DISTANT M E T A S T A S E S (M CATEGORY )
Fig. 9 - Endorectal MR scan (Courtesy Dr H. M. Pollack): T3 prostatic cancer with extension through the posterior capsule at 6 o'clock position.
Bone scintigraphy is the most sensitive imaging technique for detecting bone metastases of prostatic cancer. In recent years, its role has been assessed in conjunction with serum PSA determinations in relation to the initial staging of bone metastases, and in the follow-up of bone metastases. It may be that isotope scintigraphy is not justified in the initial staging assessment of patients with newly diagnosed prostatic cancer if the PSA level is below 20 ng/ml. Chybowski et al. [26] retrospectively reviewed 521 randomly chosen patients with newly diagnosed untreated prostatic cancer. The median serum PSA concentration in patients with a positive bone scan was 158 ng/ml, whereas men with a negative bone scan had a median serum PSA level of 11.3 ng/ml. The 99.7% negative predictive value of a low serum PSA concentration for bone scan findings was particularly noted. In 306 men with a serum PSA level of 20 ng/ml or less, only one patient (PSA 18.2 ng/ml) had a positive bone scan. Currently unpublished observations by our group would support this view that staging bone scintigraphy in a previously untreated prostatic cancer patient with a low PSA may not be necessary. CONCLUSION
Fig. I0-Endorectal MR scan (Courtesy Dr H. M. Pollack): T3 prostatic cancer with invasion of the right seminal vesicle laterally.
achieved in differentiating Stage B and C disease (Pollack HM, personal communication). However the accuracy of this technique in a group comparable to the study of Rifkin et al. [23] is not yet available.
It is likely that the investigation and assessment of patients with prostatic cancer will become an increasingly common source of uroradiological work. Recent experience has confirmed the importance of serum PSA measurements in the urological diagnosis of prostatic disease, and an understanding of these measurements will be important for the radiologist involved in the diagnosis and staging of prostatic cancer. T R U S and TRUS-guided biopsies have become accepted methods of diagnosing prostatic cancer, and TRUS provides an acceptable method of staging prostatic cancer differentiating patients who have confined (T1/T2) disease from those with more advanced disease. Improved staging accuracy is necessary prior to radical surgery and those patients considered to have cancer localized to the prostate justify further staging with magnetic resonance. Recent developments in M R techniques should further increase the accuracy of local staging of prostatic cancer. Acknowledgements.The authors wish to thank Dr H. M. Pollack, of the University of Pennsylvania, for providing Figs 8-10.
NODAL STAGING (N CATEGORY) REFERENCES
T R U S has no rote in nodal staging. Both CT and MR may be used to assess the pelvic lymph nodes and most authors claim a similar accuracy for both techniques in detecting lymphadenopathy [22,25]. Nodes, vessels and nerve bundles tend to be more difficult to differentiate with CT. Nodal staging with CT and M R is based on detection of enlarged lymph nodes and microscopic involvement of the lymph nodes cannot be detected. In these circumstances, it has been found that the results of both M R a n d C T in lymph node detection prior to radical surgery are disappointing, and Operative lymph node biopsy is still needed for full assessment prior to radical prostatectomy.
1 Whitmore WF. Hormone therapy and prostatic cancer American Journal of Medicine 1956;21:697-713. 2 Jewett HJ. The present status of radical prostatectomy for Stage A and B prostatic cancer. Urologic Clinics of North America 1975;2:105-124. 3 Wallace DM, Chisholm GD, Hendry WF. Classification for urological tumours. British Journal of Urology 1975;47:1 12. 4 Catalona WJ, Scott WW. Carcinoma of the prostate: a review. Journal of Urology 1978; 119:1-8. 5 Cantrell BB, de Klerk DP, Eggleston JC, Boittnott JK, Walsh PC. Pathological factors that influence prognosis in Stage A prostatic cancer: the influence of extent versus grade. Journal of Urology 1981,125:516-520. 6 Schr6der FH, Cooper EH, Debruyne FMJ, Denis L, Newling DWW, Pavone-Macaluso M e t al. TNM classification of genito-
STAGING PROSTATIC CANCER urinary tumours 1987 position of the E.O.R.T.C. Genitourinary group. British Journal of Urology 1988;62:502 510. 7 Byar DP, Mostofi FK. Veterans Administration Cooperative Urological Research Group (VACURG). Carcinoma of the prostateprognostic features in 208 radical prostatectomies Cancer 1972;30:5-13. 8 Wang MC, Valenzuela LD, Murphy GP, Chu TM. Purification of a human prostate specific antigen. Investigative Urology 1979;17: 159 163. 9 0 o s t e r o n K, Bogdanovicz J, Schroder FH. Evaluation of prostate specific antigen in untreated prostatic carcinoma. European Urology 1989;16:253 257. I0 Stamey TA, Kabalin JN. Prostate specific antigen in the diagnosis and treatment of adenocarcinoma of the prostate. 1. Untreated patients. Journal of Urology 1989;141:1070 1075. 11 Clements R, Etherington RJ, Griffiths G J, Peeling WB, Hughes H, Penney MD. The interrelation of serum prostatic specific antigen measurement and transrectal ultrasound in the diagnosis of benign prostatic hyperplasia and prostatic cancer. British Journal of Urology 1992;70:183 187. 12 Chan DW, Bruzek D J, Oesterling JE, Rock RC, Walsh PC. Prostate specific antigen as a marker for prostatic cancer: a monoclonal and polyclonal immuno-assay compared. Clinical Chemistry 1987; 33:1916-1918. 13 Hudson MA, Bahnson RR, Catalona WJ. Clinical use of prostatic specific antigen in patients with prostate cancer. Journal of Urology 1989; 142:1011-1017. 14 Lange PH, Ercole CJ, Lightner D J, Fraley EE, Vessella R. The value of serum prostate specific antigen determinations before and after radical prostatectomy. Journal of Urology 1989; 141:873 879. 15 Partin AW, Carter HB, Chan DW, Epstein JI, Oesterling JE, Rock RC, Weber JP, Walsh PC, Prostate specific antigen in the staging of localised prostate cancer: influence of turnout differentiation, tumour volume, and benign hyperplasia. Journal of Urology 1990;143:747 753. 16 Griffiths GJ, Clements R, Jones DJ, Roberts EE, Peeling WB, Evans KT. The ultrasound appearances of prostatic cancer with histological correlation. Clinical Radiology 1987;38:219 227.
231
17 Dahnert WF, Hamper UM, Eggleston JC, Walsh PC, Sanders RC.Prostatic evaluation by transrectal sonography with-histopathologic correlation: the echopenic appearance of early carcinoma. Radiology 1986;158:97-102. 18 McNeal JE, Redwine EA, Freiha FS, Stamey TA. Zonal distribution of prostatic adenocarcinoma: correlation with histologic pattern and direction of spread. American Journal of Clinical Pathology 1988;12:897 906. 19 Torp-Pedersen S, Lee F. Transrectal biopsy of the prostate guided by transrectal ultrasound. Urologic Clinics of North America 1989; 16:703-712. 20 Lee F, Torp-Pedersen ST, Siders DB, Littrup PJ, McLeary RD. Transrectal ultrasound in the diagnosis and staging of prostatic carcinoma. Radiology 1989;170:609 615. 21 Terris MK, McNeal JE, Stamey TA. Invasion of the seminal vesicles by prostatic cancer: detection with transrectal sonography. American Journal of Roentgenology 1990;155:811 815. 22 Friedman AC, Seidmon JE, Radecki PD, Lev-Toaff A, Caroline DF. Relative merits of MRI, transrectal endosonography and CT in diagnosis and staging of carcinoma of the prostate. Urology 1988;31:530-537. 23 Rifkin MD, Zerhouni EA, Gatsonis CA, Quint LE, Paushter DM, Epstein JI et al. Comparison of magnetic resonance imaging and ultrasonography in staging early prostate cancer: results of a multiinstitutional co-operative trial. New England Journal of Medicine 1990;323:621 626. 24 Schnall MD, Imai Y, Tomaszewski J, Pollack HM, Lenkinski RE, Kressel HY. Prostate cancer: local staging with endoreetal surface coil MR imaging. Radiology 1990;178:777 780. 25 Hricak H, Dooms GC, Brook Jeffrey R, Avallone A, Jacobs D, Benton WK et al. Prostatic carcinoma: staging by clinical assessment, CT, and MR imaging. Radiology 1987;162:331-336. 26 Chybowski FM, Larson Keller J J, Bergstrahl E J, Oesterling JE. Predicting radionuclide bone scan findings in patients with newly diagnosed, untreated prostate cancer: prostate specific antigen is supcrior to all other clinical. Journal of Urology 1991;145:313 318.
Review Staging Prostatic Cancer R. C L E M E N T S , G. J. G R I F F I T H S and W. B. P E E L I N G * Departments of Clinical Radiology and * Urology, Royal Gwent Hospital, Newport, Gwent
Prostatic carcinoma is the second most c o m m o n neoplasm in men and currently the second most c o m m o n cause of cancer death. Inevitably an ageing population with an ever increasing number of men over the age of 50 years, will mean further increases in the prevalence of the disease. Early detection and treatment are considered important if the survival rate and morbidity of prostatic cancer are to be decreased Accurate pre-operative staging of the extent of the disease is important because the mode of therapy is determined by the clinical stage. In general, patients with localized disease are either treated with localized radiotherapy or radical prostatectomy or given deferred treatment after a period of observation; patients with advanced disease are treated by hormonal manipulation and symptomatic metastatic disease may be treated by radiotherapy. Systems for clinical staging of prostatic cancer categorize patients in terms of the physical extent of their disease, so that those with a potentially curable situation can be recognized and separated from those in whom palliation is all that can be offered. Accurate staging is therefore a guide to prognosis and forms the basis upon which initial management of patients with prostatic cancer is decided. Prostatic cancer is usually staged either by the tumournode-metastases ( T N M ) system which is the method most widely used in Europe, or by a modification of the Whitmore-Jewett classification which is widely employed in the USA [1-3]. It is important to be familiar with both methods so that comparison can be made between European and American studies. The fourth revision of the T N M classification was introduced in 1987. This introduced some changes in the staging of urological tumours from the 1978 third edition that have not been acceptable to many Urologists. It is the editorial policy of the British Journal o f Urology to publish articles only using the older third edition of the classification and hence it is this version that will be described in the present article. A simplified comparison of the European and American staging system for prostatic cancer together with the instance of pelvic lymph node metastases is shown in Table 1 [4,5]. TNM
STAGING
SYSTEM
The T N M staging system is the staging system most commonly used in the UK. In the third edition of the T N M classification, Stage To represents a tumour that is not clinically appreciable on physical examination, i.e. the tumour is not palpable, the tumour tissue being found incidentally in a prostatectomy specimen that has been obtained in an operation performed for outflow tract Correspondence to: Dr R. Clements, Consultant Radiologist, Department of Clinical Radiology, Royal Gwent Hospital, Cardiff Road, Newport, Gwent NP9 2UB.
obstruction. As with the American Stage A disease this category is sub-divided into two groups, T0a and T0b which are roughly equivalent to Stage A1 and A2 respectively. T0a represents a well differentiated carcin o m a involving less than one lobe whereas T0b is a multifocal or diffuse, moderately or poorly differentiated tumour. Stage T1 is an intracapsular tumour surrounded by palpably normal prostatic tissue whereas Stage T2 is an intracapsular tumour deforming the shape of the gland leaving the lateral sulci and seminal vesicles uninvolved. A Stage T3 tumour extends beyond the capsule with or without involvement of the lateral sulci and/or the seminal vesicles. Finally, Stage T4 tumours represent locally advanced lesions that are fixed to the pelvic wall or infiltrating neighbouring structures. The N category gives the lymph node status. No patients have no evidence of regional lymph node involvement, whereas N 1 patients have evidence of involvement of a single ipsilateral regional lymph node. In N2 disease there is evidence of involvement of contralateral or bilateral or multiregional lymph nodes. N3 disease shows evidence of involvement of fixed regional lymph nodes whereas an N4 tumour shows evidence of involvement ofjuxtaregional lymph nodes. The M category is used for assessing distant metastases: in M0 disease there is no evidence of metastases, but in M1 disease metastases are present. The problems associated with the fourth edition of the T N M classification as applied to the prostate gland are described by Schr6der et al. [6]. Much attention has been focused in recent years in North America on the accurate staging of prostatic cancer prior to radical prostatectomy. This major operation has become increasingly popular in North America and in some European countries and recently has started to be performed in a few specialist centres in the UK. Meticulous histological examination of the resected specimens obtained at radical prostatectomy has allowed extremely precise correlation between the abnormality detected by clinical, serological or imaging studies and the extent of the primary tumour. Unfortunately, at present the usual clinical presentation of prostatic cancer in the Table 1 - Simplified comparison of T N M and Whitmore-Jewett classifications for staging prostatic cancer
Digital palpation Impalpable Palpable confined Palpable unconfined Any primary
TNM
Whitmore-Jewett (modified) system
Approximate% lymphnode involvement
T0a T0b T1 T2 T3 T4 N+
A1 A2 B1 B2 C1 C2 D1 D2
0 34 18 34 50 80 100
M+
226
CLINICAL RADIOLOGY
U K is with advanced disease where hormonal manipulation is the only appropriate treatment modality.
A S S E S S M E N T OF THE PRIMARY T U M O U R (T C A T E G O R Y )
Digital examination is unlikely to be superceded as the main initial method of examining the prostate to assess primary prostatic cancer but obviously the accuracy of interpretation depends on experience and the care taken by the clinician. Even so there are limitations: even the most sensitive index finger may have difficulty with prostatic cancer in its early stages. Diagnosing small prostatic cancers by palpation is notoriously inaccurate; it is well known that 50% of palpably suspicious nodules of the prostate are found to be benign [2]. Pathological studies have also shown that clinicians are p o o r at estimating the location of a cancer within the prostate probably because these lesions feel very similar to chronic prostatitis, if this is accompanied by prostatic calcification. Chronic prostatitis arises in the peripheral part of the gland and can be present alongside malignant areas [7]. Perhaps the most serious drawback of digital evaluation of prostatic cancer is the high rate of inaccuracy when staging confined (T1, T2) tumours, as m a n y capsular breaches occur anteriorly away from the examining finger. This error rate has been shown to be of the order of 25% [7]. These are serious limitations not only for clinical practice but also attract attention to the grave errors that could be introduced into clinical trials which rely on digital examination of the primary tumour. Three main areas of potential error occur in relation to staging the earlier stages of prostatic cancer: 1 Detection of impalpable subclinical cancer. 2 Recognition of cancer in small palpable nodules or areas of induration. 3 Accurate identification of those cancers that are confined within the prostatic capsule. It is important to realize that the prostatic capsule is only one or two cells thick and is deficient near the apex of the gland. It is therefore not visualized by any current imaging method. In sonographic terms, we speak of the prostatic capsule as the boundary between prostatic parenchyma and the echogenic periprostatic fat. Further investigation to assess the primary tumour of a patient with prostatic cancer is currently performed by serum tests - such as prostate specific antigen (PSA) and imaging tests such as transrectal sonography (TRUS), computed tomography (CT) and magnetic resonance imaging (MR).
S E R U M TESTS FOR THE D I A G N O S I S A N D STAGING OF PROSTATIC C A N C E R
Prostate specific antigen (PSA) is a glycoprotein protease that is produced solely by the epithelial cells of the prostate. PSA was first isolated in 1979 [8] and has recently been increasingly used in the management of prostatic disease, particularly in the management of prostatic cancer. It is now accepted to be a more sensitive serum marker for prostatic cancer than prostatic acid phosphatase [9]. PSA measurements have a number of applications in the management of patients with prostatic
cancer, a n d can be used as an aid to the staging of established disease, for monitoring the response to treatment and potentially for screening and the early detection of prostatic cancer. PSA measurements have been found of value in the initial staging of patients with prostatic cancer and may be used in conjunction with imaging studies. There are however some difficulties with the interpretation of PSA measurements of which radiologists should be aware. Firstly, two main assays are currently commercially available, one produced by Hybritech Inc. the Tandem R-PSA assay - a two site immuno-radiometric test with monoclonal antibodies and a second produced by Yang Laboratories-the Pros check PSA assay which uses a conventional radioimmunoassay technique with a polyclonal serum. There is no clinical superiority of one over the other, but the Yang method consistently results in lower values and it is important that clinical laboratories should state the name of their assay along with their results. Secondly, PSA is produced by normal and hyperplastic prostatic tissue and by prostatic cancer. Since benign prostatic hyperplasia (BPH) and prostatic cancer are most prevalent during the same age range, the influence of BPH tissue on the serum PSA concentration must be known if PSA is to be a useful marker for prostatic cancer. Both BPH and carcinoma may cause elevation of the PSA, although the elevation caused by BPH is usually slight. PSA levels are elevated approximately 0.3 ng/ml/g in benign hyperplasia compared with 3.5 ng/ml/g of cancer tissue [10]. Thus a patient with a 60 ml prostate with benign hyperplasia would produce a PSA value indistinguishable from that produced by a 30 ml patient with a 3 ml cancer. Large rises in serum PSA levels indicate prostatic malignancy and cause few problems, but there may be difficulty with interpreting smaller elevations, when there is co-existing BPH. We have recently attempted to quantify the expected element of PSA elevation caused by BPH by relating it to the volume of the gland measured ultrasonically [11]. When volume-corrected serum PSA levels from patients with M1 cancers were compared with the volume-corrected PSA determination in patients with BPH, all patients with M1 disease were clearly discriminated from the BPH pattern. When PSA from patients with M0 cancers were compared with the BPH PSA pattern, 83% of patients were discriminated from the BPH pattern by the use of volume correction. It is likely that volume correction of PSA measurements will become increasingly important in the future. Various authors have demonstrated that serum PSA concentrations increase in proportion with the advancing clinical stage of prostatic cancer. Using the Yang assay in a study of 209 patients [ 10], it was found that serum PSA was directly proportional to clinical stage and was able to distinguish between all clinical stages except between Stages B2 and B3 and Stages C and BI disease. Using the Hybritech assay, different authors [12,13] have found serum PSA measurements increased proportionally with advancing clinical stage; there was however considerable overlap amongst all clinical stages. Other studies have investigated the relationship between the pathological stage and pre-operative serum PSA level in patients undergoing radical prostatectomy and have again demonstrated a rise in PSA concentration with advancing pathological stage [14,15]. There is however overlap between different pathological stages. Thus although
STAGING PROSTATIC CANCER
Fig. 1 TRUS; Axial scan: a T2 hypoechoic cancer is present in the right peripheral zone (arrow). The adjacent capsule is intact.
there is a direct correlation with serum PSA concentration and advancing clinical and pathological stage of prostatic cancer, PSA is not sufficiently reliable to determine the stage on an individual basis. PSA alone cannot be used for local tumour staging as it does not reliably distinguish patients with extra-capsular tumour spread. It is however a simple initial test that can give an indication of the likelihood of extra-prostatic metastatic disease; its role in relation to bone scintigraphy is discussed below. IMAGING TESTS FOR DIAGNOSING PROSTATIC CANCER Any diagnostic imaging technique that attempts to rectify the shortcomings of digital assessment o f the prostate will need to fulfil the following criteria: 1 The technique should be capable of demonstrating the zonal anatomy of the prostate to provide accurate localization within the prostate of suspect areas so that they can be biopsied accurately. 2 The prostatic 'capsule' should be shown throughout its periphery and at all levels of the gland. 3 The seminal vesicles, base of bladder and the rhomboid area distal to the apex of the gland should be demonstrated. Many earlier reports of the accuracy of imaging methods in staging prostatic cancer may now be considered invalid because of suboptimal imaging techniques. With sonography, early reports were based on lower frequency probes, and with MR, reports were based on poor quality scanners or suboptimal pulse parameters. There was usually a lack of direct pathological correlation with prostatectomy specimens. TRANSRECTAL ULTRASOUND The most widely used imaging technique for the diagnosis and staging of prostatic cancer is transrectal ultrasound (TRUS). Cancer usually appears sonographi-
227
Fig. 2 - TRUS; Axial scan: a transition zone hypoechoic area is present (arrow) adjacent to the defect of the previous transurethral resection (broad arrow).
cally as a hypoechoic area but both isoechoic and hyperechoic cancers can occasionally be found [16,17]. Prostatic cancer generally occurs in the peripheral zone (Fig.l), but can also arise in the central zone and transition zone (Fig. 2) [18]. It is now accepted that prostatic cancer may be a multifocal process with separate discrete foci in different areas of the gland. In these circumstances, a TO cancer may represent a focus of cancer within the transition zone but the peripheral zone also needs to be scrutinized carefully as a small peripheral zone lesion m a y co-exist and may well be the focus that subsequently spreads outside the gland and eventually metastasizes. The diagnosis o f a hypoechoic area must be established by biopsy as there are m a n y causes of hypoechoic areas within the prostate. Biopsy of the prostate under ultrasound guidance by the transrectal route is now an easy procedure [19]. T R U S can be used to stage the extent of prostatic cancer by assessing the integrity of the capsule and the likelihood of seminal vesicle invasion. A capsular breach indicates spread of cancer into the periprostatic fat (Figs 3, 4, 5). In our experience of 121 patients with confined prostatic cancer, the site of the capsular breach is more commonly in the anterior half of the gland than the posterior half [16]. There is however difficulty with any imaging technique in assessing the likelihood of microscopic invasion of the capsule in patients in whom the capsule appears macroscopically intact. In these circumstances a peripheral zone focus adjacent to the neurovascular bundle causing a bulge of the outline (Fig. 6) should be regarded as microscopic penetration of the capsule. Such microscopic disease can be important clinically in patients being considered for radical prostatectomy and is one of the reasons that imaging techniques have been found generally to be less accurate in assessing the precise state of capsular integrity when they are compared with microscopic histological appearances of specimens obtained at radical prostatectomy. Seminal vesicle involvement by prostatic cancer may also be assessed sonographically. In early reports, the criteria of seminal vesicle involvement used was loss of the
228
CLINICAL RADIOLOGY
Fig. 3 - TRUS; Axial scan: a large T3 cancer is present mainly in the left lobe involving over 50% of the prostate gland. There is a posterolateral capsular breach (arrow).
Fig. 5 TRUS; Axial scan: a large posterior capsular breach is present (arrow) in this T3 tumour.
Fig. 4 TRUS; Saginal scan: a small T3 hypoechoic cancer with adjacent capsular irregularity is present (arrow).
angle between the seminal vesicles and its adjacent prostate and bladder base. Other authors subsequently considered a 2 cm hypoechoic area at the prostate base to be synonymous with seminal vesicle invasion and Lee and colleagues [20] paid attention to involvement of the 'invaginated extra-prostatic space' around the ejaculatory ducts as indicative of seminal vesicle involvement. The sonographic features of seminal vesicle involvement have recently been assessed fully in 95 subjects without prostatic cancer and 205 patients with seminal vesicle invasion [21] (Fig. 7). The appearances were correlated with histological examination of the seminal vesicles either from staging core biopsies or prostatectomy speci-
Fig. 6 TRUS; Axial scan: a peripheral zone cancer (large arrow) adjacent to the neurovascular bundle (arrows) is likely to have microscopic extra-capsular extension.
mens. The ultrasonic findings correlating best with tumour invasion of seminal vesicles were hyperechogenicity and a combination of two or more of the following abnormalities cystic dilatation, asymmetry, enlargement, and anterior displacement. Of the 38 patients with histologically-proven seminal vesicle invasion by prostatic cancer, 35 (92%) had an abnormal appearance of the seminal vesicles on sonography. Of 167 patients with prostatic cancer without histological involvement of the seminal vesicles, sonograms showed abnormal seminal
STAGING PROSTATIC CANCER
229
Fig. 8 Endorectal MR scan (Courtesy Dr H. I~I. Pollack): T2 cancer in left peripheral zone with no extension beyond the prostatic capsule.
Fig. 7 - TRUS; Axial scan: there is right seminal vesicle invasion. The seminal vesicle is hyperechoic and enlarged.
vesicles in 42 (25%). In 95 patients with histologicallynormal prostates and seminal vesicles, the sonograms showed abnormal seminal vesicles in 11 (12%). The other area that must be carefully assessed sonographically in relation to the prostatic capsule in patients being considered for radical surgery is the apex of the gland. This is best demonstrated by T R U S on the sagittal view where involvement of the rhomboid area formed by the distal urethra, recto-urethralis muscle, rectal wall and apex of the gland can be assessed.
CT SCANNING The prostate appears homogeneous at CT, and CT has been found to be of no value in assessing prostatic cancer confined to the gland. CT may have a limited role in the staging of prostatic cancer, particularly if T R U S and M R is not available. It is not however the optimal investigation. In a comparative study of 27 patients with biopsy-proven carcinoma of the prostate examined with TRUS, CT and MR, T R U S was found to be superior for assessing capsular disruption and M R was found to be superior to both T R U S and CT for evaluating seminal vesicle invasion. MR was also slightly better than CT for detecting lymphadenopathy [22].
MR IMAGING The zonal anatomy and intraglandular texture of the prostate may be well demonstrated with MR imaging and MR has in recent years played an increasing role in staging prostatic cancer. This has been particularly true of patients prior to radical surgery. A comparative study of MR imaging and ultrasound in staging early prostatic cancer was reported by Rifkin et al. [23]. This multicentre study involved five prestigious institutions in North America. The patients were subsequently treated by radical prostatectomy and the surgical specimens were
serially sectioned at 5 mm intervals and examined in their entirety. Two-hundred and thirty patients with clinicallylocalized prostatic cancer were evaluated pre-operatively by both techniques. M R correctly staged 77% of cases of advanced disease and 57% of cases with localized disease; T R U S correctly staged 66% of cases of advanced disease and 46% of localized disease. The figures did not vary significantly between readers, and simultaneous interpretation of MR and T R U S scans did not improve the accuracy. Various reasons were advanced for the disappointing results of this survey in comparison with previous studies but the main reason appears to be that all the patients were thought on the basis of initial assessments to have surgically resectable disease thus minimizing the possibility of including patients with obvious advanced disease which is easier to stage with imaging studies. Additionally in previous studies the entire radical prostatectomy specimens were not always 'step sectioned' histologically to correlate exactly with the imaging. None of the imaging techniques available including TRUS, CT and M R are able to identify microscopic disease. They can only detect macroscopic disease, whether or not there is an attempt to evaluate local disease, periprostatic or seminal vesicle invasion, lymph node involvement or distant spread of the tumour. There have been further improvements in M R recently with the introduction of endorectal surface coils and this technique appears to offer improvements in relation to staging when compared with MR imaging with conventional body surface coils. The endorectal coil improves the signal-to-noise ratio dramatically because of the decreased field of view and the use of thinner (3 mm) slices. The improved spatial resolution gives better delineation of the internal anatomy of the gland and should improve its accuracy in relation to staging. Schnall et al. [24] quote an accuracy of 82% in differentiating Stage B from Stage C disease with endorectal surface coils (Figs 8-10). Two recent further developments, viz. the use of fast spin echo sequences and multi-coil array imaging with one endorectal coil, and a second body surface coil on the anterior or posterior surface of the pelvis should further improve the accuracy of MR. With these techniques a staging accuracy of approximately 90% may be
230
CLINICAL RADIOLOGY
DISTANT M E T A S T A S E S (M CATEGORY )
Fig. 9 - Endorectal MR scan (Courtesy Dr H. M. Pollack): T3 prostatic cancer with extension through the posterior capsule at 6 o'clock position.
Bone scintigraphy is the most sensitive imaging technique for detecting bone metastases of prostatic cancer. In recent years, its role has been assessed in conjunction with serum PSA determinations in relation to the initial staging of bone metastases, and in the follow-up of bone metastases. It may be that isotope scintigraphy is not justified in the initial staging assessment of patients with newly diagnosed prostatic cancer if the PSA level is below 20 ng/ml. Chybowski et al. [26] retrospectively reviewed 521 randomly chosen patients with newly diagnosed untreated prostatic cancer. The median serum PSA concentration in patients with a positive bone scan was 158 ng/ml, whereas men with a negative bone scan had a median serum PSA level of 11.3 ng/ml. The 99.7% negative predictive value of a low serum PSA concentration for bone scan findings was particularly noted. In 306 men with a serum PSA level of 20 ng/ml or less, only one patient (PSA 18.2 ng/ml) had a positive bone scan. Currently unpublished observations by our group would support this view that staging bone scintigraphy in a previously untreated prostatic cancer patient with a low PSA may not be necessary. CONCLUSION
Fig. I0-Endorectal MR scan (Courtesy Dr H. M. Pollack): T3 prostatic cancer with invasion of the right seminal vesicle laterally.
achieved in differentiating Stage B and C disease (Pollack HM, personal communication). However the accuracy of this technique in a group comparable to the study of Rifkin et al. [23] is not yet available.
It is likely that the investigation and assessment of patients with prostatic cancer will become an increasingly common source of uroradiological work. Recent experience has confirmed the importance of serum PSA measurements in the urological diagnosis of prostatic disease, and an understanding of these measurements will be important for the radiologist involved in the diagnosis and staging of prostatic cancer. T R U S and TRUS-guided biopsies have become accepted methods of diagnosing prostatic cancer, and TRUS provides an acceptable method of staging prostatic cancer differentiating patients who have confined (T1/T2) disease from those with more advanced disease. Improved staging accuracy is necessary prior to radical surgery and those patients considered to have cancer localized to the prostate justify further staging with magnetic resonance. Recent developments in M R techniques should further increase the accuracy of local staging of prostatic cancer. Acknowledgements.The authors wish to thank Dr H. M. Pollack, of the University of Pennsylvania, for providing Figs 8-10.
NODAL STAGING (N CATEGORY) REFERENCES
T R U S has no rote in nodal staging. Both CT and MR may be used to assess the pelvic lymph nodes and most authors claim a similar accuracy for both techniques in detecting lymphadenopathy [22,25]. Nodes, vessels and nerve bundles tend to be more difficult to differentiate with CT. Nodal staging with CT and M R is based on detection of enlarged lymph nodes and microscopic involvement of the lymph nodes cannot be detected. In these circumstances, it has been found that the results of both M R a n d C T in lymph node detection prior to radical surgery are disappointing, and Operative lymph node biopsy is still needed for full assessment prior to radical prostatectomy.
1 Whitmore WF. Hormone therapy and prostatic cancer American Journal of Medicine 1956;21:697-713. 2 Jewett HJ. The present status of radical prostatectomy for Stage A and B prostatic cancer. Urologic Clinics of North America 1975;2:105-124. 3 Wallace DM, Chisholm GD, Hendry WF. Classification for urological tumours. British Journal of Urology 1975;47:1 12. 4 Catalona WJ, Scott WW. Carcinoma of the prostate: a review. Journal of Urology 1978; 119:1-8. 5 Cantrell BB, de Klerk DP, Eggleston JC, Boittnott JK, Walsh PC. Pathological factors that influence prognosis in Stage A prostatic cancer: the influence of extent versus grade. Journal of Urology 1981,125:516-520. 6 Schr6der FH, Cooper EH, Debruyne FMJ, Denis L, Newling DWW, Pavone-Macaluso M e t al. TNM classification of genito-
STAGING PROSTATIC CANCER urinary tumours 1987 position of the E.O.R.T.C. Genitourinary group. British Journal of Urology 1988;62:502 510. 7 Byar DP, Mostofi FK. Veterans Administration Cooperative Urological Research Group (VACURG). Carcinoma of the prostateprognostic features in 208 radical prostatectomies Cancer 1972;30:5-13. 8 Wang MC, Valenzuela LD, Murphy GP, Chu TM. Purification of a human prostate specific antigen. Investigative Urology 1979;17: 159 163. 9 0 o s t e r o n K, Bogdanovicz J, Schroder FH. Evaluation of prostate specific antigen in untreated prostatic carcinoma. European Urology 1989;16:253 257. I0 Stamey TA, Kabalin JN. Prostate specific antigen in the diagnosis and treatment of adenocarcinoma of the prostate. 1. Untreated patients. Journal of Urology 1989;141:1070 1075. 11 Clements R, Etherington RJ, Griffiths G J, Peeling WB, Hughes H, Penney MD. The interrelation of serum prostatic specific antigen measurement and transrectal ultrasound in the diagnosis of benign prostatic hyperplasia and prostatic cancer. British Journal of Urology 1992;70:183 187. 12 Chan DW, Bruzek D J, Oesterling JE, Rock RC, Walsh PC. Prostate specific antigen as a marker for prostatic cancer: a monoclonal and polyclonal immuno-assay compared. Clinical Chemistry 1987; 33:1916-1918. 13 Hudson MA, Bahnson RR, Catalona WJ. Clinical use of prostatic specific antigen in patients with prostate cancer. Journal of Urology 1989; 142:1011-1017. 14 Lange PH, Ercole CJ, Lightner D J, Fraley EE, Vessella R. The value of serum prostate specific antigen determinations before and after radical prostatectomy. Journal of Urology 1989; 141:873 879. 15 Partin AW, Carter HB, Chan DW, Epstein JI, Oesterling JE, Rock RC, Weber JP, Walsh PC, Prostate specific antigen in the staging of localised prostate cancer: influence of turnout differentiation, tumour volume, and benign hyperplasia. Journal of Urology 1990;143:747 753. 16 Griffiths GJ, Clements R, Jones DJ, Roberts EE, Peeling WB, Evans KT. The ultrasound appearances of prostatic cancer with histological correlation. Clinical Radiology 1987;38:219 227.
231
17 Dahnert WF, Hamper UM, Eggleston JC, Walsh PC, Sanders RC.Prostatic evaluation by transrectal sonography with-histopathologic correlation: the echopenic appearance of early carcinoma. Radiology 1986;158:97-102. 18 McNeal JE, Redwine EA, Freiha FS, Stamey TA. Zonal distribution of prostatic adenocarcinoma: correlation with histologic pattern and direction of spread. American Journal of Clinical Pathology 1988;12:897 906. 19 Torp-Pedersen S, Lee F. Transrectal biopsy of the prostate guided by transrectal ultrasound. Urologic Clinics of North America 1989; 16:703-712. 20 Lee F, Torp-Pedersen ST, Siders DB, Littrup PJ, McLeary RD. Transrectal ultrasound in the diagnosis and staging of prostatic carcinoma. Radiology 1989;170:609 615. 21 Terris MK, McNeal JE, Stamey TA. Invasion of the seminal vesicles by prostatic cancer: detection with transrectal sonography. American Journal of Roentgenology 1990;155:811 815. 22 Friedman AC, Seidmon JE, Radecki PD, Lev-Toaff A, Caroline DF. Relative merits of MRI, transrectal endosonography and CT in diagnosis and staging of carcinoma of the prostate. Urology 1988;31:530-537. 23 Rifkin MD, Zerhouni EA, Gatsonis CA, Quint LE, Paushter DM, Epstein JI et al. Comparison of magnetic resonance imaging and ultrasonography in staging early prostate cancer: results of a multiinstitutional co-operative trial. New England Journal of Medicine 1990;323:621 626. 24 Schnall MD, Imai Y, Tomaszewski J, Pollack HM, Lenkinski RE, Kressel HY. Prostate cancer: local staging with endoreetal surface coil MR imaging. Radiology 1990;178:777 780. 25 Hricak H, Dooms GC, Brook Jeffrey R, Avallone A, Jacobs D, Benton WK et al. Prostatic carcinoma: staging by clinical assessment, CT, and MR imaging. Radiology 1987;162:331-336. 26 Chybowski FM, Larson Keller J J, Bergstrahl E J, Oesterling JE. Predicting radionuclide bone scan findings in patients with newly diagnosed, untreated prostate cancer: prostate specific antigen is supcrior to all other clinical. Journal of Urology 1991;145:313 318.
