0022-5347/88/1404-0758$2.00/0 THE JOURNAL OF UROLOGY
Vol. 140, October Printed in U.S.A.
Copyright © 1988 by The Williams & Wilkins Co.
TRANSRECTAL ULTRASONOGRAPHY IN THE DIAGNOSIS AND STAGING OF CARCINOMA OF THE PROSTATE GERALD L. ANDRIOLE,* LOUIS R. KAVOUSSI, RALPH J. TORRENCE, HERBERT LEPOR AND WILLIAM J. CATALONA From the Division of Urologic Surgery, Washington University School of Medicine, St. Louis, Missouri
ABSTRACT
Transrectal prostatic ultrasonography is a potentially valuable means to evaluate the prostate of men with suspected carcinoma. We studied 118 patients with this modality before histological evaluation of the prostate (20 underwent radical prostatectomy, 75 core needle biopsy and aspiration cytology, and 23 transurethral resection of the prostate). Transrectal ultrasonography was more efficient than digital rectal examination in the staging of carcinoma of the prostate before radical prostatectomy. The value of transrectal ultrasonography in the diagnosis of prostatic cancer in men with an abnormal-feeling prostate on digital rectal examination is less certain, since 10 of 75 patients (13 per cent) in this group had a falsely positive scan. The predictive value of a scan positive for malignancy was 37 per cent. Further refinements in the technique of transrectal prostatic ultrasonography are needed to realize fully the diagnostic potential of this imaging modality. (J. Urol., 140: 758-760, 1988) Transrectal prostatic ultrasonography has been reported to be a valuable means to evaluate men with suspected or histologically proved carcinoma of the prostate. 1- 7 The ability of transrectal ultrasonography to identify and stage prostatic malignancies has been anticipated to be superior to that of digital rectal examination and other forms of prostatic imaging [pelvic computerized tomography (CT) and magnetic resonance imaging] owing to the fact that most prostatic tumors arise within the posterior lobe of the prostate and are hypoechoic with respect to the usual echographic pattern of the normal posterior lobe of the prostate. 7 Besides the expected enhanced sensitivity in the diagnosis of prostate cancer, ultrasonography may be superior to other forms of prostatic imaging because of its ease of performance in an office setting, reproducibility, low cost and generally acceptable patient compliance. We evaluated the usefulness oftransrectal prostatic ultrasonography in 3 specific clinical settings: 1) the staging of biopsy proved prostatic cancer before radical retropubic prostatectomy, 2) evaluation of men with an abnormal-feeling prostate on digital rectal examination and 3) evaluation of men with symptoms referable to the genitourinary system plus benign-feeling prostates. MATERIALS AND METHODS
Between July 1985 and March 1987, 224 patients at our medical center were evaluated with transrectal prostatic ultrasonography. Indications for the examination and the number of patients in each group with histological followup are shown in table 1. All ultrasonographic studies were performed in the office by a urologist formally trained in transrectal ultrasonography with a Bruel & Kjaer Model 1846 scanner equipped with a 5.5 mHz. axial transducer. All scans were performed with the patient in the left lateral decubitus position. Real-time images were viewed by more than 1 observer and 8 static images, including views of the seminal vesicles, ejaculatory ducts, prostate, apex of the prostate and any abnormal areas, were photographed. All scans were interpreted prospectively by the urologist (G. L. A. and H. L.) who performed the study. The major sonographic criterion of malignancy was the presence of a hypoechoic lesion Accepted for publication February 5, 1988. * Requests for reprints: Division of Urologic Surgery, Washington University School of Medicine, 4960 Audubon Ave., St. Louis, Missouri 63110.
in the peripheral zone of the prostate as described by Lee and associates. 5 Other sonographic findings, such as capsular asymmetry, hyperechogenicity of the peripheral zone or the presence of anechoic lesions peripherally, were not considered indicative of malignancy especially in patients who had undergone a prior prostatic operation or biopsy. A total of 118 patients underwent subsequent histological evaluation of the prostate gland: 20 via radical retropubic prostatectomy, 75 via needle aspiration cytology and core needle biopsy, and 23 via transurethral resection of the prostate. The histological and ultrasonographic findings in these 118 patients form the basis of this report. Patients with clinically localized prostatic cancer underwent nerve-sparing radical retropubic prostatectomy and bilateral pelvic lymphadenectomy. As part of the pathological evaluation a detailed description of the shape and size of the prostate, urethra, seminal vesicles, vasa deferentia and lymph nodes was recorded along with their respective weights. The surgical margins of the entire prostate and seminal vesicles were marked with india ink and immersed transiently in Bouin's fixative. This method affixes the ink permanently to the irregular surface of the extirpated prostate and ensures that no ink travels to tissue planes not representative of the true surgical margins. The entire prostate was oriented and sectioned at 5 to 7 mm. intervals. The location of the tumor and its involvement in each of the prostatic lobes were recorded, as well as the size, color and borders of the lesion. The extent of capsular and periprostatic extension if any also was noted. The urethral and bladder neck surgical margins were examined in detail for tumor involvement and they were sampled before immersion fixation. The seminal vesicles also were sectioned en bloc with the prostate and examined for involvement by tumor. After gross examination the entire prostate was immersion fixed in 10 per cent neutral buffered formalin overnight at room temperature. Approximately 10 to 20 whole mount slices of the prostate were obtained, including contiguous sections of the seminal vesicles. The urethral and bladder neck margins as well as all lateral, anterior and posterior surgical margins were sampled extensively. All calculations of sensitivity, specificity and predictive values were performed with formulas described by Wasson and associates. 8 Sensitivity was calculated by dividing the true positive values by the sum of the true positive and false negative
758
TRANSRECTAL ULTRASONOGRAPHY IN CARCINOMA OF PROSTATE TABLE 1.
Indication for transrectal prostatic ultrasonography
Staging prostate Ca before radical prostatectomy Evaluation of abnormal-feeling prostate Evaluation of benign-feeling prostate Nonprostatic indications Totals
Total No.
No. With Histological Followup
24
20
118
75
65
23
17 224
4 122
values. Specificity equals the true negative values divided by the sum of the true negative and false positive values. The predictive value of a sonogram positive for malignancy is calculated by dividing the number of true positive by the sum of the true positive and false positive values, and reflects the percentage of positive scans that were truly positive. Similarly, the predictive value of a negative scan refers to the percentage of negative scans that were truly negative and is calculated by dividing the true negative by the sum of the true negative and false negative values. RESULTS
Staging of prostatic cancer before radical prostatectomy. A total of 24 patients with clinically localized prostatic cancer (on the basis of digital rectal examination, serum acid phosphatase, bone scan and, also in 19 of 24 patients, a CT scan) underwent ultrasonography before intended radical retropubic prostatectomy. Of these patients 20 ultimately underwent removal of the prostate, since 4 at exploration were not appropriate candidates for radical prostatectomy (3 because of positive pelvic lymph nodes and 1 with a history of 2 transurethral resections of the prostate because of dense adherence of the prostate to the rectum and levator ani muscles). Of these 20 patients ultrasonography was performed before biopsy in 6 and 0 to 2 months after biopsy in 14. When these 20 patients were evaluated for the presence or absence of extracapsular extension, transrectal ultrasonography showed a sensitivity and specificity of 57 and 100 per cent, respectively (table 2). In all 4 patients whose sonogram showed extracapsular extension this condition was documented histologically (the predictive value of a sonogram suggesting extracapsular extension was 100 per cent), while in 3 whose sonograms did not show extracapsular extension the condition actually was confirmed histologically (the predictive value of a scan negative for extracapsular extension was 81 per cent). In 13 patients sonograms showed truly localized prostatic cancer as confirmed by histological findings. The over-all efficiency (the number of truly positive and truly negative scans divided by the total number of patients scanned) of transrectal ultrasonography in the local staging of prostate cancer was 85 per cent. 8 Prostatic biopsy before ultrasonography did not appear to affect adversely our ability to stage the tumor locally, since none of these patients had a false positive sonographic diagnosis of capsular extension or seminal vesicle invasion. Evaluation of the abnormal-feeling prostate before biopsy. Of 118 patients with an abnormal-feeling prostate referred for transrectal ultrasonography 75 ultimately underwent prostatic biopsy combined with aspiration cytology. Of these 75 patients 15 had had a prior needle biopsy (12) or transurethral prostatectomy (3) 9 months to 2 years before ultrasonography. Six patients had sonograms that were truly positive for malignancy, while 10 scans were falsely positive. Of the 59 sonograms that were negative for malignancy 56 were truly negative and 3 were falsely negative. The sensitivity and specificity of transrectal ultrasonography in the identification of prostatic cancer in this group are 67 and 85 per cent, respectively. Over-all, the predictive value of a scan indicating prostatic cancer was only 37 per
TABLE 2.
759
Transrectal ultrasonographic staging of prostate cancer before radical prostatectomy Extracapsular Extension on Sonography
Histological Findings
Pos. Pos. extracapsular extension Neg. extracapsular extension
Neg.
4
3
0
13
cent, while the predictive value of a scan negative for prostatic cancer was 95 per cent. If one considers only the 15 patients with prior needle biopsy or transurethral prostatectomy 4 sonograms (26 per cent) were falsely positive for malignancy. This subgroup of patients with prior prostatic biopsy, which represents only 20 per cent of the 75 patients with an abnormalfeeling prostate, accounted for 40 per cent (4 of 10) of the falsely positive scans. If these 15 patients are excluded the specifity of prostatic ultrasonography in the diagnosis of prostate cancer in patients without prior prostatic biopsy becomes 88 per cent and the predictive value of a scan positive for malignancy is 50 per cent. Evaluation of the benign-feeling prostate. A total of 65 patients with a benign-feeling prostate gland underwent transrectal ultrasonography: 7 had a history of stage Al prostate cancer diagnosed by transurethral prostatectomy 3 to 6 months before ultrasonography, 49 had clinical evidence of benign prostatic hypertrophy and 9 were asymptomatic men more than 60 years old who desired screening for prostate cancer. Of the 7 patients with stage A prostatic cancer 6 subsequently underwent either diagnostic transurethral resection (4) or core needle biopsy (2) of the prostate 4 months (5) or 6 months (1) after the initial diagnosis of carcinoma. Among these 6 patients 1 sonogram was interpreted as showing residual malignancy, which was confirmed on subsequent biopsy. The remaining 5 sonograms showed no evidence of persistent malignancy. However, 1 of these patients had residual carcinoma of the prostate documentated on transurethral biopsy. The remaining 4 patients remained free of disease for 10 to 17 months after initial diagnosis. For patients with a history of stage Al prostate cancer the sensitivity and specificity of ultrasonography in identifying residual carcinoma were 50 and 100 per cent, respectively. The predictive value of a positive sonographic finding of malignancy was 100 per cent, while the predictive value of a scan negative for residual malignancy was 80 per cent. A total of 49 patients with a clinical diagnosis of benign prostatic hypertrophy, none of whom had undergone prior prostatic biopsy, underwent transrectal ultrasonography. Of these patients 23 subsequently underwent transurethral resection of the prostate. None of the 23 patients had a sonogram that suggested prostatic cancer. One patient had a focal, well differentiated tumor (stage Al) on histological examination of the resected specimen. For patients with a clinical diagnosis of benign prostatic hypertrophy the sensitivity and specificity of transrectal ultrasonography in the identification of stage A prostatic cancer were O and 100 per cent, respectively. The predictive value of a sonogram positive for malignancy is unknown (there were no positive scans among these patients) and the predictive value of a scan negative for malignancy was 97 per cent. None of the 9 asymptomatic patients who desired screening for prostatic cancer had a sonogram suggestive of malignancy and none subsequently has undergone prostatic biopsy. DISCUSSION
The ultimate role of transrectal prostatic ultrasonography in the evaluation of patients with prostatic abnormalities is unclear. Theoretically, the facts that the majority of prostatic cancers arise in the posterior lobe of the prostate and that most are distinctly hypoechoic with respect to the normal tissues of
760
ANDRIOLE AND ASSOCIATES
Comparison of studies evaluating the ability of transrectal ultrasonography to identify extracapsular extension before radical prostatectomy
TABLE 3.
Pontes and Associates 10
Salo and Associates 11
Present Study
true pos. true neg. false pos. false neg.
13 7 2 9
12 16 1 2
4 13 0 3
% sensitivity % specificity
59 78
86 94
57 100
% predictive value of pos.
87
92
100
test % predictive value of neg. test
44
89
81
No. No. No. No.
the posterior lobe of the prostate would lead one to expect that this test would provide a relatively high sensitivity and specificity for the diagnosis and staging of prostate tumors. However, early experience has suggested that any process that alters the prostate (such as a prior prostatic operation, biopsy or the presence of prostatic inflammation) may distort the sonographic homogeneity of the posterior zone and thereby preclude identification of some small prostatic tumors, thus lowering the over-all sensitivity of transrectal ultrasonography as a diagnostic test for prostatic cancer. Furthermore, some nonneoplastic processes, such as cysts or the ejaculatory ducts, or certain periprostatic structures, s.uch as the external sphincter muscle or dilated periprostatic capsular veins, can exhibit a hypoechoic appearance on sonography that potentially can be confused with cancer. Therefore, these processes can potentially lower the over-all specificity of transrectal ultrasonography in the diagnosis and staging of prostate cancer. Because of these possibilities the actual sensitivity and specificity of transrectal ultrasonography in the identification of prostate cancer are uncertain. Our study suggests that transrectal prostatic ultrasonography as a diagnostic tool for prostate cancer currently may not be markedly superior to digital examination of the prostate. In patients with an abnormal-feeling prostate the predictive value of a sonogram positive for malignancy was only 37 per cent in the total group of patients and 50 per cent in patients without a history ofprostatic biopsy or surgery. This is not significantly different from the percentage of patients who would be expected to have prostatic cancer if core needle biopsy were performed on the basis of a suspicious-feeling prostate examination alone. Whether the predictive value of a sonogram positive for malignancy would have been greater if ultrasonographically guided biopsy or aspiration cytology9 was performed is unknown. However, all patients in this group had 2 histological cores taken from each side of the prostate after needle aspiration cytology of each lobe had been performed. Furthermore, the urologist was aware of the location of the sonographic abnormality and special efforts to biopsy that area were made. Other possible explanations for the rather poor diagnostic performance of transrectal ultrasonography include observer inexperience and/ or transducer selection. It is possible that additional scans performed with a longitudinal probe or with an axial scanner of a different frequency (for example 7 mHz.) might have decreased the number of falsely positive scans and, thus, improved the predictive value of a sonogram positive for malignancy. Transrectal ultrasonography appears to be superior to digital
examination of the prostate in the staging of proved prostate cancer in patients who are candidates for radical prostatectomy. All 20 of our patients who underwent radical prostatectomy were believed to have truly localized prostatic cancer on digital examination. Four patients had sonographic evidence of capsular disruption, while 7 ultimately had histologial evidence of extracapsular extension. Since none of our patients had digital, sonographic or histological evidence of seminal vesicle involvement the ability of ultrasonography to identify that condition cannot be determined. Our results in regard to the ability of transrectal ultrasound to identify extracapsular extension are similar to those of Pontes and associates, 10 who found that transrectal ultrasonography had a 59 per cent sensitivity and a 78 per cent specificity rate for extracapsular extension. In their series the predictive value of a scan positive for extracapsular extension was 87 per cent. A recent report by Salo and associates also noted an 86 per cent sensitivity and a 94 per cent specificity rate for ultrasonography in identifying extracapsular extension before radical prostatectomy (table 3). 11 Further investigations of transrectal prostatic ultrasonography are necessary to improve its clinical use in the evaluation of prostates with suspected carcinoma. Image enhancement via digital computer reconstruction, colorization, scanning in multiple planes with various frequencies and improved techniques of ultrasound guided needle aspiration cytology and biopsy may be helpful to realize fully the potential of this imaging modality. REFERENCES 1. Peeling, W. B., Griffiths, G. J., Evans, K. T. and Roberts, E. E.: Diagnosis and staging of prostatic cancer by transrectal ultrasonography: a preliminary study. Brit. J. Urol., 51: 565, 1979. 2. Resnick, M. I., Willard, J. W. and Boyce, W. H.: Transrectal ultrasonography in the evaluation of patients with prostatic carcinoma. J. Urol., 124: 482, 1980. 3. Peeling, W. B. and Griffiths, G. J.: Imaging of the prostate by ultrasound. J. Urol., 132: 217, 1984. 4. Fujino, A. and Scardino, P. T.: Transrectal ultrasonography for prostatic cancer: its value in staging and monitoring the response to radiotherapy and chemotherapy. J. Urol., 133: 806, 1985. 5. Lee, F., Gray, J. M., McLeary, R. D., Lee, F., Jr., McHugh, T. A., Solomon, M. H., Kumasaka, G. H., Straub, W. H., Borlaza, G. S. and Murphy, G. P.: Prostatic evaluation by transrectal sonography: criteria for diagnosis of early carcinoma. Radiology, 158: 91, 1986. 6. Rifkin, M. D., Friedland, G. W. and Shortliffe, L.: Prostatic evaluation by transrectal endosonography: detection of carcinoma. Radiology, 158: 85, 1986. 7. Diihnert, W. F., Hamper, U. M., Eggleston, J.C., Walsh, P. C. and Sanders, R. C.: Prostatic evaluation by transrectal sonography with histopathologic correlation: the echogenic appearance of early carcinoma. Radiology, 158: 97, 1986. 8. Wasson, J. H., Sox, H. C., Neff, R. K. and Goldman, L.: Clinical prediction rules. Applications and methodological standards. New Engl. J. Med., 313: 793, 1985. 9. Lee, F., Littrup, P. J., McCleary, R. D., Kumasaka, G. H., Borlaza, G. S., McHugh, T. A., Soiderer, M. H. and Roi, L. D.: Needle aspiration and core biopsy of prostate cancer: comparative evaluation with biplanar transrectal US guidance. Radiology, 163: 515, 1987. 10. Pontes, J. E., Eisenkraft, F., Watanabe, H., Ohe, H., Saitoh, M. and Murphy, G. P.: Preoperative evaluation of localized prostatic carcinoma by transrectal ultrasonography. J. Urol., 134: 289, 1985. 11. Salo, J. 0., Kivisaari, L., Rannikko, S. T. and Lehtonen, T.: Computerized tomography and transrectal ultrasound in the assessment of local extension of prostatic cancer before radical retropubic prostatectomy. J. Urol., 137: 435, 1987.