Pathogenesis and Biological Significance of Seminal Vesicle Invasion in Prostatic Adenocarcinoma

0022-534 7 /90/1436-1183$02.00/0 THE JOURNAL OF UROLOGY Copyright© 1990 by AMERICAN UROLOGICAL ASSOCIATION, INC.

Vol. 143, June

Printed in U.S.A.

PATHOGENESIS AND BIOLOGICAL SIGNIFICANCE OF SEMINAL VESICLE INVASION IN PROSTATIC ADENOCARCINOMA ARNAULD A. VILLERS,* JOHN E. McNEAL,t ELISE A. REDWINE, FUAD S. FREIHA THOMAS A. STAMEY

AND

From the Division of Urology, Stanford University School of Medicine, Stanford, California

ABSTRACT

Seminal vesicle invasion and the percentage involvement by cancer of each seminal vesicle were related to cancer volume, quantitative histological grade and presence or absence of lymph node metastases in 243 radical prostatectomy specimens. There were 47 prostates with seminal vesicle invasion. Frequency and extent of seminal vesicle invasion were strongly correlated with cancer volume, with minimal invasion noted in only 6% of the cases less than 4 cc. The relationship of seminal vesicle invasion to lymph node metastasis was statistically significant but cancer volume and histological grade were much stronger predictors of lymph node metastasis. The route of invasion from the prostate in 46 cases involved direct tumor spread into the midbase region near the ejaculatory ducts. Seminal vesicle invasion often may not be identified if the tissue nearest the ejaculatory ducts at the prostate base is not sampled. (J. Ural., 143: 1183-1187, 1990) Invasion of prostatic adenocarcinoma into the seminal vesicles generally is accepted as an index of poor prognosis. 1 • 2 Several studies have presented evidence that this prognostic association is related to a strong correlation between seminal vesicle involvement and the presence of lymph node metastases. 3· 4 However, others have observed a relatively low frequency of positive nodes in patients with seminal vesicle invasion, and a comparable effect on prognosis among patients with and without lymph node metastases. 6 No alternative explanation has been proposed for the effect of seminal vesicle involvement on survival. Uncertainty about the true biological significance of seminal vesicle invasion may result partly from the fact that its morphological features have never been studied in detail. Previous investigations have not documented the location or extent of cancer within the seminal vesicle, presence or absence of bilateral involvement, or the route by which carcinoma extended from the prostate. Typically, seminal vesicle invasion has not been related systematically to other morphological variables, such as cancer volume and histological grade, with the exception of the current Stanford radical prostatectomy series. 6 • 7 Preliminary observations from our series of prostatectomy specimens have indicated that carcinoma usually enters the seminal vesicle by growing toward and through the prostate base within the muscular wall of the ejaculatory duct. Therefore, the earliest and most extensive involvement by cancer typically is localized within the seminal vesicle wall near its point of junction with the vas deferens. This information conflicts with the description of Jewett and associates, which refers to cancer cells in the areolar tissue around the seminal vesicle. 1 Our quantitative data on cancer volume indicate that this pattern of seminal vesicle invasion is restricted to advanced tumors of larger than 12 cc that have invaded extensively outside the prostate capsule.a Minute foci of carcinoma that we have identified in the seminal vesicle wall as small as 1 or 2 cc would be expected to have different prognostic implications than the cases reported by Jewett and associates. The widely varying findings among different series undoubtedly reflect

undocumented differences in definition, observation and techniques of tissue sectioning. We elaborated on our previous quantitative observations about the natural history of seminal vesicle invasion in prostate cancer and applied them to a series that currently numbers 243 radical prostatectomies. We sought more precise information on the routes of spread from the prostate and correlated the extent of seminal vesicle involvement with frequency of lymph node metastases while making correction for cancer volume and quantitative histological grade. In addition, we considered separately those tumors that appeared to have originated in the transition zone. A substantial minority of prostate cancer cases arise within this anteromedial region, which is the site of origin of benign nodular hyperplasia. 9• 10 Previous studies have indicated that the transition zone boundary is responsible for differences in the patterns of local spread within and outside the prostate between transition zone and nontransition zone tumors. 10 These differences appear to include a relative immunity of transition zone carcinoma to the development of seminal vesicle invasion.a MATERIALS AND METHODS

Cases were selected from a series of 261 radical prostatectomies performed at our university medical center between January 1985 and March 1989. There were 18 cases excluded from the study because of prior radiation or hormonal therapy. The remaining series consisted of 243 radical retropubic prostatectomy specimens from patients 38 to 79 years old. Of these patients 20 had carcinoma demonstrated to be metastatic to the lymph nodes on permanent sections: in 17 the nodes had been negative by frozen section and the deposits were small to microscopic in extent, while 3 had small nodal metastases on frozen section but underwent prostatectomy because of local obstructive problems. Demonstration of nodal metastases at frozen section ordinarily is a contraindication for prostatectomy. Immediately after removal each specimen of prostate with the attached seminal vesicles was processed by the Stanford technique as described previously.a-10 Briefly, after surface Accepted for publication December 6, 1989. marking with dye and overnight fixation in undiluted (37%) Supported in part by The Richard M. Lucas Foundation. formalin, the prostate was sectioned at 3 mm. intervals in * Current address: Urologie, Hospital Bichat, 46 Rue Henri Huchard, transverse planes perpendicular to the rectal surface. Seminal 75018 Paris, France. t Requests for reprints: Division of Urology-S287, Stanford Univer- vesicles with vas deferens also were removed by a transverse sity Medical Center, 300 Pasteur Dr., Stanford, California 94305-5118. section taken as close as possible to the prostate base. Then, 1183

I

1184

VILLERS AND ASSOCIATES

the next transverse section, which represented the prostate base with remaining seminal vesicle tissue, was taken at a thickness of 4 to 5 mm. rather than the standard 3 mm. Then, this block of tissue, showing the prostate base and seminal vesicle, was serially subsectioned in vertical parasagittal planes spaced 3 mm. apart. In this manner the basal surface of the prostate and its relationship to the adjacent seminal vesicle tissue were optimally visualized. In some earlier specimens when this procedure had not been followed only a thin transverse section was available at the base and this often required resectioning at deeper levels to show all of the relationships. Each seminal vesicle was bisected in a coronal plane and half was submitted for microscopic study. Often, the distal tip of the seminal vesicle was amputated to accommodate the dimensions of the microscope slide. This small distal portion was not examined microscopically unless indicated by the gross or microscopic appearance of the main section. Each seminal vesicle slide showed the cut margin, which was in immediate proximity to the seminal vesicle tissue demonstrated on the prostate base sections. The outlines of carcinoma on all slides were marked in ink and transferred by tracing to a comprehensive map of the entire cancer and prostate as described previously. 8- 1° Cancer volume was calculated by computer planimetry, and the proportion of poorly differentiated elements was estimated and tabulated as a percentage of Gleason grades 4 and/or 5 elements. From the map and slides the relationship of each cancer to the transition zone boundary was determined and the tumor was classified according to whether it appeared to be of transition zone or nontransition zone origin as described previously. 10 Cancer in each seminal vesicle was carefully outlined in ink on the slide during microscopic inspection. Then, the percentage of each seminal vesicle involved by cancer was estimated by gross inspection of the outlined area of cancer on the slide in comparison to the total area of the seminal vesicle tissue on the slide. Percentage of cancer involvement was measured rather than absolute cancer volume because the relatively simple flat elongated contour of the average seminal vesicle simplified the process of estimating percentage of involvement, and the proportional number was believed to convey a more concrete visual image to the reader. For data analysis the percentage figures for the 2 seminal vesicles were added and the sums were categorized on an exponential scale (1 to 2, 2 to 4, 4 to 8, 8 to 16 and 16 to 32%, and so forth). Categories then were combined according to inspection of the data to yield a more simplified data display. The location of cancer in the seminal vesicle was noted and related to the presence of cancer in the prostate base section and other transverse prostate sections near the base. For purposes of more precise quantitation, a region surrounding the junction of seminal vesicles and vas deferens with the prostate base was defined and referred to as the midbase region of the prostate. The midbase region was defined as a block of prostatic tissue extending 1 cm. distally from the prostate base, 0.5 cm. to either side of the midline and anteroposteriorly from the rectal surface to the anterior prostate border. The midbase region would have to be traversed by any carcinoma extending into the seminal vesicles by growing along the ejaculatory ducts. Each cancer was evaluated for its extent and location of in volvement of the midbase region, and its involvement of the ejaculatory duct wall or sheath within that region. RESULTS

Seminal vesicle invasion and zone of cancer origin. Among these 243 cases 205 were of nontransition zone origin, including 201 diagnosed clinically by rectal palpation (stage B, see table). The remaining 4 tumors were not palpable and had been diagnosed by incidental discovery in transurethral resection specimens (stage A). Of 38 transition zone carcinomas 24 were diagnosed from transurethral resection specimens (stage A),

Correlation between zone of cancer origin and clinical stage in 243 radical prostatectomy specimens Zone of Ca Origin Clinical Stage

Transition Zone

Stage A (transurethral resection) Stage B: Ca palpated

Sum of Clinical Stage

Nontransition Zone

24

4

7*

201

7t

0 205

28

215 Ca not palpated Sum by zone

38 * Seven large (greater than 4 cc) anterior (transition zone) cancers had grown

posteriorly to the rectal surface. t Seven small (less than 4 cc) anterior (transition zone) cancers in which a posteriorly projecting benign hyperplastic prostatic nodule led to fortuitous needle biopsy detection.

12 >12 cc

Ix

4

Ix

X X

~

,~

10 X X X X

xi

37

w

~

:::, _.J

0

> a: 4-12 cc

12

w

6

0

z <(

X X

X X X X

X

0

117

~

X= NODES+

<4CC

7 X

0

0 <16%

>16%

S.V. INVASION FIG. 1. Frequency and extent of seminal vesicle (S. V.) invasion (0, less than 16 and greater than 16%) found in each cancer volume range (less than 4, 4 to 12 and greater than 12 cc). Number of cases in each category of cancer volume and seminal vesicle invasion is indicated by relative height of each column and number above it. Each case with lymph node metastases is indicated by X in appropriate column.

while 14 (including 7 large tumors) were palpable, having grown from the transition zone to reach the rectal surface. There also were 7 small transition zone tumors in which a posteriorly projecting benign prostatic hyperplasia nodule was palpated, leading to a fortuitously positive needle biopsy. Seminal vesicle invasion was identified in 47 of the 205 nontransition zone tumors (23%). However, none of the 38 transition zone carcinomas showed any evidence of seminal vesicle invasion and none showed any invasion of the midbase region around the ejaculatory ducts near the prostate base. When the data were analyzed according to the clinical classification the numbers remained the same; that is 47 (22%) of the 215 stage B and none of the 28 stage A tumors showed seminal vesicle invasion. The remainder of our findings are concerned exclusively with the category of nontransition zone carcinoma. Seminal vesicle invasion, cancer volume and lymph node metastasis. Among nontransition zone tumors that were less than 4 cc in volume 6% showed minimal invasion of 1 seminal

PROSTATE CANCER SEMINAL VESICLE INVASION

1185

vesicle (fig. 1). In each case 2% or less of the seminal vesicle tissue was estimated to be involved by cancer. The 3 smallest carcinomas were 1.0, 2.0 and 2.6 cc. The frequency of seminal vesicle invasion increased progressively with volume, being found in 33% of the cases between 4 and 12 cc, and 82% of z Q those larger than 12 cc. The proportion of cases with more than Cl) 16% of the seminal vesicle tissue estimated to be replaced by cancer also increased steadily across the 3 volume ranges. It ~ should be noted that among the tumors of larger than 12 cc ~ X= NODES+ UJ there still were 7 with only 1 % of involvement of seminal Cl) Ifill= SV + 12 cc The relative proportion of poorly differentiated carcinoma CANCER VOLUME (Gleason grades 4 and/or 5) also is known to increase with FIG. 2. Significance of invasion of midbase region for invasion of cancer volume. 6 •7 In our series more than 20% of the Gleason grades 4 and 5 elements already were found in 26% of the seminal vesicle (S V). Height of each column and number above indicate number of cases with or without invasion of midbase region in each of cancers of less than 4 cc. This incidence increased to 53% in 3 volume ranges. Shaded areas represent numbers of cases in that the 4 to 12 cc range and 70% for tumors of more than 12 cc. A column with seminal vesicle invasion. Each case with lymph node possible combined effect of grade and seminal vesicle invasion metastasis is marked by X in appropriate volume range, midbase on metastasis was suggested by the finding that in all volume invasion status and seminal vesicle invasion category. ranges metastasis was noted only if seminal vesicle invasion and/or more than 20% grades 4 and 5 tumor were present. and into the seminal vesicle muscular wall. In 5 of these cases Accordingly, multiple logistic regression was used to test the cancer within the midbase region was confined selectively to predictive value of cancer volume, grade and seminal vesicle the ejaculatory ducts, having invaded their wall near the veruinvasion for nodal metastasis. Each variable alone showed montanum. The remaining 38 cases reached the ejaculatory statistically significant predictive value for metastasis (p ducts by invading through the midbase region indiscriminately <0.001). The log of tumor volume was the strongest predictor along a broad front that also involved the ejaculatory duct wall. and seminal vesicle invasion was the weakest. The best comIn all 43 of these cases the replacement of seminal vesicle bination of predictors was log tumor volume plus grade. No tissue by cancer was quantitatively greatest at the proximal increase in predictive value was produced by adding seminal (prostatic) end, in the region closest to the junction of the vesicle invasion to either tumor volume or volume plus grade. seminal vesicle and vas deferens. Of these cases 3 also showed Unilateral versus bilateral seminal vesicle invasion. The fre- a second, independent smaller area of invasion by cancer into quency and extent of bilateral seminal vesicle invasion showed the tip (distal end) of the seminal vesicle. These were 3 prosprogressive increase with cancer volume in the same manner tates in which the end of the seminal vesicle farthest from the as unilateral (ipsilateral) seminal vesicle invasion. However, vas deferens was embedded in the muscular wall of the prostate the smallest of the 20 cancers with bilateral invasion was 6 cc, base far lateral to the midline. This anomaly allowed for invaas opposed to 1 cc for the smallest of the 27 tumors with sion of the seminal vesicle near the tip without extension of unilateral seminal vesicle invasion. The largest cancer with tumor outside the prostate. unilateral seminal vesicle invasion was 21 cc. Although the Only 4 of 4 7 cases remained in which invasion of the seminal frequency and extent of bilateral involvement increased stead- vesicle depended upon direct penetration of the prostate capily with a volume of greater than 6 cc, the mean average sule. These cases demonstrated 3 different routes: 1) 2 showed percentage of the seminal vesicle tissue replaced by cancer on invasion directly through the prostate base and then through the contralateral side always was approximately half the extent the external surface of the seminal vesicle into its muscular of invasion in the seminal vesicle that was ipsilateral to the wall (both penetrated the prostate base in the midbase region main tumor. Among 45 tumors of 6.0 to 15.9 cc in volume there and entered the seminal vesicle wall near the junction with the were 10 with bilateral seminal vesicle invasion (mean average vas deferens), 2) 1 showed penetration of cancer from the ipsilateral involvement was 24%, while mean involvement of midbase region directly posteriorly into Denonvilliers' fascia the contralateral side was 10%). Of 14 lesions that were 16 cc (within the fascia it had extended superiorly and then grew or larger 10 had bilateral seminal vesicle invasion (mean ipsi- anteriorly again into the adjacent seminal vesicle wall), and 3) lateral involvement was 60% and mean contralateral involve- 1 showed a solitary tiny focus of cancer embedded in the outer ment was 33%). In cases of unilateral and bilateral invasion layers of the seminal vesicle wall and also involving the perialike there was consistently an orderly pattern of tumor spread vesicular tissue. The focus was remote from any connection from the region of junction with the vas deferens (adjacent to with the primary tumor, which was a poorly differentiated the prostate base) toward the tip (distal end) of the seminal carcinoma larger than 12 cc in volume. vesicle. Thus, only 1 of 47 tumors did not involve the midbase region Routes of invasion into the seminal vesicle. In 43 of the 47 preliminary to invading the seminal vesicles. In 13 cases cancer tumors with seminal vesicle invasion cancer had invaded invaded into the midbase region but the disease had not prothrough the midbase region to reach the sheath of the ejacula- gressed to invasion of the seminal vesicles (fig. 2). There were tory ducts within the prostate and near to its base (fig. 2). In no morphological features that distinguished these cases from all but 1 of these cases cancer traversed the ejaculatory duct the cases with positive seminal vesicles. Of these tumors 6 had sheath and penetrated its muscular wall. In these 43 cases invaded the ejaculatory duct or its sheath, and 1 had lymph cancer was traced in continuity along the ejaculatory duct wall node metastases.

0

I 11

-"T.

1186

VILLERS AND ASSOCIATES

Role of Denonuilliers' fascia and capsule penetration in cancer spread. In 38 of the 59 cases with invasion of the midbase region tumor penetrated posteriorly and invaded into Denonvilliers' fascia near the midline. This pattern of spread from the midbase was noted in 36% of the cases with less than 12 cc but 72% of those with greater than 12 cc tumor volume. In only 1 of these cases, reported previously, cancer spread superiorly within the fascia and then invaded from the fascia into the seminal vesicle above the prostate base. In the remaining cases cancer sometimes was noted adjacent to the seminal vesicle in sections that also showed Denonvilliers' fascia. Among tumors of larger than 12 cc extensive capsule penetration at or near the base sometimes led to invasive cancer in fibrofatty tissue adjacent to the seminal vesicles outside Denonvilliers' fascia. This finding has been considered in a previous study on capsule penetration8 and involvement of tissues around the seminal vesicle, whether from capsule penetration or within Denonvilliers' fascia, was not tabulated in that study. For tumors of larger than 12 cc there were cases in which concomitant invasion of the seminal vesicles, prostate capsule at the base and Denonvillier's fascia coalesced into a single tumor extending above the prostate base. DISCUSSION

Seminal vesicle invasion typically is recorded as a dichotomous variable in studies relating to the natural history of prostatic carcinoma. In reality, it represents a continuum and the histologically determined extent of seminal vesicle invasion, like cancer volume and histological grade, is a morphological reflection of biological tumor progression. Among these 3 indexes of tumor progression, however, we have found that seminal vesicle invasion is the weakest predictor of cancer dissemination as measured by the presence of lymph node metastases. Multivariate logistic regression analysis on these 243 cases showed that if cancer volume or volume plus histological grade is known, extent of seminal vesicle invasion does not further increase predictive value. In support of this finding, figure 1 shows that within the volume range of 4 to 12 cc 13 of 18 tumors with seminal vesicle invasion did not have positive nodes; while among those greater than 12 cc 11 of 22 with seminal vesicle invasion showed no nodal metastases. In earlier studies when cancer volume and quantitative histological grade were not carefully determined, seminal vesicle invasion stood as a valuable predictive measure simply because of its ease of determination. Furthermore, if only cases with extensive seminal vesicle involvement were recognized the probability of metastasis was high because mainly tumors of larger than 12 cc were identified. These cases tend to be those that are identifiable grossly as a mass above the prostate, which includes cancer in the tissue around the seminal vesicle caused by capsule penetration and transport along Denonvilliers' fascia. Recognition of lesser extent of seminal vesicle invasion (less than 16%) would identify many smaller tumors with favorable prognosis but such cases undoubtedly have been overlooked when tissue at the junction of the seminal vesicle and prostate base was not examined histologically. Since the route of penetration into the seminal vesicle almost invariably traverses the midbase region, the proximity of cancer origin to the midbase probably influences the likelihood of seminal vesicle invasion. This point was demonstrated most clearly in our series by the absence of any seminal vesicle invasion in transition zone (stage A) tumors. It has been shown previously that the local spread within the prostate of these benign prostatic hyperplasia-associated tumors is limited by the boundary of the transition zone, which separates them from the ejaculatory ducts. 10 Since the ejaculatory ducts usually are completely surrounded by central zone tissue it is likely that in nontransition zone cancer only the relatively uncommon tumors of central zone origin would be able to reach the seminal vesicle while they still were as small as 1 cc. At the other end

of the continuum, cancer of more than 12 cc in volume with only 1 % involvement of 1 seminal vesicle may have originated near the apex, remote from access to the ejaculatory ducts. These uncontrolled geographic factors that may influence the probability of seminal vesicle invasion would not logically influence cancer volume or histological grade. This may explain partly why these 2 variables have potentially greater value as predictors of tumor spread to the lymph nodes. These cases represent a somewhat selected series. Except for 3 patients, those with positive lymph nodes demonstrated by frozen section at operation were excluded from radical prostectomy. This represents a large category of patients in whom morphological data on the primary cancer and seminal vesicles are unknown. If the natural history of prostate cancer represents a progressive continuum, one might expect that the morphological features of this group with positive lymph nodes would overlap our cases with positive nodes, even though our cases were biased toward less extensive nodal disease. In support of this expectation, a recent investigation has found little or no difference in recurrence rate and mortality between patients with only 1 positive node and those with extensive nodal metastasis. 11 Among our own patients 13 currently have either recurrent or persistently elevated serum prostate specific antigen and 6 of them have positive bone scans. It should be noted that our finding of the strong association of greater than 16% seminal vesicle invasion with tumors of larger than 12 cc can be translated into a form having more practical value. The majority of seminal vesicles in our series are 2.5 to 4.0 cm. long. Therefore, in a seminal vesicle with less than 16% tumor involvement cancer almost always will be confined to an area near the midline of the prostate base and within 1 cm. of the point where the seminal vesicle enters the prostate. If cancer extends more than 1 cm. along the seminal vesicle toward its tip (distal end) the presence of a tumor of larger than 12 cc is predicted. The validity of this rule of thumb is increased by the observation that most of our cases with more than 16% involvement were widely spread above that value. Therefore, neither the exact 16% value nor variations in size among different seminal vesicles diminished the validity of this finding. Dr. Lincoln Moses performed the statistical analysis of these data. REFERENCES

1. Jewett, H.J., Eggleston, J.C. and Yawn, D. H.: Radical prostatectomy in the management of carcinoma of the prostate: probable causes of some therapeutic failures. J. Urol., 107: 1034, 1972. 2. Walsh, P. C. and Jewett, H.J.: Radical surgery forprostatic cancer. Cancer, 45: 1906, 1980. 3. Middleton, R. G. and Smith, J. A., Jr.: Radical prostatectomy for stage B2 prostatic cancer. J. Urol., 127: 702, 1982. 4. Catalana, W. J., Fleischmann, J. and Menon, M.: Pelvic lymph node status as predictor of extracapsular tumor extension in clinical stage B prostatic cancer. J. Urol., 129: 327, 1982. 5. Mukamel, E., deKernion, J.B., Hannah, J. Smith, R. B., Skinner, D. G. and Goodwin, W. E.: The incidence and significance of seminal vesicle invasion in patients with adenocarcinoma of the prostate. Cancer, 59: 1535, 1987. 6. McNeal, J.E., Bostwick, D. G., Kindrachuk, R. A., Redwine, E. A., Freiha, F. S. and Stamey, T. A.: Patterns of progression in prostate cancer. Lancet, 1: 60, 1986. 7. Stamey, T. A., McNeal, J. E., Freiha, F. S. and Redwine, E.: Morphometric and clinical studies on 68 consecutive radical prostatectomies. J. Urol., 139: 1235, 1988. 8. McNeal, J. E., Villers, A. A., Redwine, E. A., Freiha, F. S. and Stamey, T. A.: Capsular penetration in prostate cancer: significance for natural history and treatment. Amer. J. Surg. Path., in press. 9. McNeal, J. E., Price, H. M., Redwine, E. A., Freiha, F. S. and Stamey, T. A.: Stage A versus stage B adenocarcinoma of the

PROSTATE CANCER SEMINAL VESICLE INVASION prostate: morphological comparison and biological significance. J. UroL, 139: 61, 1988. 10. McNeal, J. E., Redwine, K A., Freiha, F. S. and Stamey, T. A.: Zonal distribution of prostatic adenocarcinoma: correlation with histologic pattern and direction of spread. Amer. J. Surg. Path.,

l187

12: 897, 1988. 11. Gervasi, L. A, Mata, A., Easley, J. D., Wilbanks, J. H., SealeHawkins, C., Carlton, C. E., Jr. and Scardino, P. T.: Prognostic significance of lymph nodal metastases in prostate cancer. J. UroL, 142: 332, 1989.