- Email: [email protected]
Touch Preparation Cytological Evaluation of Radical Prostatectomy Specimens
0022-5347/94/1521-0124$03. 00/0 Vol. 152, 124-128, July 1994 Printed in U.S.A
THE JOURNAL OF UROLOGY Copyright © 1994 by AMERICAN UROLOGICAL AsSOCIATION, INC.
TOUCH PREPARATION CYTOLOGICAL EVALUATION OF RADICAL PROSTATECTOMY SPECIMENS MARGARET S. PEARLE, RYOICHI OYASU, DENISE HIDVEGI, DEBRA SUTKOWSKI, ALFRED RADEMAKER, RICARDO CAJULIS, JOHN T. GRAYHACK AND KEVIN T. McVARY From the Departments of Urology, Pathology and Biometry, Northwestern University Medical School, Chicago, Illinois
ABSTRACT
The failure of current histological techniques to predict local failure and disease progression after radical prostatectomy is supported by substantial evidence. Moreover, the characterization of histological findings is hampered by the lack of uniform interpretation. We report a prospective study of 92 patients undergoing radical prostatectomy for clinical stages A and B prostate cancer in which the technique of touch preparation cytological analysis of surgical margins is compared to the standard histological approach. We evaluated 47 pathological stage B, 43 stage C and 2 stage D specimens. Specimens initially assigned to pathological stage B were upstaged to stage C on review by 1 blinded pathologist in 19 of 65 cases (29% ). Overall, 15 of 4 7 histological stage B specimens (32%), 20 of 43 histological stage C specimens (47%) .and 2 of 2 histological stage D specimens (100%) had malignant cells identified on the margins by touch preparation cytology. Postoperative mean followup of 7 months (range 0.4 to 26) revealed that 7 of 14 nonstage D cancer patients (50%) with elevated serum prostate specific antigen levels had positive cytology results, including 1 with histologically confirmed organ-confined disease. Among the stage C specimens cytology was more likely to be positive if there was concomitant seminal vesicle invasion. Correlation of this information with eventual patient course during the long term' will be necessary to assess its role in patient management. KEY WORDS:
prostatic neoplasms, prostatectomy, histology, cytology, cytological techniques
Urologists rely largely on histological evaluation of prostatic capsular involvement to assess prognosis and for consideration of adjuvant radiation therapy after prostatectomy. Local failure rates of up to 40% for pathological stage B disease attest to our inadequate staging techniques. 1- 4 Whether these failures can be accounted for by missed or misinterpreted capsular perforation by the pathologist, or are due to a more fundamental limitation of histology in predicting residual disease is unclear. In an effort to address this staging inadequacy, we report our preliminary use of margin touch preparation cytological evaluation of radical prostatectomy specimens. Our goal was to improve or supplement current approaches that evaluate the excised specimens and to increase our ability to predict local failure. MATERIALS AND METHODS
a
In prospective study of 92 consecutive patients undergoing radical prostatectomy between August 1990 and February 1992 for clinically localized prostate cancer, we compared histological margins with corresponding cytological touch preparations. Along with a single blinded pathologist (R. 0.), we reexamined all histological sections of the prostate and re-staged each case with careful attention to the margins. Specimens were initially subjected in the laboratory to the touch preparation technique described, and then sent to the pathology department where the margins were stained with ink and the specimens were fixed. Apical (2 mm.) and base (3 to 4 mm.) sections were cut from the specimens transversely, and the entire apex and base were submitted for histological review. Any areas of gross tumor were also entirely submitted for examination. The remaining bulk specimen (prostate gland minus apex, base and areas of gross tumor) was halved and sectioned transversely into 2 to 3 mm. slices, of which 3 Accepted for publication December 17, 1993. Read at annual meeting of North Central Section, American Urological Association, Scottsdale, Arizona, November 3- 9, 1991.
to 4 slices (levels) were examined. Each section to be examined was cut into 4 quadrants and these histological sections were reviewed. Histological sections from each specimen were divided geographically and grouped into anterior, posterior, right lateral, left lateral, apex and base. An average of 24 histological slides were reviewed per case. Pathological stage C disease was defined as tumor on or outside the outer capsular edge (capsular perforation, fig. 1, A), on the inked margin (fig. 1, B) or in the seminal vesicles. Tumor cell penetration into but not through the capsule was considered pathological stage B (fig. 1, C). The shortest distance between the tumor front and the capsule or inked margin within each geographic region was measured with the use of light microscopy and intraoptic micrometer (1 radicle unit = 1.25 µ . at 40X). Margin touch preparation cytology studies were prepared by 2 techniques, a simple touch technique and a scraping technique. In the simple touch technique microscope slides were patted on each geographic area of the specimen and fixed. In the scraping technique the edge of a microscope slide was drawn across the 6 designated areas of the specimen (in a scraping fashion typical of touch preparation technique for lymph nodes) and fixed. These alternate techniques were performed in an attempt to differentiate contamination by prostatic secretions versus tumor cell perforation threugh the capsule. A single blinded cytopathologist (D. H.) reviewed all cytological touch preparation specimens and categorized them as malignant, benign or atypical cells (fig. 2). Inking the margins was not performed before cytological sampling for fear that we would render the cells on the microscope slides as uninterpretable by the cytologist. RESULTS
A total of 92 radical prostatectomy specimens was evaluated by touch preparation cytology. Patient age ranged from
124
l ..
•
TOUCH PREPARATION CYTOLOGICAL EVALUATION OF RADICAL PROSTATECTOMY SPECIMENS
125
FIG. 1. A, tumor cells have perforated through prostate capsule but not reached inked surgical margin. Large arrows demonstrate tumor cells in extracapsular tissue while small arrows mark tumor-free inked surgical margin. B, tumor cells (arrowheads) are noted on surgical margin. C, tumor cells invading into but not through capsule. Large arrows demonstrate prostate tumor cells within prostatic capsule. Small arrows mark outer capsular border as demonstrated by periprostatic fat. Reduced from X 186.
4 7 to 76 years (mean age 65 years). Cytology studies from 37 patients showed malignant cells, 48 benign cells and 7 atypical cells. The distribution of specimens by pathological stage is illustrated in table 1. Review by the referee pathologist resulted in reclassification of 19 stage B (3 stage Bl and 16 stage B2) specimens to pathological stage C. The distribution of specimens by cytology and pathological stage is illustrated in table 2. Overall, 15 of 47 stage B specimens (32%) and 22 of 45 stage C specimens (49%) had positive corresponding cytology results. Clearly, surface cytology smears identify a group of individuals with organ-confined cancer on histological assessment as having a high risk of extra-specimen disease. Conversely, surface cytology testing fails to provide objective evidence of extra-specimen disease in some individuals with histological findings interpreted as evidence of extra-organ or margin-positive disease. Biological assess-
ment of the significance of these discrepant findings depends on identification of individuals with recurrent disease on followup and documentation of the site of the recurrences when they are recognized. Pathological stage C specimens included those with capsular perforation, positive surgical margins and/or positive seminal vesicles. Closer evaluation of the stage C cancer patients revealed that all 12 specimens with involved seminal vesicles had extracapsular or margin-positive disease, including 8 with positive and 4 with negative cytology results. A total of 33 stage C specimens had negative seminal vesicles, including 14 with positive, 15 with negative and 4 with atypical cytology results. These data imply a greater likelihood of positive cytology findings with than without seminal vesical involvement (table 3). Followup information was available from 0.4 to 26 months postoperatively (mean 7 months). Of the patients 18 were
FIG. 2. Cells obtained by scraping method of touch preparation cytology. A , benign prostate cells. B, malignant prostate cells. Reduced from x 1,200.
126
TOUCH PREPARATION CYTOLOGICAL EVALUATION OF RADICAL PROSTATECTOMY SPECIMENS TABLE 1.
Pathological staging Initial* No.(%)
Stage
TABLE
Reviewedt No.(%)
17 (19) 49 (53) 24 (26)
14 (15) 33 (36) 43 (47) 2(2) 2 (2) Dl * Pathological staging as inferred from official pathology report. t Pathological staging as determined by review with referee pathologist (R. 0 .). Bl B2 C
followed for longer than 12 months and 3 were lost to followup. Patients were followed at 3 to 4-month intervals with physical examination and serum prostate specific antigen (PSA) level (Hybritech method) determinations. Bone scans, biopsies and additional evaluation were performed at the discretion of the attending urologist. No patient has had bone scan evidence of metastasis to date. Ten patients (excluding 1 with stage D disease) had persistent elevation of serum PSA (0.3 ng./dl. or greater, normal less than 0.3) postoperatively, of whom 6 (60%) had positive cytology findings. Of these 10 patients 8 underwent local external beam irradiation. Four patients (excluding 1 with stage D disease) had initial postoperative undetectable with subsequent increasing serum PSA levels and negative biopsies of the urethrovesical anastomosis (1 had a positive cytology result). To date, 2 of these patients have received irradiation. Three stage C cancer patients with persistently elevated or increasing serum PSA levels were treated with hormonal or immunotherapy. Both stage D cancer patients had positive cytology findings and were treated with radiation therapy. Overall, of 14 patients with elevated serum PSA levels postoperatively 7 (50%) (excluding the stage D cases) had positive cytology results (table 4). One of 4 patients with elevated postoperative serum PSA levels and pathological stage B disease had positive cytology findings. In this analysis, atypical cytology studies were considered negative to provide the most conservative estimate. We compared the 2 methods of cell sampling to determine if the scraping maneuver was superior to the simple touch technique in detecting malignant cells (table 5). The scraping method demonstrated malignant cells more frequently. Moreover, only 1 of 37 positive specimens was found by the simple touch technique alone, which suggests that a positive cytology result is not related to artifactual contamination by intraoperative spillage of malignant cells via the prostatic secretions but, instead, represents true tumor cell perforation through the prostate capsule. The shortest distance between the tumor and prostatic capsule or inked surgical margin in each geographic section was determined for all 81 cases with sufficient pathological documentation for review. Eleven cases were excluded from distance analysis because we were unable to assign histological sections to a geographic region unequivocally. However, these cases were globally restaged and were included in analyses that did not require geographic localization within the specimen. For pathological stage B cases the distance between the tumor and the outer capsular edge was always
3. Cytological status of stage C specimens
Seminal Vesical/ Extracapsular
Cytology Result Pos.
Neg.
Atypical
Totals
Pos./pos. Neg./pos. Totals
8 4 0 12 14 15 QQ 1 22 19 4 45 p = not significant Positive extracapsular disease represents either tumor at inked margin or tumor perforating through the capsule. Statistical analysis was performed using Fisher's exact test. * Includes 2 cases of positive lymph nodes that were local pathological stage C.
greater than zero. In these cases, however, the mean shortest distance from the tumor to the outer margin was 0.50 mm. for 29 specimens with 73 cytologically negative and atypical smears, and 0.80 mm. for 12 specimens with 31 cytologically positive smears (p = 0.05 by the Wilcoxon rank sum test). Using location-specific cytology, rather than global cytology for the specimen, the mean shortest distance from tumor to the outer capsule was 0.5 mm. in 96 negative and atypical smears (40 patients), and 2.84 mm. in 8 positive smears (4 patients, p = 0.04, Wilcoxon rank sum test). Additionally, there was no correlation between tumor grade and margin cytology status. DISCUSSION
This early report of an attempt to obtain specimen-related information to predict local tumor recurrence clearly shows a discrepancy between cytology and histology findings in assessing the presence of extra-specimen disease. With the limited followup available, the biological significance of these histological and cytological findings cannot be evaluated critically to date. However, we did note a trend in stage C specimens that suggests a greater likelihood of an elevated serum PSA level developing postoperatively in cytologically positive than in cytologically negative cases. The evidence that histological assessment, even with labor-intensive step-sectioning techniques and interpretation by highly skilled individuals, is unable to predict local or systemic recurrence with sufficiently high probability to guide clinical activity is substantive. Epstein stated that there is little objective evidence as to the validity of so-called positive or negative capsular margins ofresection. 5 He called for caution in diagnosing positive margins in an in-depth study of 40 radical prostatectomy specimens in which additional soft tissue was submitted separately for assessment of tumor involvement. Paulson et al presented data indicating an increased probability of local recurrence with positive surgical margins,6 while Gibbons et al reported a 25% local recurrence rate with capsular perforation and a 17% rate with margin-positive disease. 7 The length and methods of followup to assess local recurrence in these groups of patients have been limited in view of current knowledge. Local recurrence despite histologically specimen-confined
Patients with postoperative evaluated PSA classified by stage and cytological status (0.3 µg. I dl. or greater by Hybritech assay)
TABLE 4 . TABLE
2. Distribution of specimens by cytological status
Pathological Stage
Cytology* Pos.
Atypical
Neg.
No./Total Cytology(%) Totals
Bl 3 1 10 14 B2 12 2 19 33 22 19 Ct !Q 1 Totals 37 7 48 92 * Cytology is classified as positive (malignant), atypical or negative (benign). t Includes 2 cases with positive lymph nodes that were local pathological stage C.
Stage
No./Totals Pos.
Neg.
1/15 (7) 3/32 (9) 4/47 6/20 (30) 10/43 4/23 (17) 7/35 (20) 7/55 (13) Totals 14/90 Atypical cytology result was considered negative for this analysis. Stage D specimens were excluded. P value was not significant overall and within each stage. Statistical analysis was performed using Fisher's exact test. B C
TOUCH PREPARATION CYTOLOGICAL EVALUATION OF RADICAL PROSTATECTOMY SPECIMENS TABLE 5 .
Distribution of cytology results by sampling technique
Scraping Technique
Touch Technique
Neg. Pas. Pas. Pas. Pas. Neg. Neg. Neg. Atypical cytology result was considered negative for this analysis.
No.
30 6
1 55
disease is well documented. Schellhammer reported a local failure rate of 17% at 15 years for pathological stage B disease,1 while Jewett et al similarly found a local recurrence rate of 16% at 15 years after prostatectomy in 86 patients with capsule-confined disease. 2 Blute et al, in a large series of patients with localized prostate cancer, found a local recurrence rate of 4.3% when more than half of the patients were followed for more than 5 years. 3 Recently, Lightner et al biopsied the urethrovesical anastomosis in 63 patients with pathological stage B disease with no clinical evidence oflocal recurrence after radical prostatectomy and 42% had a positive biopsy.4 These observations indicate that the prevalence of local recurrence is underestimated. Why tumors recur locally after radical prostatectomy in patients with presumed margin-negative organ-confined disease on standard histological assessment represents a puzzling phenomenon that prompted this study. That patients with margin-positive stage C disease do not uniformly fail locally is similarly unexplained. Identifying patients who have a high probability of local failure compared to those in whom local recurrence is impr obable is a significant problem when considering recommendation for adjuvant therapy. An explanation, of course, could be simple human error. Capsular perforation and positive margins may be misinterpreted by the pathologist either by oversight, by inconclusive objective evidence or because of incomplete examination of the available t issue. As an example of this point, 19 of our 65 pathological stage B specimens (29%) were upstaged by our referee pathologist to stage C. Similarly, Jewett et al found 21 cases of recurrent or persistent cancer after radical prostatectomy, which on initial evaluation included 15 without capsular involvement, 5 with capsular penetration and 1 with capsular perforation.8 Reevaluation of more sections, however, demonstrated no capsular invasion in 6 cases, penetration in 9 and perforation in 6. Presently, the specimen evaluation after radical prostatectomy is done with meticulous care and good experience at several institutions. Evidence tha t these efforts will improve our ability to predict local recurrence remains elusive. Theoretically, local recurrence after radical prostatectomy with pathological stage C or, at least, margin-positive disease should approach 100% on the assumption that residual microscopic disease will eventually become clinically recognizable. Review of the literature, however, reveals a local recurrence rate of only 14 to 43% for pathological stage C disease.1 · 7 • 9 • 10 A major contributing factor t o this apparent broad range is a clear lack of consensus as to what constitutes stage C disease as reported in the literature. Penetration into the capsule, perforation t hrough the capsule and tumor at the surgical margin have often been used interchangeably without distinction between stage C and ma rgin-posit ive disease. Schellhammer reported a higher incidence of local failure for patients with seminal vesical involvement than those with only capsular involvement.1 Our data suggesting a higher likelihood of positive cytology findings in patients with seminal vesica l invasion than in those without correlate well with his clinical observations. The margins of resection from radical prostatectomies performed for prostate malign ancies are difficult to evaluate because of the relatively sparse surrounding soft tissue. Given the observations cited, histology h as clearly fallen short in predicting local control or failure. The cytological status of radical prostatectomy spec-
127
imens provides another objective observation that may potentially identify patients at increased risk of local recurrence. This technique may serve as an independent or an adjunctive observation to the traditional margin analysis. Intuitively, specimens with extracapsular disease (stage C) would be more likely than those with confined disease (stage B) to have a positive cytology result. We cannot explain the negative cytology test in t he face of clear histological margin positivity. Additionally, malignant cells were identified on the surface of glands in which tumor was histologically capsule-confined. The possibility that malignant prostat ic secretions coating the surface of the gland may account for positive touch preparation cytology warrants consideration. 11 We found only 1 instance, however, when cytology studies were positive when the microscope slide simply touched t he surface of the gland and were negative when it was drawn across the specimen, which suggests that the tumor cells adherent to the gland were perforating through the capsule and not simply bathing it (that is the product of intraoperative prostatic secretory contamination). Furthermore, while one might argue t hat more meticulous step-sectioning of the prostate might lead to finding some cases of missed capsular perforation in those with a positive cytology test, we hypothesize that individual tumor cells escape the confines of the capsule long before microscopic tumor encroaches on the margin in some circumstances. Lymphatic spread of tumor either distally or locally might also lead to some cases of missed capsular perforation. While we hypothesized an inverse relationship between tumor distance from the capsule and likelihood of a positive cytology test, t he opposite was found. These proposed renegade tumor cells responsible for positive cytology results clearly rely on factors other than sim ply tumor location to determine the predilection for capsular violation. There is a possibility that we might create false-positive histological margins by our scraping technique but our concern over rendering our cytology slides as uninterpretable prevented us from inking the margins before cytological sampling. Although less than a t otal histological review of a path ological specimen (a monumental task) could account for some cases of positive cytology results in the face of negative histological margins, we believe that our tissue processing has been accepted as able to classify pathological stage accurately.14· 15 The fact that our mean number of slides examined per case was 24 is a measure of our completeness. Additional explan ations for local failure would include occult perineural tu mor invasion or tumor spread along biopsy sites .12· 13 Clearly, these observations demonstrate that discrepant information is obtained by the u se of surface cytology evaluat ion compared to histological evaluation. The cytological observations offer a potential explanat ion for biological paradoxes, such as local recurrence in histologically organ-confined (stage B) cancer of the prostate. To date, we have 1 patient with organ-confined tumor and a positive cytology test who has h ad an elevated serum P SA level postoperatively. Whether cytological evalu ation adds critical information that will permit sufficiently accurate prediction of tumor recurrences and their sites to impact on management, either on the basis of isolated or combined observations, is unknown . The failure of our current histological techniques to do so seems supported by substantial evidence. CONCLUSIONS
Ther e exists a discrepancy between th e histological and cytological evaluation of specimen margins. Stage C positive cytology results are associated with a higher risk of elevated serum PSA level postoperatively. Upstaging of pathological stage in review by a single pathologist using a uniform crit erion is substantial. The obser vations in our preliminary report encourage pursuit of this relatively simple assessment
128
TOUCH PREPARATION CYTOLOGICAL EVALUATION OF RADICAL PROSTATECTOMY SPECIMENS
of excised radical prostatectomy specimens. Correlation of this information with eventual patient course during the long term is necessary to assess its potential role in patient management. These preliminary observations provide information of sufficient interest for us to continue to pursue this effort and to encourage others to consider it.
7.
8.
REFERENCES
1. Schellhammer, P. F.: Radical prostatectomy: patterns of local failure and survival in 67 patients. Urology, 31: 191, 1988. 2. Jewett, H.J., Bridge, R. W., Gray, G. F., Jr. and Shelley, W. M.: The radical palpable nodule of prostatic cancer. Results 15 years after radical excision. J.A.M.A., 203: 403, 1968. 3. Blute, M. L., Nativ, 0., Zincke, H., Farrow, G. M., Therneau, T. and Lieber, M. M.: Pattern of failure after radical retropubic prostatectomy for clinically and pathologically localized adenocarcinoma of the prostate: influence of tumor deoxyribonucleic acid ploidy. J. Urol., 142: 1262, 1989. 4. Lightner, D. J ., Lange, P.H., Reddy, P . K. and Moore, L.: Prostate specific antigen and local recurrence after radical prostatectomy. J. Urol., 144: 921, 1990. 5. Epstein, J. I.: Evaluation of radical prostatectomy capsular margins of resection. The significance of margins designated as negative, closely approaching, and positive. Amer. J . Surg. Path., 14: 626, 1990. 6. Paulson, D. F., Moul, J. W., Robertson, J . E. and Walther, P . J .: Postoperative radiotherapy of the prostate for patients undergoing radical prostatectomy with positive margins, seminal
9. 10. 11. 12. 13. 14. 15.
vesicle involvement and/or penetration through the capsule. J . Urol., 143: 1178, 1990. Gibbons, R. P., Cole, B. S., Richardson, R. G., Correa, R. J ., Jr., Branner, G. E., Mason, J . T., Taylor, W. J . and Hafermann, R. D.: Adjuvant radiotherapy following radical prostatectomy: results and complications. J . Urol. , 135: 65, 1986. 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. Whitmore, W. F., Jr. and MacKenzie, A. K.: Experiences with various operative procedures for the total excision of prostate cancer. Cancer, 12: 396, 1959. Bosch, R. J . L. H ., Kurth, K. H . and Schroeder, F . H .: Surgical treatment of locally advanced (T3) prostatic carcinoma: early results. J . Urol., 138: 816, 1987. Kasabian, V. S.: Does tumor spillage during prostatectomy cause local cancer recurrence? AUA Today, vol. 5, no. 9, p. 15, 1992. Hofer, D.R., Garnett, J. and Grayhack, J. T. : The role ofperineal needle biopsy on local recurrence of prostate cancer. J . Urol., part 2, 141: 282A, abstract 450, 1989. Bastacky, S. S. , Walsh, P. C. and Epstein, J . I.: Needle biopsy associated tumor tracking of adenocarcinoma of the prostate. J . Urol., 145: 1003, 1991. Epstein, J. L. : The evaluation of radical prostatectomy specimens: therapeutic and prognostic implications. Path. Ann., 26: 159, 1991. Humphrey, P . A. : Complete histologic serial sectioning of a prostate gland with adenocarcinoma. J. Urol., part 2, 149: 263A, abstract 198, 1993.
THE JOURNAL OF UROLOGY Copyright © 1994 by AMERICAN UROLOGICAL AsSOCIATION, INC.
TOUCH PREPARATION CYTOLOGICAL EVALUATION OF RADICAL PROSTATECTOMY SPECIMENS MARGARET S. PEARLE, RYOICHI OYASU, DENISE HIDVEGI, DEBRA SUTKOWSKI, ALFRED RADEMAKER, RICARDO CAJULIS, JOHN T. GRAYHACK AND KEVIN T. McVARY From the Departments of Urology, Pathology and Biometry, Northwestern University Medical School, Chicago, Illinois
ABSTRACT
The failure of current histological techniques to predict local failure and disease progression after radical prostatectomy is supported by substantial evidence. Moreover, the characterization of histological findings is hampered by the lack of uniform interpretation. We report a prospective study of 92 patients undergoing radical prostatectomy for clinical stages A and B prostate cancer in which the technique of touch preparation cytological analysis of surgical margins is compared to the standard histological approach. We evaluated 47 pathological stage B, 43 stage C and 2 stage D specimens. Specimens initially assigned to pathological stage B were upstaged to stage C on review by 1 blinded pathologist in 19 of 65 cases (29% ). Overall, 15 of 4 7 histological stage B specimens (32%), 20 of 43 histological stage C specimens (47%) .and 2 of 2 histological stage D specimens (100%) had malignant cells identified on the margins by touch preparation cytology. Postoperative mean followup of 7 months (range 0.4 to 26) revealed that 7 of 14 nonstage D cancer patients (50%) with elevated serum prostate specific antigen levels had positive cytology results, including 1 with histologically confirmed organ-confined disease. Among the stage C specimens cytology was more likely to be positive if there was concomitant seminal vesicle invasion. Correlation of this information with eventual patient course during the long term' will be necessary to assess its role in patient management. KEY WORDS:
prostatic neoplasms, prostatectomy, histology, cytology, cytological techniques
Urologists rely largely on histological evaluation of prostatic capsular involvement to assess prognosis and for consideration of adjuvant radiation therapy after prostatectomy. Local failure rates of up to 40% for pathological stage B disease attest to our inadequate staging techniques. 1- 4 Whether these failures can be accounted for by missed or misinterpreted capsular perforation by the pathologist, or are due to a more fundamental limitation of histology in predicting residual disease is unclear. In an effort to address this staging inadequacy, we report our preliminary use of margin touch preparation cytological evaluation of radical prostatectomy specimens. Our goal was to improve or supplement current approaches that evaluate the excised specimens and to increase our ability to predict local failure. MATERIALS AND METHODS
a
In prospective study of 92 consecutive patients undergoing radical prostatectomy between August 1990 and February 1992 for clinically localized prostate cancer, we compared histological margins with corresponding cytological touch preparations. Along with a single blinded pathologist (R. 0.), we reexamined all histological sections of the prostate and re-staged each case with careful attention to the margins. Specimens were initially subjected in the laboratory to the touch preparation technique described, and then sent to the pathology department where the margins were stained with ink and the specimens were fixed. Apical (2 mm.) and base (3 to 4 mm.) sections were cut from the specimens transversely, and the entire apex and base were submitted for histological review. Any areas of gross tumor were also entirely submitted for examination. The remaining bulk specimen (prostate gland minus apex, base and areas of gross tumor) was halved and sectioned transversely into 2 to 3 mm. slices, of which 3 Accepted for publication December 17, 1993. Read at annual meeting of North Central Section, American Urological Association, Scottsdale, Arizona, November 3- 9, 1991.
to 4 slices (levels) were examined. Each section to be examined was cut into 4 quadrants and these histological sections were reviewed. Histological sections from each specimen were divided geographically and grouped into anterior, posterior, right lateral, left lateral, apex and base. An average of 24 histological slides were reviewed per case. Pathological stage C disease was defined as tumor on or outside the outer capsular edge (capsular perforation, fig. 1, A), on the inked margin (fig. 1, B) or in the seminal vesicles. Tumor cell penetration into but not through the capsule was considered pathological stage B (fig. 1, C). The shortest distance between the tumor front and the capsule or inked margin within each geographic region was measured with the use of light microscopy and intraoptic micrometer (1 radicle unit = 1.25 µ . at 40X). Margin touch preparation cytology studies were prepared by 2 techniques, a simple touch technique and a scraping technique. In the simple touch technique microscope slides were patted on each geographic area of the specimen and fixed. In the scraping technique the edge of a microscope slide was drawn across the 6 designated areas of the specimen (in a scraping fashion typical of touch preparation technique for lymph nodes) and fixed. These alternate techniques were performed in an attempt to differentiate contamination by prostatic secretions versus tumor cell perforation threugh the capsule. A single blinded cytopathologist (D. H.) reviewed all cytological touch preparation specimens and categorized them as malignant, benign or atypical cells (fig. 2). Inking the margins was not performed before cytological sampling for fear that we would render the cells on the microscope slides as uninterpretable by the cytologist. RESULTS
A total of 92 radical prostatectomy specimens was evaluated by touch preparation cytology. Patient age ranged from
124
l ..
•
TOUCH PREPARATION CYTOLOGICAL EVALUATION OF RADICAL PROSTATECTOMY SPECIMENS
125
FIG. 1. A, tumor cells have perforated through prostate capsule but not reached inked surgical margin. Large arrows demonstrate tumor cells in extracapsular tissue while small arrows mark tumor-free inked surgical margin. B, tumor cells (arrowheads) are noted on surgical margin. C, tumor cells invading into but not through capsule. Large arrows demonstrate prostate tumor cells within prostatic capsule. Small arrows mark outer capsular border as demonstrated by periprostatic fat. Reduced from X 186.
4 7 to 76 years (mean age 65 years). Cytology studies from 37 patients showed malignant cells, 48 benign cells and 7 atypical cells. The distribution of specimens by pathological stage is illustrated in table 1. Review by the referee pathologist resulted in reclassification of 19 stage B (3 stage Bl and 16 stage B2) specimens to pathological stage C. The distribution of specimens by cytology and pathological stage is illustrated in table 2. Overall, 15 of 47 stage B specimens (32%) and 22 of 45 stage C specimens (49%) had positive corresponding cytology results. Clearly, surface cytology smears identify a group of individuals with organ-confined cancer on histological assessment as having a high risk of extra-specimen disease. Conversely, surface cytology testing fails to provide objective evidence of extra-specimen disease in some individuals with histological findings interpreted as evidence of extra-organ or margin-positive disease. Biological assess-
ment of the significance of these discrepant findings depends on identification of individuals with recurrent disease on followup and documentation of the site of the recurrences when they are recognized. Pathological stage C specimens included those with capsular perforation, positive surgical margins and/or positive seminal vesicles. Closer evaluation of the stage C cancer patients revealed that all 12 specimens with involved seminal vesicles had extracapsular or margin-positive disease, including 8 with positive and 4 with negative cytology results. A total of 33 stage C specimens had negative seminal vesicles, including 14 with positive, 15 with negative and 4 with atypical cytology results. These data imply a greater likelihood of positive cytology findings with than without seminal vesical involvement (table 3). Followup information was available from 0.4 to 26 months postoperatively (mean 7 months). Of the patients 18 were
FIG. 2. Cells obtained by scraping method of touch preparation cytology. A , benign prostate cells. B, malignant prostate cells. Reduced from x 1,200.
126
TOUCH PREPARATION CYTOLOGICAL EVALUATION OF RADICAL PROSTATECTOMY SPECIMENS TABLE 1.
Pathological staging Initial* No.(%)
Stage
TABLE
Reviewedt No.(%)
17 (19) 49 (53) 24 (26)
14 (15) 33 (36) 43 (47) 2(2) 2 (2) Dl * Pathological staging as inferred from official pathology report. t Pathological staging as determined by review with referee pathologist (R. 0 .). Bl B2 C
followed for longer than 12 months and 3 were lost to followup. Patients were followed at 3 to 4-month intervals with physical examination and serum prostate specific antigen (PSA) level (Hybritech method) determinations. Bone scans, biopsies and additional evaluation were performed at the discretion of the attending urologist. No patient has had bone scan evidence of metastasis to date. Ten patients (excluding 1 with stage D disease) had persistent elevation of serum PSA (0.3 ng./dl. or greater, normal less than 0.3) postoperatively, of whom 6 (60%) had positive cytology findings. Of these 10 patients 8 underwent local external beam irradiation. Four patients (excluding 1 with stage D disease) had initial postoperative undetectable with subsequent increasing serum PSA levels and negative biopsies of the urethrovesical anastomosis (1 had a positive cytology result). To date, 2 of these patients have received irradiation. Three stage C cancer patients with persistently elevated or increasing serum PSA levels were treated with hormonal or immunotherapy. Both stage D cancer patients had positive cytology findings and were treated with radiation therapy. Overall, of 14 patients with elevated serum PSA levels postoperatively 7 (50%) (excluding the stage D cases) had positive cytology results (table 4). One of 4 patients with elevated postoperative serum PSA levels and pathological stage B disease had positive cytology findings. In this analysis, atypical cytology studies were considered negative to provide the most conservative estimate. We compared the 2 methods of cell sampling to determine if the scraping maneuver was superior to the simple touch technique in detecting malignant cells (table 5). The scraping method demonstrated malignant cells more frequently. Moreover, only 1 of 37 positive specimens was found by the simple touch technique alone, which suggests that a positive cytology result is not related to artifactual contamination by intraoperative spillage of malignant cells via the prostatic secretions but, instead, represents true tumor cell perforation through the prostate capsule. The shortest distance between the tumor and prostatic capsule or inked surgical margin in each geographic section was determined for all 81 cases with sufficient pathological documentation for review. Eleven cases were excluded from distance analysis because we were unable to assign histological sections to a geographic region unequivocally. However, these cases were globally restaged and were included in analyses that did not require geographic localization within the specimen. For pathological stage B cases the distance between the tumor and the outer capsular edge was always
3. Cytological status of stage C specimens
Seminal Vesical/ Extracapsular
Cytology Result Pos.
Neg.
Atypical
Totals
Pos./pos. Neg./pos. Totals
8 4 0 12 14 15 QQ 1 22 19 4 45 p = not significant Positive extracapsular disease represents either tumor at inked margin or tumor perforating through the capsule. Statistical analysis was performed using Fisher's exact test. * Includes 2 cases of positive lymph nodes that were local pathological stage C.
greater than zero. In these cases, however, the mean shortest distance from the tumor to the outer margin was 0.50 mm. for 29 specimens with 73 cytologically negative and atypical smears, and 0.80 mm. for 12 specimens with 31 cytologically positive smears (p = 0.05 by the Wilcoxon rank sum test). Using location-specific cytology, rather than global cytology for the specimen, the mean shortest distance from tumor to the outer capsule was 0.5 mm. in 96 negative and atypical smears (40 patients), and 2.84 mm. in 8 positive smears (4 patients, p = 0.04, Wilcoxon rank sum test). Additionally, there was no correlation between tumor grade and margin cytology status. DISCUSSION
This early report of an attempt to obtain specimen-related information to predict local tumor recurrence clearly shows a discrepancy between cytology and histology findings in assessing the presence of extra-specimen disease. With the limited followup available, the biological significance of these histological and cytological findings cannot be evaluated critically to date. However, we did note a trend in stage C specimens that suggests a greater likelihood of an elevated serum PSA level developing postoperatively in cytologically positive than in cytologically negative cases. The evidence that histological assessment, even with labor-intensive step-sectioning techniques and interpretation by highly skilled individuals, is unable to predict local or systemic recurrence with sufficiently high probability to guide clinical activity is substantive. Epstein stated that there is little objective evidence as to the validity of so-called positive or negative capsular margins ofresection. 5 He called for caution in diagnosing positive margins in an in-depth study of 40 radical prostatectomy specimens in which additional soft tissue was submitted separately for assessment of tumor involvement. Paulson et al presented data indicating an increased probability of local recurrence with positive surgical margins,6 while Gibbons et al reported a 25% local recurrence rate with capsular perforation and a 17% rate with margin-positive disease. 7 The length and methods of followup to assess local recurrence in these groups of patients have been limited in view of current knowledge. Local recurrence despite histologically specimen-confined
Patients with postoperative evaluated PSA classified by stage and cytological status (0.3 µg. I dl. or greater by Hybritech assay)
TABLE 4 . TABLE
2. Distribution of specimens by cytological status
Pathological Stage
Cytology* Pos.
Atypical
Neg.
No./Total Cytology(%) Totals
Bl 3 1 10 14 B2 12 2 19 33 22 19 Ct !Q 1 Totals 37 7 48 92 * Cytology is classified as positive (malignant), atypical or negative (benign). t Includes 2 cases with positive lymph nodes that were local pathological stage C.
Stage
No./Totals Pos.
Neg.
1/15 (7) 3/32 (9) 4/47 6/20 (30) 10/43 4/23 (17) 7/35 (20) 7/55 (13) Totals 14/90 Atypical cytology result was considered negative for this analysis. Stage D specimens were excluded. P value was not significant overall and within each stage. Statistical analysis was performed using Fisher's exact test. B C
TOUCH PREPARATION CYTOLOGICAL EVALUATION OF RADICAL PROSTATECTOMY SPECIMENS TABLE 5 .
Distribution of cytology results by sampling technique
Scraping Technique
Touch Technique
Neg. Pas. Pas. Pas. Pas. Neg. Neg. Neg. Atypical cytology result was considered negative for this analysis.
No.
30 6
1 55
disease is well documented. Schellhammer reported a local failure rate of 17% at 15 years for pathological stage B disease,1 while Jewett et al similarly found a local recurrence rate of 16% at 15 years after prostatectomy in 86 patients with capsule-confined disease. 2 Blute et al, in a large series of patients with localized prostate cancer, found a local recurrence rate of 4.3% when more than half of the patients were followed for more than 5 years. 3 Recently, Lightner et al biopsied the urethrovesical anastomosis in 63 patients with pathological stage B disease with no clinical evidence oflocal recurrence after radical prostatectomy and 42% had a positive biopsy.4 These observations indicate that the prevalence of local recurrence is underestimated. Why tumors recur locally after radical prostatectomy in patients with presumed margin-negative organ-confined disease on standard histological assessment represents a puzzling phenomenon that prompted this study. That patients with margin-positive stage C disease do not uniformly fail locally is similarly unexplained. Identifying patients who have a high probability of local failure compared to those in whom local recurrence is impr obable is a significant problem when considering recommendation for adjuvant therapy. An explanation, of course, could be simple human error. Capsular perforation and positive margins may be misinterpreted by the pathologist either by oversight, by inconclusive objective evidence or because of incomplete examination of the available t issue. As an example of this point, 19 of our 65 pathological stage B specimens (29%) were upstaged by our referee pathologist to stage C. Similarly, Jewett et al found 21 cases of recurrent or persistent cancer after radical prostatectomy, which on initial evaluation included 15 without capsular involvement, 5 with capsular penetration and 1 with capsular perforation.8 Reevaluation of more sections, however, demonstrated no capsular invasion in 6 cases, penetration in 9 and perforation in 6. Presently, the specimen evaluation after radical prostatectomy is done with meticulous care and good experience at several institutions. Evidence tha t these efforts will improve our ability to predict local recurrence remains elusive. Theoretically, local recurrence after radical prostatectomy with pathological stage C or, at least, margin-positive disease should approach 100% on the assumption that residual microscopic disease will eventually become clinically recognizable. Review of the literature, however, reveals a local recurrence rate of only 14 to 43% for pathological stage C disease.1 · 7 • 9 • 10 A major contributing factor t o this apparent broad range is a clear lack of consensus as to what constitutes stage C disease as reported in the literature. Penetration into the capsule, perforation t hrough the capsule and tumor at the surgical margin have often been used interchangeably without distinction between stage C and ma rgin-posit ive disease. Schellhammer reported a higher incidence of local failure for patients with seminal vesical involvement than those with only capsular involvement.1 Our data suggesting a higher likelihood of positive cytology findings in patients with seminal vesica l invasion than in those without correlate well with his clinical observations. The margins of resection from radical prostatectomies performed for prostate malign ancies are difficult to evaluate because of the relatively sparse surrounding soft tissue. Given the observations cited, histology h as clearly fallen short in predicting local control or failure. The cytological status of radical prostatectomy spec-
127
imens provides another objective observation that may potentially identify patients at increased risk of local recurrence. This technique may serve as an independent or an adjunctive observation to the traditional margin analysis. Intuitively, specimens with extracapsular disease (stage C) would be more likely than those with confined disease (stage B) to have a positive cytology result. We cannot explain the negative cytology test in t he face of clear histological margin positivity. Additionally, malignant cells were identified on the surface of glands in which tumor was histologically capsule-confined. The possibility that malignant prostat ic secretions coating the surface of the gland may account for positive touch preparation cytology warrants consideration. 11 We found only 1 instance, however, when cytology studies were positive when the microscope slide simply touched t he surface of the gland and were negative when it was drawn across the specimen, which suggests that the tumor cells adherent to the gland were perforating through the capsule and not simply bathing it (that is the product of intraoperative prostatic secretory contamination). Furthermore, while one might argue t hat more meticulous step-sectioning of the prostate might lead to finding some cases of missed capsular perforation in those with a positive cytology test, we hypothesize that individual tumor cells escape the confines of the capsule long before microscopic tumor encroaches on the margin in some circumstances. Lymphatic spread of tumor either distally or locally might also lead to some cases of missed capsular perforation. While we hypothesized an inverse relationship between tumor distance from the capsule and likelihood of a positive cytology test, t he opposite was found. These proposed renegade tumor cells responsible for positive cytology results clearly rely on factors other than sim ply tumor location to determine the predilection for capsular violation. There is a possibility that we might create false-positive histological margins by our scraping technique but our concern over rendering our cytology slides as uninterpretable prevented us from inking the margins before cytological sampling. Although less than a t otal histological review of a path ological specimen (a monumental task) could account for some cases of positive cytology results in the face of negative histological margins, we believe that our tissue processing has been accepted as able to classify pathological stage accurately.14· 15 The fact that our mean number of slides examined per case was 24 is a measure of our completeness. Additional explan ations for local failure would include occult perineural tu mor invasion or tumor spread along biopsy sites .12· 13 Clearly, these observations demonstrate that discrepant information is obtained by the u se of surface cytology evaluat ion compared to histological evaluation. The cytological observations offer a potential explanat ion for biological paradoxes, such as local recurrence in histologically organ-confined (stage B) cancer of the prostate. To date, we have 1 patient with organ-confined tumor and a positive cytology test who has h ad an elevated serum P SA level postoperatively. Whether cytological evalu ation adds critical information that will permit sufficiently accurate prediction of tumor recurrences and their sites to impact on management, either on the basis of isolated or combined observations, is unknown . The failure of our current histological techniques to do so seems supported by substantial evidence. CONCLUSIONS
Ther e exists a discrepancy between th e histological and cytological evaluation of specimen margins. Stage C positive cytology results are associated with a higher risk of elevated serum PSA level postoperatively. Upstaging of pathological stage in review by a single pathologist using a uniform crit erion is substantial. The obser vations in our preliminary report encourage pursuit of this relatively simple assessment
128
TOUCH PREPARATION CYTOLOGICAL EVALUATION OF RADICAL PROSTATECTOMY SPECIMENS
of excised radical prostatectomy specimens. Correlation of this information with eventual patient course during the long term is necessary to assess its potential role in patient management. These preliminary observations provide information of sufficient interest for us to continue to pursue this effort and to encourage others to consider it.
7.
8.
REFERENCES
1. Schellhammer, P. F.: Radical prostatectomy: patterns of local failure and survival in 67 patients. Urology, 31: 191, 1988. 2. Jewett, H.J., Bridge, R. W., Gray, G. F., Jr. and Shelley, W. M.: The radical palpable nodule of prostatic cancer. Results 15 years after radical excision. J.A.M.A., 203: 403, 1968. 3. Blute, M. L., Nativ, 0., Zincke, H., Farrow, G. M., Therneau, T. and Lieber, M. M.: Pattern of failure after radical retropubic prostatectomy for clinically and pathologically localized adenocarcinoma of the prostate: influence of tumor deoxyribonucleic acid ploidy. J. Urol., 142: 1262, 1989. 4. Lightner, D. J ., Lange, P.H., Reddy, P . K. and Moore, L.: Prostate specific antigen and local recurrence after radical prostatectomy. J. Urol., 144: 921, 1990. 5. Epstein, J. I.: Evaluation of radical prostatectomy capsular margins of resection. The significance of margins designated as negative, closely approaching, and positive. Amer. J . Surg. Path., 14: 626, 1990. 6. Paulson, D. F., Moul, J. W., Robertson, J . E. and Walther, P . J .: Postoperative radiotherapy of the prostate for patients undergoing radical prostatectomy with positive margins, seminal
9. 10. 11. 12. 13. 14. 15.
vesicle involvement and/or penetration through the capsule. J . Urol., 143: 1178, 1990. Gibbons, R. P., Cole, B. S., Richardson, R. G., Correa, R. J ., Jr., Branner, G. E., Mason, J . T., Taylor, W. J . and Hafermann, R. D.: Adjuvant radiotherapy following radical prostatectomy: results and complications. J . Urol. , 135: 65, 1986. 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. Whitmore, W. F., Jr. and MacKenzie, A. K.: Experiences with various operative procedures for the total excision of prostate cancer. Cancer, 12: 396, 1959. Bosch, R. J . L. H ., Kurth, K. H . and Schroeder, F . H .: Surgical treatment of locally advanced (T3) prostatic carcinoma: early results. J . Urol., 138: 816, 1987. Kasabian, V. S.: Does tumor spillage during prostatectomy cause local cancer recurrence? AUA Today, vol. 5, no. 9, p. 15, 1992. Hofer, D.R., Garnett, J. and Grayhack, J. T. : The role ofperineal needle biopsy on local recurrence of prostate cancer. J . Urol., part 2, 141: 282A, abstract 450, 1989. Bastacky, S. S. , Walsh, P. C. and Epstein, J . I.: Needle biopsy associated tumor tracking of adenocarcinoma of the prostate. J . Urol., 145: 1003, 1991. Epstein, J. L. : The evaluation of radical prostatectomy specimens: therapeutic and prognostic implications. Path. Ann., 26: 159, 1991. Humphrey, P . A. : Complete histologic serial sectioning of a prostate gland with adenocarcinoma. J. Urol., part 2, 149: 263A, abstract 198, 1993.