CAN PERINEURAL INVASION ON PROSTATE NEEDLE BIOPSY PREDICT PROSTATE SPECIFIC ANTIGEN RECURRENCE AFTER RADICAL PROSTATECTOMY?

0022-5347/99/1621-0103/0 THEJOURNAL OF UROLOGY Copyright 0 1999 by AMERICAN UROWICALSOC CIA TI ON, h c .

Vol. 162, 103-106, July 1999

Printed in U S A .

CAN PERINEURAL INVASION ON PROSTATE NEEDLE BIOPSY

PREDICT PROSTATE SPECIFIC ANTIGEN RECURRENCE AFTER RADICAL PROSTATECTOMY? ALEXANDRE DE LA TAILLE, MARK A. RUBIN, EMILIA BAGIELLA, CARL A. OLSSON, RALPH BUTTYAN, TATJANA BURCHARDT, CHARLES KNIGHT, KATHLEEN M. O'TOOLE ANII AARON E. JSATZ" From the Squier Urological Clinic, College of Physicians and Surgeons and Ilepartmcnt of Biosta/r.stics. Srliwl of Public IJcalih, Columbia University, and Departments of Urology arid Pathology. Coliinibia-Presb~~lt.rian Mi,dical Ceritcr. N t w York, Neicq York

ABSTRACT

Purpose: We evaluated the role of perineural invasion identified on prostate needle biopsy as a predictor of prostate specific antigen (PSA) recurrence after radical prostatectomy. Materials and Methods: Between 1993 and 1998 radical prostatectomy was performed in 319 consecutive patients. Prostate needle biopsies were reviewed in all cases. We compared perineural invasion with other preoperative parameters, including digital rectal examination, PSA and biopsy Gleason score, for the ability to predict PSA recurrence with recurrence defined as any serum PSA level greater than 0.2 ng./ml. Results: Perineural invasion was identified on 77 of 319 preoperative prostate biopsies (24%). There was PSA recurrence in 46 patients (14.4%)a t a mean followup of 25.4 months (range 0.2 to 62.1).Perineural invasion statistically correlated with PSA recurrence. Kaplan-Meier analysis revealed disease-free survival rates of 24 versus 64%when perineural invasion was and was not present in the prostate biopsy (p = 0.0003, log rank 12.92). Multivariate analysis demonstrated that perineural invasion (p = 0.012) and PSA (p = 0.005) were independent preoperative predictive factors of PSA recurrence. When perineural invasion was compared with postoperative parameters, including disease stage, surgical margins and seminal vesicle invasion, it was not a n independent predictor because it closely correlated with tumor stage. Conclusions: Perineural invasion on preoperative prostate needle biopsy is a strong independent predictor of PSA recurrence in patients in whom prostate cancer was treated with radical prostatectomy . KEY WORDS: prostate; prostatic neoplasms; adenocarcinoma;biopsy, needle; prostate-specific antigen

Operations for curing clinically localized prostate cancer are based on the premise that all cancer cells are located within the tissue to be excised and complete excision must be properly accomplished. Despite current staging modalities, including digital rectal examination, preoperative serum prostate specific antigen (PSA) determination, transrectal ultrasound, pelvic computerized tomography and radionuclide bone scan, between 26 and 68% of patients have extraand between 23 and 53% prostatic disease have serum PSA recurrence, that is an increasing PSA level, by 10 years of f o l l o ~ u p . ~When - ~ prostate cancer has penetrated the capsule, there is a high risk of disease progression.'.' Thus, preoperatively identifying patients with extraprostatic extension is a challenge when treating prostate cancer. Perineural invasion is a well-known feature of prostatic a d e n o c a r c i n ~ r n a . l ~Villers - ~ ~ et a1 found that capsule penetration was selectively localized to the area where nerves

penetrate the prostate capsule, indicating that perineural invasion was the method of extraprostatic spread."' Byar and Mostofi identified perineural invasion in 84% of radical prostatectomy specimens examined but invasion did not predict patient survival at 5 or 7 years." Bastacky et a1 hypothesized that perineural invasion on diagnostic needle biopsy is a marker of extraprostatic extension.15 More recently, Egan and Bostwick reported that perineural invasion has no independent predictive value for extraprostatic extension, seminal vesicle invasion or pathological stage.I6 To determine the usefulness of perineural invasion for predicting PSA recurrence aRer radical prostatectomy we reviewed its presence on prostate needle biopsy in 319 consecutive patients with prostate cancer treated with radical prostatectomy at our institution. Comparisons were made with other preoperative parameters. MATERIALS AND M E T H O D S

From January 1993 t o March 1998 we entered into our Accepted for publication January 29, 1999. Supported by National Institutes of Health, National Cancer In- study 319 patients with clinically localized stage T1 to T2c stitute Grant ROI CA 70769, MSD-Merck (Bourse de I'Association prostate cancer in whom radical prostatectomy was planned. Francaise d'Urologk-MSD)and Exchange Visitor Program Lavoisier One surgeon (C. A. 0.)performed all digital rectal examina(Ministere des Aff-aires Etran eres France). Requests for reprints: Co?um$ia-Presbyterian Medical Center, tions and radical prostatectomies. All specimens were anaThe Atchlev Pavilion. 11th Floor, Room 1153, 161 Fort Washington lyzed in the department of pathology at our institution. Each Ave., New York, New York 10032. of the 319 radical prostatectomy specimens was weighed, Editor's Note: This article is the fifth of 5 published in this inked and fmed in 10%neutral buffered formalin. ARer fixissue for which category 1 CME credits can be earned. In- ation a thin piece of tissue was shaved from the proximal and structions for obtaining credits are Oven with the questions distal surgical margins (prostatic base and apex), and on pages 192 and 193. 103

104

1

CAN PERINEURAL INVASION PREDICT PROSTATE SPECIFIC ANTIGEN RECURRENCE

transected into multiple sections. The remaining prostate gland was inspected for any area of grossly recognizable tumor. In most cases gross tumor was not evident and 4 or 5 representative sections with inked peripheral margins were submitted for pathological evaluation from the right and left lobes sequentially from anterior to posterior. The seminal vesicles were submitted at the junction where they enter the prostate gland. Extraprostatic extension was diagnosed when tumor was seen in the periprostatic soft tissues or when it penetrated through a fibromuscular capsule to the other side. Seminal vesicle invasion was diagnosed when tumor was seen within the muscular wall of the seminal vessel. Tumor in the soft tissue surrounding the seminal vesicle was considered extraprostatic extension. A positive surgical margin was identified by carcinoma abutting an ink marking. Surgical margin status was not used to determine final pathological stage. All pelvic lymph nodes were evaluated for metastatic disease and all cases were assigned a Gleason score. Needle core biopsies were examined by 2 pathologists (M. A. R. andor K. M. 0.).The presence or absence of perineural invasion was reported in all cases. Even if only a single tumor focus was identified, Gleason score was also reported. Patients were seen every 4 months during the initial 2 years after radical prostatectomy and every 6 to 8 months thereafter. At each visit patients underwent a Hybritech* serum PSA assay. PSA recurrence was defined as 1 or more serum PSA measurements of 0.2 ng./ml. or greater. We used computer software for statistical analysis with p = 0.05 considered significant. All data were summarized using the mean plus or minus standard deviation for continuous variables and frequency tables for categorical variables. We performed univariate analysis with the chi-square test and Kaplan-Meier survival analysis. Cox proportional hazards regression was done for multivariate analysis. Time to recurrence was defined as the interval from surgery to PSA recurrence in patients with PSA recurrence and as the interval from surgery to the last followup in those with no biochemical recurrence. RESULTS

Included in our study were 319 consecutive men undergoing radical prostatectomy. Table 1 lists patient characteristics. Of the 319 patients digital rectal examination was abnormal in 224 (70%)and normal in 95 (30%).Serum PSA was greater than 10 ng./ml. in 61 cases (19%).Needle biopsies ranged from 1 to 12 cores with a mean Gleason score of 6.1. Gleason score was 5 or 6 in 184 biopsies (58%)and 7 to 9 in 114 (36%). Final pathological analysis of the 319 patients revealed stage pT2 disease in 218 (68%)and stage pT3 disease in 101 (32%).Final pathological staging revealed organ confined prostate cancer in 72 and 67% of the 95 men with nonpalpable and the 224 with palpable disease, respectively. Of the 205 patients in whom Gleason score was 5 or 6 on biopsy 159 (78%)had organ confined disease, in contrast to 59 of the 114 (52%)in whom Gleason score was 7 to 9 (chi-square test p <0.001). In 2 cases node positive disease was identified on

of 319 consecutive study patients who

Av.

Serum PSA (ngJml.) Gleason score: Biopsy Prostate specimen Followup (mos.)

Multiple factors are involved when assessing the prognosis in prostate cancer, including clinical examination, radiological studies and biopsy findings. Prostate needle biopsy provides certain data on the tumor. Digital rectal examination frequently under stages tumor extent, and magnetic resonance imaging and ultrasound have added little t o the accuracy of preoperative staging.l7 Preoperative interpretation of serum PSA is confounded by the volume of benign prostatic hyperplasia and degree of tumor differentiation.’ Tumor

1

0.9 0.8

0.5 0.4

underwent radical retropubic prostatectomy

Pt.age

DISCUSSION

8

* Hybritech, San Diego, California. TABLE1. Demographics

final pathological staging. Frozen sections were negative in all cases. Of the 319 patients 46 (14.4%)had serum PSA recurrence at a mean followup of 25.4 months (range 0.2 to 62.1). Disease was stage pT2 and pT3 in 13 of 218 patients (6%) and 33 of 101 (338), respectively. Mean interval between radical prostatectomy and PSA recurrence was 15.9 months (median 12.6, range 0.5 to 56). Perineural invasion on prostate needle biopsy statistically correlated with PSA recurrence. Kaplan-Meier analysis showed a statistical difference in recurrence-free survival in patients with and without perineural invasion (24 versus 64%,p = 0.0003, log rank 12.92, see figure). For perineural invasion p value was best compared with other preoperative parameters for predicting PSA recurrence (table 2). Digital rectal examination was not a predictive factor (p not significant). Needle biopsy Gleason score greater than 7 correlated with PSA recurrence. Disease-free survival was 62 versus 74% in patients in whom Gleason score was greater versus less than 7 (p = 0.0092). When we combined 2 preoperative parameters, such as perineural invasion on needle biopsy and Gleason score greater than 7, to determine whether the difference would increase, the log rank value was the same for perineural invasion only and the best combination of parameters (perineural invasion and Gleason score). However, for most postoperative parameters log rank and p values were greater than for perineural invasion or other preoperative parameters (table 2). Multivariate analysis was performed to test the independence of perineural invasion from other parameters. Table 3 shows the results of comparing perineural invasion with other preoperative parameters. Serum PSA (p = 0.005) and perineural invasion (p = 0.012) were the best independent predictors of PSA recurrence, in contrast to digital rectal examination (p = 0.302) and biopsy Gleason score greater than 7 (p = 0.088). Results were similar when we developed the Cox proportional hazards model using serum PSA as a discontinuous value of greater than or less than 10 ng./ml.

?

p=00003

0.2

SD (range)

61.4 z 5.84 (40.0-75.6) 8.02? 4.98 (0.5-35.0) 6.15f 1.02 (2.0-9.0) 6.4z 0.96 (2.0-9.0) 25.4 15.’ ‘o.2-62.1)

’

1I.

0.3

0

0

10

20

30

40

50

60

70

Followup (months)

Kaplan-Meier analysis of erineural invasion ( PNI ) and recurrence-free survival after rarfical prostatectomy (log rank test).

CAN PERINEURAL INVASION PREDICT PROSTATE SPECIFIC ANTIGEN RECURRENCE TABLE2 . Kaplan-Meier analysis

of

disease-free survival according to preoperative and postoperative parameters % Disease-Free

Log Rank (univariate analysis)

P Value

62/74 63/81 34/78 24/64 42/57

6.18 5.4 0.02 12.92 12.92

0.0092 0.02 0.87 0.0003 0.0003

35/56

15.4

o.oO01

Survival Preop. parameters: Gl&on score greaterfiess than 7 PSA greaterfiess than 10 Digital rectal examination abnormaVnormal Perineural invasion posheg. Perineural invasion pos. + Gleason score greater than 7/perineural invasion neg. + Gleason score less than 7 Perineural invasion pos. + PSA greater than 10 ngJml./perineuralinvasion neg. + PSA less than 10 ngfml. PSA greater than 10 ng./ml. + Gleason score greater than 7PSA less than 10 ngJml. + Gleason score less than 7 Postop. parameters: Gleason score greaterfiess than 7 Stage pT3/pT1 to 2 Extraprostatic extension pos./neg. Seminal vesicle invasion posheg. Surrrical m a r ~ nw h e n .

TABLE3. Multivariate analysis comparing perineural invasion to other preoperative parameters for predicting PSA recurrence

Biopsy Gleason score 1.72(0.983.22) greater than 7 Serum PSA* 1.06 (1.02-1.10) Abnormal digital 0.70 (0.361.38) rectal examination Perineural invasion 2.25 (1.19-4.23) * Considered as continuous variable.

105

0.088 0.005 0.302

0.012

grade or Gleason score on needle biopsy is a strong predictor of the Gleason score of the radical prostatectomy specimen. For example, Steinberg et al found that 87.5%of cases in which needle biopsy Gleason score was greater than 7 had similar high grade tumor in the radical prostatectomy specimen." They noted that needle biopsy Gleason score less than 7 was less predictive, since Gleason score of the tumor in the corresponding radical prostatectomy specimen was less than 7 in 63.9%of such cases. To evaluate the ability of perineural invasion on needle biopsy to predict tumor stage Bastacky et a1 evaluated 302 needle biopsies and found perineural invasion in 20%.15The positive predictive value was 93% for detecting extraprostatic extension when perineural invasion was present. The absence of perineural invasion was not as predictive, since only 49% of the cases without perineural invasion involved organ confined disease. Egan and Bostwick reported 51% sensitivity and 70%specificity16versus 27 and 96%reported by Bastacky et al,15 and 41 and 83% in our current study, respectively. Ravery et al noted that 72.7%of patients with versus 45.9%without perineural invasion on biopsy had capsular penetration (p <0.05).19 Ukimura et al observed that 60.8%of patients with versus 20.5% without perineural invasion had extraprostatic disease (p = 0.031).'0 We recently evaluated perineural invasion as a predictor of stage, seminal vesicle invasion, extracapsular extension and positive surgical margins, and determined that perineural invasion is an independent preoperative predictor of PSA recurrence.21 In that study the positive predictive value of perineural invasion was 57%. Of the 77 patients with perineural invasion 40 had stage pT3 disease. A potential reason for this lower positive predictive value than in other studies is that the whole prostate was not submitted for pathological examination. In some other series with higher positive values the whole prostate was submitted and, thus, focal areas of extraprostatic extension that may have been missed in the current study may have been identified in cases with perineural invasion on needle biopsy. Perineural invasion was found to be predictive with partial submission of the prostate.

55/72

3.91

48/61 24/90 26/85

17.26 29.53 19.29 21.17 10.66

19L57.5

61/73

0.048 o.oO01 o.oO01

o.oO01

o.ow1

0.0011

We demonstrated that perineural invasion is a predictor of PSA recurrence on the univariate level. Epstein recently compared 78 men in whom radical prostatectomy was performed by a single surgeon and in whom perineural invasion was identified by biopsy between April 1984 and February 1995 with 78 contemporary controls who did not have perineural invasion?' At followup of 3.3 and 4.0 years in patients with and without perineural invasion, respectively, there was no sigdicant difference in actuarial recurrence rates. Epstein concluded that perineural invasion does not independently influence the likelihood of PSA recurrence.% Univariate analysis revealed PSA recurrence in 27 versus 10%of our patients with versus without perineural invasion on biopsy (p = 0.001). Multivariate analysis showed that perineural invasion is an independent preoperative predictor of PSA recurrence. We previously reported that perineural invasion is closely associated with tumor stage.2' When we included disease stage in the multivariate analysis, perineural invasion lost its predictive value. Gleason score of the prostate specimen (p = 0.024) and disease stage (p = 0.0015) were the best independent predictors of PSA recurrence compared with perineural invasion (p = 0.168) and positive surgical margins (p = 0.724). Nevertheless, perineural invasion on needle biopsy remains a useful guide for urologists. When perineural invasion conelates with extraprostatic disease in the region of the neurovascular bundle, this information may be helpful for directing the surgeon while resecting the neurovascular bundle on the side of perineural invasion. Doing so may result in lower margin positivity due to wide excision of the neurovascular bundle. Epstein studied this concept and found that excision of the neurovascular bundle in radical prostatectomy cases with perineural invasion decreases the incidence of positive margins from 51.3 to 33.8%.22 CONCLUSIONS

Perineural invasion on preoperative needle biopsy was a preoperative predictor of PSA recurrence in patients with prostate cancer treated with radical prostatectomy. Multivariate analysis showed that perineural invasion is an independent predictor of PSA recurrence only when stage is not included. In our experience perineural invasion on needle biopsy represents a risk of extraprostatic extension as well as PSA recurrence. In such cases radical prostatectomy would not be considered curative and other therapy may be added. We believe that perineural invasion on prostate needle biopsy should be considered an important parameter in patients with prostate cancer and possibly a poor prognostic factor for PSA recurrence.

106

CAN PERINEURAL INVASION PREDICT PROSTATE SPECIFIC ANTIGEN RECURRENCE

13. McNeal, J . E. and Yemoto, C. E.: Spread of adenocarcinoma within prostatic ducts and acini: morphologic and clinical car1. Partin, A. W., Yoo, J., Carter, H. B., Pearson, J. D., Chan, D. W., relations. Amer. J. Surg. Path., 2 0 802, 1996. Epstein. J. I. and Walsh, P. C.: The use of prostatic specific antigen, clinical stage and Gleason score to predct patholog- 14. Villers, A. A., McNeal, J. E., Redwine, E. A., Freiha, F. S. and Stamey, T. A,: The role of penneural space invasion in the ical stage in men with localized prostate cancer. J . Urol., 1 5 0 local spread of prostatic adenocarcinoma. J . Urol., 1 4 2 763, 110, 1993. 1989. 2. Schellhammer, P. F.: Radical prostatectomy. Patterns of local 15. Bastacky, S. I., Walsh, P. C. and Epstein, J. I.: Relationship failure and survival in 67 patients. Urology, 31: 191, 1988. between perineural tumor invasion on needle biopsy and rad3. Ohori, M., Wheeler, T. M., Kattan, M. W., Goto, Y. and Scardino, ical prostatectomy capsular penetration in clinical stage B P. T.: Prognostic significance of positive surgical margins in adenocarcinma of the prostate. Amer. J. Surg. Path., 17: 336, radical prostatectomy specimens. J . Urol., 154: 1818, 1995. 1993. 4. Partin, A. W., Pound, C. R., Clemens, J . Q.,Epstein, J . I. and - A. J. and Bostwick. D. G.: Prediction of extramostntir ~.-__ Walsh, P. C.: Serum PSA after anatomic radical prostatec- 16. Eaan. extension of prostate cancer based on needle biopsy findings: tomy. The Johns Hopkins experience after 10 years. Urol. Clin. perineural invasion lacks significance on multivariate analyN. Amer., 2 0 713, 1993. sis. Amer. J . Surg. Path., 21: 1496, 1997. 5. Trapasso, J. G., deKernion, J . B., Smith, R. B. and Dorey, F.: The incidence and significance of detectable levels of serum pros- 17. Rifkin, M. D., Zerhouni, E. A,, Gatsonis, C. A., Quint, L. E., Paushter, D. M., Epstein, J. I., Hamper, U., Walsh, P. C. and tate specific antigen after radical prostatectomy. J. Urol., 1 5 2 McNeil, B. J.: Comparison of magnetic resonance imaging and 1821, 1994. ultrasonography in staging early prostate cancer. Results of a 6. Zincke, H., Oesterling, J. E., Blute, M. L., Bergstralh, E. J., multi-institutional cooperative trial. New Engl. J. Med., 323: Myers, R. P. and Barrett, D. M. Long-term (15 years) results 621, 1990. aker radical prostatectomy for clinicaily localized (stage T2c or 18. Steinberg, D. M., Sauvageot, J.,Piantadosi, S . and Epstein, J. I.: lower) prostate cancer. J . Urol., part 2, 1 5 2 1850, 1994. Correlation of prostate needle biopsy and radical prostatec7. Pound, C. R., Partin, A. W., Epstein, J . I. and Walsh, P. C.: tomy Gleason grade in academic and community settings. Prostate-specific antigen after anatomic radical retropubic Amer. J. Surg. Path., 21: 566, 1997. prostatectomy. Patterns of recurrence and cancer control. 19. Ravery, V., Boccon-Gibod, L. A., Dauge-Geffroy, M. C., Urol. Clin. N Amer., 2 4 395, 1997. Billebaud, T., Delmas, V., Meulemans, A., Toublanc, M. and 8. Ayala, A. G., Ro, J . Y., Babaian, R., Troncoso, P. and Grignon, Boccon-Gibod, L.: Systematic biopsies accurately predict exD. J.: The prostatic capsule: does it exist? Its importance in the tracapsular extension of prostate cancer and persistenthecurstaging and treatment of prostatic carcinoma. Amer. J. Surg. rent detectable PSA after radical prostatectomy. Urology, 44: Path., 1 3 21, 1989. 371, 1994. 9. Epstein, J. I., Carmichael, M. J., Pizov, G. and Walsh, P. C.: Influence of capsular penetration on progression following rad- 20 Ukimura, O., Troncoso, P., Ramirez, E. I. and Babaian, R. J.: Prostate cancer staging: correlation between ultrasound deterical prostatectomy: a study of 196 cases with long-term folmined tumor contact length and pathologically confirmed exlowup. J. Urol., 1 5 0 135, 1993. traprostatic extension. J . Urol., 1 5 9 1251, 1998. 10. Byar, D. P. and Mostofi, F. K.: Carcinoma of the prostate: prognostic evaluation of certain pathologic features in 208 radical 21. de la Taille, A., Katz, A. E., Bagiella, E., Olsson, C. A., OToole, K. M. and Rubin, M. A.: Perineural invasion on prostate needle prostatectomies. Examined by step-section technique. Cancer, biopsy: an independent predictor of final pathology, unpub3 0 5, 1972. 11. Epstein, J . I.: Diagnostic criteria of limited adenocarcinoma of lished data. 22 Epstein, J . I.: The role of perineural invasion and other bioDsv the prostate on needle biopsy. Hum. Path., 2 6 223, 1995. . _ 12 Kahler, J . E.: Carcinoma of the Drostate gland: a Datholoeic -characteristics as prognostic markers for localized prostate study. J. Urol., 41: 224, 1939. cancer. Sem. Urol. Oncol., 1 6 124, 1998. REFERENCES

.

1