|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
© 2000 American Society for Clinical Oncology Utilizing Predictions of Early Prostate-Specific Antigen Failure to Optimize Patient Selection for Adjuvant Systemic Therapy TrialsFrom the Departments of Radiation Oncology, Urology, and Pathology, Brigham and Women’s Hospital; Departments of Radiation Oncology and Medical Oncology, Dana-Farber Cancer Institute, Boston; Department of Mathematical Sciences, Worcester Polytechnic Institute, Worcester, MA; and the Departments of Radiation Oncology, Urology, and Pathology, Hospital of the University of Pennsylvania, Philadelphia, PA. Address reprint requests to Anthony V. D’Amico, MD, PhD, Department of Radiation Therapy, Brigham and Women’s Hospital, Harvard Medical School, 75 Francis St, L-2 Level, Boston, MA 02115; email adamico{at}jcrt.harvard.edu
PURPOSE: Prostate-specific antigen (PSA) failure within 2 years after radical prostatectomy (RP) has been shown to be a clinically significant predictor of distant failure. This study was performed to estimate 2-year PSA failure rates on the basis of readily available clinical and pathologic factors to identify patients for whom effective adjuvant systemic therapy is needed. PATIENTS AND METHODS: A Cox regression multivariable analysis was used to determine whether the percentage of positive prostate biopsies, PSA level, and the pathologic findings at RP in 1,728 men provided clinically relevant information about PSA outcome after RP. A bootstrapping technique with 2,000 replications was used to provide 95% confidence intervals for the predicted 2-year PSA failure rates, which were determined on the basis of the independent clinical and pathologic predictors of PSA outcome.
RESULTS: The independent predictors of time to PSA failure included a percentage of positive prostate biopsies of greater than 34% (P CONCLUSION: Patients who are most likely to benefit from effective adjuvant systemic therapy after RP can be identified using readily available clinical and pathologic data.
PREVIOUS RESULTS support the value of grouping the pretreatment prostate-specific antigen (PSA) level, biopsy Gleason score, and 1992 American Joint Commission (AJCC) clinical T-stage to improve postoperative PSA outcome prediction over that possible with any single clinical parameter.1 More recently, a fourth parameter, the percentage of positive prostate biopsies, has been suggested to provide clinically relevant information regarding PSA outcome after radical prostatectomy (RP) in intermediate-risk patients for whom the PSA level, biopsy Gleason score, and 1992 AJCC T-stage fail to provide a clinically relevant stratification of PSA outcome.2 Although these pretreatment prognostic factors can be of value for physicians counseling patients with clinically localized prostate cancer about outcome after RP, pathologic data from the RP specimen provides additional information regarding PSA outcome not reflected in the clinical pretreatment data. Ideally, combining the independent pretreatment clinical and postoperative pathologic predictors of PSA outcome would optimize the prediction of postoperative PSA outcome. Previous investigators have attempted such a synthesis using the pretreatment PSA level, pathologic stage, margin status, and prostatectomy Gleason score.3,4 However, to date, no investigator has evaluated whether the percentage of positive prostate biopsy data adds independent prognostic information regarding PSA outcome when the preoperative PSA level and pathologic findings at RP are controlled for. Therefore, this article attempts to answer whether the percentage of positive prostate biopsy information provides a more specific identification of men at high risk for early (< 2 years) PSA failure after RP. Realizing that early (< 2 years) PSA failure after RP and the subsequent development of distant metastases have been recently correlated,5 there would be great value in optimizing the identification of men who have undergone RP who are at high risk for early PSA failure. Specifically, this more precise prediction of PSA outcome would improve the identification of those men who have the most to gain from the use of effective adjuvant systemic therapy.
Patient Population Seventeen hundred twenty-eight men treated with an RP and bilateral pelvic lymph node dissection at the Hospital of the University of Pennsylvania or at the Brigham and Women’s Hospital between 1989 and 1998 who had PSA detected or clinically palpable disease comprised the study population. Table 1 lists the preoperative clinical and postoperative pathologic characteristics of the 1,728 study patients stratified by the established pretreatment risk group for PSA outcome after RP on the basis of the PSA level, biopsy Gleason score, and 1992 AJCC clinical T-stage.
Preoperative Staging Each staging evaluation included a medical history and physical examination that included a digital rectal exam (DRE), serum PSA determination, computed tomography of the pelvis or an endorectal and pelvic coil magnetic resonance imaging scan of the prostate and pelvis, bone scan, and a transrectal ultrasound-guided needle biopsy of the prostate with Gleason score histologic grading.6 A sextant biopsy was performed using a 18-gauge Tru-Cut needle (Travenol Laboratories, Deerfield, IL) via a transrectal approach in the majority of patients (65%); however, 16% had had four or five biopsies performed, and 19% had had more than six biopsies performed. The clinical stage was obtained from the DRE findings, using the 1992 AJCC staging system.7 Radiologic and biopsy information were not used to determine clinical stage. The PSA was obtained on an ambulatory basis before radiologic studies and biopsies were performed. All PSA measurements were made using the Hybritech (San Diego, CA), Tosoh (Foster City, CA), or Abbot (Chicago, IL) assays,8 and the limit of detection was 0.1 ng/mL.
Pathologic Processing
Follow-Up
Statistical Analysis A previously established clinically significant stratification of the percentage of positive prostate biopsies (< 34%, 34% to 50%, and > 50%) in predicting postoperative PSA outcome2 was used in this study. The preoperative PSA level was grouped as 10 ng/mL or less, greater than 10 to 20 ng/mL, and greater than 20 ng/mL to allow for comparison with all other published reports examining the value of this variable in predicting PSA outcome. The prostatectomy Gleason score was evaluated as a categorical variable that could take on integer values between 2 and 6 versus 7 versus between 8 and 10. The 1997 AJCC pathologic stage was grouped as T2 (organ-confined) versus T3a (extracapsular extension) versus T3b (seminal vesicle invasion). Finally, margin status was treated as a categorical variable and grouped as positive or negative. PSA failure was scored when two consecutive detectable PSA values were obtained after an undetectable value (ie, PSA < 0.2 ng/mL) to eliminate the possibility of a false-positive result. The time of the first detectable value defined PSA failure, and time zero was defined as the day of surgery. If a PSA did not reach an undetectable level by the third postoperative month, then PSA failure was defined to be at time equal to zero. RPs were aborted for patients found to have positive pelvic lymph nodes at the time the frozen-section study was performed; such patients were therefore excluded from the study. Androgen-suppression therapy was initiated for men whose pelvic lymph nodes were negative at the time of frozen-section study but positive at the time the final sections were studied. Therefore, these men were also excluded from the time to postoperative PSA failure analysis. PSA-failure–free survival was estimated using an actuarial calculation according to the method of Kaplan and Meier.10 The probability of 2-year PSA failure was calculated from the coefficients of the Cox model. The 95% confidence intervals for the 2-year PSA failure probabilities were calculated using a bootstrapping procedure11 with 2,000 replications. The test for linear trend between pretreatment clinical predictors and pathologic outcomes was performed using a Cochran-Armitage test.12
Impact of Pretreatment Risk Group on the Percentage of Positive Prostate Biopsies and Pathologic Findings Clinical and pathologic characteristics are listed in Table 1 for the 1,728 study patients, who are stratified by the pretreatment clinical risk group for postoperative PSA outcome. There was a statistically significant decrease in the pathologic organ confinement rate (POC = .001) and negative surgical margins (Pmargin = .001) with increasing risk group. Conversely, the finding of ECE (PECE = .001) or SVI (PSVI = .001) increased significantly as the risk group increased. The proportion of patients with less than 34% or more than 50% of positive prostate biopsies decreased (P = .001) and increased (P = .001), respectively, with increasing pretreatment risk group. In addition, the proportion of patients with prostatectomy Gleason scores of eight to ten increased (P = .001), whereas those with prostatectomy Gleason scores of two to six decreased (P = .001) as the pretreatment risk group advanced.
Time to PSA Failure Analysis
Tables 3, 4 and 5 display compilations of the 2-year PSA failure rates and the respective 95% confidence intervals. The tables are stratified by the percentage of positive prostate biopsy group (< 34%, >34% to 50%, > 50%) and the pathologic stage (organ-confined and margin-negative, organ-confined and margin-positive, ECE- and margin-negative, ECE- and margin-positive, seminal vesicle invasion– and margin-negative, seminal vesicle invasion– and margin-positive). Values are also differentiated by the categories of the preoperative PSA level (  10 ng/mL, 10 to 20 ng/mL, and > 20 ng/mL), as shown in Tables 3, 4, and 5, respectively.
Identifying Patients at High Risk for Early PSA Failure As noted in Tables 3, 4, and 5, 2-year PSA failure rates ranged from 1% to 99%. Defining what 2-year PSA failure rate constitutes high risk remains arbitrary, but for the purpose of interpreting the results of this study, we selected failure rates of greater than 50%, which represented the majority of patients. It should be noted, however, that the data in Tables 3 through 5 are constructed in such a way to allow for any 2-year cutoff that a physician may select. Using a greater than 50% 2-year PSA failure as the standard for high risk, only patients with SVI and at least 34% positive prostate biopsies who had a preoperative PSA level of 10 ng/mL or less were at high risk of PSA failure within 2 years after RP. However, a prostatectomy Gleason score of 7 or higher, with greater than 50% positive prostate biopsies placed patients with ECE at a greater than 50% risk of early PSA failure, despite a preoperative PSA level of 10 ng/mL or less. All patients with a PSA level higher than 10 to 20 ng/mL had a more than 50% risk of PSA failure within 2 years if they had at least 50% positive prostate biopsies. The only exceptions were in those men with organ-confined and margin-negative disease who had prostatectomy Gleason scores of 7 or lower or men with ECE, but margin-negative, disease and prostatectomy Gleason scores of 6 or lower. Most patients with a preoperative PSA level higher than 20 ng/mL who had either ECE or SVI had a greater than 50% PSA failure rate by 2 years unless they had also had fewer than 34% positive prostate biopsies. Therefore, the additional information provided by the percentage of positive prostate biopsy information provided a more specific identification of men at high risk for early PSA failure when combined with the preoperative PSA level and RP findings.
The natural history of progression to metastatic disease after PSA failure following RP has recently been described.5 Although the median time to the development of metastatic disease after PSA failure was 8 years for the entire patient cohort, select patients could be identified on the basis of postoperative factors who were at the highest risk of developing distant failure after PSA failure. Specifically, from Fig 3B in5, 50% of men who manifested PSA failure within 2 years after RP went on to develop distant failure within 5 years. By 10 years, 75% of these men had developed distant failure. Considering that the median time to death from prostate cancer was 5 years after distant failure, identifying men after RP at high risk of developing early PSA failure ( 2 years) would select a cohort of men in whom the majority could benefit from the use of effective adjuvant systemic therapy. This study attempted to combine both clinical and pathologic factors that were significant independent predictors of time to postoperative PSA failure to identify men who were at high risk (> 50%) of PSA failure within 2 years after RP. The clinical utility of a previously defined stratification for the percentage of positive prostate biopsies2 was examined to ascertain whether this parameter added clinically meaningful information after the preoperative PSA level and the pathologic findings at RP were controlled for. Although whole-mount–determined tumor volume has been shown by Stamey et al13 to be a more powerful predictor of postoperative PSA outcome than the percentage of positive prostate biopsies, whole-mounting the RP specimen is not a standard practice. The percentage of positive prostate biopsy information, however, is available for most patients, and, therefore, its ability to predict PSA outcome after RP was assessed in this study to determine whether it offered a more generalized use for these data.
The results of this study suggest that the percentage of positive prostate biopsies provides additional information for identifying patients at high risk for early ( It is interesting to note that even after the pathologic findings at RP were accounted for, both the serum PSA level and the percentage of positive prostate biopsy information added significantly to the prediction of PSA outcome after RP. Considering that the PSA level reflects both tumor grade14 and volume,13 whereas the percentage of positive biopsies is a measure—albeit a weak one—of tumor volume,15 how can they both be surrogate measures of tumor volume and still both be independent of each other in predicting PSA outcome? Perhaps one of these parameters also reflects biologic aggressiveness and the propensity to metastasize. In a future study, this question will be addressed by performance of a time to distant failure analysis that will evaluate the parameters found in the study by Pound et al5 to be predictive of time to distant failure (ie, time to PSA failure, PSA doubling time, and prostatectomy Gleason score) with the preoperative PSA level, the percentage of positive prostate biopsy, and pathologic RP data. Although this study has focused on the readily available clinical and pathologic factors, other prognostic factors that have been shown by previous investigators to predict for time to postoperative PSA failure need to be studied using multivariable methodology to further optimize patient selection for national adjuvant therapy trials. These factors include diagnostic imaging modalities (eg, color Doppler imaging16 and endorectal coil magnetic resonance imaging17 ), molecular markers (eg, p27,18 p53 and apoptotic index,19 laminin receptor,20 and chromogranin A21 ), and histopathologic parameters (eg, ploidy22 ). Further investigation in staging and outcome prediction is still needed and currently ongoing. Finally, it is important to note that the end point evaluated in this study (2-year PSA failure rates) does not guarantee distant recurrence nor does PSA failure at a later time exclude eventual distant recurrence. Selecting the 2-year PSA end point has only been suggested5 to maximize the prediction of development of a distant recurrence, given time.
In conclusion, one half of all patients with clinically localized prostate cancer who sustain PSA failure within 2 years after RP progress to metastatic disease within 5 years after such PSA failure.5 This study suggested that the percentage of positive prostate biopsies added additional information to the preoperative PSA level and pathologic findings at RP in identifying patients at high risk (> 50%) for early (
1. D’Amico AV, Whittington R, Broderick GA, et al: Biochemical outcome after radical prostatectomy, external beam radiation therapy, or interstitial radiation therapy for clinically localized prostate cancer. JAMA 280: 969-974, 1998 2. D’Amico AV, Whittington R, Malkowicz SB, et al: PSA outcome after radical prostatectomy for patients with clinically localized prostate cancer: The clinical utility of the percent positive biopsies. Proc Am Soc Clin Oncol 18: 310a, 1999 (abstr 1192)
3.
Kattan MW, Wheeler TM, Scardino PT: Postoperative nomogram for disease recurrence after radical prostatectomy. J Clin Oncol 17: 1499-1504, 1999 4. D’Amico AV, Whittington R, Malkowicz SB, et al: The combination of the preoperative prostate-specific antigen level and postoperative pathologic findings to predict PSA outcome in men with clinically localized prostate cancer. J Urol 160: 2096-2101, 1998[Medline]
5.
Pound CR, Partin AW, Eisenberger MA, et al: Natural history of progression after PSA elevation following radical prostatectomy. JAMA 281: 1591-1596, 1999 6. Gleason DF, the Veterans Administration Cooperative Urological Research Group: Histologic grading and staging of prostatic carcinoma, in Tannenbaum M (ed): Urologic Pathology. Philadelphia, PA, Lea & Febiger, 1977, pp 171-187 7. Beahrs OH, Henson DE, Hutter RVP: American Joint Committee on Cancer: Manual for Staging Cancer (ed 4). Philadelphia, PA, Lippincott, 1992 8. Oesterling JE, Jacobsen SJ, Klee GG, et al: Free, complexed and total serum prostate specific antigen: The establishment of appropriate reference ranges for their concentrations and ratios. J Urol 154: 1090-1095, 1995[Medline] 9. Cox DR: Regression models and life tables. J R Stat Soc B 34: 187-200, 1972 10. Kaplan EL, Meier P: Non-parametric estimation from incomplete observations. J Am Stat Assoc 53: 457-489, 1958 11. Efrin R, Tibsherani R: An Introduction to the Bootstrap. New York, NY, Chapman and Hall, 1993 12. Armitage P: Test for linear trend in proportions and frequencies. Biometrics 11: 375-379, 1955
13.
Stamey TA, McNeal JE, Yemoto CM, et al: Biological determinants of cancer progression in men with prostate cancer. JAMA 281: 1395-1400, 1999 14. Aihara M, Lebovitz RM, Wheeler TM, et al: Prostate specific antigen and Gleason grade: An immunohistochemical study of prostate cancer. J Urol 151: 1558-1564, 1994[Medline] 15. Gardner TA, Lemer ML, Schlegel PN, et al: Microfocal prostate cancer: Biopsy cancer volume does not predict actual tumour volume. Br J Urol 81: 839-843, 1998[Medline] 16. Ismail MT, Petersen RO, Alexander AA, et al: Tumor vascularity assessed by color doppler imaging in prostate cancer: Prognostic Implications. J Urol 159: 288, 1998 (abstr 1110)
17.
D’Amico AV, Whittington R, Malkowicz SB, et al: A critical appraisal of the role of endorectal coil magnetic resonance imaging in the prediction of pathologic stage and postoperative prostate specific antigen failure in patients with clinically organ confined prostate cancer. J Clin Oncol 14: 1770-1776, 1996 18. Yang RM, Naitoh J, Murphy M, et al: Low P27 expression predicts poor disease-free survival in patients with prostate cancer. J Urol 159: 941-944, 1998[Medline] 19. Stapleton AM, Zbell P, Kattan MW, et al: Assessment of the biologic markers p53, Ki-67, and apoptotic index as predictive indicators of prostate carcinoma recurrence after surgery. Cancer 82: 168-174, 1998[Medline]
20.
Waltregny D, de Leval L, Menard S, et al: Independent prognostic value of the 67-kd laminin receptor in human prostate cancer. J Natl Cancer Inst 89: 1224-1227, 1997 21. Theodorescu D, Broder SR, Boyd JC, et al: Cathepsin D and chromogranin A as predictors of long term disease specific survival after radical prostatectomy for localized carcinoma of the prostate. Cancer 80: 2109-2113, 1997[Medline] 22. Lerner SE, Blute ML, Bergstralh EJ, et al: Analysis of risk factors for progression in patients with pathologically confined prostate cancers after radical prostatectomy. J Urol 156: 137-141, 1996.[Medline] Submitted December 21, 1999; accepted March 8, 2000. This article has been cited by other articles:
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||
|
|||||||||||
|
|
Copyright © 2000 by the American Society of Clinical Oncology, Online ISSN: 1527-7755. Print ISSN: 0732-183X
|
TOP
ABSTRACT
INTRODUCTION
