ADULT UROLOGY TIME TO SECOND PROSTATE-SPECIFIC ANTIGEN FAILURE IS A SURROGATE ENDPOINT FOR PROSTATE CANCER DEATH IN A PROSPECTIVE TRIAL OF THERAPY FOR LOCALIZED DISEASE ANTHONY L. ZIETMAN, KATHERINE C. DALLOW, PATRICIA A. McMANUS, NIALL M. HENEY, AND WILLIAM U. SHIPLEY ABSTRACT Objectives. The most relevant endpoint in comparing the efficacy of curative therapies for prostate cancer is cancer-specific death. Prospective trials need to mature for at least a decade to yield meaningful cancer death data due to the long natural history of the disease and the use of salvage androgen suppression. This delay may be long enough that the tested treatments are outdated by the time of reporting; thus, there is a need for reliable early surrogate endpoints for cancer survival. Methods. This report evaluates 202 patients entered into a single institution prospective randomized study for T3-4 prostate cancer. Patients were accrued between 1982 and 1992 and received radical irradiation to either a standard dose of 67.2 Gy or a higher dose of 75.6 Gy. Median follow-up was 5.4 years. A total of 76 men have received androgen suppression or orchiectomy for salvage following relapse. Of this group, 35 experienced a second relapse heralded by a rise in the serum prostate-specific antigen (PSA). Results. The median survival from the time of second biochemical relapse (defined as a progression with a rise in serum PSA more than 10% above the nadir after androgen suppression) was 27 months. Kaplan-Meier analysis projected a 0% survival for this group at 4 years. All those dying after second biochemical failure died of the prostate cancer. Conclusions. Second PSA failure (or PSA progression on hormonal therapy) has potential as a surrogate for impending cancer death and its use as an endpoint in prospective studies could allow earlier reporting by 2 to 4 years. UROLOGY* 47: 236-239, 1996. P rostate cancer is the most common malignancy among men and the second most common cause of male cancer death. It is a condition that arises in an older age group and has a long and unpredictable natural history Trials attempting to assess the benefits of various therapeutic strate- gies have foundered on the long observation peri- ods necessary to determine surviva1.l In some cases the therapy has already become outmoded by the time the trial matures.2*3 The recent massive increase in prostate cancer case detection coupled with an increasing num- ber of available therapies creates an urgent need to evaluate outcome more rapidly. The serum marker prostate-specific antigen (PSA) has aided the oncologist in this regard by detecting relapse Presented at the American Society of Clinical Oncology Annual Meeting: May 1995, Los Angeles, California. From the Departments of Radiation Oncology and IJrology, and the Massachusetts General Hospital Cancer Center, Harvard Medical School, Boston, Massachusetts Reprint requests: Anthony L. Zietman, M.D., Department of Radiation Oncology, Massachusetts General Hospital Cancer Center, Boston, MA 02114 Submitted: August 22, 1995, accepted (with revisions): October 16, 1995 236 months or even years before it may be clinically appreciable. l Before reaching the endpoint of greatest inter- est, cancer-specific death, men with localized dis- ease must not only fail their primary therapy, but also fail after an attempt at salvage by androgen suppression. The presumption that first failure af- ter radical therapy correlates in a strict fashion with ultimate cancer death is thus both a loose and a distant one. It is our hypothesis that the easily measured event of second PSA failure (or the time of PSA failure on progression after the initiation of androgen suppression) strictly corre- lates with subsequent cancer death and, conse- quently, may be used as a surrogate endpoint. This would allow for the earlier completion and analy- sis of prospective trials. The use of second PSA failure as an endpoint would be limited if it was linked too closely in time to death, and we there- fore propose that this distance can be extended by the early detection of second failure using serum PSA rather than waiting for clinical or radiologic progression. PSA already has proven useful in the early detection of prostate cancer and of first fail- ure following initial therapy. A rise in serum PSA often precedes these clinical events by years. UROLOGY@ 47 (21, 1996