Association Between T-Stage and Cancer-Specific Mortality among Men with High-Risk Prostate Adenocarcinoma

Abstract

Risk group stratification identifies men with high-risk prostate adenocarcinoma (Gleason 8-10, PSA > 20 ng/mL, or cT3-T4) who benefit from more intense treatment. However, many patients are classified as high-risk with only one high-risk feature. Although many trials that are the basis for 2-3 years of androgen deprivation therapy (ADT) studied patients with mostly cT3/T4 disease, PSA-detected, non-palpable T1c disease is currently the most commonly detected. We assessed whether within localized high-risk prostate cancer, survival differences existed based on T-stage. Data from men with high-risk N0M0 prostate adenocarcinoma from the 2004 to 2015 Surveillance, Epidemiology, and End Results database were used. Ten-year PCSM was estimated using Fine-Gray competing-risk regression. The independent variable of interest was clinical T-stage (cT1/2 vs. cT3/4); adjustments were made for PSA, Gleason score, treatment with radical prostatectomy (RP) and radiation (RT), and relevant demographic factors. A second analysis assessed PCSM only among men who received RP, with pathologic T-stage as co-variate (pT2 vs. pT3/4). The average of 12 prostate cancer prognostic signatures, including Decipher prostate cancer signature, compared genomic risk score by pT-stage. For 63,319 men included (median follow-up 6.4 years, median age 68 years), 56,322 (89.0%) had cT1/2, and 6,997 (11.1%) had cT3/4 disease. Ten-year PCSM for men with cT1/2 disease was 12.5%, and for men with cT3/4 disease was 17.5% (competing risk regression HR 0.55, 95CI 0.51—0.60, P<0.001). Greater PSA and greater Gleason score were associated with greater PCSM; receipt of RP, RT, being married, greater income, and later year of diagnosis were associated with lower PCSM. African American and Other race were also associated with lower PCSM. Of the 20,409 men who received radical prostatectomy, 9,953 (48.8%) had pT2 disease and 10,456 (51.2%) had pT3/4 disease. Ten-year PCSM for post-RP men with pT2 disease was 3.5%, and for men with pT3/4 disease was 11.9% (competing risk regression HR 0.29, 95CI 0.24—0.35, P<0.001). Greater PSA, Gleason score, receipt of radiation, and earlier year of diagnosis were associated with worse PCSM. Within a high-risk prospective RP cohort (n = 10,419) from the Decipher GRID registry (NCT02609269), patients with pT3/4 had significantly higher genomic risk scores than patients with pT2 (p<0.001). Men with cT1/2 disease had improved cancer-specific outcomes compared to men with cT3/4. Among men treated with RP, pT2 disease was associated with lower PCSM compared to pT3/4. Genomic risk scores validated these findings. Therefore, treatment paradigms given clinically relevant sub-classification of high-risk tumors based on size merit further study, with a focus on shorter-duration ADT among men with high-risk but T1/2 disease. As cT1c disease is most commonly diagnosed, more men with smaller tumors should be included in trials.