Mortality and Additional Treatment Rates in Pathologically High-Risk Prostate Cancer With Prostate-Specific Antigen Persistence at Robot-Assisted Radical Prostatectomy: Long-Term Report From Single Tertiary Referral Center.

Background: Long-term cancer control efficacy of robotic-assisted laparoscopic prostatectomy (RALP) in men with pathologically high-risk prostate cancer and prostate-specific antigen (PSA) persistence remains poorly addressed in the literature. Our aim was to evaluate long-term survival and additional treatment (AT) rates in these individuals.

Methods: We included 803 patients who underwent RALP for pathologically high-risk PCa (pT ≥ 3a, pN0-1 or GG ≥ 4) between 2001 and 2022 at a single tertiary referral center (Henry Ford Hospital, Detroit). Patients without adequate information about PSA persistence were excluded from the analysis (n = 128). Kaplan-Meier curves estimated AT free-survival (ATFS) and all-cause mortality (ACM) free-survival, whereas the competing risk method was used to estimate cancer-specific mortality (CSM) free-survival, after stratification according to PSA persistence. Competing risk and Cox regression models tested the impact of PSA persistence on three endpoints: AT rates, CSM, and ACM.

Results: Our final cohort consisted of 675 who underwent RALP for pathologically high-risk PCa, 187 (27.7%) of whom had PSA persistence. The median age at surgery was 64 years (IQR 59-68), and the median follow-up duration was 75 months (IQR 33-125). Patients with PSA persistence were more likely to have higher PSA values at surgery (8 vs. 7 ng/mL, p < 0.001), pT3b-4 PCa (62.5% vs. 39.9%, p < 0.001), pN1 PCa (55.6% vs. 35.7%, p < 0.001), and positive surgical margins (PSMs) (65.2% vs. 43.4%, p < 0.001). Moreover, patients in the PSA persistence group had higher proportion undergoing only hormone therapy (HT) (24.1% vs. 11.9%, p < 0.001) and radiotherapy (RT) plus HT (50.8% vs. 31.1%, p < 0.001), reporting higher median PSA values at RT (0.6 vs. 0.2 ng/mL, p < 0.001), compared to patients with undetectable PSA. At 10 years after RALP, CSM-FS and ACM-FS were 79.7% versus 90.3% (Gray-test p-value = 0.001) and 72.1% versus 79.6% (log-rank p-value = 0.013), for persistent versus undetectable PSA, respectively. The 10-year rates of ATFS were 6.6% versus 33.2% (log-rank p-value < 0.0001), for persistent versus undetectable PSA, respectively. At MVA, persistent PSA was associated with AT (HR: 3.05, p < 0.001), but not with CSM (HR: 1.49, p = 0.2) or ACM (HR: 1.09, p = 0.9).

Conclusion: Patients with pathologically high-risk PCa and PSA persistence after RALP, despite being at greater hazard of AT (HT and/or RT), did not have less favorable cancer control outcomes at 10 years than their counterparts with undetectable PSA levels. Our report provides the longest follow-up after RALP for this subset of patients, making it a valuable resource for counseling patients on the long-term oncologic outcomes of this procedure and postoperative adjuvant/salvage therapies.