Early PSA density kinetics predicts biochemical and local failure following extreme hypofractionated radiotherapy in intermediate-risk prostate cancer

•PSA density kinetics is a proxy endpoint of biochemical freedom from relapse.•3-month PSA density is predictive of long-term freedom from biochemical relapse.•3–6 months PSA density decay slope also predicts.•PSA outcome PSA density kinetics may serve as a tool to validate classifiers of radioresistance. The present study explores PSA density (PSA-D) as predictor of biochemical and local failure in organ-confined prostate cancer at 3–6 months after hypofractionated stereotactic ablative radiotherapy (SABR). A cohort of 219, hormone-naïve, NCCN intermediate-risk prostate cancer patients were derived from a phase 2 study of 5 × 9 Gy prostate cancer SABR. PSA-D was calculated at 3 and 6 months by dividing serum PSA by the MR-derived prostate CTV, while the slope of the 3–6 months curve was used to express the kinetics of PSA-D decay. The median follow-up was 60.3 (range 46–76) months, and the actuarial 7-year bRFS was 98.0% for intermediate-risk favorable (FIR) patients versus 84.5% for the unfavorable (UIR) subgroup (P = 0.02). Fourteen patients developed a Phoenix-defined biochemical PSA relapse (bRFS) at a median of 34.2 months, 11 confirmed with 68Ga-PSMA PET/CT scan that revealed tracer uptake at the site of dominant intraprostatic pretreatment lesion in 8 patients. The 3-month PSA-D concertation (cut-off 0.08 ng/ml2) and 3–6 months decay slope (cut-off −0.45) values were predictive of long-term bRFS. A dual adverse PSA-D permutation was detected in 25/148 UIR patients, exhibiting 47.5% 7-year bRFS compared with 94.1% for the remaining 123 UIR patients with favorable PSA-D kinetics (P = 0.0006). Intraprostatic local relapse in patients with a 3-month PSA-D > 0.080 ng/ml2 was 11.0% vs. 1.7% for patients with PSA-D ≤ 0.080 ng/ml2 (P = 0.01) and 2.3% vs. 4.3%, respectively, for nodal progression (P = 0.68). Early post-treatment PSA-D kinetics transcends pre-treatment risk stratification of tumor relapse and adds a nuance in the biological characterization of intermediate-risk prostate cancer phenotypes. The dual adverse PSA-D algorithm may serve as a tool to validate current search of classifiers of radioresistance in prostate cancer with therapeutic implications.