Genomic risk classifier performance in PET-guided post-prostatectomy patients: Secondary analysis of a randomized trial.

386 Background: Genomic risk assessment with genomic classifiers (GC) have enhanced risk stratification in post-prostatectomy patients, however, validation studies have relied on conventional imaging workup to guide secondary radiation treatment (SRT). Our hypothesis is that GC will remain prognostic in a population receiving PET-guided SRT. Methods: We performed a protocol-specified whole-transcriptome assay (Decipher GC, Veracyte) on prostatectomy specimens from patients enrolled on a randomized trial studying event-free survival [EFS: biochemical/clinical/radiologic progression or systemic therapy initiation] of 18F-fluciclovine or 68Ga-PSMA-11 guided SRT with incorporated PET-guided dose escalation. Treatment volumes and prescription doses were rigidly defined on protocol: no uptake/prostate bed (PB) only - PB XRT (64.8-70.2Gy at 1.8Gy/fx); pelvic nodal (PLN) +/- PB uptake – PLN (45-50.4Gy @ 1.8Gy/fx) + PB XRT; extrapelvic – no XRT. Sites of PET uptake received simultaneous integrated boosts (SIB) of 74-76Gy at 2Gy/fx in PB and 54-56Gy at 2Gy/fx in PLN. GCs were categorized as low (0-0.44), intermediate (0.45-0.6) and high (>0.6) as reported for commercial testing. An optimal cutoff maximizing separation above (AOC) vs below (BOC) was identified using a bias-adjusted log-rank test. Z-test was performed at specified time points to assess EFS between cohorts per study protocol. Results: Of 140 patients enrolled on trial, 69 (49.3%) had prostatectomy specimens available for GC analysis with 44 (63.8%) high, 11 (15.9%) intermediate and 14 (20.3%) low scores. Androgen deprivation therapy (ADT) was given in 4/14 (28.6%) low, 8/11 (72.7%) intermediate, and 36/44 (81.8%) high GC patients based on clinical characteristics. Minimum follow-up was 2.00 years (median 2.9 yrs, range: 2.0-5.0 yrs). High GC was associated with worse 4-year EFS compared to low GC (56.5% vs 84.6%, p<0.01). Optimal GC cutoff was 0.84 with 28 AOC (23/28 with ADT) and 41 BOC (25/41 with ADT). Score AOC was associated with worse 3-year (69.6% vs 89.3%, p<0.01) and 4-year (46.4% vs 70.2% p=0.01) EFS compared to BOC. Use of SIB was associated with improved EFS in score BOC at 2 and 3-years post treatment (81.8% vs 92.3% at both time points, p=0.03); however, SIB was not associated with improved EFS in score AOC at 2 years or beyond. Conclusions: GC remained prognostic for EFS in post-prostatectomy patients treated with PET-guided radiation when assessed at commercial and optimal cutoff values. Dose escalation to sites of PET uptake was associated with improved EFS at 2+ years in BOC (<0.84) patients. Creation of an integrated radiogenomic model based on PET findings and GC is forthcoming.