Background: The SPPORT trial demonstrated that salvage whole pelvic radiotherapy (sWPRT) after radical prostatectomy (RP) reduced prostate cancer progression compared to salvage prostate bed radiotherapy (sPBRT), at the expense of greater toxicity. To identify pts most appropriate for sWPRT, we studied the incidence of and risk factors associated with radiologic relapse after sPBRT on 18F-DCFPyL PET, a sensitive new imaging modality. Methods: We analyzed 538 18F-DCFPyL PET/CTs performed at our large academic institution and screened for pts who had biochemical recurrence (BCR), defined as PSA > 0.2 ng/mL, after receiving sPBRT for persistent or recurrent PSA. Risk factors including pathologic T stage (pT), pathologic Gleason score (pGS), seminal vesicle invasion (SVI), positive margins, perineural invasion (PNI), PSA at RP (PSA1), prior to RT (PSA2), and at time of scan (PSA3), and PSA doubling time after sPBRT (PSADT) were obtained. Associations between risk factors and radiographic relapse were quantified with logistic regression, with a p < 0.05 defined as significant. Results: 38 pts with BCR after sPBRT were identified. Median values were: pt age = 69.5 years (interquartile range 64 – 74), PSA1 = 7.83 ng/mL (4.7 – 11.5), PSA2 = 0.26 (0.1 – 0.8), PSA3 = 0.71 (0.5 – 2.2), PSA DT = 9.8 (5.3 – 18.2). Pt characteristics were: pT (pT2, 46% of pts; pT3a/b, 54%), pGS ( < 7, 11%; 3+4, 32%; 4+3, 39%; > 7, 18%), 19% w/ SVI, 68% w/ positive margins, 93% w/ PNI and 80% (of evaluable pts) w/ high-risk decipher score. 42% of pts received salvage ADT with sPBRT. 68% of pts had detectable disease on 18F-DCFPyL PET. Radiologic relapse in pelvic LNs was present in 47% of all pts, and 32% of pts had relapse only in the pelvic nodes. 26% of pts had distant metastases (DM) on imaging. Only 13% of pts had DM without pelvic LN involvement. Between pts with pelvic LN only relapse (n = 12) and pts without radiographic disease (n = 12), age, pT, pGS, SVI, margins, ADT, PSA1-3 and PSADT did not differ on univariate or multivariate analyses. On the univariate analysis between pts with DM (n = 10) and without DM (n = 28), only PSA3 (p < 0.05) was associated with radiographic relapse, and pT trended toward a negative association (p< 0.10) with DM. Only PSA3 (p < 0.10) trended toward association with DM in a multivariate analysis using age, pT, margins, ADT, and PSA2-3. Conclusions: Pts with biochemical relapse after sPBRT had frequent high-risk pathology. Aside from PSA3, no risk factors were predictive of detectable pelvic nodal or DM disease. However, most pts with detectable disease had pelvic LN involvement, suggesting that a substantial portion of these high-risk pts may have benefited from whole-pelvic RT. Further data on risk factors will help inform selection of pts appropriate for salvage whole pelvic RT.