Tumor-informed ctDNA as a biomarker of outcomes following stereotactic body radiotherapy in metastatic urothelial carcinoma.

Abstract

e16590 Background: Stereotactic body radiotherapy (SBRT) is used in metastatic urothelial carcinoma (mUC) to defer or augment systemic therapy, but biomarkers predicting outcomes after SBRT are lacking. Tumor-informed circulating tumor DNA (ctDNA) assays have prognostic value after cystectomy. We evaluated ctDNA for predicting outcomes following SBRT in patients (pts) with mUC. Methods: We retrospectively identified pts with mUC treated with SBRT who underwent tumor-informed ctDNA testing (Signatera, Natera) before and/or after SBRT between January 2023 and September 2025. Pts with < 3 months (mos) follow-up were excluded. Variables included number and location of metastatic sites, SBRT dose and fractionation, number of sites treated, and oligoprogressive (OP) versus oligometastatic presentation. OP disease was defined as ≤5 new lesions while on systemic therapy; pts with any number of metastases were eligible. ctDNA was assessed within 3 mos prior to SBRT and within 1 mos after SBRT. The primary endpoint was progression-free survival (PFS) from SBRT completion to radiographic progression or last follow-up. The secondary endpoint was time to next systemic therapy (TNST). Univariate Cox proportional hazards frailty models accounted for clustering of multiple lesions per patient. Results: Twenty-one pts underwent 34 SBRT courses targeting 72 metastases. Most pts were male (90.5%), had bladder primary tumors (80.0%), and had a median age of 74 years (IQR 66–78). Variant histology was present in 28.6%, and 70.8% had OP disease at SBRT. More than one-third (38.0%) had received ≥2 prior lines of systemic therapy. The median number of sites treated per SBRT course was 5 (IQR 3–7). Median SBRT dose was 35 Gy (range 18–54 Gy) in 5 fractions. Commonly treated sites included lymph nodes (47%), lung (19%), and bone (19%). At a median follow-up of 13.7 mos (IQR 9.5–15.2), median PFS was 2.9 mos (IQR 2.2–4.6). Median ctDNA at relapse was 7.02 MTM/mL. Subsequent management included new systemic therapy (38.9%) or repeat SBRT (37.7%). Median TNST was 3.4 mos (IQR 2.7–6.4). Baseline ctDNA was detectable prior to SBRT in 97.0% of courses, and ctDNA declined within 1 mos post-SBRT in 50% of cases. On univariate analysis, Hispanic ethnicity (p = 0.03), greater number of metastatic sites (p = 0.03), and OP disease (p < 0.01) were associated with shorter PFS. Pts with increased ctDNA within 1 mos after SBRT had significantly shorter PFS (1.3 vs 4.6 mos, p < 0.01) and TNST (2.6 vs 5.6 mos, p = 0.01). Absolute ctDNA levels at any time point and relative ctDNA changes were not associated with outcomes. Conclusions: Early post-SBRT increases in ctDNA are associated with inferior PFS and shorter time to next systemic therapy in pts with mUC, particularly those with OP disease. These findings suggest ctDNA kinetics may aid risk stratification after SBRT and warrant validation in larger cohorts.