Use of baseline and serial ctDNA dynamics to predict outcomes in patients treated with first-line tebentafusp, including those who were and were not treated beyond progression.

9536 Background: Tebentafusp (tebe), a bispecific soluble T cell receptor specific for a gp100 peptide, is licensed for the treatment of metastatic uveal melanoma (mUM) in HLA-A*02:01+ individuals. As part of the pivotal, Phase 3 IMCgp100-202 study, 378 mUM pts were randomized to receive tebe or investigators choice. Approximately 60% of patients (pts) with metastatic UM in the IMCgp100-202 study had detectable baseline ctDNA. We have reported the strong relationship between change in ctDNA at 9 weeks (wks) and overall survival (OS). Additionally, radiographic treatment response is a poor predictor of who is benefitting from therapy. We hypothesized that ctDNA dynamics would help identify those patients (pts) who benefit from treatment beyond progression (TBP). Methods: Sera (n=202 pts) collected at baseline and wk 9 on tebe were analyzed for ctDNA using a targeted mPCR-NGS assay for 15 genes including GNAQ, GNA11, SF3B1, PLBC4, EIF1AX, and CYSLTR2. RECISTv1.1 was assessed by investigators; TBP was allowed per protocol. OS was analyzed in subsets of pts who were not TBP, and those who were TBP and treated with tebe for 16 wks; the approximate time point where TBP pts would have had confirmatory imaging, although confirmatory scan data is not available. At baseline, pts were determined to be ctDNA detectable or undetectable. At wk 9, pts were determined to have cleared ctDNA, < or ≥ 68% (0.5 log-fold) reduction, < or ≥ 50% reduction. Data cut-off was July 2023. Results: Of 202 pts randomized to tebe and with ctDNA at baseline and wk 9, 126 (62%) received TBP (80 <16 wks, 46 ≥ 16 wks). Undetectable ctDNA at baseline was associated with improved OS regardless of TBP, while TBP was associated with improved OS in those with baseline detectable ctDNA (hazard ratio, HR, 0.59, 95% confidence interval, CI, 0.4-0.88). Subgroup analysis of those with baseline detectable ctDNA demonstrated improved OS in only those who received TBP for ≥ 16 wks (HR 0.33; 95% CI 0.21-0.51). Clearance of ctDNA at wk 9 was associated with better OS than in those who did not have clearance. TBP was associated with better OS in pts who did not have clearance (HR 0.48, 95% CI 0.30-0.79), but not in those with clearance (HR 0.87, 95% CI 0.43-1.78). Finally, TBP was associated with improved outcomes in pts with ≥50% and ≥68% reduction, compared to those without reduction, HR 0.63 (95% CI 0.4-0.97) and 0.62 (95% CI 0.39-0.99), respectively. Conclusions: Baseline and wk 9 ctDNA detection predicts OS in mUM pts treated with first line tebe. The utility of ctDNA predicting pts most likely to benefit from TBP is less clear, although TBP seems to be more beneficial in those with reduction of ctDNA at wk 9. Optimization of time points, assay, and integration of ctDNA data with radiomics are necessary to understand the true utility of ctDNA in helping select which pts should receive TBP. Clinical trial information: NCT03070392 .