Prognostic significance of ctDNA in patients with metastatic uveal melanoma

Background: Circulating tumor DNA (ctDNA) is an emerging prognostic biomarker in solid tumors. However, the independent prognostic value in metastatic uveal melanoma (mUM) remains incompletely defined. We evaluated the prognostic significance of baseline ctDNA as measured by the Signatera assay, assessed the relationship to known prognostic factors such as LDH and liver metastases, and the impact of ctDNA clearance while on treatment. Methods: We retrospectively evaluated patients with mUM who had ctDNA collected within 30 days prior to treatment initiation. Kaplan-Meier (KM) and Cox proportional-hazard models were used to evaluate OS and PFS. Baseline ctDNA values were categorized into 3 tiers: Not-detected (ND), Low (<5 MTM/mL), and High (>5 MTM/mL). Multivariate Cox models were used to adjust for LDH (log-transformed) and liver metastases. Clearance of ctDNA was compared to persistently positive ctDNA. Patients treated with front-line tebentafusp were assessed as a prespecified subgroup. Results: 41 patients had baseline ctDNA measurements taken prior to frontline treatment; 26 received front-line tebentafusp and 66% had detectable ctDNA. Higher ctDNA levels correlated with elevated LDH. In the full cohort, detectable ctDNA was associated with shorter PFS (HR 3.07, 95% CI 1.14-8.30; p=0.027). In a 3-tiered stratification, high ctDNA (>5 MTM/mL) was predictive of shorter PFS (HR 4.65, 95% CI 1.62-13.4; p=0.004) and approached significance for OS (HR 4.60, 95% CI 0.97-21.8; p =0.055). On KM analysis, there was significant separation between the groups for both PFS (p=0.0048) and OS (p=0.011). As a continuous variable, higher ctDNA values (log-transformed) strongly predicted both PFS (HR 2.23, 95% CI 1.54-3.23; p<0.001) and OS (HR 3.26, 95% CI 1.79-5.92; p<0.001). In multivariate analysis, ctDNA remained independently prognostic of PFS (HR 1.75, 95% CI 1.06-2.87; p=0.028). Clearance of ctDNA was associated with improved PFS (HR 0.25, CI 95% 0.08-0.76; p=0.015) and OS, with no deaths in the clearance group (HR not calculable). In the tebentafusp-only subgroup, baseline ctDNA remained significant for OS (HR 2.47, 95% CI 1.30-4.66; p=0.005) and PFS (HR 1.83, 95% CI 1.22-2.74l; p=0.003). Clearance in this subgroup remained associated with improved PFS (HR 0.21, CI 95% 0.06-0.79; p=0.021) and OS (no deaths). Multivariate model trends were consistent with the full cohort but were limited by sample size. Conclusions: Baseline and on treatment ctDNA measured using the Signatera assay are prognostic in patients with mUM. ctDNA results remained prognostic in the tebentafusp subgroup and after adjusting for LDH and liver metastases.