Molecular testing in primary advanced or recurrent endometrial cancer: A cost-effectiveness analysis.

5598 Background: The Proactive Molecular Risk Classifier for Endometrial Cancer (ProMisE) algorithm is a classification scheme for endometrial cancer (EC) based on sequential testing for DNA mismatch repair deficiency (dMMR), POLE exonuclease domain mutations, and p53 mutations. The cost-effectiveness of ProMisE vs no molecular testing to inform initial systemic treatment choice in patients with stage III/IV primary advanced/recurrent EC (pA/rEC) was assessed from payer and societal perspectives. Methods: A hybrid model comprising a decision tree for molecular classification (ProMisE vs no testing) followed by partitioned survival models (PSM; progression-free disease, progressed disease, and death) was developed. Lifetime costs and outcomes following first-line systemic treatments were estimated using this model. Patients in the no-testing arm were assigned to receive carboplatin-paclitaxel (CP), dostarlimab + CP, pembrolizumab + CP, or hormonal therapy (everolimus/letrozole). Patients in the ProMisE arm were assigned to the same treatments or bevacizumab + CP, according to their molecular profile. Survival functions were derived from published trial results. EQ-5D-5L utility values were sourced from the RUBY trial (NCT03981796). Costs were sourced from US-focused databases and publicly available literature. The base case presented a US third-party payer perspective. A scenario analysis presented a modified societal perspective. Model outcomes were total costs, total life-years (LYs), total quality-adjusted LYs (QALYs), and the incremental cost-effectiveness ratio (ICER) of ProMisE vs no testing. An annual discount rate of 3% per year was applied to future costs and outcomes. Uncertainty was evaluated using one-way (OWSA) and probabilistic sensitivity analyses (PSA). Results: From a third-party payer perspective, total LYs and total QALYs were 5.36 and 4.08, respectively, with ProMisE vs 3.83 and 2.89 with no testing. Total costs were $233,989 with ProMisE vs $155,305 with no testing. Thus, incremental LYs and QALYs were 1.53 and 1.19 greater with ProMisE vs no testing; incremental costs over a lifetime were $78,684 greater with ProMisE. Assuming a cost-effectiveness threshold of $150,000 per QALY gained, ProMisE was cost-effective with an ICER of $66,321 per QALY gained compared with no testing. In the OWSA, ProMisE remained cost-effective over all parameter ranges (±10%). In the PSA, ProMisE was below the threshold of $150,000/QALY for 97.7% of iterations. From a societal perspective, lifetime costs were lower (−$22,973) and QALYs greater with ProMisE vs no testing. Conclusions: ProMisE testing is cost-effective vs no testing when using a $150,000/QALY-gained threshold. Given the heterogeneity of molecular subtypes in stage III/IV pA/rEC, molecular testing enables personalized treatment that is clinically meaningful and high value from payer and societal perspectives.