Multi-Combo Comparator.

Four combinations, three indications.

Benefit density.

The comparator's "best combo" star (★) marks the combination with the highest benefit density in the selected indication for the input population profile.

Benefit density = months PFS gained per 1% fatal risk. Tie-breaker priority: (1) fewest Gr≥3 events, (2) lowest discontinuation rate, (3) longest mOS, (4) familiarity / payer access.

This is a single-number summary that captures the safety/efficacy trade-off in a way that scales with both severity and duration. A regimen that gains 14 mo PFS at 1% fatal risk (CLEAR-like) beats one that gains 6 mo at 0.8% fatal risk — and the comparator surfaces that judgment automatically.

RCC 1L — the cleanest use case.

RCC 1L is the only indication with three approved IO+TKI combinations. Indirect comparison via the framework provides decision support where no head-to-head trial exists. Pembro+Lenva, Nivo+Cabo, and Pembro+Axi can be evaluated for the same patient profile, and the comparator surfaces the trade-offs explicitly.

For endometrial 2L and HCC 1L, the comparator's value is narrower — endometrial has only one approved IO+TKI; HCC has one approved IO+anti-VEGF (atezo+bev) and one negative trial (pembro+lenva via LEAP-002).

What the comparator isn't.

• Not a substitute for randomized trial. No head-to-head IO+TKI trial exists. The comparator is an indirect-comparison framework, not RCT evidence.
• Efficacy comparisons across non-randomized cohorts are exploratory. Different trial populations, different statistical priors, different follow-up duration. Safety side is more comparable than efficacy side because the safety mechanisms (irAE + VEGFR) are class-conserved.
• Patient-level decisions should incorporate factors not in the comparator — payer access, prior treatment history, patient preference, comorbidity stack beyond what's in the input set.