LIPI strongly stratifies survival in ES-SCLC but does not predict chemo-IO benefit

In a post hoc pooled analysis of the Phase 3 CASPIAN (durvalumab + platinum/etoposide) and IMpower133 (atezolizumab + carboplatin/etoposide) trials, the Lung Immune Prognostic Index (LIPI; good = dNLR<3 & LDH<ULN; intermediate = one factor; poor = dNLR≥3 & LDH≥ULN) was evaluated as a prognostic tool in first-line extensive-stage SCLC (n=1140; good 34%, intermediate 49%, poor 17%; 708 received chemo-IO and 432 chemo alone). Median OS fell stepwise by LIPI group—14.6 mo (good) vs 10.9 mo (intermediate) vs 8.4 mo (poor), p<0.0001—and remained independently prognostic in multivariable models for OS (HR poor vs good 1.76, 95%CI 1.45–2.15; p<0.001) and PFS (HR 1.59, 95%CI 1.33–1.90; p<0.001). This sits on top of the established first-line chemo-IO benchmark in ES-SCLC (e.g., IMpower133 OS 12.3 vs 10.3 mo; HR 0.70; CASPIAN durvalumab+EP OS 12.9 vs 10.5 mo; HR 0.71), suggesting LIPI is more about baseline risk than picking an IO “winner” (no significant treatment-by-LIPI interaction was seen), and “poor LIPI” patients also enriched for adverse features (e.g., more liver mets, lower albumin).