Computational modeling of comprehensive genomic profiling to predict chemo-immunotherapy benefit in early stage NSCLC

Background: Immune checkpoint inhibition (ICI), alone and in combination with chemotherapy (ICI+C), has transformed the treatment landscape for non-small cell lung cancer (NSCLC). We have previously reported results from the myCare-040 study[1], where we validated an algorithm capable of distinguishing advanced NSCLC patients with favorable ICI+C benefit from those with no benefit. To understand whether the underlying molecular mechanisms used by the algorithm are conserved across disease stages, we have evaluated the algorithm in a cohort of patients with early stage NSCLC receiving adjuvant ICI or ICI + C. Design: The ∆TRI algorithm uses Cellworks' computational model of a patient's tumor genomics to predict biomarker changes related to disease progression and potential benefit from ICI+C therapy. The previously validated ∆TRI and clinical threshold (16) were evaluated in 51 non-squamous, early stage NSCLC patients (Stage I=20, Stage II=12, Stage IIIA=19) receiving adjuvant ICI or ICI+C ,with complete clinical and genomic information (Foundation One CDx) derived from the nationwide (US-based) de-identified ConcertAI Genomics360 database. Results: Patients in the ∆TRI High Benefit Group (∆TRI ≥ 16, n = 11), had an incremental benefit in median OS of 19.4 months with the addition of chemotherapy to ICI (logrank p = 0.057, median OS ICI = 7 months vs ICI+C = 26.6 months). In contrast, patients in the ∆TRI No Benefit Group (∆TRI < 16, n = 40) showed no improvement in OS when receiving ICI+C (logrank p = 0.84, median OS ICI = 13 months vs ICI+C = 9 months). A likelihood ratio test of interaction between the linear ∆TRI and treatment (ICI versus ICI+C) was significant (LR p = 0.038). Cut-point optimization for the early stage population (∆TRI= 9) improved the logrank statistics in the High Benefit Group (∆TRI ≥ 9; logrank p = 0.003). Conclusions: Although developed and validated in patients with advanced NSCLC, the ∆TRI also predicted incremental chemotherapy benefit in an real-world cohort of patients with early stage NSCLC receiving adjuvant ICI or ICI+C. Further work is needed to understand how these observations could be translated into clinical use. 1 Aggarawal et al, WCLC 2025