Predictive and prognostic impacts of SCLC comprehensive index (SCI) in extensive-stage small-cell lung cancer (ES-SCLC) treated with chemo-immunotherapy.

8100 Background: First-line (1L) chemo-immunotherapy is standard of care for ES-SCLC with improved survival outcomes (IMpower133 and CASPIAN). However, there were no reliable biomarkers associated with survival outcomes in these patients (pts). This study aimed to develop a SCI signature using RNA profiling with the nCounter system to predict chemo-immunotherapy outcomes. Methods: The SCI genes were selected based on IMpower133 transcriptomic data (Cancer Cell 2024;42:429) by K-means clustering to determine optimal cut-offs and risk grouping. A gene scoring system was developed with weights being assigned to genes linked to better survival rates. The validation cohort consisted of 93 ES-SCLC pts who received 1L chemo-immunotherapy (etoposide, carboplatin, and atezolizumab) at Seoul National University Hospital (SNUH). NanoString nCounter analysis was performed on all FFPE samples and RNA-seq was validated on 40 samples. Cox proportional hazards models were used for univariable and multivariable analyses. Results: The SCI was developed using genes related to neural (N=7), epithelial-to-mesenchymal transition (N=5), tumor-associated macrophages (N=5), and the T-cell inflamed signature (TIS) (N=18). The SCI signature also included molecular subtypes (N=3), targetable genes (N=4) and 5 housekeeping genes. Our validation cohort included 93 pts with mean age of 69 years and male-to-female ratio of 7.5:1. The median progression-free survival (PFS) and overall survival (OS) were 5.7 months and 12.9 months, respectively. SCLC molecular subtypes were as follows: SCLC-ASCL1 (39%) -NEUROD1 (24%) -POU2F3 (3%) -YAP1 (32%) and -Inflamed (TIS) (9%). The SCI model using 47 genes stratified pts into high- (N=29), intermediate- (N=48), and low- (N=16) risk groups with median PFSs of 4.8, 5.9, and 9.9 months and median OSs of 8.1, 14.3, and 24.4 months, respectively. In this cohort, the high-risk group showed significantly worse PFS (HR=4.68, P < 0.001) and OS (HR=5.03, P < 0.001) compared to the low-risk group. Similarly, in the IMpower133 cohort, the high-risk group demonstrated poorer outcomes with PFS (HR=2.33, P = 0.001) and OS (HR=3.47, P < 0.001) compared to the low-risk group. Conclusions: The SCI 47-gene panel based on IMpower133 transcriptome was validated through nCounter analysis system and effectively stratified ES-SCLC pts into distinct risk groups with strong predictive and prognostic capacities. It provides a practical biomarker for guiding immunotherapy in pts with ES-SCLC. Correlative analyses of nCounter with RNA-seq and AI-powered TIL will be presented.