A Computational Histology Artificial Intelligence Prognostic Biomarker in Non‑Small Cell Lung Cancer Using the Cancer Genome Atlas
CHAI‑positive patients had significantly worse overall survival than CHAI‑negative patients.
The study evaluated whether a computational histology artificial intelligence (CHAI) biomarker, derived solely from diagnostic hematoxylin‑and‑eosin (H&E) whole‑slide images, provides prognostic information independent of established clinicopathologic factors in non‑small cell lung cancer (NSCLC). Using TCGA LUAD and LUSC cohorts, 914 patients were stratified, split into a development (30%) and a validation (70%) set, and a CHAI histologic signature was trained and validated. The primary endpoint was overall survival (OS), and the secondary endpoint was progression‑free interval (PFI). In the validation cohort (n=644), CHAI‑positive patients had significantly worse OS than CHAI‑negative patients (3‑year OS: 53% vs. 69%). Multivariable Cox models adjusted for age, sex, smoking status, AJCC stage, histologic subtype, and genomic alterations (EGFR, KRAS, BRAF, ERBB2) confirmed the prognostic value of CHAI. The CHAI biomarker demonstrated independent prognostic significance for both OS and PFI.
The CHAI biomarker provides independent prognostic information in NSCLC, potentially guiding risk‑based clinical decision‑making and identifying patients who may benefit from intensified therapy.
Evidence level: Állatkísérletes. Állatmodellben vizsgálták.
Related signals
- Telisotuzumab Vedotin + Osimertinib for Non‑Small Cell Lung Cancer
- Ramucirumab plus erlotinib in patients with untreated, EGFR‑mutated, advanced non‑small‑cell lung cancer (RELAY): a randomised, double‑blind, placebo‑controlled, phase 3 trial
- Recent advancements in lung cancer research: a narrative review
- Advances in Non-Small Cell Lung Cancer: Current Insights and Future Directions
- Structured proxy features for multimodal NSCLC survival prediction from pretreatment CT