A DEEP learning model offering a powerful tool for predicting how patients with advanced non-small cell lung cancer (NSCLC) will respond to immune checkpoint inhibitors (ICIs), potentially improving treatment personalization, has been newly developed.
In a multicenter cohort study, researchers trained and validated a deep learning-based model using whole slide hematoxylin and eosin-stained images from 958 NSCLC patients treated with ICIs. The study included data from one U.S. center (614 patients) and three European Union centers (344 patients), spanning from August 2014 to December 2022. The model’s ability to differentiate responders from non-responders was measured against established biomarkers, including programmed death-ligand 1 (PD-L1), tumor mutational burden (TMB), and tumor-infiltrating lymphocytes (TILs).
The model demonstrated strong predictive accuracy, achieving an area under the receiver operating characteristic curve (AUC) of 0.75 (95% CI, 0.64-0.85) in internal testing and 0.66 (95% CI, 0.60-0.72) in the validation cohort. It was an independent predictor of both progression-free survival (hazard ratio, 0.56; 95% CI, 0.42-0.76; P < .001) and overall survival (hazard ratio, 0.53; 95% CI, 0.39-0.73; P < .001). The deep learning model outperformed TILs and TMB and was comparable to PD-L1 (AUC, 0.67; 95% CI, 0.60-0.74), with a 10-percentage point improvement in specificity. When combined with PD-L1, the model’s performance improved further (AUC, 0.70; 95% CI, 0.63-0.76), achieving a response rate of 51%, compared to 41% for PD-L1 (≥50%) alone. These findings highlight the potential of AI-driven approaches to refine treatment decisions for NSCLC, offering a more precise method for identifying patients most likely to benefit from ICIs. If integrated into clinical practice, this deep learning model could improve patient outcomes by guiding more effective immunotherapy strategies. Reference: Rakaee M et al. Deep Learning Model for Predicting Immunotherapy Response in Advanced Non-Small Cell Lung Cancer. JAMA Oncol. 2025 Feb 1;11(2):109-118. Anaya Malik | AMJ
