Integrating radiomics and real-world data to predict immune-checkpoint inhibitors efficacy in advanced Non-Small Cell Lung Cancer

Background and purpose Immunotherapy (IO) revolutionized the prognosis of patients with Non-Small Cell Lung Cancer (NSCLC). However, identifying optimal candidates for this treatment remains challenging. Based on a previous study suggesting the potential predictive power of radiomics in predicting clinical outcomes in different clinical settings, we aimed to assess the capability of radiomics in predicting IO efficacy in advanced NSCLC patients treated with immunotherapy. Materials and Methods 375 advanced NSCLC patients treated with IO-based regimens from April 2013 to May 2022 were enrolled. Primary lung lesions were segmented and radiomics feature extracted. Using Clinical Benefit Rate (CBR) and survival status at 6 and 24 months (OS6 and OS24) as endpoints, ML classifiers were trained and then evaluated on a test set. Results Model achieving the highest prediction performance predicting long-term survival (OS24), reached an accuracy of 0.71 and AUC of 0.79 on test set with the combination of 20 radiomics features and real-world data (RWD). Combining radiomics with RWD features consistently allowed to outperform the standard predictive biomarker, PD-L1, for the majority of outcomes. Conclusions We identified a radiomics and RWD-based signature able to predict prognosis of NSCLC patients treated with IO therapy. If validated, this model could support oncologists in making prognostications.