What question did this study set out to answer?

This research aims to determine whether radiomics and deep learning models can predict outcomes for advanced non-small cell lung cancer in patients treated with immunotherapy and stereotactic radiotherapy.

May 24, 2026Open Access

Deep learning and radiomics models in patients with advanced non-small cell lung cancer treated with immunotherapy combined with stereotactic radiotherapy

Key Points

This research aims to determine whether radiomics and deep learning models can predict outcomes for advanced non-small cell lung cancer in patients treated with immunotherapy and stereotactic radiotherapy.
Utilized data from randomized trials: PEMBRO-RT, NIVORAD, and MDACC
Developed random forest and deep learning models to assess predictive outcomes based on radiomics features
Conducted external validation of models on different trial data sets.
Deep learning model achieved an AUC of 0.92 for predicting per-lesion progressive disease, surpassing the random forest model's AUC of 0.57.
Improvement in overall survival prediction with a concordance index of 0.67 after adding clinical features.
External validation resulted in reduced AUCs, indicating lower predictive performance in different datasets.

Abstract

Quantitative imaging is an emerging field that may allow prediction of oncological outcomes. We investigate whether radiomics and deep learning can predict outcomes in metastatic non-small cell lung cancer utilizing randomized trials of PD-1 inhibitors + /− stereotactic ablative body radiotherapy: PEMBRO-RT(NCT02492568), NIVORAD(ACTRN12616000352404) and MDACC(NCT02444741). A random forest model developed on PEMBRO-RT using radiomics features had an AUC of 0.57 for prediction of per-lesion progressive disease on immunotherapy compared to an AUC of 0.92 for a deep learning model. A random forest survival model using radiomics features for overall survival (progression free survival) had a concordance index of 0.63(0.59) and improved to 0.67(0.65) by adding clinical features, including PD-L1 and treatment arm. Validation on NIVORAD and MDACC revealed reduced AUCs. Overall, a deep learning compared to a radiomics model demonstrated excellent predictive value for per-lesion progressive disease for patients on immunotherapy. Models had reduced performance on external validation. Research improving generalizability is required for clinical translation.

Mark Helpful

Bookmark

Relay

View Full Paper