Integrated Multi-Omics Model Predicts Whole-Brain Radiotherapy Response in Small Cell Lung Cancer Brain Metastases

Clinical Summary:

• Design/Population: This retrospective multi-omics study evaluated 144 patients with small cell lung cancer (SCLC) brain metastases treated with whole-brain radiotherapy, integrating clinical, radiomic, and dosiomic features to develop models predicting short-term treatment response.
• Key Outcomes: Concurrent chemoradiotherapy and conformal boost radiotherapy were independently associated with response to whole-brain radiotherapy. A hybrid model integrating clinical, radiomic, and dosiomic features demonstrated the strongest predictive performance and outperformed models based on individual feature sets.
• Clinical Relevance: These findings suggest that combining clinical, radiomic, and dosiomic data may improve prediction of response to whole-brain radiotherapy in patients with SCLC brain metastases and support further validation of multi-omics approaches for individualized radiotherapy planning.

Results from a retrospective multi-omics analysis demonstrated that integrating clinical, radiomic, and dosiomic features improved prediction of response to whole-brain radiotherapy in patients with small cell lung cancer (SCLC) brain metastases.

“Dosiomics and radiomics elaborate the low-and high-order features extracted from images to predict clinical outcomes,” stated Yifan Lei, MD, Kunming Medical University, Yunnan, China, and coauthors. “The study seeks to develop accurate machine learning models to predict the radiotherapy response of [whole-brain radiotherapy].”

In this study, investigators analyzed 144 patients with SCLC brain metastases treated with whole-brain radiotherapy between January 2020 and June 2024. Radiomic features were extracted from pretreatment CT images, while dosiomic features were derived from radiotherapy dose distributions. Machine learning models incorporating clinical, radiomic, dosiomic, and combined multi-omics features were developed to predict short-term treatment response.

Patients were categorized as responders (complete or partial response; n = 74) or nonresponders (stable or progressive disease; n = 70). The primary objective was to compare the predictive performance of individual and integrated models, with additional external validation and development of a nomogram for individualized response prediction.

Multivariable analysis identified concurrent chemoradiotherapy (P = .042), conformal boost radiotherapy (P = .027), and selected radiomic and dosiomic features as independent predictors of treatment response. Following feature selection, three dosiomic and four radiomic features were incorporated into the final prediction model.

The integrated multi-omics model achieved the highest predictive performance, with an area under the receiver operating characteristic curve (AUC) of 0.792 in the training cohort and 0.711 in the external validation cohort. Nomograms derived from the model also demonstrated favorable calibration and potential clinical applicability.

 “The integration of clinical parameters with dosiomics and radiomic features in a multi-omics framework demonstrates enhanced predictive accuracy for assessing whole-brain radiotherapy outcomes in [SCLC],”concluded Dr Lei et al. “This comprehensive approach may facilitate clinical decision-making by enabling more precise treatment customization and individualized therapeutic strategies.”

Source:

Lei Y, Bai H, Gong C, et al. Multi‐omics predicts radiotherapy response in small cell lung cancer patients receiving whole brain irradiation. J Appl Clin Med Phys. Published online: January 22, 2026. doi: 10.1002/acm2.70466