Radiomic-Based Prediction of Lesion-Specific Systemic Treatment Response in Metastatic Disease.

Intensity values in CT scans and their corresponding spatial distribution convey important information.A model to predict lesion-specific response to systemic treatment using image-derived features is proposed.Up to a 5-fold increase in predictive capacity compared to a no-skill classifier was obtained, with AUPRC of 0.79 for the most precise model (FDR = 0.01).Assessing treatment response on a lesion-level acknowledges biological diversity within metastatic subclones, which could facilitate management strategies involving selective ablation of resistant clones in the setting of systemic therapy.