Habitat-based CT radiomics model predicts complete response after neoadjuvant chemoimmunotherapy in NSCLC

A habitat-based CT radiomics model predicted pathological complete response after neoadjuvant chemoimmunotherapy for non-small cell lung cancer with AUCs of 0.74 (training), 0.71 (internal), and 0.67 (external) — outperforming whole-tumor models.

• Multicenter retrospective study of 287 NSCLC patients: training (n=147), internal validation (n=63), external test (n=77).
• SHAP analysis identified features from the high-entropy habitat subregion as most predictive.
• Combined model (habitat + whole-tumor) did not improve upon the habitat-based model alone.

Automated summary

RadPigeon summaries are original and for information only. They are not clinical advice.