CT radiomics model predicts NSCLC treatment response in distinct therapy cohorts

CT radiomics-based SVM model predicted early treatment response in advanced NSCLC, with AUC 0.90 (95% CI 0.82-0.97) for chemo-immunotherapy and 0.84 (95% CI 0.73-0.95) for immunotherapy, externally validated.

• In the internal cohort (n=90), the support vector machine (SVM) model achieved a negative predictive value of 94% for chemo-immunotherapy and positive predictive value of 80% for immunotherapy alone.
• External validation in 40 patients from a second center confirmed generalizability.
• CT radiomic features from baseline scans may serve as imaging biomarkers for treatment stratification.

Automated summary

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