Imaging-based PRediction of Eligibility for ChemoImmunotherapy in reSEctable NSCLC, iPRECISE

This is a prospective observational study designed to evaluate imaging biomarkers in patients with resectable non-small cell lung cancer who receive neoadjuvant chemoimmunotherapy before planned surgery. Neoadjuvant chemoimmunotherapy has become an important treatment option for resectable non-small cell lung cancer. Although pathologic complete response and major pathologic response are associated with favorable outcomes, reliable noninvasive methods for predicting these responses remain limited. Imaging biomarkers derived from computed tomography may provide a quantitative and repeatable way to assess tumor burden, nodal disease, treatment response, and treatment-related lung toxicity over time. In this study, participants will undergo chest computed tomography before starting neoadjuvant chemoimmunotherapy and again after completion of neoadjuvant treatment before surgery. If disease progression is clinically suspected before planned surgery, an additional computed tomography scan may be performed and used for response assessment. Computed tomography scans will be acquired using a standardized imaging protocol. In addition to conventional computed tomography reconstruction, high-resolution deep learning-based computed tomography reconstruction will be applied after image acquisition. The study will evaluate imaging features of the primary tumor, lymph nodes, suspected extranodal extension, adjacent structure involvement, radiologic treatment response, and possible immune-related pneumonitis. Imaging findings will be compared with clinical data, molecular and immunohistochemical findings, and surgical pathology findings, including pathologic complete response, major pathologic response, final pathologic stage, extranodal extension, and spread through air spaces when available. The study will also explore whether imaging biomarkers are associated with progression-free survival, overall survival, treatment-related pneumonitis, and whether high-resolution deep learning-based CT reconstruction improves imaging biomarker development compared with conventional reconstruction. The anticipated enrollment is 150 participants.