Baseline Radiomic Signature to Estimate Overall Survival in Patients With NSCLC.

Area of Science:

• Radiology
• Oncology
• Medical Imaging

Background:

• Non-small cell lung cancer (NSCLC) treatment response varies significantly.
• Predictive biomarkers for overall survival (OS) are crucial for treatment selection.
• Radiomics offers a quantitative approach to extract imaging features for survival prediction.

Purpose of the Study:

• To develop and validate a baseline radiomic signature for predicting OS in NSCLC patients.
• To assess the association of the radiomic signature with OS in patients treated with nivolumab or chemotherapy.

Main Methods:

• A radiomic signature was developed using baseline computed tomography (CT) images from 758 NSCLC patients treated with nivolumab.
• Radiomic variables were extracted from segmented tumor lesions.
• The signature's association with OS was evaluated in validation (nivolumab) and test (chemotherapy) sets.

Main Results:

• A radiomic signature combining spatial correlation, liver tumor volume, and mediastinal lymph node tumor volume was identified.
• The signature predicted 3-month OS with 86% sensitivity and 77.8% specificity at a 0.55 threshold.
• The signature was significantly associated with OS (p < 0.0001) in both nivolumab- and chemotherapy-treated patients.

Conclusions:

• The developed radiomic signature provides an early prediction of OS for NSCLC patients.
• This imaging biomarker can offer valuable prognostic information.
• The signature may support risk stratification in clinical trials for NSCLC.