Preoperative CT-based radiomic prognostic index to predict the benefit of postoperative radiotherapy in patients with...

Area of Science:

• Radiology and Oncology
• Medical Imaging Analysis
• Cancer Prognostics

Background:

• The efficacy of postoperative radiotherapy (PORT) in non-small cell lung cancer (NSCLC) is debated.
• Identifying NSCLC patients who may benefit from PORT is crucial for personalized treatment.
• Preoperative imaging features may offer prognostic value.

Purpose of the Study:

• To develop and validate radiomic prognostic indexes (RPI and LRPI) using preoperative CT scans.
• To identify NSCLC patients who could benefit from PORT based on these indexes.
• To assess the predictive performance of LRPI compared to traditional staging methods.

Main Methods:

• Radiomic texture features were extracted from preoperative CT scans of NSCLC patients.
• A least absolute shrinkage and selection operator-Cox model was used for feature selection and RPI construction.
• The lymph-radiomic prognostic index (LRPI) was developed by combining RPI with positive lymph node count.

Main Results:

• The RPI demonstrated prognostic value for overall survival (OS) across six independent cohorts.
• LRPI showed superior prediction of OS compared to pT stage, pT+PLN, and RPI alone.
• Patients categorized as moderate-risk by LRPI significantly benefited from PORT (HR=0.60, p=0.02).

Conclusions:

• Preoperative CT-based radiomic and lymph-radiomic prognostic indexes (RPI and LRPI) were successfully developed.
• These indexes can predict overall survival in NSCLC patients.
• LRPI effectively identifies NSCLC patients who benefit from postoperative radiotherapy.