Enhancing Nasopharyngeal Carcinoma Survival Prediction: Integrating Pre- and Post-Treatment MRI Radiomics with...

Area of Science:

• Oncology
• Radiology
• Medical Imaging

Background:

• Nasopharyngeal carcinoma (NPC) frequently recurs despite high remission rates, causing significant morbidity.
• Predicting NPC survival is crucial for effective treatment planning and patient management.

Purpose of the Study:

• To develop a prediction model for NPC survival using pre- and post-treatment MRI radiomics and clinical data.
• To focus on 3-year progression-free survival (PFS) as the primary outcome.
• To create a tool for individualized treatment planning and patient counseling.

Main Methods:

• Retrospective data collection from 276 NPC patients across three hospitals.
• Extraction of 3404 radiomics features from pre- and post-treatment contrast-enhanced T1-weighted MRI scans.
• Application of feature selection, Cox proportional hazards models, ROC analysis, Kaplan-Meier method, and nomogram construction.

Main Results:

• A synergistic combination of pre- and post-treatment clinical and radiomics data proved crucial for NPC survival prediction.
• The developed prediction model demonstrated robust performance with AUCs of 0.66 (training), 0.717 (testing), and 0.827 (validation).
• The model accurately prognosticates patient outcomes, indicating its potential clinical utility.

Conclusions:

• A novel and effective prediction model for NPC survival was developed, integrating pre- and post-treatment clinical and MRI radiomics data.
• The constructed nomogram offers a valuable tool for clinicians in individualized treatment planning and patient counseling.
• This approach holds significant implications for advancing NPC research and patient care.