Radiomics-Based Prediction of Lymphedema after Radiotherapy in Breast Cancer: Integrating Clinical and Dosimetric...

Area of Science:

• Oncology
• Medical Imaging
• Radiotherapy

Background:

• Arm lymphedema is a frequent complication for breast cancer patients post-radiotherapy.
• Current prediction models use clinical and dosimetric factors but have limitations.

Purpose of the Study:

• To develop and evaluate a predictive model for arm lymphedema incorporating radiomic features.
• To improve the accuracy of lymphedema risk prediction in breast cancer patients undergoing postoperative radiotherapy (PORT).

Main Methods:

• A predictive model was developed using clinical, dosimetric, and radiomic features from 532 breast cancer patients.
• Radiomic features were extracted from CT scans of specific anatomical regions relevant to PORT.
• Least absolute shrinkage and selection operator regression was used for feature selection and model building.

Main Results:

• The combined model (clinical, dosimetric, radiomic) demonstrated superior predictive performance (AUC: 0.779) compared to models using only clinical/dosimetric (AUC: 0.717) or only radiomic features (AUC: 0.708).
• The combined model achieved higher accuracy, sensitivity, and specificity.
• Statistical analysis confirmed the significant outperformance of the combined model over the clinical/dosimetric model.

Conclusions:

• Integrating radiomic features with clinical and dosimetric data enhances lymphedema prediction accuracy in breast cancer patients receiving PORT.
• This advanced predictive model holds potential for guiding personalized treatment strategies and improving patient outcomes.