Radiomics Analysis for Predicting Medication-Related Osteonecrosis of the Jaw Using Computed Tomography.

Area of Science:

• Oral and Maxillofacial Radiology
• Medical Imaging
• Oncology

Background:

• Medication-related osteonecrosis of the jaw (MRONJ) is a serious complication associated with bone resorption inhibitors.
• Accurate prediction of MRONJ risk before treatment initiation is crucial for patient management.

Purpose of the Study:

• To evaluate the efficacy of pretreatment radiomics in predicting the risk of developing medication-related osteonecrosis of the jaw (MRONJ).

Main Methods:

• Retrospective analysis of 94 patients who underwent MDCT before bone resorption inhibitor therapy.
• Extraction and quantification of radiomic features from segmented cancellous bone and surrounding cortical bone.
• Development and evaluation of machine learning models (Random Forest, SVM, MLP) using sensitivity, specificity, and AUC.

Main Results:

• Models incorporating cancellous and cortical bone achieved high predictive performance, with AUCs up to 0.931.
• Machine learning models demonstrated significant accuracy in differentiating patients who developed MRONJ from those who did not.
• The inclusion of cortical bone in segmentation improved prediction accuracy compared to cancellous bone alone.

Conclusions:

• Pretreatment radiomics analysis of MDCT images shows strong potential for predicting MRONJ risk.
• Radiomics offers a non-invasive tool for early identification of patients at high risk for MRONJ.
• Further validation may integrate radiomics into clinical decision-making for patients on bone resorption inhibitors.