Radiomics-based classification of medication-related osteonecrosis of the jaw using panoramic radiographs

• DOI: 10.1007/s11282-025-00826-1
• Publication Date: 2025-05-05T11:45:49Z
• Authors: Masaru Konishi; Hiromi Nishi; Hiroyuki Kawaguchi; Naoya Kakimoto
• Keywords: Mandible (arthropod mouthpart); Mental foramen; Mandibular nerve

Abstract Objectives Medication-related osteonecrosis of the jaw (MRONJ) caused by bone resorption inhibitors is difficult to treat and reduces patient’s quality life. We aimed classify likelihood MRONJ development using panoramic radiographs taken prior inhibitor administration. Methods included patients who underwent radiographic evaluation at Hiroshima University Hospital Thirty-two with MRON mandible (16 men 16 women) 57 without (23 34 were selected. The was segmented from mental foramen anterior mandibular angle notch on before treatment. image features within this region extracted quantified. Overall, 13 shape, 18 histogram-based, 75 texture-based, 744 wavelet extracted. Least absolute shrinkage selection operator regression used select relevant data. Support vector machine (SVM) neural network multilayer perceptron (MLP) as machine-learning models. sensitivity, specificity, area under curve (AUC) receiver operating characteristic curves evaluate diagnostic performances. Results SVM model achieved a sensitivity 0.667, specificity 0.833, an AUC 0.903 in test group. Meanwhile, MLP 0.750, Conclusion Radiomics analysis administration can differentiate between those MRONJ.