Abstract The present study aims to evaluate the effectiveness of retrospective positron emission tomography/magnetic resonance (PET/MR) fusion, in conjunction with the Neck Imaging Reporting and Data System (NI-RADS) classification system, to improve diagnostic accuracy and facilitate the detection of recurrence in patients with oral cancer. This retrospective study analyzed 106 patients with surgically treated oral cavity malignancies followed by adjuvant chemotherapy and/or radiotherapy and under surveillance for 9 to 12 months. In all patients, PET data was fused with postcontrast magnetic resonance imaging (MRI) images using automated software, and various MRI sequences were analyzed. Statistical analysis was performed using GraphPad Prism (version 10.2). Majority of patients were male (90.6%), aged 51 to 60 years (30.2%), with single lesions (96.2%). Histopathology confirmed 70% malignant and 30% benign lesions. MRI had high concordance with pathology in NI-RADS grade I (90.5%) but struggled with higher grades. PET showed similar trends, with increasing discordance in intermediate-risk cases. PET/MR fusion outperformed both modalities, achieving the highest concordance rates, especially in grades I (88.2%), II (93.3%), and IV (100%). PET/MR fusion demonstrated superior diagnostic performance with 96.1% sensitivity, 87.88% specificity, and 93.64% accuracy. Receiver operating characteristic analysis confirmed its superiority (area under the ROC curve: 0.948). PET/MR fusion offers superior detection rates for recurrent tumors in oral cancer surveillance, in comparison to PET and MRI alone with NI-RADS category I and II, while MRI with diffusion-weighted imaging excels in identifying perineural spread and locoregional extension. By leveraging both modalities, it minimizes false negatives and facilitates earlier interventions, improving patient outcomes.
Shah et al. (Fri,) studied this question.