Predictive Value of 18 F-FDG PET/MRI Parameters for MYCN Amplification in High-Risk Neuroblastoma

Background To investigate the clinical value of ¹⁸ F-fluorodeoxyglucose positron emission tomography/magnetic resonance imaging ( ¹⁸ F-FDG PET/MRI) parameters in predicting the MYCN gene amplification status in patients with high-risk neuroblastoma (HR-NB). Methods A retrospective analysis was conducted on 72 HR-NB patients who underwent ¹⁸ F-FDG PET/MRI examinations at our institution between December 2018 and December 2024. Based on MYCN genetic testing results, the patients were classified into the MYCN amplification group (MNA) and the non-amplification group (MYCN-NA). The clinical data of the patients and the imaging characteristics of the primary tumors were collected. The GE post-processing workstation was used to identify lesions, and quantitative parameters on PET/MRI images were semi-automatically extracted. Multivariable logistic regression analysis was employed to screen for independent predictive factors. Diagnostic performance was assessed using Receiver Operating Characteristic (ROC) curves by calculating the Area Under the Curve (AUC), sensitivity, and specificity. Calibration plot and Decision Curve Analysis (DCA) were used to evaluate the calibration and clinical utility of the models, respectively. A combined model was visualized using a nomogram. Results Multivariate logistic regression analysis identified tumor necrosis (P = 0.039, OR = 5.52; 95% CI: 1.091–27.916), age (P = 0.042, OR = 0.959; 95% CI: 0.920–0.999), and Total Lesion Glycolysis (TLG) (P = 0.011, OR = 1.004; 95% CI: 0.982–1.008) as independent predictive factors of MYCN amplification in HR-NB. ROC curve analysis demonstrated that the diagnostic performance of the combined model had superior diagnostic performance (AUC: 0.858, 95% CI: 0.756–0.929, Sensitivity: 0.724, Specificity: 0.930) compared to using necrosis alone (AUC: 0.648, 95% CI: 0.526–0.757, Sensitivity: 0.621, Specificity: 0.674), age alone (AUC: 0.724, 95% CI: 0.606–0.823, Sensitivity: 0.517, Specificity: 0.930), or TLG alone (AUC: 0.791, 95% CI: 0.679–0.878, Sensitivity: 0.724, Specificity: 0.837). Calibration curves and DCA further confirmed the optimal clinical utility of the combined model. Conclusion The prediction model integrating tumor necrosis, TLG, and age can effectively and non-invasively predict the MYCN amplification status in HR-NB patients, exhibiting good diagnostic efficacy and clinical application potential.