68Ga DOTATATE PET/MR Imaging of Well-Differentiated Primary and Metastatic Liver Neuroendocrine Tumors

Purpose: To evaluate the potential of integrated multiparametric 68 Ga DOTATATE PET/MR imaging for assessing liver lesions of well-differentiated neuroendocrine tumors (NETs) and to identify imaging parameters predictive of primary tumor localization. Patients and Methods: This retrospective study involves patients with well-differentiated NETs who underwent 68 Ga DOTATATE PET/MRI between September 2018 and November 2024. Inclusion criteria required histopathologically proven NETs with 68 Ga DOTATATE-avid liver metastases and complete multiparametric MRI sequences. PET and MRI-derived variables, including SUV max , ADC min , T/L ratios, and tumor volume (log-transformed tumor volume: LOGVOL), were analyzed. Linear mixed-effects models and logistic regression analyses were performed to identify relationships between imaging features and tumor characteristics. ROC analyses were conducted to evaluate the accuracy of primary tumor origin predictions. Results: Of 43 imaging sessions, 14 patients (7 male, 7 female; median age 59 y) with 181 lesions met the inclusion criteria. SUV max was significantly correlated with LOGVOL and contrast enhancement parameters (eg, WO liver ). Linear mixed-effects models revealed that LOGVOL and WO liver were independent predictors of SUV max . In the binomial regression analysis, tumor precontrast T1 intensity, T/L art , and T/L ven were significant factors in differentiation between pancreatic and gastrointestinal (GIS) NET metastases, with pancreatic tumors demonstrating higher T/L art and GIS tumors exhibiting higher T/L ven and T1 intensity. Logistic regression achieved an AUC of 0.911, with a sensitivity of 86% and specificity of 76%. Conclusions: 68 Ga DOTATATE PET/MRI effectively integrates metabolic and anatomical imaging for characterizing NET liver metastases. Parameters such as LOGVOL and WO liver independently predict SUV max , while precontrast T1 intensity, T/L art , and T/L ven assist in differentiating pancreatic from GIS NETs. These findings underscore the potential of 68 Ga DOTATATE PET/MRI in personalized NET management and suggest avenues for further research to confirm these results.