Objective: To validate whether PET/MR could provide a semi-quantitative measurement (SUV(max)) comparable to that produced by PET/CT in normal organs.
Methods: 277 subjects underwent an ordinary ¹⁸F-FDG PET/CT followed by a PET/MR scan with a 25-45 min interval. Region of interest (ROIs) were drawn in 4 reference normal organs/tissues in both MRAC-PET and CTAC-PET images and the liver and erector spinae in the dual-time point PET/CT images.
Results: 259 malignant and 21 benign lesions, pathologically confirmed, were detected in the 220 subjects. SUV(max) derived from PET/CT (SUV(max)-CT) and PET/MR (SUV(max)-MRI) was highly correlated over the reference organ ROIs (r = 0.62-0.73), except lung (r = 0.44). The SUV(max)-MRI was significantly lower than the respective SUV(max)-CT in all 4 organs and after delay-correction in liver and muscle.
Conclusion: The results indicate that PET/MR can provide reliable measurement in physiological organs.