Brian Stewart Pugmire MD, Presenter: Nothing to Disclose

Alexander Ramos Guimaraes MD, PhD, Abstract Co-Author: Speakers Bureau, Siemens AG Expert Witness, Rice, Dolan, Kershaw

Ruth Lim MD, Abstract Co-Author: Consultant, Alexion Pharmaceuticals, Inc Officer, New England PET Imaging System

Howard Jeffrey Weinstein MD, Abstract Co-Author: Nothing to Disclose

Ciprian Catana MD, PhD, Abstract Co-Author: Nothing to Disclose

Michael Stanley Gee MD, PhD, Abstract Co-Author: Nothing to Disclose

Combined PET/MRI is a promising new imaging modality in children, particularly pediatric cancer patients. Data regarding pediatric PET/MRI are scarce. In this pilot study, we report our initial experience with PET/MRI in young cancer patients.

Patients ≤19 years of age with known malignancy and not requiring sedation were eligible for this study. After completion of a clinically indicated 18F-FDG PET/CT, whole body PET/MRI was performed on a Siemens mMR with no additional radiotracer. Attenuation correction for the PET/MRI was performed using Dixon in-/opposed-phase sequences. PET/MRI SUVmax and DWI ADCmin values were calculated for lesions visualized on PET/CT using co-localized 2D ROIs. Test performance characteristics of PET/MRI parameters for detecting malignant lesions were calculated using lesion PET/CT SUVmax >2.5 as a reference standard for malignancy. Statistical analyses of correlation between PET/CT and PET/MRI SUVmax values were performed using linear regression.

4 patients were enrolled with a total of 5 paired PET/CT and PET/MRI exams. Mean delay from tracer injection to PET/CT and PET/MRI was 70 and 175 minutes, respectively, with an average PET/MRI scan time of 35 minutes. Mean total effective dose for PET/CT was 12.94mSv, including 7.2mSv (56%) from CT. A total of 7 malignant and 19 benign lesions were included for analysis. There was significant correlation between PET/CT and PET/MR SUVmax for all lesions (r2=0.95, p<0.00001), as well as between ADCmin and PET/CT SUVmax (r2=0.201, p=0.036). The sensitivity, specificity, PPV, and NPV of PET/MRI for detection of malignant lesions were 100%. Using a threshold of <1.0x10-3mm2/sec, PET/MRI DWI exhibited sensitivity and NPV of 100% with 94.7% sensitivity and PPV 80%. 1 indeterminate liver lesion on PET/CT was diagnosed as a complex cyst on PET/MRI.

Our early experience suggests that PET/MRI derived SUVmax and ADC values are sensitive and specific for detection of malignant lesions compared with PET/CT reference. Substitution of PET/MRI for PET/CT would result in significant radiation dose reduction and may help characterize more indeterminate lesions. Studies with more subjects are needed to confirm these findings.

Combined FDG-PET/MRI shows great promise as an imaging modality in children due to its similar cancer detection rates compared to FDG-PET/CT with reduced radiation exposure.

Pugmire, B, Guimaraes, A, Lim, R, Weinstein, H, Catana, C, Gee, M, Early Experience of Combined 18F-FDG PET/MRI in Pediatric Cancer Patients.  Radiological Society of North America 2014 Scientific Assembly and Annual Meeting, - ,Chicago IL. http://archive.rsna.org/2014/14045515.html