Comparison between PET/MR and PET/CT in evaluation of oncological patients

Objective To verify the feasibility of the integrated PET/MR for oncological applications by comparing PET/MR with PET/CT in terms of lesion detection and quantitative measurement. Methods A total of 277 patients (165 males, 112 females, average age (52.9±12.6) years) voluntarily participated in this same-day PET/CT and PET/MR comparative study. The time interval between the two studies was 15-35 min. PET/CT images were acquired and reconstructed following standard protocols. PET/MR covered the body trunk with a sequence combination of transverse T1 weighted imaging (WI) 3D-volumetric interpolated breath-hold, T2WI turbo spin echo with fat saturation, diffusion-weighted imaging, and simultaneous PET acquisition. PET images were reconstructed by vender-provided attenuation correction methods. The results of PET/CT and PET/MR were regarded as positive if any modality (CT, PET or MRI) was positive. SUVmax was obtained by the manually drawn ROI. Detection rates were compared with χ2 test and SUVmax from the two modalities was analyzed with Spearman correlation analysis. Results A total of 353 lesions were detected in 220 patients. Compared to PET/CT, PET/MR revealed 30 additional true-positive lesions, while missed 6. The detection rates between PET/CT and PET/MR were significantly different (P<0.05). The lesion-based and patient-based consistency was 89.8% (317/353) and 85.9% (189/220), respectively. There were significant correlations of SUVmax between PET/MR and PET/CT for lesions(rs=0.91, P<0.01) and for normal tissues(rs=0.62-0.76, all P<0.01). Conclusions With reference to PET/CT, integrated PET/MR may provide comparable semi-quantitative measurements of pathological lesions as well as normal tissues. Integrated PET/MR may be more effective to detect lesions in abdomen and pelvis. Key words: Tomography, emission-computed; Tomography, X-ray computed; Magnetic resonance imaging; Image fusion; Comparative study