Quantitative Comparison of Misregistration in Abdominal and Pelvic Organs Between PET/MRI and PET/CT: Effect of Mode of Acquisition and Type of Sequence on Different Organs

Objective: The objective of our study was to quantitatively compare misregistration in selected abdominopelvic organs between PET/CT acquisitions and simultaneous and sequential PET/MRI acquisitions.

Subjects and methods: PET/MR images of 15 healthy volunteers and seven patients with bladder cancer were acquired. Ten clinical PET/CT studies acquired during the same time frame of body mass index-matched control subjects were chosen. PET/MRI and PET/CT registration of selected abdominopelvic organs was measured and compared.

Results: The overall mean misregistration with PET/MRI was significantly higher than that with PET/CT (p < 0.001). Sequential PET/MRI acquisition was significantly inferior to PET/CT (p = 0.02), whereas there was no significant difference between simultaneous PET/MRI acquisition and PET/CT (p = 0.38). Simultaneous PET/MRI acquisition was significantly better than sequential PET/MRI acquisition (p < 0.001). The mean misregistration for all organs with the T1-weighted volumetric interpolated breath-hold examination (VIBE) sequence (2.39 cm) was significantly inferior to PET/CT (p < 0.001). Although the T2-weighted HASTE breath-hold sequence was significantly inferior to PET/CT (p = 0.04), the T2 HASTE non-breath-hold sequence and T2 STIR sequence (0.18 cm) were significantly superior to both PET/CT and the T1 VIBE sequence (p < 0.001). Within the same sequence (T1 VIBE breath-hold sequence), the mean misregistrations with sequential and simultaneous PET/MRI acquisitions were both significantly greater than with PET/CT (p < 0.001), whereas simultaneous PET/MRI acquisition was superior to sequential PET/MRI acquisition (p < 0.001).

Conclusion: In the abdominopelvic organs, sequentially obtained PET/MRI data have significantly higher misregistration than both PET/CT data and simultaneously acquired PET/MRI data. Simultaneously obtained PET/MRI data are statistically noninferior to PET/CT. Nonradial T1 VIBE has the highest misregistration, whereas T2 STIR and T2 HASTE non-breath-hold are significantly better than both PET/CT and T1 VIBE.