An Optimized Binning and Probabilistic Slice Sharing Algorithm for Motion Correction in Abdominal DW-MRI
Diffusion-weighted magnetic resonance imaging (DW-MRI) is a powerful, non-invasive tool for detecting and characterizing abdominal lesions to facilitate early diagnosis, but respiratory motion during a scan reduces image quality and accuracy of quantitative biomarkers. Respiratory binning, which groups image slices into motion phase bins based on a navigator signal, can help mitigate motion artifacts. However, in DW-MRI, the standard binning technique often generates volumes with missing slices along the superior-inferior axis. Thus, longer scans are required to obtain volumes without gaps. In this study, we proposed a new binning technique to minimize missing slices without increasing scan time. We first designed an algorithm using dynamic programming and prefix sum approaches to optimize the initial binning of MR images. Then, we developed a probabilistic refinement phase, selecting some slices to belong in two neighboring bins to further reduce missing slices. We tested our two-phase technique on free-breathing abdominal DW-MRI scans from eight subjects, including one with tumors. The proposed technique significantly reduced missing slices compared to standard binning (p<1.0*10-15), yielding an average reduction of 81.74+/-7.58%. Our technique also reduced motion artifacts, improving the conspicuity of malignant lesions. Apparent Diffusion Coefficient (ADC) maps generated from free-breathing scans corrected using the proposed technique had lower intra-subject variability compared to ADC maps from uncorrected free-breathing and shallow-breathing scans (p<0.001). Additionally, ADC maps from shallow-breathing scans were more consistent with corrected free-breathing maps rather than uncorrected free-breathing maps (p<0.01). The proposed technique corrects for motion while simultaneously reducing missing slices, allowing for shorter acquisition times compared to standard binning.