Single slice thigh CT muscle group segmentation with domain adaptation and self-training

Abstract

Thigh muscle group segmentation is important for assessing muscle anatomy, metabolic disease, and aging. Many efforts have been put into quantifying muscle tissues with magnetic resonance (MR) imaging, including manual annotation of individual muscles. However, leveraging publicly available annotations in MR images to achieve muscle group segmentation on single-slice computed tomography (CT) thigh images is challenging. Approach:We propose an unsupervised domain adaptation pipeline with selftraining to transfer labels from three-dimensional MR to single CT slices. First, we transform the image appearance from MR to CT with CycleGAN and feed the synthesized CT images to a segmenter simultaneously. Single CT slices are divided into hard and easy cohorts based on the entropy of pseudo-labels predicted by the segmenter. After refining easy cohort pseudo-labels based on anatomical assumption, self-training with easy and hard splits is applied to fine-tune the segmenter.Results: On 152 withheld single CT thigh images, the proposed pipeline achieved a mean Dice of 0.888 (0.041) across all muscle groups, including gracilis, hamstrings, quadriceps femoris, and sartorius muscle.Conclusions: To our best knowledge, this is the first pipeline to achieve domain adaptation from MR to CT for thigh images. The proposed pipeline effectively and robustly extracts muscle groups on two-dimensional single-slice CT thigh images.

Deep conditional generative model for longitudinal single-slice abdominal computed tomography harmonization

“…5 Existing longitudinal CT scans of the abdomen from the Baltimore longitudinal study of aging (BLSA) dataset 6 provide a valuable opportunity to characterize the relationship between body composition changes and age-related disease, cognitive disease, and metabolic health. [7][8][9][10][11][12] To minimize radiation exposure for longitudinal imaging and potential risk associated with contrast administration, two-dimensional (2D) non-contrast axial single-slice CT is taken as opposed to three-dimensional (3D) volumetric CT commonly acquired in clinical practice. However, it is difficult to locate the same cross-sectional location in longitudinal imaging, and thus there is substantial variation in the organs and tissues captured in different years, as shown in Fig.…”