Development and Validation of a Deep Learning System for Segmentation of Abdominal Muscle and Fat on Computed Tomography

Abstract

Objective: We aimed to develop and validate a deep learning system for fully automated segmentation of abdominal muscle and fat areas on computed tomography (CT) images. Materials and Methods: A fully convolutional network-based segmentation system was developed using a training dataset of 883 CT scans from 467 subjects. Axial CT images obtained at the inferior endplate level of the 3rd lumbar vertebra were used for the analysis. Manually drawn segmentation maps of the skeletal muscle, visceral fat, and subcutaneous fat were created to serve as ground truth data. The performance of the fully convolutional network-based segmentation system was evaluated using the Dice similarity coefficient and cross-sectional area error, for both a separate internal validation dataset (426 CT scans from 308 subjects) and an external validation dataset (171 CT scans from 171 subjects from two outside hospitals). Results: The mean Dice similarity coefficients for muscle, subcutaneous fat, and visceral fat were high for both the internal (0.96, 0.97, and 0.97, respectively) and external (0.97, 0.97, and 0.97, respectively) validation datasets, while the mean cross-sectional area errors for muscle, subcutaneous fat, and visceral fat were low for both internal (2.1%, 3.8%, and 1.8%, respectively) and external (2.7%, 4.6%, and 2.3%, respectively) validation datasets. Conclusion: The fully convolutional network-based segmentation system exhibited high performance and accuracy in the automatic segmentation of abdominal muscle and fat on CT images.

Assessment of Myosteatosis on Computed Tomography by Automatic Generation of a Muscle Quality Map Using a Web-Based Toolkit: Feasibility Study

“…The selected CT image was saved as a digital imaging and communications in medicine (DICOM) file and was uploaded to our web-based toolkit in a drag-and-drop manner. First, we performed automatic segmentation of the abdominal compartments using a predeveloped deep learning model based on a fully convolutional network [ 10 ]. This model was reported to segment abdominal body compartments into the total abdominal muscle area (TAMA), subcutaneous fat area, and visceral fat area with a dice similarity coefficient of 0.97 and mean cross-sectional area error of 2.26% [ 10 ].…”

Preoperative nutritional risk index and postoperative one-year skeletal muscle loss can predict the prognosis of patients with gastric adenocarcinoma: a registry-based study

“…Preoperative CT was checked within 1 month before surgery on average. Body composition was assessed with abdominopelvic CT using an automated artificial intelligence software (AID-U™, iAID Inc., Seoul, Korea), which was developed using a fully convolutional network (FCN) segmentation technique [16]. A specialized abdominal radiologist (K.W.K), who was blinded to the clinical information, selected the axial CT slice at the L3 vertebral inferior endplate level in a semi-automatic manner with the aid of coronal reconstructed images.…”

Technical Note: Automatic segmentation of CT images for ventral body composition analysis

“…Park et al (2020) 22 Weston et al (2018) 19 Lee et al (2017) 13 Wang et al (2017) 20 This study 150 2D slices were used to evaluate the body compartment segmentation results. Even though the total number of slices for evaluation is similar to the comparable studies, using significantly more slices for a more thorough evaluation might be necessary in the future work.…”