CT analysis of thoracolumbar body composition for estimating whole-body composition

Abstract

Background To evaluate the correlation between single- and multi-slice cross-sectional thoracolumbar and whole-body compositions. Methods We retrospectively included patients who underwent whole-body PET–CT scans from January 2016 to December 2019 at multiple institutions. A priori-developed, deep learning-based commercially available 3D U-Net segmentation provided whole-body 3D reference volumes and 2D areas of muscle, visceral fat, and subcutaneous fat at the upper, middle, and lower endplate of the individual T1–L5 vertebrae. In the derivation set, we analyzed the Pearson correlation coefficients of single-slice and multi-slice averaged 2D areas (waist and T12–L1) with the reference values. We then built prediction models using the top three correlated levels and tested the models in the validation set. Results The derivation and validation datasets included 203 (mean age 58.2 years; 101 men) and 239 patients (mean age 57.8 years; 80 men). The coefficients were distributed bimodally, with the first peak at T4 (coefficient, 0.78) and the second peak at L2-3 (coefficient 0.90). The top three correlations in the abdominal scan range were found for multi-slice waist averaging (0.92) and single-slice L3 and L2 (0.90, each), while those in the chest scan range were multi-slice T12–L1 averaging (0.89), single-slice L1 (0.89), and T12 (0.86). The model performance at the top three levels for estimating whole-body composition was similar in the derivation and validation datasets. Conclusions Single-slice L2–3 (abdominal CT range) and L1 (chest CT range) analysis best correlated with whole-body composition around 0.90 (coefficient). Multi-slice waist averaging provided a slightly higher correlation of 0.92.

Development and Validation of an Artificial Intelligence Preoperative Planning and Patient-Specific Instrumentation System for Total Knee Arthroplasty

“…The convolutional neural network segmentation model treats each patient's CT image as an array unit, utilizing GPU acceleration to simultaneously process multiple layers of data. In contrast, manual segmentation requires layer-by-layer segmentation, resulting in lower efficiency [30,31]. Unlike traditional 2D convolutional neural networks, the 3D convolutional neural network employed in this study accounts for spatial and contextual information between CT slices.…”

Association between body fat decrease during the first year after diagnosis and the prognosis of idiopathic pulmonary fibrosis: CT-based body composition analysis

“…We used this measurement approach instead of using whole-body composition results obtained on chest CT because not all patients had exactly the same CT scan coverage. Indeed, multi-slice averaged body composition analysis results at the level of T12-L1 are known to be well correlated with whole-body composition [ 31 ]. Therefore, in this study, the subcutaneous and visceral fat areas were summed and defined as body fat area to represent whole-body fat.…”

A new attempt at full-scale jump connectivity and Transformer

“…On the other hand, the studies by Arayne et al [38] and Koitka et al [39] emphasized the consistency and possible clinical application of body composition analysis at different lumbar levels (e. g., T4 and L3). Furthermore, the study of Hong et al [40] further demonstrated the accuracy of CT analysis of the thoracolumbar spine in the estimation of whole body composition, while the work of Cespedes et al [41] demonstrated the value of an automated approach in the study of cancer prognosis. Finally, Lee et al [42] used an advanced 3D Unet technique to provide an accurate and e cient method for body composition segmentation for whole-body PET-CT images.…”