A deep learning model for CT-based 3D pectoralis muscle segmentation outperforms standard 2D techniques in assessing muscle loss among COPD patients.
Key Details
- 1Researchers compared automated 3D pectoralis muscle volume (PMV) segmentation to traditional 2D muscle area (PMA) analysis using CT scans.
- 2The study included 1,235 participants (634 with COPD, 601 controls) from the CanCOLD cohort.
- 3The U-Net model achieved Dice similarity coefficients of 0.94 (training/validation), 0.93 (internal test), and 0.92 (external test) for PMV segmentation.
- 4Both PMA and PMV were significantly lower in COPD patients (p < 0.05), but PMV correlated more strongly with FEV1, Dlco, and VO2 clinical measures.
- 5Authors recommend PMV as a surrogate marker for muscle mass and comorbid sarcopenia in COPD and call for normative reference studies.
Why It Matters
Automating muscle mass assessment from routine chest CTs may enable early identification of sarcopenia in COPD patients, facilitating targeted interventions and improved prognostication. This highlights an expanding role for AI in quantitative radiology beyond traditional pulmonary imaging.
Source
AuntMinnie
Related News
•AuntMinnie
Study: Computer Vision Models Best LLMs in Chest CT Breast Abnormality Detection
Computer vision models (CVMs) surpass large language models (LLMs) in accurately labeling incidental breast abnormalities on chest CT scans.
•Cardiovascular Business
Radiology Maintains Lead in FDA-Cleared AI Algorithms, Cardiology Follows
Radiology remains the top specialty for FDA-cleared AI, with cardiology as a strong second, particularly in cardiovascular imaging.
•AuntMinnie
Deep Learning Models Rival Radiologists for Pancreatic Cancer Detection on CT
Deep-learning models achieved comparable or superior accuracy to experienced radiologists in detecting pancreatic cancer on CT scans, especially for small tumors.