Asan Medical Center (AMC) and medical AI company Coreline Soft have developed an artificial intelligence technology that aims to enhance the diagnosis of sarcopenia.

Utilizing CT scans, this technology automatically measures psoas muscle volume, providing results 48 times faster than traditional manual methods.

Sarcopenia, a condition marked by the decline in skeletal muscle mass and strength due to aging, is linked to decreased physical function and other health conditions such as diabetes and cardiovascular diseases. It can also lead to severe issues like falls and fractures.

With an increasing elderly population, sarcopenia has been classified as a disease in Korea since 2021.

Professor Kim Ji-wan’s team at AMC used Coreline’s AI solution, AVIEW, to create a method for accurate and quick psoas muscle volume diagnosis.

The study, titled "Development and validation of a reliable method for automated measurements of psoas muscle volume in CT scans using deep learning-based segmentation: a cross-sectional study," suggests that this approach could enhance early detection and treatment of sarcopenia.

The research used CT scan data from 520 participants to develop an AI model based on nnU-Net for automatic segmentation and volume measurement of the psoas muscle.

This method, implemented in Coreline’s AVIEW solution, proved to be 48 times faster than manual segmentation. The AI’s accuracy, measured by a Dice score of 0.927, indicates high efficiency and precision.

The research team at AMC noted that this study demonstrates the potential of medical AI models using diverse CT scan data.

They emphasized the reliability of the results in large-scale studies, which could enable rapid analysis of psoas muscle volume in a significant number of Korean subjects, which could be clinically useful for establishing normal and abnormal ranges.

The team also stressed that reanalyzing CT datasets initially intended for other purposes could facilitate further research and additional diagnoses of sarcopenia.

"This research could serve as a cornerstone for sarcopenia research and diagnosis,” Professor Kim said. “If applied in clinical practice, it could provide patients with more accurate and rapid diagnoses, significantly improving the quality of medical services."

Doctor Choi Woo-rim, the first author of the paper, remarked, "The value of AI technology is maximized when it achieves significant results in real-world medical applications.”

This study is expected to make important contributions not only to the diagnosis of sarcopenia but also to the diagnosis and research of various muscle-related diseases, Choi added.

The research results were published in the BMJ Open.