Tuesday 08 April 2025
A team of researchers has developed an innovative tool that uses artificial intelligence to analyze abdominal CT scans and identify specific tissues, such as muscle and fat, in patients with gastrointestinal cancer. The tool is designed to help doctors quickly and accurately assess the condition of their patients, which could lead to better treatment outcomes.
The new tool combines two key components: a multi-view localization model and a high-precision segmentation model based on a 2D neural network. The first component uses multiple views of the CT scan to detect the abdominal region and extract slices of the tissue. This is done by analyzing the spacing between pixels in the image, allowing the algorithm to pinpoint the location of the abdominal cavity.
The second component is responsible for segmenting the extracted slices into different types of tissues, such as muscle and fat. This is achieved through a process called deep learning, which involves training an artificial neural network on a large dataset of labeled images. The network learns to recognize patterns in the images and identify specific features that distinguish one type of tissue from another.
The researchers tested their tool on a dataset of 1,230 CT scans and found that it was able to accurately segment muscle, subcutaneous fat, and visceral fat with high precision. They also found that the tool was able to localize the abdominal region with high accuracy, even in cases where the patient’s body composition was abnormal.
One of the key advantages of this new tool is its ability to automate a time-consuming and labor-intensive process. Current methods for analyzing abdominal CT scans require manual annotation by specialists, which can be both costly and inefficient. The new tool can process images much faster than humans, and with higher accuracy, making it an attractive option for busy hospitals and clinics.
The potential benefits of this technology are significant. Accurate assessment of abdominal tissue composition could help doctors identify patients who are at risk of poor outcomes, and develop targeted treatment plans to improve their chances of survival. Additionally, the tool could be used to monitor the effectiveness of treatment over time, allowing doctors to adjust their approach as needed.
The researchers plan to continue refining their tool in the coming months, with the goal of making it available for clinical use within the next year or two. If successful, this technology has the potential to revolutionize the way doctors diagnose and treat gastrointestinal cancer, improving patient outcomes and saving lives.
Cite this article: “AI-Powered Tool Revolutionizes Abdominal CT Body Composition Analysis in Gastrointestinal Cancer Patients”, The Science Archive, 2025.
Artificial Intelligence, Abdominal Ct Scans, Muscle, Fat, Gastrointestinal Cancer, Segmentation Model, Deep Learning, Neural Network, Localization Model, Medical Imaging Analysis.
