A multicenter study led by Sapienza University of Rome reports the clinical value of AI-based fat tissue analysis

July 1, 2026 | Medical Newspaper

MEDICAL IP announced that an international study using its AI-powered body composition platform, DeepCatch C, has been published in La Radiologia Medica. The multicenter study, led by researchers at Sapienza University of Rome, analyzed abdominal CT scans from 293 patients with non-metastatic colorectal cancer.

The study examined not only the amount of visceral and subcutaneous fat, but also its tissue characteristics. DeepCatch C automatically quantified average fat attenuation on CT, measured in Hounsfield units (HU). The researchers reported that visceral fat attenuation differed between patients who died from cancer-related causes and those who survived: the average was −77 HU in the former group and −82 HU in the latter.

The findings suggest that measurements derived from existing CT scans may provide additional information relevant to cancer prognosis, beyond conventional body composition measures. Because the analysis uses scans already acquired for clinical care, it may offer this information without requiring an additional scan.

The study adds to research exploring how AI-based analysis of medical images can uncover clinically relevant information that may not be apparent from routine visual assessment. MEDICAL IP plans to expand its imaging-based analysis and research collaborations across CT, X-ray, and MRI.