A routine CT scan already performed for other reasons could soon reveal hidden inflammation in belly fat that signals early risk of diabetes and heart disease. Doctors currently have no way to detect this dangerous visceral fat inflammation—it does not show up in blood tests or standard obesity measures. This blind spot means patients receive incomplete risk assessments, unnecessary tests, and drugs that may be wasted or misdirected. The researchers have discovered that artificial intelligence can extract new biomarkers from existing CT scans, validated against tissue biopsies, to quantify this inflammation directly. If the technology, called FatHealth, passes clinical testing on 20,000 NHS scans over the next 18 months, it could transform cardiometabolic risk assessment without adding any new scans, costs, or appointments. The AI would plug into current clinical pathways, generating risk data from scans already performed for other purposes. This could help target powerful new drugs to the patients who will benefit most, reduce wasted treatments, and catch diabetes and heart disease risk earlier—all from imaging data that already exists but currently goes unread.
View original technical description
Obesity, diabetes and cardiometabolic disease are global health economic burdens. Depot-specific adipose tissue (AT) inflammation, particularly in visceral AT, is an indicator of cardiometaboilic risk and a potential therapeutic target. However, visceral AT inflammation is not identifiable by any current blood or imaging biomarkers and does not correlate with simple indicators of obesity. This major gap in our diagnostic approach gives incomplete or even misleading information, wastes tests and drug treatments, and fails to take full advantage of powerful new drugs, that cannot be optimally targeted. We discovered new AI imaging biomarkers, derived from routine clinical CT scans, that provide highly quantitative readouts of depot-specific AT inflammation, validated against the pivotal molecular signatures of cardiometabolic risk, from transcriptional analysis of AT biopsies. Caristo Diagnostics has developed a clinically-applicable technology, FatHealth, that quantifies cardiometabolic risk from routine CT scans, acquired in everyday clinical practice. We will work with collaborating Universities and NHS Hospitals to develop FatHealth into a carefully tested commercial product. In the next 18 months we will analyse 20,000 CT scans (from existing NHS imaging resources) to refine the technical aspects of FatHealth across different CT platforms and scan types and validate the ability of FatHealth to stratify cardiometabolic risks using detailed metabolic phenotyping. We will develop an automated, regulatory-approved, web-based portal to deliver FatHealth for use in healthcare systems. We will work with clinical NHS organisations and patient groups to evaluate the clinical effectiveness of FatHealth in people and patients at risk of diabetes. We have formed collaborations with GPs in Primary Care and with the Academic Health Science Networks (AHSNs). FatHealth will fit seamlessly into current clinical pathways to transform cardiometabolic risk assessment, by generating new data from analysis of CT scans that are already widely performed in the NHS and other healthcare systems globally.
Plain English summaries and category classifications on this site are generated by AI and may not perfectly reflect the original research.
Is something wrong? Let us know