Innovative system enables clinicians to better assess health risks of obese patients by accurately identifying and quantifying body fats
UCF researchers have developed the first automated system to provide accurate, quantitative data regarding a person’s white and brown body fat (adipose tissue) levels. The innovation enables clinicians to better assess possible health issues and risks associated with abdominal obesity, one of the most prevalent health conditions today. Using imaging scans and a novel computerized automatic detection (CAD) system, the invention provides the distribution of a person’s white adipose tissue (WAT) and brown adipose tissue (BAT) at the whole body, body region and organ levels.
More than 70 percent of adults in the United States and about 30 percent worldwide are either overweight or obese, carrying excessive amounts of WAT and little or no BAT. Though BAT is a “healthier” fat and a natural defense against hypothermia and obesity, WAT can increase the risk of cardiovascular diseases, diabetes and certain types of cancer. To assess a patient’s risks, clinicians currently use manual interactions and subjective processes. This includes conventional measuring methods, such as waist circumference, waist to hip ratio and body mass index. The results are usually inefficient and inaccurate. In contrast, the UCF patient-specific adiposity analysis system is fully automated and data-driven. By evaluating detailed radiology scans, the system can accurately quantify a patient’s fat tissue and provide meaningful data regarding health risks. For example, the system can discern and quantify the two types of WAT (subcutaneous and visceral fat) found in abdominal obesity even though the two share similar intensity characteristics and are widely connected.
Technical Details
The UCF system detects, distinguishes, and quantifies WAT and BAT from an imaging scan such as a PET/CT (positron emission tomography / computed tomography) or a PET/MRI (magnetic resonance imaging) scan. Various algorithms and equations support the system’s two main modules. The first module identifies, separates and quantifies the WAT (both subcutaneous and visceral) in the abdominal and thorax regions. Based on deep learning features, the module uses a novel algorithm to differentiate between the two regions. A second module detects and separates BAT by using another algorithm. Structure cues automatically detect the body region and constrain the algorithm to focus only on potential BAT regions (head, neck, and thorax). An optional third module automatically detects specific organs containing adipose tissue. As required, the modules can operate independently or in combination with each other. In some embodiments, the first and second modules operate together, while other embodiments use all three modules. The system also includes a method of creating a risk profile based on the quantitative data.
Benefits
- Faster and more accurate than other methods
- Enables clinicians to quantify white and brown fat during routine patient exams
- Data-driven and easily adjusts for different personalized parameters of each patient
- Can be extended for use as a mobile app that performs fat quantification using the images taken from a camera.
Applications
- Biomedical imaging and radiology
- Hospitals and health centers