University of Pittsburgh researchers have developed a novel tool to classify vascular calcification as clinical risk. Designed to be an efficient and comprehensive tool that includes a machine learning-based framework, the tool would allow clinicians to stratify patients with vascular calcifications into risk categories.
Description
Vascular calcifications are strongly associated with major adverse cardiovascular events (MACE) such as stroke or myocardial infarction, the leading cause of death globally. Recent research has uncovered that not all vascular calcifications pose the same risk and classifying these remains an unmet clinical need. This tool would identify patients most at risk of MACE which could lead to targeted treatment for high-risk patients and prevent low-risk patients undergoing unnecessary and potentially harmful treatments, improving patient outcomes.
Applications
1. Cardiovascular diseases including cranial aneurysms 2. Myocardial infarction 3. Stroke
Advantages
Current risk assessment tools, coronary artery calcium (CAC) scores, assume a linear relationship with the extent of calcification. This assumption fails to consider how medication and exercise can increase CAC scores while lowering cardiovascular risk, and incorrectly assumes all calcifications pose equal risk. Recent research established that not all calcifications increase the risk of MACE. Macrocalcifications were found to stabilize plaques in coronary arteries, but those with jagged edges could be detrimental and lead to tissue failure. Relationships between high densities of microcalcifications and lower levels of collagen fibers were also discovered, suggesting microcalcifications lower the load-bearing capacity of affected tissue potentially leading to rupture. Additional studies have reported the environment of the calcification, atherosclerotic calcifications (e.g., embedded in lipid pools) or non-atherosclerotic calcifications also impact on the risk of rupture. This novel risk assessment tool assesses the risk from calcification based on the phenotype and overcomes many failings of traditional CAC scoring.
Invention Readiness
A prototype risk assessment tool has been developed. Training used a small number of high-resolution ?-CT images, and this tool has been applied to thousands of images where it detected and classified microcalcifications based on key features. It is suitable for high throughput use with calcification phenotyping in less than seven hours. This novel algorithm could lead to more accurate and individualized risk assessment of patients at risk of CVD.
IP Status
Patent Pending