AI-powered SPOT imaging for enhanced myocardial scar detection and quantification

Comprehensive Summary

Late gadolinium enhanced (LGE) cardiac magnetic resonance imaging, is used to assess and characterize myocardial injuries and patients. These images can also be used to detect the likelihood of a ventricular arrhythmia, allowing for prevention and treatment of cardiac issues to occur earlier. The introduction of black blood imaging allowed for improved contrast between the blood and scar, however it was seen that there is a loss of anatomical information with this method. The aim of this study was to develop an AI model that uses both black blood LGE imaging and bright blood imaging to provide clear anatomical detail along with high quality contrast between scar and blood to help in accurate detections of a myocardial injury. This model is called SPOT. SPOT also quantifies the scar information using segmentation methods in Ai to provide a fully automated analysis of the myocardial scars. This imaging technology was tested and validated using simulations, animal models, and was tested on 450 patients with suspected heart disease. When testing this program on the patients, there was an 11.2% in detection of the LGE segments using the SPOT model compared to normal bright-blood imaging. The SPOT program was successful in enhancing scar detection and localization, improved reproducibility of LV wall, and automated quantification of myocardial scars. Additionally, it is able to details such as scar size allowing for better predictions of cardiac diseases.

Outcomes and Implications

Currently, LGE models are not as strong in the contrast, making it hard to distinguish between the blood and scar tissue. Generally, the process of analyzing this data is time consuming for physicians; but when the imaging is not clear, this process becomes much harder and can have variability depending on who is analyzing the imaging. The use of the SPOT model, developed in this study, can help make this process more efficient and even more accurate. However, this model does not account for quantification of fat presence which is a good indicator of cardiomyopathies. The authors suggest several revisions to SPOT in order to visualize fibrofatty infiltration. More testing would need to be done, in order to ensure its accuracy.