Integrated machine learning method for coronary heart disease screening based on two-dimensional spot tracking technology

Abstract

The invention relates to an integrated machine learning method for coronary heart disease screening based on a two-dimensional spot tracking technology. The method comprises the following steps: step 1, selecting a plurality of suspicious coronary heart disease objects; step 2, performing ultrasonic examination on all the objects, and retaining corresponding section images one by one; step 3, performing offline analysis on all the section images through EchoPAC; step 4, obtaining feature quantification and clinical feature classification from a two-dimensional spot tracking cardiogram; step 5, applying t inspection to compare the difference of quantitative features of coronary heart disease positive in a case group and coronary heart disease negative in a control group; applying principal component analysis PCA to 17 fragments of peak shrinkage strain PSS, strain rate SSR and peak time TP; and step 6, reducing result error amplification caused by classification errors through a two-step stacking method. According to the method, the integrated machine learning method is adopted, 19 popular classification methods are integrated, and a model is established to provide an optimal diagnosis prediction effect.

Description

technical field [0001] The invention relates to the technical field of medical imaging analysis and machine learning under computer graphics, in particular to an integrated machine learning method for coronary heart disease screening based on two-dimensional speckle tracking technology. Background technique [0002] Coronary heart disease is a disease sweeping the world, with tens of millions of deaths every year. Early screening and control of coronary heart disease can significantly reduce patient mortality and improve prognosis. There are already many clinical examination methods for early screening of coronary heart disease, but there are still patients with coronary heart disease who are missed. In addition to electrocardiogram, there is still a lack of simple and efficient auxiliary screening methods. Coronary angiography or coronary CTA is the gold standard for diagnosing coronary heart disease, and these methods can visually display coronary stenosis. However, this...

AI Technical Summary

Problems solved by technology

When the modeling data is highly heterogeneous and noisy, the results of different methods vary

Because underlying assumptions are often difficult to test, it can be difficult to judge which approach is most appropriate

Furthermore, there is no single classification method that provides satisfactory results

Images