Increased cardiac sphericity or roundness, even in normal hearts can be used to predict the risk of cardiomyopathy and related outcomes
US researcher using deep learning-enabled image segmentation of cardiac magnetic resonance imaging data, have found that increases in the left ventricle (LV) sphericity index in healthy hearts, predicts the risk for cardiomyopathy and related outcomes such as atrial fibrillation.
Dilation of cardiac chambers and or a decline in systolic function are often used as key indicators of disease and which can be assessed using conventional imaging modalities to quantify such changes. Moreover, deep neural networks have shown a great potential in image pattern recognition and automated methods achieve a performance on par with human experts in analysing cardiovascular magnetic resonance images and deriving clinically relevant measures. Cardiomyopathies of different aetiologies can often result in a similar end-stage phenotype of a more round, spherical ventricle. In fact, in patients with cardiac diseases, a greater sphericity of the left ventricle, has, for example, been shown to be an independent predictor of 10-year survival following an acute myocardial infarction. In the current study, researchers thought that even among those with normal heart function, there was likely to be variation in cardiac sphericity, in particular, sphericity of the left ventricle and that this may serve as marker of cardiac risk, especially among those with an underlying genetic risk.
Using automated deep-learning segmentation of cardiac magnetic resonance imaging (MRI) data, the researchers estimated and analysed the sphericity index in patients who were part of the UK Biobank database but excluded those with either abnormal left ventricular size or systolic function.
Cardiac sphericity and risk of cardiomyopathy
In a total of 38,897 participants, the researchers calculated that for one standard deviation increase in the sphericity index, or roundness of the heart, there was an associated 47% increased incidence of cardiomyopathy (hazard ratio, HR = 1.47, 95% CI 1.10 – 1.98, p = 0.01). In addition, the same increase in the sphericity index, was associated with a 20% increased incidence of atrial fibrillation (HR = 1.20, 95% CI 1.11 – 1.28, p < 0.001) and which was independent of clinical factors and traditional magnetic resonance imaging (MRI) measurements. In contrast, similar increases in the sphericity index were non-significantly associated with the risk of both heart failure (p = 0.3) and cardiac arrest (p = 0.70).
The team also identified four loci associated with sphericity at genome-wide significance and concluded that the variation in left ventricular sphericity in otherwise normal hearts, predicts the risk for cardiomyopathy and related outcomes and is caused by non-ischaemic cardiomyopathy.
Citation
Vukadinovic M et al. Deep learning-enabled analysis of medical images identifies cardiac sphericity as an early marker of cardiomyopathy and related outcomes. Med 2023
