Machine Learning Algorithms to Distinguish Myocardial Perfusion SPECT Polar Maps

Frontiers in Cardiovascular Medicine

Christiane Wiefels¹  Benjamin J. W. Chow¹  Erito Marques de Souza Filho²  Lucas Nunes Dalbonio de Carvalho²  Claudio Tinoco Mesquita³  Fernando de Amorim Fernandes⁴  Flávio Luiz Seixas⁵  Evandro Tinoco Mesquita⁶  Ronaldo Altenburg Gismondi⁶  Tadeu Francisco dos Santos⁶  Alair Augusto Sarmet M. D. dos Santos⁶ 

[1] Department of Cardiac Image, University of Ottawa Heart Institute, Ottawa, ON, Canada;Department of Languages and Technologies, Universidade Federal Rural do Rio de Janeiro, Rio de Janeiro, Brazil;Department of Nuclear Medicine, Hospital Pró-Cardíaco, Americas Serviços Medicos, Rio de Janeiro, Brazil;Department of Nuclear Medicine, Hospital Universitário Antônio Pedro/EBSERH, Universidade Federal Fluminense, Rio de Janeiro, Brazil;Institute of Computing, Universidade Federal Fluminense, Rio de Janeiro, Brazil;Post-graduation in Cardiovascular Sciences, Universidade Federal Fluminense, Niterói, Rio de Janeiro, Brazil;

关键词: machine learning;    polar maps;    myocardial perfusion imaging (MPI);    coronary artery disease;    random forest;   

DOI  :  10.3389/fcvm.2021.741667

来源: DOAJ

【 摘 要 】

Myocardial perfusion imaging (MPI) plays an important role in patients with suspected and documented coronary artery disease (CAD). Machine Learning (ML) algorithms have been developed for many medical applications with excellent performance. This study used ML algorithms to discern normal and abnormal gated Single Photon Emission Computed Tomography (SPECT) images. We analyzed one thousand and seven polar maps from a database of patients referred to a university hospital for clinically indicated MPI between January 2016 and December 2018. These studies were reported and evaluated by two different expert readers. The image features were extracted from a specific type of polar map segmentation based on horizontal and vertical slices. A senior expert reading was the comparator (gold standard). We used cross-validation to divide the dataset into training and testing subsets, using data augmentation in the training set, and evaluated 04 ML models. All models had accuracy >90% and area under the receiver operating characteristics curve (AUC) >0.80 except for Adaptive Boosting (AUC = 0.77), while all precision and sensitivity obtained were >96 and 92%, respectively. Random Forest had the best performance (AUC: 0.853; accuracy: 0,938; precision: 0.968; sensitivity: 0.963). ML algorithms performed very well in image classification. These models were capable of distinguishing polar maps remarkably into normal and abnormal.

【 授权许可】

Unknown