Research Progress of Machine Learning and Deep Learning in Intelligent Diagnosis of the Coronary Atherosclerotic Heart Disease

Table 4

A summary of AI applications in CHD.

Fields	Paper	Algorithm	Measure	Value	Calculate time or cost
Intelligent diagnosis model	Kathleen et al. [19]	Adaptive boosting algorithm	Accuracy	96.72%	—
	Hassannataj et al. [20]	RF	Accuracy	90.50%	—
	Beunza et al. [21]	CNN	AUC	0.71	More than 10 minutes
	Tan et al. [23]	LSTM	Accuracy	99.85%	Approximately 51 s to run a single epoch
CCTA	Cao et al. [24]	DT	Average quality score		Within 2 minutes
	Kang et al. [25]	SVM	AUC	0.94	Less than 1 second
	Muhammad et al. [26]	SVM	DSC	83.2%	—
	Zreik et al. [27]	CNN	Accuracy	77%	—
	Kumamaru et al. [28]	DL	AUC	0.78	A few seconds
	Zreik et al. [30]	CNN SVM	AUC		—
	Hamersvelt et al. [31]	CNN	AUC	0.76	—
CAG	Cho et al. [34]	XG boost	AUC	0.87	—
	Yang et al. [35]	CNN	F1	0.917	36236 seconds of training time
IVUS	Lucas et al. [36]	SVM RF	Jaccard measure		—
	Wang et al. [37]	RF	Accuracy	91.47%	—
	Jun et al. [38]	CNN	AUC	0.911	3,584 CUDA cores and 12GB of GPU memory
	Yang et al. [39]	DPU-net	Jaccard measure	0.869	Run in 0.03 seconds
	Lee et al. [40]	CNN	AUC	0.84-0.87	—
IVOCT	Kolluru et al. [41]	DT	Accuracy		Under 4 seconds when run on a standard 12-core CPU
	Lee et al. [43]	CNN	Sensitivity/specificity	85.1%/94.2%	0.27 seconds of each image
	Xu et al. [44]	CNN	Accuracy	76.39%	—
MRI	Benedikt et al. [45]	Decision forest	Accuracy	91.8%	—
	Baessler et al. [46]	DL	AUC	0.92	—
Functional diagnosis of CHD	Coenen et al. [55]	ML	—	—	—
	Doeberitz et al. [56]	ML	—	—	—
	Kishi et al. [59]	DL	—	—	minutes of average analysis time
	Doeperitz et al. [60]	DL	Accuracy	92%	—
	Yu et al. [66]	DL	Accuracy	90.5%	Median analysis time is 102 seconds