Distinguish the Value of the Benign Nevus and Melanomas Using Machine Learning: A Meta-Analysis and Systematic Review

Table 1

Summary of specificity and -index in subgroup analysis.

E	Model	c-statistic		Sensitivity and specificity
		Number	95% CI	Number	SEN (95% CI)	SPE (95% CI)
1	DCNN	33	0.85 (0.81, 0.89)	45	0.84 (0.82, 0.86)	0.85 (0.81, 0.88)
2	RF	3	0.98 (0.97, 0.99)	3	0.89 (0.84, 0.93)	0.94 (0.84, 0.98)
3	SVM	2	0.93 (0.86, 1.00)	4	0.91 (0.80, 0.96)	0.94 (0.89, 0.97)
4	Other DL	6	0.93 (0.88, 0.99)	4	0.94 (0.87, 0.97)	0.71 (0.56, 0.86)
5	Other ML	2	0.90 (0.84, 0.96)	4	0.74 (0.67, 0.80)	0.87 (0.82, 0.91)
Overall		46	0.87 (0.84, 0.90)	61	0.85 (0.82, 0.87)	0.86 (0.82, 0.88)

In the original study, other DL was described as a combination of various artificial neural networks; other ML contained KNN and LR.