Differentiation of the Follicular Neoplasm on the Gray-Scale US by Image Selection Subsampling along with the Marginal Outline Using Convolutional Neural Network

Table 9

Comparison result of diagnostic performance with other USFNA method [5] for follicular thyroid neoplasm.

(%)	FS (Frozen Section)	USFNA	Our proposed
Sensitivity	80.0 (24/30)	84.2 (48/57)	71.05 (27/38)
Specificity	96.3 (77/80)	52.2 (36/69)	93.19 (178/191)
PPV	88.9 (24/27)	59.3 (48/81)	67.49 (27/40)
NPV	92.8 (77/83)	80.0 (36/45)	89.89 (178/189)
Accuracy	91.8 (101/110)	66.7 (84/126)	89.52 (205/229)