Skip to main content

Table 3 Diagnostic performances of CNN and endoscopist A in the per-image analysis

From: Convolutional neural network-based system for endocytoscopic diagnosis of early gastric cancer

 

CNN

Endoscopist A

P value

Accuracy, % (95% CI)

83.2 (79.8–86.2)

82.3 (78.8–85.4)

0.75

Sensitivity, % (95% CI)

76.4 (71.3–81.0)

86.6 (82.3–90.2)

0.0014

Specificity, % (95% CI)

92.3 (88.2–95.4)

76.6 (70.7–81.9)

 < 0.001

PPV, % (95% CI)

93.0 (89.2–95.8)

83.1 (78.6–87.0)

 < 0.001

NPV, % (95% CI)

74.6 (69.2–79.5)

81.1 (75.3–86.0)

0.089

  1. CNN, convolutional neural network; CI, confidence interval; PPV, positive predictive value; NPV, negative predictive value; AUC, area under the curve
Back to article page

BMC Gastroenterology

ISSN: 1471-230X

Contact us