Preliminary analysis of predicting the first recurrence in patients with neovascular age-related macular degeneration using deep learning

Table 2 Classification performance of the convolutional neural network models

	OCT at baseline		OCT at after the loading phase
	Entire region	Fluid region	Entire region	Fluid region
AUC	0.572 ± 0.043	0.600 ± 0.030	0.716 ± 0.027	0.725 ± 0.012
Accuracy	0.528 ± 0.020	0.572 ± 0.021	0.667 ± 0.016	0.679 ± 0.016
Sensitivity	0.579 ± 0.179	0.566 ± 0.113	0.608 ± 0.083	0.621 ± 0.068
Specificity	0.472 ± 0.172	0.578 ± 0.146	0.731 ± 0.094	0.742 ± 0.075
PPV	0.544 ± 0.02	0.599 ± 0.049	0.717 ± 0.052	0.725 ± 0.036
NPV	0.521 ± 0.035	0.557 ± 0.019	0.638 ± 0.022	0.649 ± 0.022
F1 score	0.545 ± 0.088	0.573 ± 0.039	0.651 ± 0.036	0.665 ± 0.029

OCT optical coherence tomography, AUC area under the curve, PPV positive predictive value, NPV negative predictive value

ISSN: 1471-2415