Table 2.

Performance for the field-of-view-level classification models with the internal (1st cohort) and external (2nd cohort) testing sets.

Architecture	Precision	Sensitivity	F1-score	AUC
Internal test
ConvNeXtTiny	0.79 ± 0.01	0.80 ± 0.01	0.80 ± 0.01	0.91 ± 0.01
ResNet50	0.82 ± 0.01	0.77 ± 0.01	0.79 ± 0.01	0.90 ± 0.01
DenseNet121	0.83 ± 0.02	0.74 ± 0.03	0.77 ± 0.02	0.89 ± 0.01
UNI2-h	0.75 ± 0.02	0.70 ± 0.02	0.72 ± 0.01	0.84 ± 0.01
External test
ConvNeXtTiny	0.78 ± 0.01	0.69 ± 0.01	0.73 ± 0.01	0.83 ± 0.01
ResNet50	0.80 ± 0.01	0.64 ± 0.01	0.68 ± 0.01	0.79 ± 0.01
DenseNet121	0.81 ± 0.04	0.64 ± 0.02	0.68 ± 0.01	0.82 ± 0.01
UNI2-h	0.81 ± 0.01	0.51 ± 0.02	0.62 ± 0.01	0.75 ± 0.01