Table 2.

Performance comparison of ResNet and ViT, each trained on base (unexpanded images) and Gen2-expanded datasets. Standard deviation ranges are bracketed (−1 SD, + 1 SD).

Model Performance
Model	Dataset	AUC	Pooled AUC	Accuracy	Recall	Specificity
ResNet	Base	0.68 (0.54, 0.81)	0.63	0.61 (0.55, 0.67)	0.48 (0.47, 0.50)	0.97 (0.94, 1.00)
ResNet	Gen 2	0.69 (0.58, 0.80)	0.67	0.66 (0.60, 0.71)	0.64 (0.54, 0.73)	0.67 (0.54, 0.81)
ViT	Base	0.70 (0.64, 0.76)	0.67	0.62 (0.57, 0.67)	0.55 (0.48, 0.62)	0.55 (0.48, 0.62)
ViT	Gen 2	0.79 (0.68, 0.90)	0.79	0.71 (0.61, 0.81)	0.68 (0.58, 0.78)	0.68 (0.58, 0.78)