Table 1.

Composition of datasets used in the development and testing of the deep-learning algorithm.

	Whole Dataset		Training Set		Internal Test Set		External Test Set
	Number of Images	Number of Patients	Number of Images	Number of Patients	Number of Images	Number of Patients	Number of Images	Number of Patients
Overall	2899	846	2590	762	309	85	206	197
Mucosa-confined lesions	1900	580	1693	522	207	58	126	119
Low-grade dysplasia	727	233	630	205	97	28	68	66
High-grade dysplasia	421	131	390	123	31	8	21	21
EGC	752	230	673	205	79	25	37	37
Submucosa-invaded lesion	999	270	897	243	102	27	80	78
EGC	282	81	242	71	40	10	23	23
AGC	717	189	655	172	62	17	57	55

EGC, early gastric cancer; AGC, advanced gastric cancer.