Table 1:

Multi-class tissue classification performance of baseline, SOTA, pathology foundation, and custom models trained on HMU, NCT, and STARC-9 for seven common tissue types (ADI, LYM, MUS, MUC, NCS, TUM, NOR) evaluated on the STANFORD-CRC-HE-VAL-LARGE dataset. The highest accuracy models for each dataset are highlighted in bold.

Model	Precision			Recall			F1-macro			Accuracy			No. of params.
	NCT	HMU	STARC-9	NCT	HMU	STARC-9	NCT	HMU	STARC-9	NCT	HMU	STARC-9
Baseline models
ResNet50 [10]	84.08	87.81	98.92	62.59	85.71	98.64	63.17	86.00	98.78	62.59	85.71	98.64	24 M
EfficientNet-B7 [38]	89.99	88.65	99.11	82.47	87.45	98.80	84.55	87.87	98.95	82.47	84.45	98.80	64 M
ViT-base [8]	92.71	91.57	98.49	84.25	90.29	98.09	87.30	90.87	98.28	84.25	90.29	98.09	86 M
SOTA models
DeiT-B [41]	94.28	90.97	98.99	81.63	90.05	98.65	85.35	90.40	98.81	81.63	90.05	98.65	86 M
Swin Trans-base [21]	90.11	93.17	99.09	79.05	91.88	98.80	82.52	92.46	98.94	79.05	91.88	98.79	87 M
KimiaNet [32]	87.25	88.60	99.03	71.53	86.67	98.72	71.53	87.04	98.87	68.69	86.67	98.72	7M
ConvNeXT-base [22]	91.95	92.09	99.01	82.82	91.07	98.36	85.56	91.50	98.68	82.82	91.07	98.36	88 M
Pathology foundation models
CTransPath [44]	90.11	93.17	99.34	79.05	91.88	99.00	82.52	92.46	99.16	79.05	91.88	99.00	87 M
HiPT [6]	90.92	93.21	98.64	74.51	91.99	98.32	77.41	92.54	98.47	74.51	91.99	98.32	86 M
ProvGigPath [45]	89.43	91.47	98.74	74.18	90.60	98.37	78.40	90.92	98.55	74.18	90.60	98.37	305 M
PathDino [1]	92.93	91.19	98.67	77.35	89.64	98.37	81.71	90.22	98.51	77.35	89.64	98.37	22 M
CONCH [25]	91.53	91.41	98.56	75.69	90.02	98.19	78.08	90.52	98.37	75.69	90.02	98.19	87 M
UNI [7]	94.55	93.03	98.67	80.43	91.80	98.25	84.42	92.36	98.45	80.43	91.80	98.26	88 M
Virchow [42]	92.51	92.35	98.63	79.02	91.23	98.28	82.05	91.69	98.45	79.02	91.23	98.28	305 M
VIM4PATH [28]	91.51	92.66	98.53	75.41	91.50	98.27	79.10	92.01	98.40	75.41	91.50	98.29	86 M
Customized models (trained from scratch)
CNN	83.97	78.45	98.10	64.21	68.10	97.81	68.12	66.39	97.93	64.21	68.10	97.81	3.9 M
Histo-ViT	86.17	76.45	96.88	69.48	67.16	96.32	72.01	67.77	96.52	69.48	67.16	96.32	86 M