WSI_SOTA

The state-of-the-art performance on WSI using deep learning.Here we divide the wsi tasks into three parts:classification,segmentation and detection.

Classification

Dataset	WSI number	Dataset link	Code link	Paper	Year	Method	Accuracy(slide)	F1	AUC(slide)	Precision	Recall
TCGA-NSCLC	1053	https://portal.gdc.cancer.gov/repository	https://github.com/cvlab-stonybrook/local learning wsi	Gigapixel Whole-Slide Images Classification using Locally Supervised Learning (arxiv.org)	2022	Locally Supervised Learning	87.85		93.77
TCGA-RCC	939	https://portal.gdc.cancer.gov/repository	https://github.com/cvlab-stonybrook/local learning wsi	Gigapixel Whole-Slide Images Classification using Locally Supervised Learning (arxiv.org)	2022	Locally Supervised Learning	91.40		97.60
LKS	684	https://github.com/cradleai/LKS-Dataset	https://github.com/cvlab-stonybrook/local learning wsi	Gigapixel Whole-Slide Images Classification using Locally Supervised Learning (arxiv.org)	2022	Locally Supervised Learning	89.76
TCGA-GBM、TCGA-LGG	736	https://portal.gdc.cancer.gov/repository	https://github.com/CityUAIM-Group/MultiModal-learning	Discrepancy and Gradient-Guided Multi-modal Knowledge Distillation for Pathological Glioma Grading	2022	Multi-modal Knowledge Distillation	76.78		92.35
TCGA-LMS	85		https://github.com/machiraju-lab/UA-CNN	Uncertainty Aware Sampling Framework of Weak-Label Learning for Histology Image Classification			0.83 ± 0.09	0.77 ± 0.10		0.75 ± 0.10	0.83 ± 0.0
CAMELYON16	400	https://camelyon17.grand-challenge.org/Data/	https://github.com/miccaiif/DGMIL	DGMIL: Distribution Guided Multiple Instance Learning for Whole Slide Image Classification	2022	MIL	0.8018		0.8368
CAMELYON16	400	https://camelyon17.grand-challenge.org/Data/	https://github.com/TencentAILabHealthcare/ReMix	ReMix：General and Efficient Framework for Multiple Instance Learning Based Whole Slide Image Classification	2022	MIL	0.9543			0.9639	0.9410
CAMELYON16	401	https://camelyon17.grand-challenge.org/Data/	https://github.com/PhilipChicco/FRMIL	Feature Re-calibration Based Multiple Instance Learning for Whole Slide Image Classification	2022	MIL	0.8910		0.8950
TCGA Lung	1054	https://portal.gdc.cancer.gov/repository	https://github.com/miccaiif/DGMIL	DGMIL: Distribution Guided Multiple Instance Learning for Whole Slide Image Classification	2022	MIL	0.9200		0.9702
UniToPatho	292	https://ieee-dataport.org/open-access/unitopatho	https://github.com/TencentAILabHealthcare/ReMix	ReMix：General and Efficient Framework for Multiple Instance Learning Based Whole Slide Image Classification	2022	MIL	0.8068			0.7820	0.8094

Segmentation

Dataset	WSI number	Dataset link	Code link	Paper	Year	Method	Accuracy	F1	mIoU	Precision	Recall	Dice	AJI	Hausdorff	PQ
GlaS	16	https://warwick.ac.uk/fac/cross_fac/tia/data/glascontest/	https://github.com/xmed-lab/OEEM	Online Easy Example Mining for Weakly-supervised Gland Segmentation from Histology Images (arxiv.org)	2022	CAM pseudo mask		87.35	77.56			87.36
Kumar (nucleus)	30	https://monuseg.grand-challenge.org/Data/	https://github.com/hust-linyi/insmix	InsMix: Towards Realistic Generative Data Augmentation for Nuclei Instance Segmentation	2022	Data Augmentation						82.65	63.31
DeepLIIF	1667	https://deepliif.org/	https://github.com/loic-lb/Unsupervised-Nuclei-Segmentation-using-Spatial-Organization-Priors	Unsupervised Nuclei Segmentation Using Spatial Organization Priors	2022	Unsupervised GAN	79.65(Semantic)			87.89(Semantic)	66.30(Semantic)	"69.81(Semantic)
63.47(Object)"	41.91(Object)	14.79(Object)
Warwick HER2	84	https://warwick.ac.uk/fac/cross_fac/tia/data/her2contest/	https://github.com/loic-lb/Unsupervised-Nuclei-Segmentation-using-Spatial-Organization-Priors	Unsupervised Nuclei Segmentation Using Spatial Organization Priors	2022	Unsupervised GAN	81.64(Semantic)			75.71(Semantic)	64.34(Semantic)	"58.46(Semantic)
58.65(Object)"	39.07(Object)	6.12(Object)
TNBC->KIRC/STAD	50(TNBC)/486(KIRC)/99(STAD)	"https://zenodo.org/record/1175282#.Y2Ux2_dByiN
https://portal.gdc.cancer.gov/projects/TCGA-KIRC
https://www.cancerimagingarchive.net/"	https://github.com/YashSharma/MaNi	MaNi: Maximizing Mutual Information for Nuclei Cross-Domain Unsupervised Segmentation	2022	Unsupervised Cross-Domain						"0.733(KIRC)
0.776(TCIA)"
STAD->KIRC/TNBC	50(TNBC)/486(KIRC)/99(STAD)	"https://zenodo.org/record/1175282#.Y2Ux2_dByiN
https://portal.gdc.cancer.gov/projects/TCGA-KIRC
https://www.cancerimagingarchive.net/"	https://github.com/YashSharma/MaNi	MaNi: Maximizing Mutual Information for Nuclei Cross-Domain Unsupervised Segmentation	2022	Unsupervised Cross-Domain						"0.727(KIRC)
0.821(TCIA)"
CoNSep->PanNuke	41(CoNSep)/481(PanNuke)	"https://warwick.ac.uk/fac/cross_fac/tia/data/hovernet/
https://jgamper.github.io/PanNukeDataset/"	https://github.com/YashSharma/MaNi	MaNi: Maximizing Mutual Information for Nuclei Cross-Domain Unsupervised Segmentation	2022	Unsupervised Cross-Domain						0.74	0.534		0.477
CAMELYON16	400		https://github.com/miccaiif/DGMIL	DGMIL: Distribution Guided Multiple Instance Learning for Whole Slide Image Classification	2022	MIL
TCGA Lung	1054	https://portal.gdc.cancer.gov/repository	https://github.com/miccaiif/DGMIL	DGMIL: Distribution Guided Multiple Instance Learning for Whole Slide Image Classification	2022	MIL

Detection

Dataset	WSI number	Dataset link	Code link	Paper	Year	Method	AUC
TCGA-CRC、TCGA-STAD	896+2675	https://portal.gdc.cancer.gov/projects/TCGA-STAD	/	Joint Region-Attention and Multi-scale Transformer for Microsatellite Instability Detection from Whole Slide Images in Gastrointestinal Cancer	2022	Multi-scale transformer	0.921

Contributors

Yuan Zhang，Hanhuan Cui，Jingjing Pei