The DOI matches its Crossref record; 186 of the 186 references it cites were found in a registry or an index.
Towards a general-purpose foundation model for computational pathology
Open accessAbstract
Quantitative evaluation of tissue images is crucial for computational pathology (CPath) tasks, requiring the objective characterization of histopathological entities from whole-slide images (WSIs). The high resolution of WSIs and the variability of morphological features present significant challenges, complicating the large-scale annotation of data for high-performance applications. To address this challenge, current efforts have proposed the use of pretrained image encoders through transfer learning from natural image datasets or self-supervised learning on publicly available histopathology datasets, but have not been extensively developed and evaluated across diverse tissue types at scale. We introduce UNI, a general-purpose self-supervised model for pathology, pretrained using more than 100 million images from over 100,000 diagnostic H&E-stained WSIs (>77 TB of data) across 20 major tissue types. The model was evaluated on 34 representative CPath tasks of varying diagnostic difficulty. In addition to outperforming previous state-of-the-art models, we demonstrate new modeling capabilities in CPath such as resolution-agnostic tissue classification, slide classification using few-shot class prototypes, and disease subtyping generalization in classifying up to 108 cancer types in the OncoTree classification system. UNI advances unsupervised representation learning at scale in CPath in terms of both pretraining data and downstream evaluation, enabling data-efficient artificial intelligence models that can generalize and transfer to a wide range of diagnostically challenging tasks and clinical workflows in anatomic pathology. Pretrained using over 100,000 diagnostic histopathological slides across 20 major tissue types, a self-supervised model is shown to outperform existing baselines across various clinically relevant computational pathology tasks.
Overview
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“By increasing the patch size from the widely used 256 × 256 pixels, we effectively reduce the sequence length by four without impacting the representation quality due to higher resolution patch input, leveraging the robustness of the patch-level foundation models in generalizing to higher resolutions9,10,87.”
From A multimodal whole-slide foundation model for pathology (2025) · cites this paper as [9]
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“we present Virchow, the largest foundation model for computational pathology to date.”
From A foundation model for clinical-grade computational pathology and rare cancers detection (2024) · cites this paper as [41]
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“To overcome these limitations, several research groups have alternatively proposed the use of weakly-supervised deep learning models, which can be trained to perform various downstream computational pathology tasks such as tumor subtyping and prognosis32–37 without any region or pixel-level annotations.”
From Predicting the tumor microenvironment composition and immunotherapy response in non-small cell lung cancer from digital histopathology images (2024) · cites this paper as [32]
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“UNI [191], Prov-GigaPath [192], and Virchow [193] were pretrained solely on pathology image data to perform cancer classification, diagnosis, and prognosis.”
From Challenges in AI-driven Biomedical Multimodal Data Fusion and Analysis (2025) · cites this paper as [191]
Experimental — accuracy not yet measuredThe cited source says what the claim attributes to it.
In the citing paper
“The vision module is based on the patch encoder CTransPath17 for the small and medium models, and UNI18 for the large model.”
From Generating dermatopathology reports from gigapixel whole slide images with HistoGPT (2025) · cites this paper as [18]
The cited source says what the claim attributes to it.
In the citing paper
“We found out that the UNI foundation model32, when paired with multi-task AB-MIL, achieved the best predictive performance for the various TME signature prediction tasks.”
From Predicting the tumor microenvironment composition and immunotherapy response in non-small cell lung cancer from digital histopathology images (2024) · cites this paper as [32]
In the citing paper
“In fact GMA is widely used to assess the performance of pathology foundation models23,25,26.”
From A clinical benchmark of public self-supervised pathology foundation models (2025) · cites this paper as [23]
In the citing paper
“Self-supervised learning has become a widely used approach for pretraining foundation models on thousands to tens of thousands of images without annotations or labels, marking an important paradigm shift in the field.20,21”
From Use of artificial intelligence for liver diseases: A survey from the EASL congress 2024 (2024) · cites this paper as [20]
In the citing paper
“An increasing number of published studies show the potential of foundation models20,24; however, they are currently underrepresented in the field of liver diseases.4”
From Use of artificial intelligence for liver diseases: A survey from the EASL congress 2024 (2024) · cites this paper as [20]
In the citing paper
“These works demonstrate that image features produced with self-supervised learning of pathology images outperform image features trained on natural images and that performance improves with scale.”
From A foundation model for clinical-grade computational pathology and rare cancers detection (2024) · cites this paper as [41]
In the citing paper
“Despite the challenges in collecting large-scale datasets in the pathology domain, recent pioneering works have utilized datasets ranging from 30,000 to 400,000 WSIs to train foundation models ranging in size from 28 million to 307 million parameters35–42”
From A foundation model for clinical-grade computational pathology and rare cancers detection (2024) · cites this paper as [41]
In the citing paper
“Built on rich knowledge representations, these models have shown impressive performance across medical fields such as ophthalmology20, radiology21 and pathology22–25.”
From A Multimodal Vision Foundation Model for Clinical Dermatology (2025) · cites this paper as [23]
In the citing paper
“Specialty-specific foundation models20,21,23 typically focus on single imaging modalities”
From A Multimodal Vision Foundation Model for Clinical Dermatology (2025) · cites this paper as [23]
In the citing paper
“The scaling behavior observed in PanDerm’s performance aligns with recent foundation model trends20,22,23,59”
From A Multimodal Vision Foundation Model for Clinical Dermatology (2025) · cites this paper as [23]
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
Deep Residual Learning for Image Recognition
He, K., Zhang, X., Ren, S., & Sun, J. (2016) · 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2016.90 · Cited by 235,474
Scikit-learn: Machine Learning in Python
Pedregosa, F., Varoquaux, G., Gramfort, A., Michel, V., Thirion, B., Grisel, O., Blondel, M., Müller, A., Nothman, J., Louppe, G., et al. (2012) · arXiv · doi:10.48550/arxiv.1201.0490 · Cited by 71,016
ImageNet: A large-scale hierarchical image database
Deng, J., Dong, W., Socher, R., Li, L.-J., Li, K., & Fei-Fei, L. (2009) · 2009 IEEE Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2009.5206848 · Cited by 65,333
ImageNet Large Scale Visual Recognition Challenge
Russakovsky, O., Deng, J., Su, H., Krause, J., Satheesh, S., Ma, S., Huang, Z., Karpathy, A., Khosla, A., Bernstein, M., et al. (2015) · International Journal of Computer Vision · doi:10.1007/s11263-015-0816-y · Cited by 39,380
Swin Transformer: Hierarchical Vision Transformer using Shifted Windows
Liu, Z., Lin, Y., Cao, Y., Hu, H., Wei, Y., Zhang, Z., Lin, S., & Guo, B. (2021) · 2021 IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv48922.2021.00986 · Cited by 35,534
PyTorch: An Imperative Style, High-Performance Deep Learning Library
Paszke, A., Gross, S., Massa, F., Lerer, A., Bradbury, J., Chanan, G., Killeen, T., Lin, Z., Gimelshein, N., Antiga, L., et al. (2019) · arXiv · doi:10.48550/arxiv.1912.01703 · Cited by 31,473
Language Models are Few-Shot Learners
Brown, T. B., Mann, B., Ryder, N., Subbiah, M., Kaplan, J., Dhariwal, P., Neelakantan, A., Shyam, P., Sastry, G., Askell, A., et al. (2020) · arXiv · doi:10.48550/arxiv.2005.14165 · Cited by 23,372
Advances in neural information processing systems 7
Cios, K. J., & Shields, M. E. (1997) · Neurocomputing · doi:10.1016/s0925-2312(97)90018-7 · Cited by 22,093
Least squares quantization in PCM
Lloyd, S. (1982) · IEEE Transactions on Information Theory · doi:10.1109/tit.1982.1056489 · Cited by 15,836
Goldberger, A. L., Amaral, L. A. N., Glass, L., Hausdorff, J. M., Ivanov, P. C., Mark, R. G., Mietus, J. E., Moody, G. B., Peng, C.-K., & Stanley, H. E. (2000) · Circulation · doi:10.1161/01.cir.101.23.e215 · Cited by 14,809
Research works printed but not linked: 13
- Chen, T., Kornblith, S., Norouzi, M. & Hinton, G. A simple framework for contrastive learning of visual representations. In International Conference on Machine Learning, 1597–1607 (PMLR, 2020).
- Chen, R. J. & Krishnan, R. G. Self-supervised vision transformers learn visual concepts in histopathology. In Learning Meaningful Representations of Life, NeurIPS 2021 (2022).
- Dosovitskiy, A. et al. An image is worth 16×16 words: transformers for image recognition at scale. In International Conference on Learning Representations (2021).
- Xiang, J. & Zhang, J. Exploring low-rank property in multiple instance learning for whole slide image classification. In The Eleventh International Conference on Learning Representations (2022).
- Wang, X. et al. SCL-WC: cross-slide contrastive learning for weakly-supervised whole-slide image classification. Advances in Neural Information Processing Systems 35, 18009–18021 (2022).
- Zhou, J. et al. iBOT: image BERT pre-training with online tokenizer. In International Conference on Learning Representations (2022).
- Devlin, J., Chang, M.-W., Lee, K. & Toutanova, K. BERT: pre-training of deep bidirectional transformers for language understanding. In Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and Short Papers) (2018).
- Sablayrolles, A., Douze, M., Schmid, C. & Jégou, H. Spreading vectors for similarity search. In International Conference on Learning Representations (2019).
- Touvron, H., Vedaldi, A., Douze, M. & Jegou, H. Fixing the train–test resolution discrepancy. In Advances in Neural Information Processing Systems, Vol. 32 (eds Wallach, H. et al.) (Curran Associates, 2019).
- Dao, T., Fu, D. Y., Ermon, S., Rudra, A. & Ré, C. FlashAttention: fast and memory-efficient exact attention with IO-awareness. In Advances in Neural Information Processing Systems (2022).
- Loshchilov, I. & Hutter, F. Decoupled weight decay regularization. In International Conference on Learning Representations (2019).
- Sarıyıldız, M. B., Kalantidis, Y., Alahari, K. & Larlus, D. No reason for no supervision: improved generalization in supervised models. In The Eleventh International Conference on Learning Representations (2023).
- Yang, J., Chen, H., Yan, J., Chen, X. & Yao, J. Towards better understanding and better generalization of low-shot classification in histology images with contrastive learning. In International Conference on Learning Representations (2021).
Cited by
Sorted by citations
Items 701–800 of 1,003
A Multimodal Fusion Framework Employing Full-Stage Attention for Survival Prediction
Liu, A., Guo, J., Wang, Y., Liu, J., Tian, H., Qin, T., Xu, H., & Wei, H. (2025) · Lecture notes in computer science · doi:10.1007/978-981-95-3052-6_15 · Cited by 1
Patherea: Cell detection and classification for the 2020s
Štepec, D., Jerše, M., Đokić, S., Jeruc, J., Zidar, N., & Skočaj, D. (2025) · Medical Image Analysis · doi:10.1016/j.media.2025.103868 · Cited by 1
Yang, D., Tang, C., Liu, Z., Guo, J., Tan, T., Wu, Y., Lu, X.-J., Yang, F., Luo, H., Liu, Z., et al. (2025) · Information Fusion · doi:10.1016/j.inffus.2025.103963 · Cited by 1
Collaborative optimization for whole slide image classification
Zheng, T., Yao, H., Jiang, K., & Xiao, Y. (2025) · Knowledge-Based Systems · doi:10.1016/j.knosys.2025.114936 · Cited by 1
Rana, P., Gide, T. N., Adegoke, N. A., Mao, Y., Berkovsky, S., Coiera, E. W., Wilmott, J. S., & Liu, S. (2025) · IEEE Journal of Biomedical and Health Informatics · doi:10.1109/jbhi.2025.3623576 · Cited by 1
“Accordingly,inadditiontoas-sessing the impact of the GCT by comparing COMIL with and without it using ResNet-101 features, we evaluated seven other encoders, ResNet-101/50/18 [28] and DenseNet-201/161/121 [29], UNI [30], and Virchow [31], on internal and external test sets.”
PathoEye: A deep learning framework for the whole-slide image analysis of skin tissue
Lin, Y., Lin, F., Zhang, Y., Wen, J., Li, G., Zeng, X., Sun, H., Jiang, H., Lin, J., Yan, T., et al. (2025) · Computational and Structural Biotechnology Journal · doi:10.1016/j.csbj.2025.11.052 · Cited by 1
“Most existing methods require substantial, manually curated WSIs, which limits their applications in clinical diagnosis [22, 23,27].”
Guan, M., Sun, Y., Belete, M. K., Muthuswamy, A., Farma, M., Kaufmann, J., Lechpammer, M., Sridhar, S., Higgs, B. W., & Si, H. (2026) · Computational and Structural Biotechnology Journal · doi:10.34133/csbj.0086 · Cited by 1
“CTransPath (Swin Transformer) [ 25 ], Histossl (iBOT) [ 26 ], Virchow (DINOv2) [ 27 ], PLIP (Contrastive Learning) [ 28 ], and RETCCL (Clustering-guided Contrastive Learning) [ 29 ] specialize in tasks such as biomarker classification and image retrieval, demonstrating marked advancements in…”
Zhang, H., Wang, T., Feng, J., Wu, J., He, J., Liu, Y., Song, Q., & Xu, B. (2025) · npj Precision Oncology · doi:10.1038/s41698-025-01199-8 · Cited by 2
Yu, B., Liu, J., & Wang, H. (2025) · Biometrics · doi:10.1093/biomtc/ujaf149 · Cited by 1
“Among those models, the UNI model of Chen et al. (2024) demonstrated the state-of-the-art (SOTA) prediction accuracy on a number of important benchmark datasets.”
ECP-KD: Efficient Computational Pathology Heterogeneous Model Fusion Using Knowledge Distillation
Ni, T., Zhu, H., Han, Y., Lai, C., Chen, J., & Pan, Y. (2025) · Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC) · doi:10.1109/embc58623.2025.11254751 · Cited by 1
“Recent development in pre-trained computational pathology models, such as CLAM [1], UNI [2], CONCH [3], and CONCH1.”
InfluentialResearch progress of universalvision foundation models in computational pathology
Fu, Z., Da, Q., Wang, W., Wang, C., Zhong, D., Wang, S., & Wang, L. (2025) · Chinese Science Bulletin (Chinese Version) · doi:10.1360/csb-2025-5045 · Cited by 1
Saeidi, N., Torbati, N., Woitek, R., & Mahbod, A. (2025) · Bioengineering · doi:10.3390/bioengineering12121332 · Cited by 1
“For pre-trained models, we employ the same CNNs and ViT-Base as baseline models, along with three pathology foundation models: CONCH [39], UNI [38], and UNI-2 [38,40].”
InfluentialSubtype classification of gastric spindle cell tumors in whole slide images
Lee, H., Lee, Y. J., Kim, E., Lee, J., Ahn, S., & Lee, S. H. (2025) · Computers in Biology and Medicine · doi:10.1016/j.compbiomed.2025.111410 · Cited by 1
Ji, X., Zelić, R., Aspegren, O. P., Mulliqi, N., Fiorentino, M., Giunchi, F., Molinaro, L., Boman, S. E., Szolnoky, K., Liu, L. X., et al. (2025) · BMJ Open · doi:10.1136/bmjopen-2025-111361 · Cited by 2
Silver, F. H., Deshmukh, T., & Nadiminti, H. (2025) · Journal of Dermatology Research · doi:10.46889/jdr.2025.6317 · Cited by 1
Chinese expert consensus on precision diagnosis and treatment of refractory tumors
Wang, L., Zhang, Y., Shang, Y., Lu, Y., Hu, H., Wang, G., & Wang, H. (2026) · Critical Reviews in Oncology/Hematology · doi:10.1016/j.critrevonc.2025.105109 · Cited by 1
Silver, F. H., Deshmukh, T., & Kollipara, G. (2026) · Onco · doi:10.3390/onco6010003 · Cited by 1
“It is pretrained using more than 100 million images across 20 major tissue types used to facilitate AI model training [15].”
Liu, Y., Hu, B., Hao, X., Huang, J., & Huang, J. (2026) · Biomedical Signal Processing and Control · doi:10.1016/j.bspc.2026.109527 · Cited by 1
Cazzaniga, G., Carbone, M., Barretta, R., Casati, G., Vatrano, S., Gambaro, G., Vischini, G., Capelli, I., Mignani, R., Pasquinelli, G., et al. (2026) · Scientific Reports · doi:10.1038/s41598-026-35466-w · Cited by 1
Kaczmarzyk, J., Alsten, S. C. V., Cozzo, A. J., Gupta, R., Koo, P. K., Troester, M. A., Hoadley, K. A., & Saltz, J. H. (2026) · npj Digital Medicine · doi:10.1038/s41746-025-02334-2 · Cited by 2
Knowledge-guided multi-modality transformer for multi-label genetic mutation prediction
Huang, G., Wu, C., Li, M., Chang, X., Sun, Y., Xing, L., Liang, X., Lin, L., Yang, G., & Zhao, S. (2026) · Pattern Recognition · doi:10.1016/j.patcog.2026.113047 · Cited by 1
Yao, J., Li, S., Liang, P., Xu, X., Elder, D. E., & Huang, Z. (2026) · npj Digital Medicine · doi:10.1038/s41746-026-02357-3 · Cited by 4
“However, it can also be extended to image-only models such as H-Optimus-1 46 , UNI-2 31 , or Virchow-2 47 .”
Large-scale spatial profiling of the tumor microenvironment
Kilgallon, A., Tubelleza, R., & Kulasinghe, A. (2026) · Cell · doi:10.1016/j.cell.2025.12.028 · Cited by 1
Ding, M., Chen, W., Luo, X., Zhong, H., & Shen, L. (2026) · Pattern Recognition · doi:10.1016/j.patcog.2026.113126 · Cited by 1
Feng, Y., Yi, J., Xia, M., Zhang, X., Cai, W., & Zhang, X. (2026) · IEEE Transactions on Medical Imaging · doi:10.1109/tmi.2026.3657524 · Cited by 1
“Notably, pathological whole slide images (WSIs) di ff er from ordinary images in that they reach gigapixel scale and show high morphological heterogeneity across regions [21], [22].”
InfluentialBedőházi, Z., Biricz, A., Kilim, O., Foster, N., Gregus, B., Tõkés, A., Pollner, P., Csabai, I., & Knudsen, B. S. (2026) · Journal of Pathology Informatics · doi:10.1016/j.jpi.2026.100644 · Cited by 2
“Tiles from the BRACS dataset were used toevaluateUNI,UNI-FT,and ResNet-50-FT underidenticalconditions using a cross-validation approach.”
InfluentialBarua, B., Chyrmang, G., Bora, K., Kakoti, L., Ahmed, G. N., Neog, D. R., & Saikia, M. J. (2026) · Journal Of Big Data · doi:10.1186/s40537-025-01358-5 · Cited by 1
“Our proposed Keratin-Net achieved AUC: 0.99 and F1: 94.48 ± 0.01% , aligning with foundation models such as UNI [19] (94.28 ± 0.14”
InfluentialMoMIL: Multi-order enhanced multiple instance learning for computational pathology
Zhang, Y., Zhang, X., Wang, J., Liang, B., Yang, Y., & Tong, C. (2026) · Image and Vision Computing · doi:10.1016/j.imavis.2026.105918 · Cited by 1
Self-citationZhen, S., Zhang, X., Li, Q., Liu, Z., Wang, D., Xi, S., Shi, S., Tian, G., Yang, J., Zhao, H., et al. (2026) · Discover Oncology · doi:10.1007/s12672-026-04498-z · Cited by 1
“These patches were pre-trained using the Universal Self-supervised model (UNI model) [36], which was trained on large-scale pathological data.”
Beyond H&E: Unlocking Pathological Insights with Polarization Imaging
Du, Y., Zhuang, J., Zheng, X., Cong, J., Guo, L., He, C., Luo, L., & Li, X. (2025) · IEEE International Conference on Bioinformatics and Biomedicine (BIBM) · doi:10.1109/bibm66473.2025.11356760 · Cited by 1
“Chaoyang Dataset [39] [39] 83.40 76.54 94.51 SSAT [5] 82.52 - - ViT-DAE [32] 80.12 78.60 - INCV [2] 80.34 74.11 92.63 2HSL [9] 82.90 - - OUSM [33] 80.53 73.70 93.69 SELF [22] 80.49 75.31 93.99 DiRL [15] 83.60 - 95.01 SH-PEFT [19] 84.80 80.61 - UNI [3] 85.51 trained on top of the frozen encoder.”
InfluentialZero-Shot Grounding of Histological Images Using Single-Cell Transcriptomic Profiles
Wu, C., Zuo, M., & Xie, L. (2025) · IEEE International Conference on Bioinformatics and Biomedicine (BIBM) · doi:10.1109/bibm66473.2025.11356619 · Cited by 1
“For the image encoder, we use the state-of-the-art histology image encoder UNI [25] to generate image features.”
ECG-Expert-QA: A Benchmark for Evaluating Medical Large Language Models in ECG
Wang, X., Kang, J., Han, P., Liu, R., Gong, L., & Fan, F. (2025) · IEEE International Conference on Bioinformatics and Biomedicine (BIBM) · doi:10.1109/bibm66473.2025.11356744 · Cited by 2
GeneQuery: Generalized Gene Expression Prediction from Histology Images via Image-Gene QA
Xiong, Y., Liu, L., Cui, Y., Wu, S., Liu, X., Chan, A. B., & Xue, C. J. (2025) · IEEE International Conference on Bioinformatics and Biomedicine (BIBM) · doi:10.1109/bibm66473.2025.11356542 · Cited by 1
Tang, E., Xia, H., Yuan, Z., Zhao, Y., Wang, S., Ye, Z., Gao, S., Zhou, Z., Zhao, Y., Zeng, J., et al. (2026) · Biomedicines · doi:10.3390/biomedicines14020348 · Cited by 1
“Employing the CLAM architecture with attention-based learning, our model identified diagnostically critical subregions in WSIs and refined feature representation through instance-level clustering [30].”
Skrede, O.-J., Pradhan, M., Isaksen, M. X., Hveem, T. S., Vlatkovic, L., Nesbakken, A., Lindemann, K. Y., Kristensen, G. B., Kasius, J. C., Zeimet, A. G., et al. (2026) · npj Precision Oncology · doi:10.1038/s41698-026-01311-6 · Cited by 1
“Recently, we have seen foundation models 13 published for computational pathology, which are trained by self-supervised learning on histological images from multiple cancertypes 14 – 20 .”
Elshaer, Z., Jamal, A., & Rashed, E. A. (2026) · Biomedical Signal Processing and Control · doi:10.1016/j.bspc.2026.109741 · Cited by 1
Transcriptomic-guided whole-slide image classification for molecular subtype identification
Wang, W., Zhang, X., & Xiong, Y. (2026) · PLoS Computational Biology · doi:10.1371/journal.pcbi.1013950 · Cited by 1
Reinisch, W., Rittscher, J., Iacucci, M., Travis, S. P. L., & Friedrich, M. (2026) · Gastroenterology · doi:10.1053/j.gastro.2025.12.002 · Cited by 2
Peroz, M., Roussot, N., Ilie, A., Rageot, D., Dérangère, V., Truntzer, C., & Ghiringhelli, F. (2026) · Frontiers in Immunology · doi:10.3389/fimmu.2026.1750816 · Cited by 1
Gyarteng, E. S. A., XiaoLing, Z., & Ukwuoma, C. C. (2026) · Information Processing & Management · doi:10.1016/j.ipm.2026.104679 · Cited by 1
Bahcekapili, A., Arslan, D., Ozdemir, U., Ozkirli, B., Akbaş, E., Acar, A., Akar, G. B., He, B., Xu, S., Çağlar, Ü. M., et al. (2025) · IEEE International Conference on Image Processing Workshops (ICIPW) · doi:10.1109/icipw68931.2025.11385990 · Cited by 1
Hao, X., Xu, H., Zhang, J., Xu, Q., Pölönen, I., & Cong, F. (2026) · Medical Image Analysis · doi:10.1016/j.media.2026.104005 · Cited by 1
Zhang, C., Qu, L., Liu, X., & Song, Z. (2026) · Computerized Medical Imaging and Graphics · doi:10.1016/j.compmedimag.2026.102733 · Cited by 1
From Cosmos to Clinic: Interpretable Spatial Statistics for Histopathology Prognosis
Barna, A., Kilim, O., & Csabai, I. (2025) · IEEE International Conference on Computer Vision Workshops · doi:10.1109/iccvw69036.2025.00100 · Cited by 1
“Each patch P k is processed through a foundation model encoder ϕ : R 256 × 256 × 3 → R d (e.g., UNI [5]) to produce high-dimensional feature embeddings h k = ϕ ( P k ) ∈ R d where d = 1024 .”
Towards Robust Multimodal Representation: A Unified Approach with Adaptive Experts and Alignment
Moradinasab, N., Sengupta, S., Liu, J., Syed, S., & Brown, D. E. (2025) · IEEE International Conference on Computer Vision Workshops · doi:10.1109/iccvw69036.2025.00130 · Cited by 1
“MoSARe receives 2048 sampled patch embeddings along with GMM means per slide, normalizes them, and applies ABMIL to the sampled patch embed-dings to derive a global representation.”
Computational pathology in bladder cancer: A scoping review
Superdock, M., Wobker, S. E., Carmicheal, I., & Kim, W. Y. (2026) · Bladder Cancer · doi:10.1177/23523735251413333 · Cited by 2
“19 Emerging approaches such asimproved architecturesforwhole-slide image analysis (HIPT, 20 GigaPath 21 ) or foundation models for digitalpathology(UNI, 22 CONCH, 23 Prov-GigaPath 21 ) may alleviate some of these issues for weakly supervised models.”
Niu, Y., Liu, J., Zhan, Y., Shi, J., Zhang, D., Reinius, M., Machado, I. P., Crispin‐Ortuzar, M., Wu, J., Li, C., et al. (2026) · Medical Image Analysis · doi:10.1016/j.media.2026.104008 · Cited by 6
Predicting Stereotactic Body Radiation Therapy Response Using an AI-Based Tumor Vessel Biomarker
Park, J. H., Lim, J. H., KIM, S., Kim, C.-H., You, S., Choi, J.-S., Chang, J. W., Park, D. H.-B., Lee, M.-W., Kim, S., et al. (2026) · Technology in Cancer Research & Treatment · doi:10.1177/15330338261428377 · Cited by 1
Butt, N. A., Sarwat, D., Noya, I. D., Pifarré, K. T., Samee, N. M. A., & Ashraf, I. (2026) · Scientific Reports · doi:10.1038/s41598-026-39196-x · Cited by 2
“Developed by the Mahmood Lab at Harvard/BWH, UNI2 is a Vision Transformer (ViT-H/14) trained on over 200 million pathology images from more than 350,000 diverse whole slide images covering H&E and IHC staining 23 .”
GE-ViP: Graph-enhanced vision pipeline for weakly supervised histopathology slide analysis
Rahman, M., Ahad, J. I., Hassan, M. M., Moursalin, G., & Momen, S. (2026) · Neurocomputing · doi:10.1016/j.neucom.2026.133230 · Cited by 1
Cheng, Y., Azouzi, N., Laurent-Bellue, A., Guo, Z., Chung, T., Zeng, Q., Ghodsifard, A., Albrecht, T., Roessler, S. K., Boulagnon‐Rombi, C., et al. (2026) · Annals of Oncology · doi:10.1016/j.annonc.2026.02.018 · Cited by 1
A content-aware variable-rate framework for pathology learned image compression (PathoLIC)
Li, W., Li, Y., Chen, H., Yang, L., Wu, L., Li, Z., Ke, J., & Shen, D. G. (2026) · Medical Image Analysis · doi:10.1016/j.media.2026.104018 · Cited by 1
A Contrastive Variational AutoEncoder for NSCLC Survival Prediction with Missing Modalities
Zanitti, M., Miskovic, V., Trovò, F., Pedrocchi, A., Shen, M., Tun, Y. K., Prelaj, A., & Kosta, S. (2025) · IEEE International Conference on Big Data (BigData) · doi:10.1109/bigdata66926.2025.11401431 · Cited by 1
“To extract patch features, we selected the UNI foundation model [17].”
PTCMIL: multiple instance learning via prompt token clustering for whole slide image analysis
Zhao, B., Kim, S., Chen, H., Zhou, C., Gao, Z., Wang, G., & Li, X. (2026) · Medical Image Analysis · doi:10.1016/j.media.2026.104013 · Cited by 4
Kaleem, M. B., Li, H., Zhu, Z., Chen, X., Fan, Y., Yan, L., & Liu, W. (2026) · IEEE Transactions on Transportation Electrification · doi:10.1109/tte.2026.3673041 · Cited by 1
“These models have demonstrated strong predictive and generalization capabilities across various fields, including remote sensing [27] and computational biology, such as in cell-type annotation, transfer learning in genomics, and foundation models for single-cell data [28], [29], [30], [31].”
Convpaint—Interactive pixel classification using pretrained neural networks
Hinderling, L., Schwob, R., Witz, G., Stojiljković, A., Dobrzyński, M., Vladymyrov, M., Frei, J., Grädel, B., Frismantiene, A., & Pertz, O. C. (2026) · Cell Reports Methods · doi:10.1016/j.crmeth.2026.101335 · Cited by 14
Petäinen, L., Väyrynen, J. P., Böhm, J., Ruusuvuori, P., Ahtiainen, M., Elomaa, H., Karjalainen, H., Kastinen, M., Tapiainen, V. V., Äijälä, V. K., et al. (2026) · Computer Methods and Programs in Biomedicine · doi:10.1016/j.cmpb.2026.109317 · Cited by 2
BioFuse: an embedding fusion framework for biomedical foundation models
Hossain, M. N., & Harris‐Birtill, D. (2026) · PLoS ONE · doi:10.1371/journal.pone.0320989 · Cited by 2
Morphology-based cytogenetic risk prediction in multiple myeloma from bone marrow smears
Gao, M., Yan, X., Chen, L., Ouyang, G., Liu, Y., & Chen, S. (2026) · Frontiers in Oncology · doi:10.3389/fonc.2026.1790130 · Cited by 1
“…pooling of DinoBloom embeddings, (3) attention-based MIL with ImageNet-pretrained ResNet50 features, (4) attention-based MIL with UNI embeddings (25), a general histopathology foundation model pretrained on over 100 million pathology image tiles, (5) attention-based MIL with CONCH embeddings…”
Deng, X. G., Zhu, T., Wang, Y., Hu, B., Yuan, H., Yu, D., Zeng, F., Huang, Z., Jia, X., Gao, Y., et al. (2026) · Journal of Translational Medicine · doi:10.1186/s12967-026-07954-0 · Cited by 1
Bridging the gap between tumor evolution and clonality diagnostics in multiple lung carcinomas
Janssen, J., Allgäuer, M., Ylstra, B., Draijer, S., Kim, Y., Budczies, J., Kazdal, D. N., Stenzinger, A., Kirchner, M., & Radonic, T. R. (2026) · Seminars in Cancer Biology · doi:10.1016/j.semcancer.2026.03.003 · Cited by 1
Dantone, M., Lacsamana, M., Zeng, K., Kenny, P. A., Geras, K. J., & Witowski, J. S. (2026) · Diagnostics · doi:10.3390/diagnostics16071023 · Cited by 1
Biologically-inspired semi-supervised semantic segmentation for biomedical imaging
Ciampi, L., Lagani, G., Amato, G. M., & Falchi, F. (2026) · Neural Networks · doi:10.1016/j.neunet.2026.108925 · Cited by 4
A Spatial‐Mechanistic Design Map for Microenvironment‐Responsive Nanomaterials in Solid Tumors
Hamad, R. S., Saber, S., Sameh, A., Elmorsy, E. A., Farrag, A. A., Eissa, H., El-kott, A. F., AlShehri, M. A., Alfarteesh, H. A., Eltantawy, W., et al. (2026) · The FASEB Journal · doi:10.1096/fj.202600821r · Cited by 1
Sajjad, U., Akbar, A., Su, Z., Leyva, A., Knight, D. A., Frankel, W. L., Gürcan, M. N., Chen, W., & Niazi, M. K. K. (2026) · Cancers · doi:10.3390/cancers18071150 · Cited by 9
“More recently, foundation models trained using self-supervised learning on millions of WSIs have emerged as a dominant paradigm in computational pathology [20–25].”
InfluentialFully Automated Stain Quantification Framework for IHC Whole Slide Images in Breast Cancer
Yin, T., Lifrange, F., Denis, Z., Caluwé, A. D., Buisseret, L., Catteau, X., Legros, C., Reynaert, N., & Dhont, J. (2026) · Technology in Cancer Research & Treatment · doi:10.1177/15330338251407734 · Cited by 1
“14 To address these limitations, we present and publicly release a fully automated, compartment-speci fi c (ie, tumor and stroma) H-scoring framework for IHC analysis, built on well-established DL architectures and validated across three different biomarker stains.”
PRET is a few-shot system for pan-cancer recognition without example training
Li, Y., Ning, Z., Xiang, T., Zhang, Q., Lin, Z., Yi, M., Feng, F., Zeng, B., Qian, X., Sun, L., et al. (2026) · Nature Cancer · doi:10.1038/s43018-026-01141-2 · Cited by 4
“No statistical methods were used to pre-determine sample sizes but our sample sizes are similar to those reported in previous publications11,16,41.”
InfluentialGAMMA-Net: Gated Attention Multi-scale Modular Architecture
Arora, M., Negi, A., Thakur, A., Zahra, S. R., & Gupta, A. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-16370-7_20 · Cited by 1
Lu, S., Pang, Y., Wen, H., Hu, L., Zhou, X., Yang, X., Qi, X., Liu, C., Liu, J., Qi, P., et al. (2026) · npj Precision Oncology · doi:10.1038/s41698-026-01405-1 · Cited by 1
Li, W., & Tan, X. (2026) · Journal of Biomedical Informatics · doi:10.1016/j.jbi.2026.105044 · Cited by 1
SSMamba: A self-supervised hybrid state space model for pathological image classification
Chai, E., Chen, S., Zhang, T., Li, X., & Cui, T. (2026) · Medical Image Analysis · doi:10.1016/j.media.2026.104080 · Cited by 1
“The comparative baselines include ViT [46], Swin-Transformer (Swin) [47], MAE [48], VMamba [49], UNI [24], UNI2 [24], MUSK [28], CONCH [29], CTransPath [23], Prov-GigaPath (Gigapath) [26], and Virchow2 [25].”
Qu, H., Li, J., Gao, Y., Xu, X.-M., Zhang, X.-B., & Yang, S.-D. (2026) · Frontiers in Artificial Intelligence · doi:10.3389/frai.2026.1807400 · Cited by 1
Wang, C., Lai, P.-J., Chu, T.-C., Wu, T.-K., Liang, C.-H., & Chao, T.-K. (2026) · npj Digital Medicine · doi:10.1038/s41746-026-02604-7 · Cited by 1
“This is signi fi cantly faster than using UNI (21.49 h ), Virchow2 (40.5 h), or GigaPath (34.98 h).”
InfluentialClinical Implementation of Artificial Intelligence Systems in Medical Imaging and Radiotherapy
Vienna International Centre, 100, P. B., Vienna, A.-1., Austria, 2600-0, T. +. (., & 2600-7, F. +. (. (2026) · IAEA human health series · doi:10.61092/iaea.9olq-aqa9 · Cited by 2
Cen, M., Zhuang, Z., Zhang, Y., Zeng, M., Magnier, B., Yu, L., Zhang, H., & Wang, L. S. (2025) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51701.2025.02261 · Cited by 1
“Due to the huge size of WSIs, end-to-end training is impractical, so MIL typically follows the two-stage pipeline: (1) feature extraction using pretrained models such as ViT [9], ctranspath [37], or UNI [7] and (2) feature aggregation and prediction.”
Conditional Visual Autoregressive Modeling for Pathological Image Restoration
Liu, Z. T., Xu, Z., Ma, J., Li, W., Wang, R., Du, B., & Chen, H. (2025) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51701.2025.01656 · Cited by 2
“High-quality pathological images play a crucial role in accurate diagnosis, effective treatment planning, and ongo-* ing patient care [4, 7, 38].”
Ma, Y., An, B., Shen, A., Yuan, M., Duan, M., & Wang, M. (2025) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51701.2025.02187 · Cited by 1
“As the emergence of a series of pathology foundation models [9, 16, 30], the instance feature extractor has been largely improved.”
Qiu, Z., Chao, H., Lin, T., Chang, W., Yang, Z., Jiao, W., Shen, Y., Zhang, Y., Yang, Y., Liu, W., et al. (2025) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51701.2025.02111 · Cited by 1
“Two-stage MIL: Two-stage MIL methods first segment WSIs into fixed-size patches using tools like CLAM [23] and extract features with pre-trained networks (e.g., ResNet [13], ViT [1], or other foundation models [5, 24, 38, 43]).”
Bedau, T., Harder, C., Al-Shughri, A., Wang, Y., Pryalukhin, A., Eich, M., Lyu, S. I., Büttner, R., Quaas, A., & Tolkach, Y. (2026) · Communications Medicine · doi:10.1038/s43856-026-01601-x · Cited by 3
“Aparticularlyactiveareaofresearchhasbeenthedevelopmentoflarge histopathology foundation models, 6 – 10 , which provide feature embeddings for downstream deep learning tasks.”
Ji, J., Zhang, X., Hua, M., Wang, M., Li, H., Zheng, X., Wang, L., Wang, H., Song, Y., He, J., et al. (2026) · npj Precision Oncology · doi:10.1038/s41698-026-01449-3 · Cited by 1
“1b) 44 .”
Artificial intelligence in cancer diagnostics: a primer for clinicians and scientists
Rizwan, A. M., Makhlouf, H. R., Bharti, S., Dey, S. M., Ossandon, M. R., Robustelli, C. E., Samant, M., Robles, A. I., Sorg, B. S., & Harris, L. N. (2026) · JNCI Journal of the National Cancer Institute · doi:10.1093/jnci/djag143 · Cited by 2
Lecuelle, J., Truntzer, C., Basile, D., Laghi, L., Greco, L., Ilie, A., Rageot, D., Huppé, T., Emile, J., Bibeau, F., et al. (2026) · OncoImmunology · doi:10.1080/2162402x.2026.2668797 · Cited by 1
“26 For the CNN-10cat model, 512 features per tile were collected, and 1024 for the UNI model.”
A straightforward framework to harmonize computational pathology
Dy, A., & Lennerz, J. K. (2026) · Journal of Pathology Informatics · doi:10.1016/j.jpi.2026.100676 · Cited by 1
“To illustrate this con fl ation, we reviewed several widely cited pathology foundation models, 1,2,4,6,7,9,11,12 and how the indication “ colorectal cancer ” was represented in their training datasets.”
Yin, W., Peng, Q., Meng, F., Wan, Y., Zhang, W., & Zhou, Y. (2026) · Nature Computational Science · doi:10.1038/s43588-026-00992-0 · Cited by 1
Ikinci, M., Toma, L., Loeffler, K. U., Ussem, L., Süsskind, D., Weller, J. M., Yeganeh, Y., Herwig‐Carl, M. C., & Albarqouni, S. (2026) · IEEE International Symposium on Biomedical Imaging (ISBI) · doi:10.1109/isbi61048.2026.11515389 · Cited by 1
“Recent advances in deep learning, particularly foundation models trained on large-scale pathology datasets [7, 8, 9, 10], have substantially improved performance in tissue classification, biomarker identification, and prognostic prediction.”
Alasi, H., Yilmaz, I., Reynolds, J., Agarwal, S., Nassar, A., Zubair, A., Horbinski, C., Dangott, B. J., & Akkus, Z. (2026) · Bioengineering · doi:10.3390/bioengineering13050577 · Cited by 1
“Over the past several years, multiple PFMs have been released by both academic and industry teams, reflecting growing interest in scalable and generalizable approaches to histopathology analysis [21,26–31 , 43–”
Zhong, J., Zhang, M., Liu, L., Liu, T., Jiang, J., Piao, Y., Xu, R., Tian, F., Sun, W., Bi, H., et al. (2026) · Medical Image Analysis · doi:10.1016/j.media.2026.104138 · Cited by 1
Ebbert, J. L., & Corte, D. D. (2026) · AI in Medicine · doi:10.3390/aimed1020014 · Cited by 1
Foundation Models in Cancer Pathology: Techniques, Applications, and Future Directions
Zhang, B., Cui, V. Y., Wu, T., Su, G., Huang, Q., Shang, B., Kong, M., Yu, W., Yuan, Y., & Tu, H. (2026) · Research · doi:10.34133/research.1332 · Cited by 1
Cai, L., Meng, B., Huang, J., Ding, G., Ju, M., Wang, W.-W., Shi-jie, D., Lai, L., Wang, J., Yang, C.-X., et al. (2026) · LabMed discovery. · doi:10.1016/j.lmd.2026.100130 · Cited by 1
Self-citationIdaji, M. J., Hense, J., Neuhauser, T., Krause, A., Luo, Y., Eberle, O., Schnake, T., Ciernik, L., Jafari, F. R., Vahidimajd, R., et al. (2026) · Medical Image Analysis · doi:10.1016/j.media.2026.104148 · Cited by 2
“To investigate whether using different feature extractors impacts MIL explanations, we included two state-of-the-art models, namely UNI2-h 5 [64] and Virchow2 [68] in our experiments.”
Chuter, B., Emmert, N., Kim, M. Y., Stiemke, A. B., Dave, N., Herrin, J., Zhou, Z., Wall, G., Lu, L., Williams, R. W., et al. (2026) · Bioengineering · doi:10.3390/bioengineering13060647 · Cited by 1
Gui, K., Li, Q., Xie, Z. Y., Zhan, T., Long, X., Yang, G., Luo, K., & Liu, H. (2026) · Progress in Biomedical Engineering · doi:10.1088/2516-1091/ae7860 · Cited by 1
“Commonly used encoders include PLIP [94], Prov-GigaPath [47], and UNI [95], which are specifically designed or adapted for pathology images and demonstrate strong generalization ability across datasets.”
Jiang, X., Li, P., Chen, Z., Yao, H., Jia, H., Wang, T., & Tang, X. (2026) · Frontiers in Immunology · doi:10.3389/fimmu.2026.1841261 · Cited by 1
“The emergence of computational pathology, driven by foundational AI models such as UNI and CONCH and more recent pathology foundation models (12 – 16), has initiated a paradigm shift in digital histopathology.”
Schwarzmann, L., Stöhr, R., Emons, J., Pöschke, P., Seitz, K., Rübner, M., Beckmann, M. W., Diederich, M. B., Unser, I., Hartmann, A., et al. (2026) · npj Precision Oncology · doi:10.1038/s41698-026-01522-x · Cited by 1
“We benchmarked the feature extraction with CTransPath 33 and the foundation model UNI 32 , demonstrating robust results across different scanner hardware and quanti fi ed performance variation by additional stain normalization methods.”
InfluentialTiwari, A. K., & Bhandari, A. K. (2026) · Intelligence-Based Medicine · doi:10.1016/j.ibmed.2026.100404 · Cited by 2
Torbati, N., Meshcheryakova, A., Woitek, R., Hatamikia, S., Mechtcheriakova, D., & Mahbod, A. (2026) · Machine Learning with Applications · doi:10.1016/j.mlwa.2026.100933 · Cited by 1
Intelligent three-dimensional imaging in pathology: Applications and developments
Si, W., Zhao, F., Kuang, D., Zhao, Y., Liu, Y., Sun, M., Duan, Y. Q., Fei, P., & Wang, G. (2026) · Chinese Medical Journal · doi:10.1097/cm9.0000000000004176 · Cited by 1
Zhang, X., Wang, T., Yan, C., Najdawi, F., Zhou, K., Ma, Y., Cheung, Y.-M., Yeung, M. C. F., & Malin, B. (2026) · Nature Biomedical Engineering · doi:10.1038/s41551-026-01694-8 · Cited by 1
“The ABMIL model architecture and training configuration followed those established in UNI 28 and CONCH 29 .”
Ye, Y., Yuan, J., Tang, J., Wan, P., Sun, L., Sheng, J., Zhang, D., & Shao, W. (2026) · IEEE Transactions on Image Processing · doi:10.1109/tip.2026.3705155 · Cited by 1
“Moreover, recent pathology foundation models have shown substantially advancement for zero-shot pathological image analysis, the studies in [28], [51], and [52] leverage pathology image-text pairs to build pathology foundation models for multi-modal representations.”
Metadata
- Authors
- Chen, R. J., Ding, T., Lu, M. Y., Williamson, D. F. K., Jaume, G., Song, A. H., Chen, B., Zhang, A., Shao, D., Shaban, M., Williams, M., Oldenburg, L., Weishaupt, L. L., Wang, J. J., Vaidya, A., Le, L. P., Gerber, G., Sahai, S., Williams, W., & Mahmood, F.
- Affiliations
- Broad Institute (US)
- Brigham and Women's Hospital (US)
- Harvard University (US)
- Corresponding author
- Mahmood, Faisal
- ORCID
- Venue
- Nature Medicine
- Published
- 2024-03
- Volume
- 30
- Issue
- 3
- Pages
- 850-862
- ISSN
- 1078-8956 (ISSN-L)
- 1546-170X
- PMC
- PMC11403354
- OpenAlex
- W4392947521
- Semantic Scholar
- cac496ddebbff6bbba78242780cde01bd962bfc9
- Cited by
- 1,560
- References
- 170
- Citations by year
- 2026: 789 · 2025: 613 · 2024: 154
- FWCI
- 229.15
- Type
- journal article
- Language
- en
- Open access
- green
- www.ncbi.nlm.nih.gov
- Licence
- Locations
- Nature Medicine (publishedVersion)
- PubMed (publishedVersion)
- PubMed Central (acceptedVersion)
- Funders
- National Human Genome Research Institute (T32HG002295)
- National Cancer Institute (T32CA251062)
- National Institute of General Medical Sciences (R35GM149270)
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › AI in cancer detection
- Health Sciences › Medicine › Radiology, Nuclear Medicine and Imaging › Radiomics and Machine Learning in Medical Imaging
- Physical Sciences › Computer Science › Artificial Intelligence › Domain Adaptation and Few-Shot Learning
- Keywords
- computational pathology, whole slide images, self-supervised learning, foundation models, unsupervised feature learning, tissue classification, few-shot learning, cancer subtyping, histopathology, transfer learning
- MeSH terms
- Artificial Intelligence, Workflow
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
