Magnification Prior: A Self-Supervised Method for Learning Representations on Breast Cancer Histopathological Images

Chhipa, Prakash Chandra; Upadhyay, Richa; Pihlgren, Gustav Grund; Saini, Rajkumar; Uchida, Seiichi; Liwicki, Marcus

doi:10.1109/wacv56688.2023.00274

Cited by 21 publications

(6 citation statements)

References 50 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“… a , Average accuracy comparison at the patient level for different resolutions. The compared models include CNN models (Deep 31 , SW 31 , GLPB 48 , RPDB 49 ), weakly supervised models (MIL-NP 32 , MILCNN 32 ), and self-supervised model (MPCS-RP 38 ). b , Average accuracy comparison at the image level for different resolutions.…”

Section: Resultsmentioning

confidence: 99%

“…On the other hand, image-level accuracy does not consider patient-level details. It is calculated as the proportion of correctly classified images among all images, regardless of the patients they belong to 38 . Let N p represents the number of pathological images owned by patient p. For each patient, N rec is the number of pathological images that are correctly classified.…”

Section: Methodsmentioning

confidence: 99%

See 1 more Smart Citation

A foundation model for generalizable cancer diagnosis and survival prediction from histopathological images

Yang,

Wei,

Liang

et al. 2024

Preprint

View full text Add to dashboard Cite

Computational pathology, utilizing whole slide image (WSI) for pathological diagnosis, has advanced the development of intelligent healthcare. However, the scarcity of annotated data and histological differences hinder the general application of existing methods. Extensive histopathological data and the robustness of self-supervised models in small-scale data demonstrate promising prospects for developing foundation pathology models. In this work, we propose the BEPH (BEiT-based model Pre-training on Histopathological image), a general method that leverages self-supervised learning to learn meaningful representations from 11 million unlabeled histopathological images. These representations are then efficiently adapted to various tasks, including patch-level cancer recognition, WSI-level cancer classification, and survival prediction for multiple cancer subtypes. Experimental results demonstrate that our model consistently outperformsseveral comparative models, even with limited training data reduced to 50%. Especially when the downstream structure is the same, the model can improve ResNet and DINO by up to a maximum increase of 8.8% and 7.2% (WSI level classification), and 6.44% and 3.28% on average (survival prediction), respectively. Therefore, BEPH offers a universal solution to enhance model performance, reduce the burden of expert annotations, and enable widespread clinical applications of artificial intelligence. The code and models can be obtained at https://github.com/Zhcyoung/BEPH. And currently, online fine-tuning of WSI classification tasks is available for use on http://yulab-sjtu.natapp1.cc/BEPH.

show abstract

Section: Resultsmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%

A foundation model for generalizable cancer diagnosis and survival prediction from histopathological images

Yang,

Wei,

Liang

et al. 2024

Preprint

View full text Add to dashboard Cite

show abstract

“…The success of these approaches depends on the nature of the ML task and the validity of assumptions underlying these approaches. Recently, self‐supervised learning methods [ 96 , 97 , 98 , 99 ] that exploit supervisory signals in the data itself with the help of domain‐specific as well as domain‐agnostic tasks have proven to be successful for effective tumor detection with limited available annotations. However, development of truly generalizable weakly supervised or self‐supervised approaches remains an open problem [ 28 ].…”

Section: Limitations Challenges and Recommendationsmentioning

confidence: 99%

Unleashing the potential of AI for pathology: challenges and recommendations

Asif

Rajpoot

Graham

et al. 2023

The Journal of Pathology

View full text Add to dashboard Cite

Computational pathology is currently witnessing a surge in the development of AI techniques, offering promise for achieving breakthroughs and significantly impacting the practices of pathology and oncology. These AI methods bring with them the potential to revolutionize diagnostic pipelines as well as treatment planning and overall patient care. Numerous peer‐reviewed studies reporting remarkable performance across diverse tasks serve as a testimony to the potential of AI in the field. However, widespread adoption of these methods in clinical and pre‐clinical settings still remains a challenge. In this review article, we present a detailed analysis of the major obstacles encountered during the development of effective models and their deployment in practice. We aim to provide readers with an overview of the latest developments, assist them with insights into identifying some specific challenges that may require resolution, and suggest recommendations and potential future research directions. © 2023 The Authors. The Journal of Pathology published by John Wiley & Sons Ltd on behalf of The Pathological Society of Great Britain and Ireland.

show abstract

“…There are a number of studies in the literature which have been applied to these datasets. For instance, many researchers have utilised BreakHis dataset to test their networks [11,12,19,20,[26][27][28][45][46][47][48][49][50][51][52]. Zhou et al [45] proposed a novel resolution adaptive network (RAN) to classify different forms of breast cancer.…”

Section: Introductionmentioning

confidence: 99%

Enhancing Breast Cancer Histopathological Image Classification using Attention-Based High Order Covariance Pooling

Waqas,

Ahmed,

Maul

et al. 2023

Preprint

View full text Add to dashboard Cite

The type of cancer that affects female patients most frequently is breast cancer. Computer-aided diagnosis, these days, proves to be helpful for many diseases including breast cancer. Deep learning based approaches have yielded encouraging results in many computer vision applications. In the case of breast cancer histopathological image classification, many methods have been proposed but the accuracy is still insufficient for full deployment. Due to this reason, we propose a deep learning based approach, DHA-Net which consists of an attention module and a high-order pooling layer incorporated into a ResNet-18 architecture. DHA-Net was applied to three well-known datasets in the field of breast cancer histopathological image classification, i.e. BreakHis, BACH2018 and a closely related Kaggle-Breast cancer histopathology dataset. In this area of medical imaging, experiments have shown that DHA-Net works better than other existing techniques.

show abstract

Magnification Prior: A Self-Supervised Method for Learning Representations on Breast Cancer Histopathological Images

Cited by 21 publications

References 50 publications

A foundation model for generalizable cancer diagnosis and survival prediction from histopathological images

A foundation model for generalizable cancer diagnosis and survival prediction from histopathological images

Unleashing the potential of AI for pathology: challenges and recommendations

Enhancing Breast Cancer Histopathological Image Classification using Attention-Based High Order Covariance Pooling

Contact Info

Product

Resources

About