Robust and Efficient Medical Imaging with Self-Supervision

Azizi, Shekoofeh; Culp, Laura; Freyberg, Jan; Mustafa, Basil; Baur, Sebastien; Kornblith, Simon; Chen, Ting; MacWilliams, Patricia; Mahdavi, Sara; Wulczyn, Ellery; Babenko, Boris; Wilson, Megan; Loh, Aaron; Chen, Po-Hsuan Cameron; Liu, Yuan; Bavishi, Pinal; McKinney, Scott Mayer; Winkens, Jim; Roy, Abhijit Guha; Beaver, Zach; Ryan, Fiona; Krogue, Justin D.; Etemadi, Mozziyar; Telang, Umesh; Liu, Yun; Peng, Lily; Corrado, Greg S.; Webster, Dale R.; Fleet, David J.; Hinton, Geoffrey E.; Houlsby, Neil; Karthikesalingam, Alan; Norouzi, Mohammad; Natarajan, Vivek

doi:10.48550/arxiv.2205.09723

Cited by 14 publications

(19 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…where (e 2 i ) shows that sum of squared residuals and (y i − ȳ) 2 shows total sum squared. R 2 is commonly used in clinical studies to assess how well a model explains and predicts future outcomes [19].…”

Section: Resultsmentioning

confidence: 99%

“…While some approaches have designed domainspecific pretext tasks [5,54,68,69]], others have adjusted well-known self-supervised learning methods to medical data [25,30,53,66]. Very recently [2] has applied SimCLR on a combination of unlabeled ImageNet dataset and task specific medical images for medical image classification; their experiments and improved performance suggest that pre-training on ImageNet is complementary to pre-training on unlabeled medical images.…”

Section: Related Workmentioning

confidence: 99%

“…Despite demand, the use of self-supervised approaches in the medical image domain has received limited attention. Only few studies have investigated the impact of self-supervised learning in the medical image analysis domain for limited applications including classification [2,25,32,34,53,68] and segmentation [5,9,50,54].…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Cross-Domain Self-Supervised Deep Learning for Robust Alzheimer's Disease Progression Modeling

Dadsetan¹,

Hejrati²,

Wu³

et al. 2022

Preprint

View full text Add to dashboard Cite

Developing successful artificial intelligence systems in practice depends both on robust deep learning models as well as large high quality data. Acquiring and labeling data can become prohibitively expensive and time-consuming in many real-world applications such as clinical disease models. Self-supervised learning has demonstrated great potential in increasing model accuracy and robustness in small data regimes. In addition, many clinical imaging and disease modeling applications rely heavily on regression of continuous quantities. However, the applicability of selfsupervised learning for these medical-imaging regression tasks has not been extensively studied. In this study, we develop a cross-domain self-supervised learning approach for disease prognostic modeling as a regression problem using 3D images as input. We demonstrate that self-supervised pre-training can improve the prediction of Alzheimer's Disease progression from brain MRI. We also show that pretraining on extended (but not labeled) brain MRI data outperforms pre-training on natural images. We further observe that the highest performance is achieved when both natural images and extended brain-MRI data are used for pre-training.

show abstract

Section: Resultsmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

Cross-Domain Self-Supervised Deep Learning for Robust Alzheimer's Disease Progression Modeling

Dadsetan¹,

Hejrati²,

Wu³

et al. 2022

Preprint

View full text Add to dashboard Cite

show abstract

“…Although we have tried our best to include a diverse set of algorithms and datasets in our benchmark, it is certainly not exhaustive. There are methods to promote fairness from other perspectives, e.g., self-supervised learning may be more robust (Liu et al, 2021;Azizi et al, 2022). Also, datasets from other medical data modalities (e.g., cardiology, digital pathology) should be added.…”

Section: Relation Of Domain Generalization and Fairnessmentioning

confidence: 99%

MEDFAIR: Benchmarking Fairness for Medical Imaging

Zong¹,

Yang²,

Hospedales³

2022

Preprint

View full text Add to dashboard Cite

A multitude of work has shown that machine learning-based medical diagnosis systems can be biased against certain subgroups of people. This has motivated a growing number of bias mitigation algorithms that aim to address fairness issues in machine learning. However, it is difficult to compare their effectiveness in medical imaging for two reasons. First, there is little consensus on the criteria to assess fairness. Second, existing bias mitigation algorithms are developed under different settings, e.g., datasets, model selection strategies, backbones, and fairness metrics, making a direct comparison and evaluation based on existing results impossible. In this work, we introduce MEDFAIR, a framework to benchmark the fairness of machine learning models for medical imaging. MEDFAIR covers eleven algorithms from various categories, nine datasets from different imaging modalities, and three model selection criteria. Through extensive experiments, we find that the under-studied issue of model selection criterion can have a significant impact on fairness outcomes; while in contrast, state-ofthe-art bias mitigation algorithms do not significantly improve fairness outcomes over empirical risk minimization (ERM) in both in-distribution and out-of-distribution settings. We evaluate fairness from various perspectives and make recommendations for different medical application scenarios that require different ethical principles. Our framework provides a reproducible and easy-to-use entry point for the development and evaluation of future bias mitigation algorithms in deep learning. Code is available at https://github.com/ys-zong/MEDFAIR.

show abstract

“…Self-supervised models can be more robust to dataset-level distribution shift [41] and have better transfer learning performance [42] than their supervised counterparts. The benefits of transfer learning using SSL on domain-specific data have been shown for a variety of x-ray and histology slide image tasks [43]. Finally, and possibly the most compelling, is that SSL enables learning with much more abundant unlabeled data, addressing the data scarcity challenge directly.…”

mentioning

confidence: 99%

Self-Supervised Learning with Limited Labeled Data for Prostate Cancer Detection in High Frequency Ultrasound

Wilson¹,

Mahdi²,

Jamzad³

et al. 2022

Preprint

View full text Add to dashboard Cite

Deep learning-based analysis of highfrequency, high-resolution micro-ultrasound data shows great promise for prostate cancer detection. Previous approaches to analysis of ultrasound data largely follow a supervised learning paradigm. Ground truth labels for ultrasound images used for training deep networks often include coarse annotations generated from the histopathological analysis of tissue samples obtained via biopsy. This creates inherent limitations on the availability and quality of labeled data, posing major challenges to the success of supervised learning methods. On the other hand, unlabeled prostate ultrasound data are more abundant. In this work, we successfully apply selfsupervised representation learning to micro-ultrasound data. Using ultrasound data from 1028 biopsy cores of 391 subjects obtained in two clinical centres, we demonstrate that feature representations learnt with this method can be used to classify cancer from non-cancer tissue, obtaining an AUROC score of 91% on an independent test set. To the best of our knowledge, this is the first successful end-to-end self-supervised learning approach for prostate cancer detection using ultrasound data. Our method outperforms baseline supervised learning approaches, generalizes well between different data centers, and scale well in performance as more unlabeled data are added, making it a promising approach for future research using large volumes of unlabeled data.

show abstract

Robust and Efficient Medical Imaging with Self-Supervision

Cited by 14 publications

References 20 publications

Cross-Domain Self-Supervised Deep Learning for Robust Alzheimer's Disease Progression Modeling

Cross-Domain Self-Supervised Deep Learning for Robust Alzheimer's Disease Progression Modeling

MEDFAIR: Benchmarking Fairness for Medical Imaging

Self-Supervised Learning with Limited Labeled Data for Prostate Cancer Detection in High Frequency Ultrasound

Contact Info

Product

Resources

About