An interpretable classifier for high-resolution breast cancer screening images utilizing weakly supervised localization

Shen, Yiqiu; Wu, Nan; Phang, Jason; Park, Jungkyu; Liu, Kangning; Tyagi, Sudarshini; Heacock, Laura; Kim, Sungheon; Moy, Linda; Cho, Kyunghyun; Geras, Krzysztof J.

doi:10.1016/j.media.2020.101908

Cited by 122 publications

(144 citation statements)

References 39 publications

Supporting

Mentioning

132

Contrasting

Unclassified

Order By: Relevance

“…More data enhancement strategies including better CA methods need further exploration to make the best use of pre‐trained models and thus boost performance. In the current implementation of GL1 and GL2, the default hyper‐parameter settings as suggested in Shen et al 14 . and Shen et al 15 . were used.…”

Section: Discussionmentioning

confidence: 99%

“…Finally, the fusion module aggregates the global information (generated by the global module) and the local details (generated by the local module) to make a prediction on the existence of malignant lesions in a mammogram. The details of GL2 can be referred to in Shen et al 15 …”

Section: Methodsmentioning

confidence: 99%

“…We explored three of such pre‐trained models ‡ with ResNet22 and ResNet18 16 used in global and local modules, corresponding to three different hyper‐parameter settings, that is GL2‐ResNet22/18‐Setting 1, GL2‐ResNet22/18‐Setting 2 and GL2‐ResNet22/18‐Setting 3. To re‐train them on our own training data, we resized images to 2944 × 1920 (to enable the use of the pre‐trained model), employed the pre‐trained model to initialise the model to be re‐trained, and used the default data pre‐processing operations and hyper‐parameter settings suggested in Shen et al 15 …”

Section: Methodsmentioning

confidence: 99%

See 2 more Smart Citations

Evaluation of deep learning‐based artificial intelligence techniques for breast cancer detection on mammograms: Results from a retrospective study using a BreastScreen Victoria dataset

et al. 2021

View full text Add to dashboard Cite

Introduction This study aims to evaluate deep learning (DL)‐based artificial intelligence (AI) techniques for detecting the presence of breast cancer on a digital mammogram image. Methods We evaluated several DL‐based AI techniques that employ different approaches and backbone DL models and tested the effect on performance of using different data‐processing strategies on a set of digital mammographic images with annotations of pathologically proven breast cancer. Results Our evaluation uses the area under curve (AUC) and accuracy (ACC) for performance measurement. The best evaluation result, based on 349 test cases (930 test images), was an AUC of 0.8979 [95% confidence interval (CI) 0.873, 0.923] and ACC of 0.8178 [95% CI 0.785, 0.850]. This was achieved by an AI technique that utilises a certain family of DL models, namely ResNet, as its backbone, combines the global features extracted from the whole mammogram and the local features extracted from the automatically detected cancer and non‐cancer local regions in the whole image, and leverages background cropping and text removal, contrast adjustment and more training data. Conclusion DL‐based AI techniques have shown promising results in retrospective studies for many medical image analysis applications. Our study demonstrates a significant opportunity to boost the performance of such techniques applied to breast cancer detection by exploring different types of approaches, backbone DL models and data‐processing strategies. The promising results we have obtained suggest further development of AI reading services could transform breast cancer screening in the future.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%

See 1 more Smart Citation

Evaluation of deep learning‐based artificial intelligence techniques for breast cancer detection on mammograms: Results from a retrospective study using a BreastScreen Victoria dataset

et al. 2021

View full text Add to dashboard Cite

show abstract

“…In our work, we parameterize f agg as the top t% pooling proposed by Shen et al [57]. Namely, we define the aggregation function as…”

Section: B) a Summary Of The Acquisition Devices Is Shown In Tablementioning

confidence: 99%

Artificial Intelligence System Reduces False-Positive Findings in the Interpretation of Breast Ultrasound Exams

Shen

Shamout

Oliver

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

Ultrasound is an important imaging modality for the detection and characterization of breast cancer. Though consistently shown to detect mammographically occult cancers, especially in women with dense breasts, breast ultrasound has been noted to have high false-positive rates. In this work, we present an artificial intelligence (AI) system that achieves radiologist-level accuracy in identifying breast cancer in ultrasound images. To develop and validate this system, we curated a dataset consisting of 288,767 ultrasound exams from 143,203 patients examined at NYU Langone Health, between 2012 and 2019. On a test set consisting of 44,755 exams, the AI system achieved an area under the receiver operating characteristic curve (AUROC) of 0.976. In a reader study, the AI system achieved a higher AUROC than the average of ten board-certified breast radiologists (AUROC: 0.962 AI, 0.924±0.02 radiologists). With the help of the AI, radiologists decreased their false positive rates by 37.4% and reduced the number of requested biopsies by 27.8%, while maintaining the same level of sensitivity. To confirm its generalizability, we evaluated our system on an independent external test dataset where it achieved an AUROC of 0.911. This highlights the potential of AI in improving the accuracy, consistency, and efficiency of breast ultrasound diagnosis worldwide.

show abstract

“…

Deep learning (DL) has been applied with success in proofs of concept across biomedical imaging, including across modalities and medical specialties [1][2][3][4][5][6][7][8][9][10][11][12][13][14][15][16][17] . Labeled data is critical to training and testing DL models, and such models traditionally require large amounts of training data, straining the limited (human) resources available for expert labeling/annotation.

…”

mentioning

confidence: 99%

ENRICHing Medical Imaging Training Sets Enables More Efficient Machine Learning

Chinn

Arora

Arnaout

et al. 2021

Preprint

View full text Add to dashboard Cite

Deep learning (DL) has been applied with success in proofs of concept across biomedical imaging, including across modalities and medical specialties1-17. Labeled data is critical to training and testing DL models, and such models traditionally require large amounts of training data, straining the limited (human) resources available for expert labeling/annotation. It would be ideal to prioritize labeling those images that are most likely to improve model performance and skip images that are redundant. However, straightforward, robust, and quantitative metrics for measuring and eliminating redundancy in datasets have not yet been described. Here, we introduce a new method, ENRIch (Eliminate Needless Redundancy in Imaging datasets), for assessing image dataset redundancy and test it on a well benchmarked medical imaging dataset3. First, we compute pairwise similarity metrics for images in a given dataset, resulting in a matrix of pairwise similarity values. We then rank images based on this matrix and use these rankings to curate the dataset, to minimize dataset redundancy. Using this method, we achieve similar AUC scores in a binary classification task with just a fraction of our original dataset (AUC of 0.99 +/- 1.35e-05 on 44 percent of available images vs. AUC of 0.99 +/- 9.32e-06 on all available images, p-value 0.0002) and better scores than the same sized training subsets chosen at random. We also demonstrate similar Jaccard scores in a multi-class segmentation task while eliminating redundant images (average Jaccard index of 0.58 on 80 percent of available images vs. 0.60 on all available images). Thus, algorithms that reduce dataset redundancy based on image similarity can significantly reduce the number of training images required, while preserving performance, in medical imaging datasets.

show abstract

An interpretable classifier for high-resolution breast cancer screening images utilizing weakly supervised localization

Cited by 122 publications

References 39 publications

Evaluation of deep learning‐based artificial intelligence techniques for breast cancer detection on mammograms: Results from a retrospective study using a BreastScreen Victoria dataset

Evaluation of deep learning‐based artificial intelligence techniques for breast cancer detection on mammograms: Results from a retrospective study using a BreastScreen Victoria dataset

Artificial Intelligence System Reduces False-Positive Findings in the Interpretation of Breast Ultrasound Exams

ENRICHing Medical Imaging Training Sets Enables More Efficient Machine Learning

Contact Info

Product

Resources

About