Our Approach
Baselines:
Convolutional neural network (CNN) models may benefit from the ability of radiology experts to extract the "gist" of abnormality from mammograms. Prior research has demonstrated that radiologists can classify a mammogram as normal or abnormal at better than chance levels after less than a second's exposure to the image. In this work, we combine radiologist gist inputs into pre-trained machine learning models to validate that integrating human gist perception with a CNN model can achieve an AUC (area under the curve) of up to 0.899 – statistically significantly higher than either the gist perception of radiologists or the model without gist input.
Baselines:
Human Gist Processing Augments Deep Learning Breast Cancer Risk Assessment
Skylar W. Wurster, Arkadiusz Sitek, Jian Chen, Karla Evans, Gaeun Kim, and Jeremy Wolfe
Comming soon
[1] Yala, A., Lehman, C., Schuster, T., Portnoi, T. & Barzilay, R. A deep learning mammography-based model for improved breast cancer risk prediction. Radiology 182716, DOI: 10.1148/radiol.2019182716 (2019).
[2] Sitek, A. & Wolfe, J. M. Assessing cancer risk from mammograms: Deep learning is superior to conventional risk models. Radiology 292, 67–68, DOI: 10.1148/radiol.2019190791 (2019).
[3] Oliva, A. & Torralba, A. Building the gist of a scene: The role of global image features in recognition. Prog. brain research 155, 23–36, DOI: 10.1016/S0079-6123(06)55002-2 (2006).
[4] Evans, K. K., Georgian-Smith, D., Tambouret, R., Birdwell, R. L. & Wolfe, J. M. The gist of the abnormal: Above-chance medical decision making in the blink of an eye. Psychon. bulletin & review 20, 1170–1175, DOI: https://doi.org/10.3758/s13423-013-0459-3 (2013).
[5] Evans, K. K., Haygood, T. M., Cooper, J., Culpan, A.-M. & Wolfe, J. M. A half-second glimpse often lets radiologists identify breast cancer cases even when viewing the mammogram of the opposite breast. Proc. Natl. Acad. Sci. 113, 10292–10297, DOI: 10.1073/pnas.160618711310.1073/pnas.1606187113 (2016).
[6] Evans, K. K., Culpan, A.-M. & Wolfe, J. M. Detecting the “gist” of breast cancer in mammograms three years before localized signs of cancer are visible. The Br. J. Radiol. 92, 20190136, DOI: 10.1259/bjr.20190136 (2019).
[7] Brennan, P. C. et al. Radiologists can detect the ‘gist’ of breast cancer before any overt signs of cancer appear. Sci. Reports 8, DOI: https://doi.org/10.1038/s41598-018-26100-5 (2018).
[8] Deng, J. et al. ImageNet: A large-scale hierarchical image database. In IEEE Conference on Computer Vision and Pattern Recognition, 248–255, DOI: 10.1109/CVPR.2009.5206848 (2009).
[9] Ke, G. et al. LightGBM: A highly efficient gradient boosting decision tree. Adv. Neural Inf. Process. Syst. 3149–3157 http://papers.nips.cc/paper/6907-lightgbm-a-highly-efficient-gradient-boosting-decision-tree.pdf (2017).
[10] Akselrod-Ballin, A. et al. Predicting breast cancer by applying deep learning to linked health records and mammograms. Radiology 292, DOI: 10.1148/radiol.2019182622 (2019).
[11] Rampun, A., Morrow, P. J., Scotney, B. W. & Winder, J. Fully automated breast boundary and pectoral muscle segmentation in mammograms. Artif. Intell. Medicine 79, 28–41, DOI: 10.1016/j.artmed.2017.06.001 (2017).
