• Thomas Schaffter, BuistDiana S MDSMKaiser Permanente Washington Health Research Institute, Seattle, Washington., Christoph I Lee, Yaroslav Nikulin, Dezso Ribli, Yuanfang Guan, William Lotter, Zequn Jie, Hao Du, Sijia Wang, Jiashi Feng, Mengling Feng, Hyo-Eun Kim, Francisco Albiol, Alberto Albiol, Stephen Morrell, Zbigniew Wojna, Mehmet Eren Ahsen, Umar Asif, Antonio Jimeno Yepes, Shivanthan Yohanandan, Simona Rabinovici-Cohen, Darvin Yi, Bruce Hoff, Thomas Yu, Chaibub NetoEliasEComputational Oncology, Sage Bionetworks, Seattle, Washington., Daniel L Rubin, Peter Lindholm, Laurie R Margolies, Russell Bailey McBride, Joseph H Rothstein, Weiva Sieh, Rami Ben-Ari, Stefan Harrer, Andrew Trister, Stephen Friend, Thea Norman, Berkman Sahiner, Fredrik Strand, Justin Guinney, Gustavo Stolovitzky, and the DM DREAM Consortium, Lester Mackey, Joyce Cahoon, Li Shen, Jae Ho Sohn, Hari Trivedi, Yiqiu Shen, Ljubomir Buturovic, Jose Costa Pereira, Jaime S Cardoso, Eduardo Castro, Karl Trygve Kalleberg, Obioma Pelka, Imane Nedjar, Krzysztof J Geras, Felix Nensa, Ethan Goan, Sven Koitka, Luis Caballero, David D Cox, Pavitra Krishnaswamy, Gaurav Pandey, Christoph M Friedrich, Dimitri Perrin, Clinton Fookes, Bibo Shi, Gerard Cardoso Negrie, Michael Kawczynski, Kyunghyun Cho, Can Son Khoo, Joseph Y Lo, A Gregory Sorensen, and Hwejin Jung.
• Computational Oncology, Sage Bionetworks, Seattle, Washington.
• JAMA Netw Open. 2020 Mar 2; 3 (3): e200265.

ImportanceMammography screening currently relies on subjective human interpretation. Artificial intelligence (AI) advances could be used to increase mammography screening accuracy by reducing missed cancers and false positives.ObjectiveTo evaluate whether AI can overcome human mammography interpretation limitations with a rigorous, unbiased evaluation of machine learning algorithms.Design, Setting, And ParticipantsIn this diagnostic accuracy study conducted between September 2016 and November 2017, an international, crowdsourced challenge was hosted to foster AI algorithm development focused on interpreting screening mammography. More than 1100 participants comprising 126 teams from 44 countries participated. Analysis began November 18, 2016.Main Outcomes And MeasurementsAlgorithms used images alone (challenge 1) or combined images, previous examinations (if available), and clinical and demographic risk factor data (challenge 2) and output a score that translated to cancer yes/no within 12 months. Algorithm accuracy for breast cancer detection was evaluated using area under the curve and algorithm specificity compared with radiologists' specificity with radiologists' sensitivity set at 85.9% (United States) and 83.9% (Sweden). An ensemble method aggregating top-performing AI algorithms and radiologists' recall assessment was developed and evaluated.ResultsOverall, 144 231 screening mammograms from 85 580 US women (952 cancer positive ≤12 months from screening) were used for algorithm training and validation. A second independent validation cohort included 166 578 examinations from 68 008 Swedish women (780 cancer positive). The top-performing algorithm achieved an area under the curve of 0.858 (United States) and 0.903 (Sweden) and 66.2% (United States) and 81.2% (Sweden) specificity at the radiologists' sensitivity, lower than community-practice radiologists' specificity of 90.5% (United States) and 98.5% (Sweden). Combining top-performing algorithms and US radiologist assessments resulted in a higher area under the curve of 0.942 and achieved a significantly improved specificity (92.0%) at the same sensitivity.Conclusions And RelevanceWhile no single AI algorithm outperformed radiologists, an ensemble of AI algorithms combined with radiologist assessment in a single-reader screening environment improved overall accuracy. This study underscores the potential of using machine learning methods for enhancing mammography screening interpretation.

21 keywords...

hide…