A Multisite Study of a Breast Density Deep Learning Model for Full-Field Digital Mammography and Synthetic Mammography.
| Autor: | Matthews TP; Whiterabbit AI, Inc, 3930 Freedom Circle, Suite 101, Santa Clara, CA 95054 (T.P.M., S.S., B.M., J.S., M.P.S., S.P., A.L., R.M.H., N.G., D.S.); Mallinckrodt Institute of Radiology, Washington University School of Medicine, St Louis, Mo (D.M., J.G., S.M.M., R.L.W.); and Peninsula Diagnostic Imaging, San Mateo, Calif (S.C.M.)., Singh S; Whiterabbit AI, Inc, 3930 Freedom Circle, Suite 101, Santa Clara, CA 95054 (T.P.M., S.S., B.M., J.S., M.P.S., S.P., A.L., R.M.H., N.G., D.S.); Mallinckrodt Institute of Radiology, Washington University School of Medicine, St Louis, Mo (D.M., J.G., S.M.M., R.L.W.); and Peninsula Diagnostic Imaging, San Mateo, Calif (S.C.M.)., Mombourquette B; Whiterabbit AI, Inc, 3930 Freedom Circle, Suite 101, Santa Clara, CA 95054 (T.P.M., S.S., B.M., J.S., M.P.S., S.P., A.L., R.M.H., N.G., D.S.); Mallinckrodt Institute of Radiology, Washington University School of Medicine, St Louis, Mo (D.M., J.G., S.M.M., R.L.W.); and Peninsula Diagnostic Imaging, San Mateo, Calif (S.C.M.)., Su J; Whiterabbit AI, Inc, 3930 Freedom Circle, Suite 101, Santa Clara, CA 95054 (T.P.M., S.S., B.M., J.S., M.P.S., S.P., A.L., R.M.H., N.G., D.S.); Mallinckrodt Institute of Radiology, Washington University School of Medicine, St Louis, Mo (D.M., J.G., S.M.M., R.L.W.); and Peninsula Diagnostic Imaging, San Mateo, Calif (S.C.M.)., Shah MP; Whiterabbit AI, Inc, 3930 Freedom Circle, Suite 101, Santa Clara, CA 95054 (T.P.M., S.S., B.M., J.S., M.P.S., S.P., A.L., R.M.H., N.G., D.S.); Mallinckrodt Institute of Radiology, Washington University School of Medicine, St Louis, Mo (D.M., J.G., S.M.M., R.L.W.); and Peninsula Diagnostic Imaging, San Mateo, Calif (S.C.M.)., Pedemonte S; Whiterabbit AI, Inc, 3930 Freedom Circle, Suite 101, Santa Clara, CA 95054 (T.P.M., S.S., B.M., J.S., M.P.S., S.P., A.L., R.M.H., N.G., D.S.); Mallinckrodt Institute of Radiology, Washington University School of Medicine, St Louis, Mo (D.M., J.G., S.M.M., R.L.W.); and Peninsula Diagnostic Imaging, San Mateo, Calif (S.C.M.)., Long A; Whiterabbit AI, Inc, 3930 Freedom Circle, Suite 101, Santa Clara, CA 95054 (T.P.M., S.S., B.M., J.S., M.P.S., S.P., A.L., R.M.H., N.G., D.S.); Mallinckrodt Institute of Radiology, Washington University School of Medicine, St Louis, Mo (D.M., J.G., S.M.M., R.L.W.); and Peninsula Diagnostic Imaging, San Mateo, Calif (S.C.M.)., Maffit D; Whiterabbit AI, Inc, 3930 Freedom Circle, Suite 101, Santa Clara, CA 95054 (T.P.M., S.S., B.M., J.S., M.P.S., S.P., A.L., R.M.H., N.G., D.S.); Mallinckrodt Institute of Radiology, Washington University School of Medicine, St Louis, Mo (D.M., J.G., S.M.M., R.L.W.); and Peninsula Diagnostic Imaging, San Mateo, Calif (S.C.M.)., Gurney J; Whiterabbit AI, Inc, 3930 Freedom Circle, Suite 101, Santa Clara, CA 95054 (T.P.M., S.S., B.M., J.S., M.P.S., S.P., A.L., R.M.H., N.G., D.S.); Mallinckrodt Institute of Radiology, Washington University School of Medicine, St Louis, Mo (D.M., J.G., S.M.M., R.L.W.); and Peninsula Diagnostic Imaging, San Mateo, Calif (S.C.M.)., Hoil RM; Whiterabbit AI, Inc, 3930 Freedom Circle, Suite 101, Santa Clara, CA 95054 (T.P.M., S.S., B.M., J.S., M.P.S., S.P., A.L., R.M.H., N.G., D.S.); Mallinckrodt Institute of Radiology, Washington University School of Medicine, St Louis, Mo (D.M., J.G., S.M.M., R.L.W.); and Peninsula Diagnostic Imaging, San Mateo, Calif (S.C.M.)., Ghare N; Whiterabbit AI, Inc, 3930 Freedom Circle, Suite 101, Santa Clara, CA 95054 (T.P.M., S.S., B.M., J.S., M.P.S., S.P., A.L., R.M.H., N.G., D.S.); Mallinckrodt Institute of Radiology, Washington University School of Medicine, St Louis, Mo (D.M., J.G., S.M.M., R.L.W.); and Peninsula Diagnostic Imaging, San Mateo, Calif (S.C.M.)., Smith D; Whiterabbit AI, Inc, 3930 Freedom Circle, Suite 101, Santa Clara, CA 95054 (T.P.M., S.S., B.M., J.S., M.P.S., S.P., A.L., R.M.H., N.G., D.S.); Mallinckrodt Institute of Radiology, Washington University School of Medicine, St Louis, Mo (D.M., J.G., S.M.M., R.L.W.); and Peninsula Diagnostic Imaging, San Mateo, Calif (S.C.M.)., Moore SM; Whiterabbit AI, Inc, 3930 Freedom Circle, Suite 101, Santa Clara, CA 95054 (T.P.M., S.S., B.M., J.S., M.P.S., S.P., A.L., R.M.H., N.G., D.S.); Mallinckrodt Institute of Radiology, Washington University School of Medicine, St Louis, Mo (D.M., J.G., S.M.M., R.L.W.); and Peninsula Diagnostic Imaging, San Mateo, Calif (S.C.M.)., Marks SC; Whiterabbit AI, Inc, 3930 Freedom Circle, Suite 101, Santa Clara, CA 95054 (T.P.M., S.S., B.M., J.S., M.P.S., S.P., A.L., R.M.H., N.G., D.S.); Mallinckrodt Institute of Radiology, Washington University School of Medicine, St Louis, Mo (D.M., J.G., S.M.M., R.L.W.); and Peninsula Diagnostic Imaging, San Mateo, Calif (S.C.M.)., Wahl RL; Whiterabbit AI, Inc, 3930 Freedom Circle, Suite 101, Santa Clara, CA 95054 (T.P.M., S.S., B.M., J.S., M.P.S., S.P., A.L., R.M.H., N.G., D.S.); Mallinckrodt Institute of Radiology, Washington University School of Medicine, St Louis, Mo (D.M., J.G., S.M.M., R.L.W.); and Peninsula Diagnostic Imaging, San Mateo, Calif (S.C.M.). |
|---|---|
| Jazyk: | angličtina |
| Zdroj: | Radiology. Artificial intelligence [Radiol Artif Intell] 2020 Nov 04; Vol. 3 (1), pp. e200015. Date of Electronic Publication: 2020 Nov 04 (Print Publication: 2021). |
| DOI: | 10.1148/ryai.2020200015 |
| Abstrakt: | Purpose: To develop a Breast Imaging Reporting and Data System (BI-RADS) breast density deep learning (DL) model in a multisite setting for synthetic two-dimensional mammographic (SM) images derived from digital breast tomosynthesis examinations by using full-field digital mammographic (FFDM) images and limited SM data. Materials and Methods: A DL model was trained to predict BI-RADS breast density by using FFDM images acquired from 2008 to 2017 (site 1: 57 492 patients, 187 627 examinations, 750 752 images) for this retrospective study. The FFDM model was evaluated by using SM datasets from two institutions (site 1: 3842 patients, 3866 examinations, 14 472 images, acquired from 2016 to 2017; site 2: 7557 patients, 16 283 examinations, 63 973 images, 2015 to 2019). Each of the three datasets were then split into training, validation, and test. Adaptation methods were investigated to improve performance on the SM datasets, and the effect of dataset size on each adaptation method was considered. Statistical significance was assessed by using CIs, which were estimated by bootstrapping. Results: Without adaptation, the model demonstrated substantial agreement with the original reporting radiologists for all three datasets (site 1 FFDM: linearly weighted Cohen κ [κ Conclusion: A BI-RADS breast density DL model demonstrated strong performance on FFDM and SM images from two institutions without training on SM images and improved by using few SM images. Supplemental material is available for this article. Published under a CC BY 4.0 license. Competing Interests: Disclosures of Conflicts of Interest: T.P.M. Activities related to the present article: disclosed no relevant relationships. Activities not related to the present article: disclosed money paid to author for employment from Whiterabbit; disclosed stock/stock options in Whiterabbit; disclosed money to author’s institution for patents related to Whiterabbit. Other relationships: disclosed no relevant relationships. S.S. Activities related to the present article: disclosed no relevant relationships. Activities not related to the present article: disclosed payment to author for employment from Whiterabbit. Other relationships: disclosed no relevant relationships. B.M. Activities related to the present article: disclosed no relevant relationships. Activities not related to the present article: disclosed money paid to author for employment from Whiterabbit; disclosed stock/stock options in Whiterabbit; disclosed money to author’s institution for patents related to Whiterabbit. Other relationships: disclosed no relevant relationships. J.S. Activities related to the present article: disclosed no relevant relationships. Activities not related to the present article: disclosed money paid to author for board membership from Whiterabbit; disclosed employment from Whiterabbit; disclosed stock/stock options from Whiterabbit. Other relationships: disclosed no relevant relationships. M.P.S. disclosed no relevant relationships. S.P. Activities related to the present article: disclosed no relevant relationships. Activities not related to the present article: disclosed money paid to author for employment from Whiterabbit. Other relationships: disclosed no relevant relationships. A.L. disclosed no relevant relationships. D.M. disclosed no relevant relationships. J.G. disclosed no relevant relationships. R.M.H. Activities related to the present article: disclosed no relevant relationships. Activities not related to the present article: disclosed payment to author for employment from Whiterabbit; disclosed stock/stock options from Whiterabbit. Other relationships: disclosed no relevant relationships. N.G. Activities related to the present article: disclosed no relevant relationships. Activities not related to the present article: disclosed payment to author for employment from Whiterabbit; disclosed stock/stock options from Whiterabbit. Other relationships: disclosed no relevant relationships. D.S. disclosed no relevant relationships. S.M.M. Activities related to the present article: disclosed money to author’s institution from Whiterabbit as a grant. Activities not related to the present article: disclosed no relevant relationships. Other relationships: disclosed no relevant relationships. S.C.M. disclosed no relevant relationships. R.L.W. Activities related to the present article: disclosed grant to author’s institution from Whiterabbit. Activities not related to the present article: disclosed possibility of royalties if the FDA approves a related product; disclosed stock options for Washington University from Whiterabbit. Other relationships: disclosed no relevant relationships. (2020 by the Radiological Society of North America, Inc.) |
| Databáze: | MEDLINE |
| Externí odkaz: |
|