TY - JOUR
T1 - A Deep Learning Decision Support Tool to Improve Risk Stratification and Reduce Unnecessary Biopsies in BI-RADS 4 Mammograms
AU - Ezeana, Chika F.
AU - He, Tiancheng
AU - Patel, Tejal A.
AU - Kaklamani, Virginia
AU - Elmi, Maryam
AU - Brigmon, Erika
AU - Otto, Pamela M.
AU - Kist, Kenneth A.
AU - Speck, Heather
AU - Wang, Lin
AU - Ensor, Joe
AU - Shih, Ya Chen T.
AU - Kim, Bumyang
AU - Pan, I. Wen
AU - Cohen, Adam L.
AU - Kelley, Kristen
AU - Spak, David
AU - Yang, Wei T.
AU - Chang, Jenny C.
AU - Wong, Stephen T.C.
N1 - Funding Information:
P.M.O. No relevant relationships. K.A.K. No relevant relationships. H.S. No relevant relationships. L.W. No relevant relationships. J.E. No relevant relationships. Y.C.T.S. Research grants from the National Cancer Institute on topics unrelated to this manuscript. B.K. No relevant relationships. I.W.P. No relevant relationships. A.L.C. Grants or contracts from Novartis and Acrotech. K.K. No relevant relationships. D.S. No relevant relationships. W.T.Y. Research grant with Clarity sponsorship for “Evaluation of individual level breast cancer risk prediction in a cohort of patients with a personal history of breast cancer using a novel software as a medical device;” royalties or licenses from Elsevier. J.C.C. Support from the Breast Cancer Research Foundation (BCRF); philanthropic support from M. Neal and R. Neal; National Cancer Institute/National Institutes of Health grant number U01 CA268813; grants or contracts from the Cancer Prevention and Research Institute of Texas (grant no. RP220650); payment or honoraria from Duke and NUS Singapore (August 7, 2022); sole inventor on patent application number 10420838 entitled “Methods for treating cancer using iNOS-inhibitory compositions” held by Houston Methodist Hospital; participation on Merck Triple Negative Breast Cancer Advisory Board (December 13, 2022), Lilly Loxo Advisory Board (December 8, 2022), and BCRF Annual Meeting (October 26, 2022). S.T.C.W. No relevant relationships.
Funding Information:
Supported by the Ting Tsung & Wei Fong Chao Family Foundation, the John S. Dunn Research Foundation, the Breast Cancer Research Foundation, and The National Institutes of Health/National Cancer Institute grant no. 1R01CA251710.
Publisher Copyright:
© 2023, Radiological Society of North America Inc.. All rights reserved.
PY - 2023/11
Y1 - 2023/11
N2 - PURPOSE: To evaluate the performance of a biopsy decision support algorithmic model, the intelligent-augmented breast cancer risk calculator (iBRISK), on a multicenter patient dataset.MATERIALS AND METHODS: iBRISK was previously developed by applying deep learning to clinical risk factors and mammographic descriptors from 9700 patient records at the primary institution and validated using another 1078 patients. All patients were seen from March 2006 to December 2016. In this multicenter study, iBRISK was further assessed on an independent, retrospective dataset (January 2015-June 2019) from three major health care institutions in Texas, with Breast Imaging Reporting and Data System (BI-RADS) category 4 lesions. Data were dichotomized and trichotomized to measure precision in risk stratification and probability of malignancy (POM) estimation. iBRISK score was also evaluated as a continuous predictor of malignancy, and cost savings analysis was performed.RESULTS: The iBRISK model's accuracy was 89.5%, area under the receiver operating characteristic curve (AUC) was 0.93 (95% CI: 0.92, 0.95), sensitivity was 100%, and specificity was 81%. A total of 4209 women (median age, 56 years [IQR, 45-65 years]) were included in the multicenter dataset. Only two of 1228 patients (0.16%) in the "low" POM group had malignant lesions, while in the "high" POM group, the malignancy rate was 85.9%. iBRISK score as a continuous predictor of malignancy yielded an AUC of 0.97 (95% CI: 0.97, 0.98). Estimated potential cost savings were more than $420 million.CONCLUSION: iBRISK demonstrated high sensitivity in the malignancy prediction of BI-RADS 4 lesions. iBRISK may safely obviate biopsies in up to 50% of patients in low or moderate POM groups and reduce biopsy-associated costs.
Keywords: Mammography, Breast, Oncology, Biopsy/Needle Aspiration, Radiomics, Precision Mammography, AI-augmented Biopsy Decision Support Tool, Breast Cancer Risk Calculator, BI-RADS 4 Mammography Risk Stratification, Overbiopsy Reduction, Probability of Malignancy (POM) Assessment, Biopsy-based Positive Predictive Value (PPV3)
Supplemental material is available for this article. Published under a CC BY 4.0 license.See also the commentary by McDonald and Conant in this issue.
AB - PURPOSE: To evaluate the performance of a biopsy decision support algorithmic model, the intelligent-augmented breast cancer risk calculator (iBRISK), on a multicenter patient dataset.MATERIALS AND METHODS: iBRISK was previously developed by applying deep learning to clinical risk factors and mammographic descriptors from 9700 patient records at the primary institution and validated using another 1078 patients. All patients were seen from March 2006 to December 2016. In this multicenter study, iBRISK was further assessed on an independent, retrospective dataset (January 2015-June 2019) from three major health care institutions in Texas, with Breast Imaging Reporting and Data System (BI-RADS) category 4 lesions. Data were dichotomized and trichotomized to measure precision in risk stratification and probability of malignancy (POM) estimation. iBRISK score was also evaluated as a continuous predictor of malignancy, and cost savings analysis was performed.RESULTS: The iBRISK model's accuracy was 89.5%, area under the receiver operating characteristic curve (AUC) was 0.93 (95% CI: 0.92, 0.95), sensitivity was 100%, and specificity was 81%. A total of 4209 women (median age, 56 years [IQR, 45-65 years]) were included in the multicenter dataset. Only two of 1228 patients (0.16%) in the "low" POM group had malignant lesions, while in the "high" POM group, the malignancy rate was 85.9%. iBRISK score as a continuous predictor of malignancy yielded an AUC of 0.97 (95% CI: 0.97, 0.98). Estimated potential cost savings were more than $420 million.CONCLUSION: iBRISK demonstrated high sensitivity in the malignancy prediction of BI-RADS 4 lesions. iBRISK may safely obviate biopsies in up to 50% of patients in low or moderate POM groups and reduce biopsy-associated costs.
Keywords: Mammography, Breast, Oncology, Biopsy/Needle Aspiration, Radiomics, Precision Mammography, AI-augmented Biopsy Decision Support Tool, Breast Cancer Risk Calculator, BI-RADS 4 Mammography Risk Stratification, Overbiopsy Reduction, Probability of Malignancy (POM) Assessment, Biopsy-based Positive Predictive Value (PPV3)
Supplemental material is available for this article. Published under a CC BY 4.0 license.See also the commentary by McDonald and Conant in this issue.
KW - AI-augmented Biopsy Decision Support Tool
KW - BI-RADS 4 Mammography Risk Stratification
KW - Biopsy-based Positive Predictive Value (PPV3)
KW - Biopsy/Needle Aspiration
KW - Breast
KW - Breast Cancer Risk Calculator
KW - Mammography
KW - Oncology
KW - Overbiopsy Reduction
KW - Precision Mammography
KW - Probability of Malignancy (POM) Assessment
KW - Radiomics
UR - http://www.scopus.com/inward/record.url?scp=85177745324&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85177745324&partnerID=8YFLogxK
U2 - 10.1148/ryai.220259
DO - 10.1148/ryai.220259
M3 - Article
C2 - 38074778
AN - SCOPUS:85177745324
SN - 2638-6100
VL - 5
SP - e220259
JO - Radiology: Artificial Intelligence
JF - Radiology: Artificial Intelligence
IS - 6
M1 - e220259
ER -