Overview

Schema Version

https://atlas.rsna.org/schemas/2025-11/dataset.json

Name

ChiMEC: Chicago Multiethnic Epidemiologic Breast Cancer Cohort

Link

https://pubs.rsna.org/doi/10.1148/ryai.220299

Indexing

Keywords: Mirai, mammography, breast cancer risk prediction, deep learning, BI-RADS, breast density, African American, BRCA mutation, case-control, external validation

Content: BR, OI, RS

RadLex: RID35976, RID34265, RID10357, RID5675, RID34240

SNOMED: 254837009, 109889007

Author(s)

Olasubomi J. Omoleye

Anna E. Woodard

Frederick M. Howard

Fangyuan Zhao

Toshio F. Yoshimatsu

Yonglan Zheng

Alexander T. Pearson

Maksim Levental

Benjamin S. Aribisala

Kirti Kulkarni

Gregory S. Karczmar

Olufunmilayo I. Olopade

Hiroyuki Abe

Dezheng Huo

Organization(s)

The University of Chicago

Lagos State University

Contact

Corresponding author: Dezheng Huo (email as listed: ude.ogacihcu.dsb@ouhd)

Funding

University of Chicago Comprehensive Cancer Care Center Spotlight grant (6-9398-9660); Susan G. Komen (SAC210203, TREND21675016); NIH/NCI (P20CA233307); Breast Cancer Research Foundation (BCRF-21-071).

Ethical review

Health Insurance Portability and Accountability Act compliant; approved by the University of Chicago Institutional Review Board.

Comments

Retrospective case-control screening mammography dataset from a high-risk, racially diverse cohort used to externally evaluate the Mirai deep learning model for 1–5 year breast cancer risk prediction.

Date

Published: 2023-07-26

References

[1] Omoleye OJ, Woodard AE, Howard FM, Zhao F, Yoshimatsu TF, Zheng Y, Pearson AT, Levental M, Aribisala BS, Kulkarni K, Karczmar GS, Olopade OI, Abe H, Huo D. "External Evaluation of a Mammography-based Deep Learning Model for Predicting Breast Cancer in an Ethnically Diverse Population". Radiology: Artificial Intelligence. 2023-11-01. doi:10.1148/ryai.220299. PMID: 38074785. PMCID: PMC10698602.

Dataset

Motivation

To externally evaluate the Mirai mammography-based DL model in a high-risk, racially diverse cohort and compare performance with BI-RADS and breast density.

Sampling

Consecutive women evaluated in the Cancer Risk Clinic or Breast Center since 1992; screening mammograms and clinicopathologic data from 2006–2020 included; retrospective case-control selection.

Partitioning scheme

Case-control structure with time-to-event labeling at 1–5 year horizons; some analyses excluded cases with time to cancer <6 months.

Missing information

Imaging resolution and exact image file counts not reported; genetic testing available for a subset of patients only.

Relationships between instances

Multiple examinations per patient; each examination comprises four standard mammographic views (bilateral CC and MLO).

External data

None reported; all images and clinical data sourced from University of Chicago Medicine records (2006–2020).

Confidentiality

De-identified patient data were used; HIPAA compliant and IRB approved.

Re-identification

De-identified prior to analysis; images with burned-in annotations were excluded.

Sensitive data

Clinical outcomes (pathology-confirmed cancer), BI-RADS assessments, breast density; subset includes genetic testing (e.g., BRCA) information.