Related Experiment Video
Updated: Jul 6, 2025
Competing-Risk Nomogram for Predicting Cancer-Specific Survival in Multiple Primary Colorectal Cancer Patients after Surgery
Published on: September 27, 2024
Breast Cancer Method of Detection: 5-Year Outcomes Across a Multisite Health Care Enterprise
Alan Zhu1, Bhavika K Patel2, Aditya Khurana3
1Mayo Clinic Alix School of Medicine, Phoenix, Arizona.
Purpose:
To investigate the feasibility and accuracy of radiologists categorizing the method of detection (MOD) when performing image-guided breast biopsies.
Methods:
This retrospective, observational study was conducted across a health care enterprise that provides breast imaging services at 18 imaging sites across four US states. Radiologists used standardized templates to categorize the MOD, defined as the first test, sign, or symptom that triggered the subsequent workup and recommendation for biopsy. All image-guided breast biopsies since the implementation of the MOD-inclusive standardized template-from October 31, 2017 to July 6, 2023-were extracted. A random sample of biopsy reports was manually reviewed to evaluate the accuracy of MOD categorization.
Results:
A total of 29,999 biopsies were analyzed. MOD was reported in 29,423 biopsies (98.1%) at a sustained rate that improved over time. The 10 MOD categories in this study included the following: 15,184 mammograms (51.6%); 4,561 MRIs (15.5%); 3,473 ultrasounds (11.8%); 2,382 self-examinations (8.1%); 2,073 tomosynthesis studies (7.0%); 432 clinical examinations (1.5%); 421 molecular breast imaging studies (1.4%); 357 other studies (1.2%); 338 contrast-enhanced digital mammograms (1.1%); and 202 PET studies (0.7%). Original assignments of the MOD agreed with author assignments in 87% of manually reviewed biopsies (n = 100, 95% confidence interval: [80.4%, 93.6%]).
Conclusions:
This study demonstrates that US radiologists can consistently and accurately categorize the MOD over an extended time across a health care enterprise.
