Related Experiment Video
Updated: Jul 8, 2025
13:44
Detection of Architectural Distortion in Prior Mammograms via Analysis of Oriented Patterns
Published on: August 30, 2013
42.9K
Impact of Artificial Intelligence-driven Quality Improvement Software on Mammography Technical Repeat and Recall
Peter R Eby1, Linda M Martis1, Jeremy T Paluch1
1From the Department of Radiology, Virginia Mason Franciscan Health, 1100 9th Ave, Seattle, WA 98101 (P.R.E., J.T.P., J.J.P.); and Volpara Health Technologies, Wellington, New Zealand (L.M.M., A.H.L.C.).
Radiology. Artificial Intelligence
|December 11, 2023
Summary
Artificial intelligence (AI) software significantly improved mammography image quality and reduced technical repeats and recalls. This technology offers a valuable tool for enhancing patient care and diagnostic accuracy in breast imaging.
Area of Science:
- Radiology and Medical Imaging
- Artificial Intelligence in Healthcare
- Quality Assurance in Medical Diagnostics
Background:
- Poor breast positioning is a primary cause of decreased mammography sensitivity and inadequate image quality (IQ).
- Inadequate IQ can lead to repeat mammography views or patient recalls, increasing healthcare costs and patient anxiety.
- Objective evaluation of breast positioning and compression metrics is crucial for improving mammography outcomes.
Purpose of the Study:
- To assess the impact of implementing artificial intelligence (AI) software on mammography image quality (IQ).
- To determine if AI software implementation reduces rates of technical repeats and recalls (TR) in mammography.
- To evaluate the effectiveness of AI in objectively assessing breast positioning and compression metrics.
Main Methods:
- Retrospective evaluation of technical repeats and recalls (TR) for 40 technologists (198,054 images) from April 2019 to March 2022.
- Utilized AI software (Volpara Health Technologies) to analyze image quality metrics for 42 technologists (211,821 images).
- Compared baseline (April 2019-March 2020) and current (April 2021-March 2022) periods using statistical tests (Kolmogorov-Smirnov, χ², paired t tests).
Main Results:
- Technical repeats and recalls (TR) significantly reduced from 0.77% at baseline to 0.17% in the current period (P < .001).
- Overall mean image quality score improved by 6% (P = .001).
- AI software provided objective metrics for breast positioning and compression, correlating with improved IQ.
Conclusions:
- AI software implementation demonstrably improves mammography image quality.
- AI tools effectively reduce technical repeats and recalls, enhancing patient experience and resource utilization.
- AI represents a promising technology for objective quality assurance in mammography screening programs.
