AI Can Enhance Mammogram Screening, in Turn Transforming Breast Cancer Detection

News
Article

When compared to a double reading of a mammogram screening by two radiologists, a double reading by one radiologist plus AI demonstrated a 4% higher non-inferior cancer detection rate, highlighting the potential of AI in enhancing cancer detection in mammography screening.

Substituting one radiologist with artificial intelligence for an independent reading of screening mammograms resulted in a 4% higher cancer detection rate that was non-inferior when compared to the standard-of-care practice of double reading by two radiologists.

Mammography screening has been a fundamental tool of early detection of breast cancer for almost four decades. However, challenges have come up between radiologists in diagnostic accuracy, which leads to unnecessary recalls and missed cancer. There is also a global shortage of breast radiologists which increases demands for precision diagnostics from both providers and patients.

According to a study published in THE LANCET Digital Health, artificial intelligence (AI) has the potential to address these challenges.

The aim of this study was to evaluate whether AI could match or exceed the cancer detection rate of double reading by two radiologists, the current standard of care.

Researchers conducted the study, known as ScreenTrustCAD, at the Capio Sankt Göran Hospital in Stockholm, Sweden. A population-based, paired-reader, non-inferiority study included 55,581 women aged 40–74 participating in population-based breast cancer screening.

From April 1, 2021, to June 9, 2022, 269 women were diagnosed with screen-detected breast cancer based on an initial positive read. When compared to double reading by two radiologists, double reading by one radiologist plus AI demonstrated a 4% higher non-inferior cancer detection rate, highlighting the potential of AI in enhancing cancer detection in mammography screening.

Single reading by AI and triple reading by two radiologists plus AI also proved to be non-inferior to double reading by two radiologists, highlighting the versatility and reliability of AI in various screening scenarios.

Additionally, single reading by AI demonstrated a similar cancer detection rate with a lower recall rate compared to double reading by two radiologists. However, triple reading by two radiologists plus AI, while highly effective in cancer detection, came with increased costs and raised concerns about participant anxiety due to more recalls.

What sets this study apart was its strengths, which include the integration of AI into the existing screening workflow and the experience of the radiologists involved. The study's design provided flexibility in assessing different reader strategies and it accurately represented real-world conditions. However, authors note the availability of AI results during consensus discussions may have influenced radiologists' decisions, potentially leading to an underrepresentation of AI's detection capabilities.

Outside of its strengths, there are limitations. The single-arm paired design restricts future comparisons of interval cancer rates; however, researchers plan to address the issue in a 23-month follow-up study.

In addition, results may be limited as they were obtained within a specific workflow using particular equipment and an AI system. Lastly, with breast implants were excluded due to the AI system's lack of validation in this group.

The key takeaway from this study is that AI has the potential to evolve breast cancer screening. The strategy of double reading by one radiologist plus AI not only increased the cancer detection rate, but also improved efficiency. AI's sensitivity in detecting cancer, combined with the radiologists' ability to dismiss false positives during consensus discussions, played a pivotal role in achieving these results.

Recent Videos
Video 8 - "Risk and Value-Based Contracting for Prescription Drug Therapeutics and Manufacturer’s Support in Utilization"
Video 7 - "Prescribing Prescription Drug Therapeutics: Factors for Providers and Payers to Consider"
Video 9 -"Overcoming Implementation Barriers for Digital Therapeutics Adoption"
Video 6 - "Navigating Insurance Coverage for Prescription Digital Therapeutics"
Video 5 - "FDA Approval Pathway for Prescription Digital Therapeutics"
Video 8 - "Gaps in Evidence Generation for Digital Therapeutics"
Video 7 - "Adoption Lessons For Payers"
Video 4 - "Assessing the Cost-Effectiveness of Prescription Digital Therapeutics "
Video 3 - "Harnessing Prescription Drug Therapeutics as Monotherapy and Adjunct Therapy"
Video 6 - "Developing Reimbursement Models for Digital Therapeutics"
Related Content
© 2024 MJH Life Sciences

All rights reserved.