Purpose
The importance of breast screening programs is evident. However, mammography does have its limitations and a significant number of cancers are not detected in the regular screening round, but are diagnosed as interval cancers (IC) during follow-up [1]. Of these, approximately 13.6%–35% are retrospectively classified as having been missed (false-negative) in screening [1].
Current AI systems, developed for detection of breast cancer, have reported stand-alone performance comparable to humans. Several retrospective studies demonstrated the ability of AI to detect some of these false negative cancers[2,3,4],...
Methods and materials
A commercially available AI system was implemented at a large breast cancer screening centre in Hungary to analyse images of women who participated in the national breast screening program or who came to the clinic for opportunistic screening within a six month period in 2021. The AI system was employed as an extra reader (XR) in addition to standard human double reading (HDR) and itsrecommendation was based on pre-defined operating points [5].
The XR workflow involved flagging cases the AI suggested to recall which HDR...
Results
Over a 6 month period, 3746 patients were screened.
Of the cases that were not recalled by HDR, the AI flagged 396 cases for extra review by a human reader (positive discordance rate of 10.6%).
HDR had a recall rate (RR) of 6.7%, an arbitration rate of 3.0%, and a cancer detection rate (CDR) of 12.5/1000 ((47 cancer cases out of 3746 screens).
Extra human arbitration, based on XR, resulted in recalling 6 patients, all of whom were diagnosed with breast cancer.[Fig 2]
This equated...
Conclusion
This prospective real-world data provides evidence for the potential of AI to detect cancers missed by human readers and increase cancer detection rates without recalling extra false positives.
Combining the XR workflow with workflows focused on workload savings will mitigate the increased arbitration rate and optimise clinical and operational benefits.
The results provide prospective real-world evidence showing the benefit of using AI as an extra reader in breast cancer screening.
References
Houssami N, Hunter K (2017) The epidemiology, radiology and biological characteristics of interval breast cancers in population mammography screening. NPJ Breast Cancer 3:12
Park GE, Kang BJ, Kim SH, Lee J (2022) Retrospective Review of Missed Cancer Detection and Its Mammography Findings with Artificial-Intelligence-Based, Computer-Aided Diagnosis. Diagnostics (Basel) 12.: https://doi.org/10.3390/diagnostics12020387
Byng D, Strauch B, Gnas L, et al (2022) AI-based prevention of interval cancers in a national mammography screening program. Eur J Radiol 152:110321
Chorev M, Shoshan Y, Akselrod-Ballin A, et al (2020) The Case...