Current Innovation Project by

Mia, AI saving lives through early breast detection

ccorales1 — 2024-09-13 15:26:39 UTC

Source Details

ccorales1 — 2024-09-13 15:26:43 UTC

Title:

Prospective implementation of AI-assisted screen reading to improve early detection of breast cancer.

Authors & Affiliations:

Kheiron Medical Technologies, London, UK
Annie Y. Ng, Cary J. G. Oberije, Edit Karpati, Georgia Fox, Ben Glocker & Peter D. Kecskemethy
MaMMa Egészségügyi Zrt., Budapest, Hungary
Éva Ambrózay, Endre Szabó & Orsolya Serfőző
Department of Computing, Imperial College London, London, UK
Ben Glocker
University of California, Davis, Davis, CA, USA
Elizabeth A. Morris
Duna Medical Center, Budapest, Hungary
Gábor Forrai

Credible Source:

Nature Medicine Journal

Date Published

ccorales1 — 2024-09-13 15:26:46 UTC

November 16, 2023

Summary of Mia

ccorales1 — 2024-09-13 15:26:49 UTC

According to the Nature Medicine Journal, the current method of cancer screening involves 2 radiologists who read the breast cancer screening images and determine if the woman should be "recalled". A 3rd radiologist called an "arbitrator" makes the decision if there is disagreement among the 2 radiologists. The research showed implementation of the AI-assisted additional-reader workflow resulted in 24 more cancer cases detected. Data suggests that using AI as an additional reader can improve the early detection of breast cancer with relevant prognostic features, with minimal to no unnecessary recalls. Although it requires additional reads, the higher positive predictive value (PPV) suggests that it can increase screening effectiveness.

Link/Video or Photos

ccorales1 — 2024-09-13 15:26:52 UTC

https://doi.org/10.1038/s41591-023-02625-9

Pros

ccorales1 — 2024-09-13 15:26:56 UTC

Large-scale retrospective studies of the same AI system used in this assessment have demonstrated that AI as an independent second reader can offer:

-up to 45% workload savings

-offsetting the 3–11% additional arbitration reads (1–6% additional overall reading workload) for the AI-assisted additional-reader workflow while providing the benefit of increased cancer detections

-minimal to no unnecessary recalls

Opinion and Rationale

ccorales1 — 2024-09-13 15:26:58 UTC

In my opinion, AI assisted mammography reading is a helpful tool that can help identify early stages of breast cancer. Computers have the ability to identify microscope changes that the human eye cannot. There is more research needed to ensure accuracy, but the current data is promising.

Using Strengths

ccorales1 — 2024-09-13 15:27:01 UTC

My 5 strengths and what I will bring to the team:

Relator- caring, trusting and honest

Learner- catches on quickly, wants to consistently improve

Achiever- work hard, strong work ethic

Analytical- thorough, thinks things through

Responsibility- committed, accountable, independent

Teams/ Rationale

Arranger- (Abraham) flexibility, organization, aligns and realigns tasks to find most productive configuration possible. This will offset "achiever" tendencies, being too concentrated on work.

Includer- (Isabelle) accepting/inclusive of others. Will help offset "relator" traits of playing favorites and having an inner circle.

Focus- (Fernando) prioritize then act, goal setter and goal getter, disciplined. Will help offset "learner" know it all, lacks focus on results

Discipline- (Daniel) highly productive and accurate because of ability to structure. Will offset "responsibility" micromanager, obsessive, takes on too much.

Context- (Carlos) Can leverage knowledge of the past. Help offset "analytical" never satisfied with answer and asks too many questions.

References

ccorales1 — 2024-09-13 15:27:04 UTC

Ng, A.Y., Oberije, C.J.G., Ambrózay, É. et al. (2023). Prospective implementation of AI-assisted screen reading to improve early detection of breast cancer. Nat Med 29, 3044–3049. https://doi.org/10.1038/s41591-023-02625-9

Cons

ccorales1 — 2024-09-14 01:24:04 UTC

Evaluations of real-world performance of AI integrated into live clinical workflows have been limited to date. However, the current data available seems to be very promising.