Performance of #AI comparable to humans when reading test sets from a quality assurance scheme #PERFORMS used to assess readers in the UK breast screening programme. May provide a model for the testing of #AI deployed in clinical practice @radiology_rsna https://t.co/9WddLnpb7i
Performance of a Breast Cancer Detection AI Algorithm Using the Personal Performance in Mammographic Screening Scheme.
Mentions
Extended reach
Total likes
Video views
Mentions per month
Sep 2023 – Apr 2024Where it's discussed
-
News93 articles
-
X / Twitter13 posts
-
Blogs4 articles
-
YouTube2 videos
-
Facebook2 posts
AI synopsis
What people are saying
Participants show strong agreement with the article’s conclusion that AI can match human performance, with comments like ‘AI comparable to humans’ and ‘may provide a model for testing AI in practice.’ The tone suggests high interest and moderate engagement, focusing on the implications for future screening, though some concerns are raised about patient communication and ethical considerations.
Most discussions acknowledge the study's findings that AI performs comparably to human radiologists, showing general agreement on its validity.
Posts display high interest by emphasizing the importance of the research and potential applications in clinical practice.
Several comments delve into specific performance metrics, such as specificity and AI's potential, indicating active engagement.
The discussions highlight the significance of the study for future breast cancer screening, emphasizing its potential to influence clinical protocols.
Abstract in brief
- The study compares how well a commercial AI algorithm and human readers perform in detecting breast cancer using the PERFORMS mammographic screening scheme, finding no significant difference in overall accuracy (AUC) between AI and humans.
- When applying standard thresholds, the AI had higher specificity (fewer false positives) but similar sensitivity to humans; however, the study couldn't confirm they are statistically equivalent due to limited test size.
- Overall, the AI demonstrated comparable diagnostic performance to human readers in this context, suggesting potential for AI to assist or augment breast cancer screening.]
Top posts
Most liked on X and BlueskyNormally used to assess the performance of human readers, the Personal Performance in Mammographic Screening (PERFORMS) scheme was used to evaluate the performance of an AI detection algorithm in this study by Chen et al. https://t.co/SVH2ORBZou https://t.co/NEitXLJItU
Performance of a Breast Cancer Detection AI Algorithm Using the Personal Performance in Mammographic Screening Scheme | Radiology Amazing hugely important work from @DrYanChen7 and colleagues
Performance of a Breast Cancer Detection AI Algorithm Using the Personal Performance in Mammographic Screening Scheme | Radiology https://t.co/3EZ1jhrgoa https://t.co/y8MMKS5fZl
Performance of a Breast Cancer Detection AI Algorithm Using the Personal Performance in Mammographic Screening Scheme | Radiology https://t.co/3EZ1jhrgoa https://t.co/NacJXLeZkz
RT @drjonjames: Performance of #AI comparable to humans when reading test sets from a quality assurance scheme #PERFORMS used to assess rea…
All mentions
AI Algorithm Shows Promise in Breast Cancer Detection Efficiency
Researchers trained an AI to analyze mammograms, achieving accuracy comparable to human readers. The AI shows potential to assist in breast cancer screening when human double reading isn't available.
AI Performance in Breast Cancer Detection in Mammography
Dr Reni Butler interviews Dr Yan Chen about the diagnostic performance of a commercial AI algorithm in breast cancer detection using mammographic screening. The study compares AI and human reader accuracy, showing AI…
A mi me sigue preocupando no el uso de la AI y su veracidad en la detección del cáncer de mama sino la comunicación de una información tan sensible al paciente sin un circuito bien establecido con humanidad y cercanía al paciente. #radiologiademama https://t.co/0oNUIuy7iy
Performance of a Breast Cancer Detection AI Algorithm Using the Personal Performance in Mammographic Screening Scheme | Radiology https://t.co/3EZ1jhrgoa https://t.co/NacJXLeZkz
Performance of a Breast Cancer Detection AI Algorithm Using the Personal Performance in Mammographic Screening Scheme | Radiology https://t.co/3EZ1jhrgoa https://t.co/y8MMKS5fZl
Normally used to assess the performance of human readers, the Personal Performance in Mammographic Screening (PERFORMS) scheme was used to evaluate the performance of an AI detection algorithm in this study by Chen et al. https://t.co/SVH2ORBZou https://t.co/NEitXLJItU
RT @drjonjames: Performance of #AI comparable to humans when reading test sets from a quality assurance scheme #PERFORMS used to assess rea…
RT @lungradiologist: Performance of a Breast Cancer Detection AI Algorithm Using the Personal Performance in Mammographic Screening Scheme…
RT @lungradiologist: Performance of a Breast Cancer Detection AI Algorithm Using the Personal Performance in Mammographic Screening Scheme…
RT @drjonjames: Performance of #AI comparable to humans when reading test sets from a quality assurance scheme #PERFORMS used to assess rea…
Performance of a Breast Cancer Detection AI Algorithm Using the Personal Performance in Mammographic Screening Scheme | Radiology Amazing hugely important work from @DrYanChen7 and colleagues
RT @drjonjames: Performance of #AI comparable to humans when reading test sets from a quality assurance scheme #PERFORMS used to assess rea…
RT @drjonjames: Performance of #AI comparable to humans when reading test sets from a quality assurance scheme #PERFORMS used to assess rea…
Performance of #AI comparable to humans when reading test sets from a quality assurance scheme #PERFORMS used to assess readers in the UK breast screening programme. May provide a model for the testing of #AI deployed in clinical practice @radiology_rsna https://t.co/9WddLnpb7i
Performance of a Breast Cancer Detection AI Algorithm Meets/Exceeds Radiologist Reads - and, specificity was better with AI | Radiology https://t.co/aXP5geVS3Y