@CotsCardiology
Pinned post
The software used for this project was Coreline (@CorelineSoft), which also provides AI chest interpretation and adds AI chest findings to calcium scoring in one applications. Recommended by @CotsCardiology. #AI #Cardiology #CalciumScore #ChestCT #MedicalImaging #artificiallntelligenceView quoted post on X
Regarding the VA AI calcium score study (ai.nejm.org/doi/full/10.10…):
Our team at @CVRF_SoCal led one of the first FDA-cleared validation of AI coronary calcium scoring but not from chest CTs
(pubmed.ncbi.nlm.nih.gov/37139562/).
In 100 non-contrast CACs, AI matched 3 expert readers with R=0.996.
Using CAC-DRS calcification system, AI showed 86% agreement; 14% were reclassified—mostly due to underestimating RCA calcium or overestimating RV density for CAC 0–1 which can be clinically relevant to treatment.
Overall accuracy of our study: 89%, aligning with the VA study of 84% using chest CT - encouraging for wide-spread adoption.
@khurramn1 @CotsCardiology #calciumscore @Heart_SCCT @DrMarthaGulati @EricTopol @RonBlankstein @CVMGBevHills @CAIALOfficial @BudoffMd #artificalintelligence @elonmusk @WHO @BhavikPatelMD @ACCinTouchView quoted post on X
AI-CAC enables coronary calcium scoring from routine chest CTs
20K scans, 98 VA sites, 89% accuracy
From my perspective, time to connect AI outputs to action, to truly maximize implementation-->prevention
ai.nejm.org/doi/full/10.10… #DigitalHealth @HoustonMethodist @RiceUniversityView quoted post on X
