1120_Kumar.pdf

1、ARTIFICIAL INTELLIGENCE AND MEDICAL IMAGING:CURRENT TECHNOLOGY AND FUTURE APPLICATIONSShivani Kumar,MD RPVIAssistant Professor Vascular Surgery Program Director,Vascular FellowshipTufts University School of MedicineDISCLOSURESSpeaking/Consulting:Gore Medical,Cook Medical,Shockwave Medical,Medtronic

2、PI Clinical Trials:Shockwave Medical,EndologixAI IS REVOLUTIONIZING HEALTHCARE Diagnostics,Prognosis,Clinical workflow Key Drivers-Rising rates of chronic disease-Radiologist/PCP burnout and shortage-Demand for increased efficiency and accuracy Current Technology-AI automates tasks like image enhanc

3、ement,anatomical segmentation,and real-time case triageFUTURE APPLICATION-Radiomics&Radiogenomics:Extracting quantitative data to predict disease progression and treatment response.-Multimodal AI:Fusing imaging with genomics and EHRs for a holistic view.-Generative AI:Creating synthetic data to addr

4、ess data scarcity and biasCHALLENGES Algorithmic bias-Representation of individual vs population potential to exacerbate health disparities Data privacy-HIPPA Regulatory oversight-FDA The evolving human-AI relationship Accountability&Liability All critical considerationsWHAT IS ARTIFICIAL INTELLIGEN

5、CE?Broad term to enable a machine or system to perform,reason,act,adapt,interpret like a human Machine Learning-Algorithms learn from data without explicit programming-Broad,general Deep Learning-A subset of ML using multi-layered neural networks-More powerful training/predictions DEEP LEARNING Conv

6、olutional Neural Networks(CNN)-Most prevalent DL model for medical imaging-CNNs learn hierarchically,from simple shapes to complex features,to identify anomalies like tumors or fractures What Made This Possible?-Exponential increase in computing power(GPUs).-Availability of large,annotated datasets.