2023 SIIM-ACR Data Science Summit

Keynote | Richard Bruce, MD — Establishing an AI Value Proposition at Your Institution


Session 1 | AI Reimbursement: Progress and Challenges
This session focuses on the evolving payment landscape in the context of regulation, the AI market and organizational needs. AI reimbursement depends on regulatory and CMS policies and returns on investment (ROI) of radiology AI or any informatics product depends on the problem it solves. Current standard understanding of ROI is based on the radiology value chain, and where tools can help augment automate the “value.” 

      Gregory N. Nicola, MD, FACR
      Chair, Commission on Economics, American College of Radiology (ACR)
      Chief Medical Officer, NeuTigers
      Executive Leadership, Hackensack Radiology Group

      Lauren Nicola, MD
      Chair, Commission on Ultrasound, American College of Radiology
      Chief Executive Officer of Triad Radiology Associates in Winston Salem, NC,
      Chair of Radiology at Forsyth Medical Center
      Chief Medical Officer of Strategic Radiology’s Patient Safety Organization (PSO)


Session 2 | Ensuring Equitable Use of AI
This session will focus on ethical considerations and trust in algorithms which are crucial factors that need to be considered when selecting an AI system, although they may not have an immediate impact on the bottom line. Proper governance structures are necessary to address these issues. However, physicians are often unfamiliar with how to utilize these factors when selecting a system, as well as the metrics and challenges associated with each factor.

Moderator:
      Marla Sammer, MD, MHA, FAAP
      Chair, Pediatric Ai Working Group, Informatics Commission, ACR
      Vice-Chair for Clinical Affairs, Texas Children’s Department of Radiology
      Associate Professor, Baylor College of Medicine

Speakers:
      Alex Towbin, MD, FACR, FAAP
      Associate Chief Medical Information Officer
      Associate Chief, Department of Radiology at Cincinnati Children's Hospital

      B. Bersu Ozcan, MD
      Research Fellow
      Breast Imaging Division
      University of Texas Southwestern


Session 3 | Setting Up a Quality AI Program for Your Practice
This session will discuss how many radiologists have an incomplete understanding of how to govern the selection, validation, and deployment of Artificial Intelligence in radiology. Despite growing awareness of AI, most radiologists are unfamiliar with the process of acquiring AI products, which are actively marketed to them. They lack knowledge on how to source IT products and are unsure about the benefits and risks associated with AI products. Many physicians are considering adopting their first AI models, but they are unaware that the performance of these models can fluctuate and degrade over time due to changes in the real world. Therefore, they are unsure how to implement strategies for measuring the performance of AI outputs. 

Moderator:
      Howard (Po-Hao) Chen, MD, MBA
      Chief Imaging Informatics Officer, IT Medical Director for Enterprise Radiology
      Staff Radiologist in Musculoskeletal Imaging at Cleveland Clinic

Speakers:
      Katherine Andriole, PhD
      Director of Academic Research and Education at MGH & BWH Data Science Office
      Associate Professor of Radiology, Director of Imaging Informatics

      Marc Kohli, MD
      Associate Chair of Imaging Informatics at University of California, San Francisco
      Associate Professor in the Department of Radiology and Biomedical Imaging


Session 4 | ACR® & DSI Programs Promoting Quality AI in Clinical Practice
This session will help Radiologists gain knowledge about evaluating AI performance within their specific patient population and environment. Most are unaware of the AI evaluation tools that are freely available through the ACR. Moreover, they are unfamiliar with the available tools for evaluating, deploying, and monitoring AI products long-term, such as building a test dataset for the AI model, maintaining a data registry, or using data that is representative of their own patient cohorts. Many radiologist also lack understanding about how these tools can help protect data security and patient privacy in AI.