AJR Podcasts Artificial Intelligence as a Safety Net: AJR Podcast Series on Diagnostic Excellence and Error, Episode 10
21 snips
Mar 26, 2025 
Benjamin Strong, a board-certified radiologist and chief medical officer at Virtual Radiologic, joins Francis Deng to delve into the transformative role of artificial intelligence in diagnostic radiology. They discuss how AI can detect critical pathologies while enhancing quality assurance and reducing malpractice risk. Strong highlights the unique challenges of emergency radiology, the importance of a problem-first approach to AI implementation, and the significant impact of natural language processing on care quality. The conversation also touches on the future of AI, from pathology-specific to generative models.
AI Snips
Chapters
Transcript
Episode notes
Big Five From Data, Not Assumptions
- VRAD identified a 'Big Five' list of high-risk pathologies from QA and malpractice data.
- These five differ from general practice because VRAD focuses on emergency teleradiology with unique case mix.
Choose AI After Identifying Problems
- Identify the specific diagnostic problems your practice actually has before buying or building AI.
- Let your QA and caseload data drive which algorithms to prioritize rather than shopping for trendy models.
Quantify ROI With NLP And Miss Rates
- Calculate ROI using incidence, miss rate, malpractice conversion, and average indemnity for each pathology.
- Use NLP on archives to get incidence and then compute expected savings to justify AI development costs.