HitConsultant argues radiology AI pivot

Peter Nemeth used a May 18 opinion essay on HitConsultant to argue that radiology AI's next practical role may be patient explanation, not autonomous diagnosis. Nemeth, who is founder and chief executive of ReadYourLab, wrote that patients now often see imaging results before speaking with a clinician and that reports remain difficult for non-specialists to read. He tied that argument to immediate electronic access requirements under the 21st Century Cures Act and to patient demand for rapid release of results. The essay does not present a new product launch, regulatory filing or clinical trial. It is an opinion piece in HitConsultant's Thought Leaders section, and Nemeth frames it as a case for using AI as a translation layer between technical imaging data and patient understanding. ### Why is the argument centered on patient understanding instead of diagnosis? (hitconsultant.net) Nemeth wrote on May 18 that debate over AI in radiology has long focused on whether software will replace radiologists or outperform specialists. He said that framing misses what he called a more immediate opportunity: helping patients understand imaging information they already receive. (hitconsultant.net) The article says patients increasingly access results before any explanation arrives. Nemeth linked that to the Cures Act's push for immediate electronic access to many types of health information, then argued that transparency by itself does not produce understanding. ### What evidence did Nemeth use to support that case? Nemeth cited a survey of 8,139 patients in which 96% said they wanted to continue receiving immediately released test results online even if a clinician had not yet reviewed them. (hitconsultant.net) He used that figure to argue that patients favor access, even when explanation may lag. The same essay said radiology reports are rarely written for lay readers. (hitconsultant.net) Nemeth wrote that a study in a large U.S. health system found only a very small share of radiology reports were readable at the eighth-grade level, which he described as roughly the reading level of the average American adult. ### What problem in radiology workflows is he describing? (hitconsultant.net) Nemeth wrote that CT and MRI reports often combine anatomy, technical imaging language, incidental findings and differential diagnosis in a format that assumes medical training. He said that leaves many patients with confusion rather than clarity, especially in oncology, neurology and other high-stakes settings. (hitconsultant.net) One study cited in the essay found patients generally expected outpatient imaging results within one to three days, and 45% reported an emotional change while waiting, most commonly anxiety. Nemeth used that to describe the patient experience as access to complex results under stress, often before explanation arrives. ### Is this idea already showing up in products? (hitconsultant.net) PocketHealth said in February 2025 that it launched an AI-powered Image Reader designed to help patients better understand medical imaging results. The company said the feature provides visual context within CT and X-ray scans and complements a Report Reader that simplifies radiology reports in plain language. (hitconsultant.net) Dr. Bo Wang, chief AI scientist at University Health Network, said at the time that medical imaging AI had primarily focused on clinical applications but that there was also an opportunity to improve patient understanding. ### What are the limits and objections around patient-facing AI? Nemeth wrote that AI is not ready to diagnose patients independently. (hitconsultant.net) Instead, he said open medical models are becoming capable enough to help translate technical imaging data for patients, citing Google's MedGemma 1.5 model card as notable for support around three-dimensional CT and MRI volume representations. The American Medical Association's 2026 physician survey, as reported by HitConsultant in March, said nearly half of surveyed physicians strongly oppose patients using AI to interpret pathology or radiology results. The same report said 81% of doctors now use AI professionally, while liability, privacy and skill erosion remain central concerns. (hitconsultant.net) ### What happens next? HitConsultant published Nemeth's commentary on May 18, and the piece leaves the next step with developers, radiology groups and patient-facing imaging companies testing whether explanation tools can fit alongside clinician review. Named participants already visible in that discussion include ReadYourLab, PocketHealth and physician groups represented in the AMA survey. (hitconsultant.net 1) (hitconsultant.net 2)