Experts Hesitate on AI-Assisted Colonoscopies
A panel of specialists has decided to postpone a recommendation on the use of AI-assisted colonoscopies, according to new clinical guidelines released by a leading specialty society. After evaluating studies and modeling data, the American Gastroenterological Association (AGA) determined that AI reliably increases the adenoma detection rate (ADR), which identifies polyps. However, with only limited certainty, the panel estimates that AI-assisted screening would lead to a minimal reduction in colorectal cancer-related deaths (approximately 2 fewer deaths per 10,000 people over a decade). The panel opted not to issue a recommendation due to the uncertainty and trade-offs involved in utilizing new technology for disease screening.
The ADR is a crucial quality measure associated with enhanced patient outcomes. However, employing technology to boost ADR can result in more testing, procedures, and increased healthcare costs. Furthermore, the panel’s reluctance stems from uncertainty over whether AI significantly reduces cancer incidence and mortality rates.
The authors point out that the increase in ADR is “primarily attributable to detection of diminutive to small polyps with low potential for progression.” Benjamin Lebwohl and Shahnaz Sultan, authors of the guidelines, communicated that the lag time between finding these polyps and realizing their potential benefit is significant, so these data will not be available soon. Without conclusive data on mortality, evidence demonstrating AI’s ability to locate more dangerous polyps could alter the decision-makers’ stance. Interestingly, the guideline document initially recommended AI-assisted screening but was later withdrawn following feedback. The current lack of recommendation resulted from a vote due to the absence of panel consensus. The guideline will be revisited in a year or two, pending new data.
Nvidia Announces Platform for AI-Powered Robots in Healthcare
Nvidia recently hosted its GTC event, which shone a spotlight on significant players in healthcare. As the primary provider of chips used to train large language models and other AI technologies, Nvidia significantly fuels the AI innovations currently discussed. While announcements were often complex, Microsoft mentioned in its press release that it plans to run open-source generative AI models on an Nvidia and Microsoft offering.
GE HealthCare also announced its collaboration with Nvidia to “reimagine diagnostic imaging with autonomous X-ray and ultrasound solutions.” The companies aim to utilize Nvidia’s new platform for AI-powered healthcare robots and synthetic data to automate repetitive tasks performed by technologists. Addressing radiology staff shortages is another key goal.
At the TD Cowen Healthcare Conference earlier this month, Nvidia’s VP of healthcare, Kimberly Powell, outlined a vision for life science research companies, such as Genentech and Relation Therapeutics, to leverage NVIDIA’s BioNeMo platform to maximize value from data and scientific knowledge:
Every single experiment that you do, every idea your scientist has, can now be captured and codified to build upon an institutional knowledge that otherwise is maybe sitting in an electronic lab notebook somewhere. But how could you represent even those decades of electronic lab notebooks in a model, and then every single experiment you make, put it back in? And so, we’ve invented a platform called BioNeMo to serve just that.
(Brittany Trang contributed to this item.)
Medicare Advisors Suggest Telehealth’s Continued Relevance
As expected, the recent stop-gap funding bill approved by Congress included a continuation of Medicare coverage for telehealth services. The Medicare Payment Advisory Commission (MedPAC) released its March report to Congress, noting that while some adjustments to pandemic care have reverted, telehealth appears to have staying power. In a 2024 survey of beneficiaries, 33% reported using telehealth, and MedPAC plans to continue monitoring its impact as additional post-pandemic claims become available. However, data from Medicare indicates that 24% of beneficiaries used telehealth at least once in 2023.
Oura, EpiWatch, and Apple’s ResearchKit Legacy
This week, an app utilizing the Apple Watch to detect seizures received clearance from the Food and Drug Administration. EpiWatch, originally developed at Johns Hopkins, was one of the first research projects on Apple’s ResearchKit platform in 2015. It has since spun off into a company focused on care programs centered around this technology. This mirrors the trajectory of a Parkinson’s research project from 2015 that also led to several companies.
In 2016, Ricky Bloomfield, then a director of mobile technology strategy at Duke University, joined Apple. He worked there for over eight years, until recently. He has been named the new chief medical officer of Oura, a smart ring manufacturer. Oura is seeking to expand its role in the healthcare field, citing Bloomfield’s work on data interoperability at Duke as a valuable asset.
Hospital at Home Merger
Medically Home and DispatchHealth announced a merger plan. Medically Home is a key company helping health systems to provide hospital-at-home services. DispatchHealth also delivers hospital-at-home care but primarily offers same-day, in-home care to treat serious and chronic medical conditions to avoid trips to the emergency department.
Amazon Executive Departures
Following his abrupt departure from Amazon Pharmacy in February, Vin Gupta is joining consultancy Manatt, along with former CDC director Mandy Cohen, to build a public health communications practice. Earlier this month, Trent Green, CEO of Amazon’s One Medical, announced he would be leaving the company in April to become CEO of National Research Corp.
