AI flags esophageal cancer on plain CT; Rush hands chart abstraction to a vendor; AI firms sponsor MAHA Summit; Newsom vetoes prior authorization bill
A Chinese research group reported Tuesday in Nature Medicine that a model reading the esophagus on noncontrast CT outperformed all 17 radiologists in a reader study and was validated on 80,612 patients. The model flags esophageal cancer and high-grade precancer on the plain scan ordered for kidney stones and lung nodules. The validation covered 12 centers in three countries, and the model was then run prospectively on hospital scans and on 10,959 consecutive people in a low-dose lung screening program. The paper appeared one day after Alibaba released an abdominal CT model free on GitHub.
The Food and Drug Administration (FDA) press page lists nothing newer than a gene therapy approval on Thursday, and the Centers for Medicare & Medicaid Services (CMS) on Monday announced preliminary lab fee schedule rates. President Donald Trump on Friday announced an AI Force and a coming AI czar with no detail beyond a promise not to hinder the industry and a statement that the existing criminal and civil courts can handle bad behavior, Axios and Fortune reported; the announcement so far includes nothing on medicine. The Make America Healthy Again (MAHA) Summit on Sept. 29 is sponsored by OpenAI, Anthropic, Hims & Hers, Noom, Sword Health, Tempus, Prenuvo and Centene, among others, and sponsors receive dinner with the speakers, according to the summit's site and Becker's Hospital Review. In Sacramento, Gov. Gavin Newsom on Sunday vetoed a prior authorization bill that had passed 64-4 and 31-0 and has not acted on four health-AI bills, with eight days left to do so.
Elsewhere, a survey of 13 countries found that 49% of respondents believe a person with no medical training but skilled with AI could do at least one physician task as well as a physician; DexCare bought Mila Health in the second front-door agent acquisition in a week; and Becker's Hospital Review reported on how health systems triage the AI features Epic ships by the dozen, in a piece in which Rush's chief information officer said on the record that a chart-review job went to a model and is not coming back.
- Evidence Sept 22, 2026Zhou, Guo, Wang and colleagues trained a two-stage model, one stage to locate the esophagus and one to segment lesions and score malignancy, on 6,813 patients from two centers, then validated it on 80,612 patients across 12 centers in China, the Czech Republic and Australia. On an external test set of 11,466 patients from eight centers, the model found esophageal cancer with 90.0% sensitivity and 98.5% specificity and high-grade precancer with 52.5% sensitivity; on 1,607 low-dose chest CTs it ran at 88.4% sensitivity and 99.0% specificity. In a reader study it outperformed all 17 radiologists, exceeding their average by 19.1 points of sensitivity and 18.4 points of specificity, and as an assistant it raised the radiologists' cancer sensitivity from 71.9% to 85.7%. Run prospectively on hospital scans from January to April 2025, followed through July 2026, and on 10,959 consecutive asymptomatic lung screening participants, the model's flags had a positive predictive value of 42.2%, with specificity of 99.94% in the screening cohort. The authors listed the study's limits: nearly all patients were Chinese, a population in which squamous cell cancer dominates, with only preliminary validation abroad; performance was worse in women because the training set skewed male; follow-up was less than two years with imperfect compliance; and the reader study lacked a full crossover design. In the U.S. the common tumor is adenocarcinoma of the distal esophagus, which the model has not been shown to detect. The paper appeared one day after Alibaba's abdominal CT model was released free of charge.Source: Nature Medicine
- Deployment Sept 21, 2026Becker's Hospital Review asked seven health systems how they handle the AI features Epic now ships by the dozen with each upgrade. Rush's chief information officer, Jeff Gautney, said a clinical registry task that had taken five people reviewing charts under an outsourced contract of about $1 million a year now runs on Layer Health, while Epic has been "threatening for two years to deliver a product." Of the five jobs, he said: "That's not a job that's coming back." Seattle Children's runs an AI Review Board, chaired by its chief medical information officer, Clara Lin, that ranks features by expected return before anyone pilots them, because "it's impossible, obviously, to do a six-month pilot with every single one." Andrew Rosenberg, Michigan Medicine's former chief information officer, described the default: "We'll start with Epic, and even if it isn't as good, it's good enough." Sam Amirfar of Brooklyn Hospital Center said: "We prefer not to go first." The five Rush positions were administrative chart-abstraction roles at a single hospital.Source: Becker's Hospital Review
- Regulation Sept 21, 2026The summit's site lists the sponsors and partners of the one-day event at the Waldorf Astoria in Washington: GRAIL, ResMed, Advocate Health, Anthropic, Candid Health, Compass Pathways, Hims & Hers, Noom, OpenAI, Sutter Health, Centene, Prenuvo, Sword Health, Tempus, Walmart and a dozen smaller names. The event is run by MAHA Center, a nonprofit built around the movement of Robert F. Kennedy Jr., the secretary of the Department of Health and Human Services (HHS), and its speakers include Anthropic's Syed Mohiuddin, OpenAI's Felipe Millon and Dr. Jay Bhattacharya, the director of the National Institutes of Health. Sponsorship buys a private dinner with the summit's speakers, HHS officials among them, and executives from UnitedHealth, Elevance and the insurers' trade group are taking part, Becker's Hospital Review reported. The top package was pitched at $300,000 for a speaking slot, a say in panel topics and a dinner for 30 with Kennedy and Dr. Mehmet Oz, according to a report relaying Bloomberg's account; an HHS spokesperson said the department had no part in the sponsorship planning, and the price has not been independently verified. The sponsor list includes two AI model developers with clinical products, a direct-to-consumer telehealth company, two ACCESS participants, Noom and Sword Health, which now owns Headspace, a full-body MRI seller and a company that won Advanced Research Projects Agency for Health (ARPA-H) funding for autonomous cardiology; no physician society appears on it. The sponsorships come a week before comments close on the FDA's generative-AI paper and as Medicare officials discuss a payment category for AI physicians.
- Evidence Sept 16, 2026The 49% figure comes from a special analysis of the 2026 Edelman Trust Barometer written with the Yale School of Public Health and released this month, which Becker's Hospital Review reported on Monday. Edelman surveyed 12,998 people in 13 countries, about 1,000 per country including 995 Americans, between Feb. 28 and March 11. Asked which tasks a person with no medical training but skilled with AI could do as well as a trained professional, 26% chose deciding whether someone needs care, 19% chose determining treatment or medication, 19% chose performing basic procedures and 16% chose diagnosing illness. Agreement ran 59% among ages 18 to 34, 52% among ages 35 to 54 and 35% among those over 55, and 56% among the university-educated against 42% among those without a university education. The same analysis found that only 51% of people who receive conflicting recommendations always follow their doctor's. Edelman is a public relations firm and the barometer is its product; the analysis was written with Yale and its method is published. The question measured what respondents believe an AI-skilled layperson could do, not what such a person can do.
- Deployment Sept 21, 2026DexCare, a patient-access platform that says it has handled more than 10 million bookings for 57 million patients at systems including Kaiser Permanente, Piedmont, Texas Health Resources and Tampa General, announced the acquisition on Monday; terms were not disclosed. Mila's agents make the outbound calls and texts that a call center used to make, to schedule visits, prepare patients for them and follow up afterward, and the company said the agents produce a 54% patient response rate on outreach, cut no-show rates in half and lift revenue 18%; the figures are the company's own, with no published study behind them. Tampa General's Peter Chang said DexCare codified 260 scheduling policies for his system and found more than 50 gaps in them. Fierce Healthcare reported two similar moves last week: Hello Patient bought Converse Health on Sept. 17 to add referral and fax intake, authorization paperwork and registration to its patient-conversation agents, and Luma Health's Navigator agent now rebooks missed appointments and closes care gaps on its own. Zocdoc has also opened its rails to Gemini and Amazon.
- Regulation Sept 20, 2026Gov. Gavin Newsom's Sunday legislative update lists 83 bills signed and 24 vetoed. Among the vetoes is AB 539, by Assemblymember Pilar Schiavo, which would have required health plans to honor a prior authorization for at least a year, or for the full course of a shorter treatment; it cleared the Senate 31-0 and the Assembly 64-4, and the California Hospital Association supported it. AB 539 is not an AI bill. AB 1979, AB 2575, SB 903 and SB 503 appear on neither Sunday's list nor Monday's, when the governor signed seven data-center bills on water and power reporting and said nothing about health. All four AI bills passed with a handful of votes against them, and all four are opposed by hospitals, insurers or technology companies, or all three. Among the questions the four bills would settle is whether an AI can direct a medical assistant; the governor has eight days left to act.