Daily Briefing for October 5, 2026

Show Notes

CPT panel adds AI codes, rejects three; Medicare contractors propose no coverage for AI brain MRI tools; review maps rural surgical telehealth; eye AI lifts referrals

The voices on this episode are generated, and so is the script, from the day's post at Physician in the Loop. Every fact in the post links to its source.

Read the post and its sources

The American Medical Association's CPT Editorial Panel created a Category I code for AI arrhythmia analysis and three Category III AI codes, and rejected three other AI proposals, according to a summary released Oct. 2.

The new temporary codes cover AI identification of ST-elevation heart attacks, a breast cancer risk score from screening mammograms and analysis of digitized prostate cancer slides, and the summary lists Jan. 1, 2027 as their effective date. The arrhythmia code takes effect in January 2028, according to the summary. Separately, two Medicare contractors, Palmetto GBA and Noridian Healthcare Solutions, proposed to stop Medicare coverage of AI software that measures brain structures on MRI in 20 states, saying the evidence of clinical utility is insufficient.

The weekend was otherwise quiet. The Federal Register's listing of CMS documents that mention artificial intelligence showed no new entries since Sept. 24. The other items are a scoping review of telemedicine in rural surgical and anesthesia care and a Shanghai study of a language model used in primary eye care.

In this episode

• CPT panel creates Category I code for AI arrhythmia analysis, adds three AI Category III codes and rejects liver, LVEDP and skin-lesion AI
• Palmetto and Noridian propose Medicare noncoverage for AI brain MRI quantification software in 20 states
• Review of 98 studies finds rural surgical and anesthesia telemedicine centers on specialist access, preoperative triage and less travel
• Shanghai eye-care language model raises referral adherence in 170-person community study

What is Physician in the Loop?

How technological innovation is changing medicine for patients and the people who practice on the front lines.

Voice 1: This is Physician in the Loop, the daily briefing on AI in medicine. It's Monday, October 5th, 2026.

Voice 1: The American Medical Association's CPT panel just created 4 new AI billing codes, and rejected 3 others outright.

Voice 2: Which one actually matters most, the one that got real staying power?

Voice 1: Noninvasive AI arrhythmia analysis, code 93XX7. It's Category I, the permanent kind, not the temporary codes everything else gets. Also today: Medicare contractors proposing to stop paying for AI brain MRI tools, a review of rural surgical telehealth, and a Shanghai eye-care chatbot study.

Voice 2: The other 3 are all temporary, Category III, effective January 2027. One reads digitized prostate cancer slides to predict progression, 2 more pull a breast malignancy risk score out of screening mammograms, and a fourth flags ST elevation heart attacks and their equivalents, as an add-on code.

Voice 1: The AMA describes those as “a temporary set of codes for emerging technologies, services and procedures.” The arrhythmia code is permanent, but it doesn't take effect until January 2028, a year after these ones.

Voice 2: So the permanent code arrives a year late. Odd ordering.

Voice 1: And on the rejection side, 3 proposals got turned down outright: multiparametric liver tissue quantification, left ventricular end-diastolic pressure measurement, and electrical impedance spectroscopy of skin lesions.

Voice 2: What happened to the other 3 that were floated?

Voice 1: Withdrawn before any vote: acoustic and ECG analysis, algorithmic CT angiography, and ultrasound assessment of middle ear effusion.

Voice 2: Does the summary say anything about what any of this will actually pay?

Voice 1: Nothing on payment rates. The panel meets again in February 2027.

Voice 2: 4 new codes and not a dollar figure in sight.

Voice 1: 2 Medicare contractors went the opposite direction this week, on a completely different kind of AI tool. Palmetto and Noridian are proposing to stop covering AI software that measures brain structures on MRI, across 20 states combined.

Voice 2: What's their reasoning?

Voice 1: Noridian's proposal is blunt: “There is insufficient evidence of clinical utility or validity, and use of this technology is considered investigational and therefore not covered.” It names tools like NeuroQuant and icobrain, and covers 13 Western states plus 3 Pacific territories.

Voice 2: Palmetto covers the other 7, Alabama, Georgia, North and South Carolina, Tennessee, Virginia and West Virginia. Their complaint is more technical: no shared measurement standards across platforms, not enough evidence of improved outcomes, proprietary datasets that don't interchange, and limited diversity in the training data.

Voice 1: Is this a new worry, or has it come up with these contractors before?

Voice 2: 2 other contractors, CGS Administrators and National Government Services, posted proposals with the same title back in September 2025 and January 2026. So this is round three of the same argument.

Voice 1: When do comments close?

Voice 2: November 14th, with open meetings on October 22nd for Noridian and October 26th for Palmetto.

Voice 1: Three rounds of the same proposal and it still hasn't settled anything.

Voice 2: Same question, asked a third time now.

Voice 1: A scoping review out today looked at what telemedicine actually does for rural surgical and anesthesia care. Researchers led from the University of Calgary's Cumming School of Medicine screened almost 10,000 records and landed on 98 studies, some going back to 1992.

Voice 2: What did it find for rural care specifically?

Voice 1: Telemedicine there centered on specialist access, preoperative triage and less travel. Urban studies leaned more toward postoperative follow-up and specialist consultation.

Voice 2: And across both settings, short-term clinical outcomes looked similar to in-person care, with high patient satisfaction and lower travel costs. Video and telephone are still the main modes, though app and web platforms have grown since 2020.

Voice 1: Any AI in any of this?

Voice 2: None. The review doesn't address artificial intelligence at all, and the authors are calling for more comparative research on costs, access and outcomes.

Voice 1: Good to have the baseline, before AI gets layered on top of it.

Voice 2: Last one's out of Shanghai, a language model for primary eye care, called EyeSeek.

Voice 1: Who tested it, and how?

Voice 2: Researchers from Tongji University and Shanghai Eye Hospital split 170 people in community settings, 84 using the model and 86 getting care without it.

Voice 1: And the group that used it?

Voice 2: Higher adherence to referrals, with a P value of 0.037, plus improved health literacy. Readability analysis showed its answers adjusted down to grade 3 through grade 8 reading levels.

Voice 1: Does it say whether the model is actually accurate at screening, next to an ophthalmologist's own assessment?

Voice 2: It doesn't report that comparison at all. The study was prospective but not randomized, and it ran at one center, which the authors themselves say limits how far it generalizes.

Voice 1: Sources and corrections are at physicianintheloop.org. This briefing was read by AI voices.