One of the major topics at AMA is how the AMA CPT will handle coding and payment for software-dominant services. The AMA held an hour call on this earlier in the year, and an hour session at the September AMA CPT meeting.
Update: AMA will now hold an all-day workshop on Coding & Payment for Algorithmic Services on Monday, December 8, 2025 (9-3 central time). On-site registration is sold out but video registration is open.
https://www.ama-assn.org/member-benefits/events/coding-payment-algorithmic-services
###
###
AMA meetings tend to have multi-layered confidentiality restrictions, but I was able to write a block on the topic on September 10, because I attended a public genomics conference in August where an AMA speaker talked about the AMA's options, a public talk with no associated confidentiality contracts.
https://www.discoveriesinhealthpolicy.com/2025/08/digital-pathology-ai-and-coding-why.html
https://www.discoveriesinhealthpolicy.com/2025/09/update-on-ama-digital-and-ai-coding.html
You can also find non-confidential information via the AMA webpages for "Digital Medicine Payment Advisory Group" DMPAG and "Digital Medicine Coding Committee" DMCC.
https://www.ama-assn.org/practice-management/digital-health/digital-medicine-payment-advisory-group
One of the key ideas is to make a new section of AMA CPT - not PLA codes and not Category I or III codes - that would be a sort of registry of SaMD services (in pathology, radiology, cardiology, etc). A lot of AMA attention goes to the fact there are new software services that don't include physician work, which have historically been very rare codes.
However, in pathology and laboratory mediciene probably far more than 90% of the codes don't involve "physician work" (even "whole genome" or "comprehensive genomic procedures" on 400 tumor genes.) And CMS has put all the PLA codes that involve whole slide imaging and AI on the Clinical Lab Fee Schedule, even some with up-close CMS review as ADLTs. So my perspective is that CMS can ALREADY handle the coding and regulations for pathology AI within existing coding and regulations.
##
For interesting thought capital on AI reimbursement and adoption from Parikh, Warshawsky, and others, blog here.
For my ideas on why existing coding and regulations work (or could work) for AI pathology, blog here.
