You'll be excused for getting lost in the chain of events around MolDx and coverage for transplant rejection by means of donor-DNA assays. In January 2026, Craig-Hallum (investors) downgraded CareDx stock over continuing LCD uncertainty.
This led to several news reports, but over the past 30 days, it looks like the stock has traded in a narrow band from $19 to $20. We'll hear more at JP Morgan next week.
What Happened Over Time?
MolDx has covered donor-DNA assays for transplant rejection, available from both CareDx and Natera, for a number of years.
Back around March 2023, MolDx (meaning the Noridian MAC, which handles California claims) made some changes in claims processing that caused a stock drop at CareDx. (Here.) Later, MolDx put out a slightly revised LCD, with statements that the changes were only clarifying. However, there was another round of agitation. [My blog at the time, here.] Eventually, that "proposed LCD" was retired without finalization. However, in summer 2025, MolDx put out a more heavily revised LCD, as a "new LCD" with a new tracking number. The comment period ran for 45 days. Unlike comments on NCDs, comments on LCDs are not published by CMS. [My blog at the time, here.]
Since the "new LCD" in summer 2025 was lliterally "new," it's harder to compare the impact of the older and newer version easily, like you can when one can be redlined against the other.
Along the way there were at least 4 or 5 pieces in the Wall Street Journal, like this one.
So now we're in the ongoing wait for a "new final LCD" on donor DNA transplant testing. That's where the "new news" comes in.
What Happened This Week?
Investment group Craig-Hallum downgraded CareDx from "buy" to "hold." See Investing.com here (email registration required). See a short piece at StreetInsider.com here. See a subscription article at 360Dx here.
Some text from StreetInsider.com:
The analyst comments "With CDNA shares having fully recovered to where they were before MolDX issued its restrictive draft LCD back in July, the risk-reward appears less compelling. At this point, 2026 revenue and longer-term growth will be highly dependent on the shape of the final LCD. We can conceive of a scenario where next year’s numbers are too high and longer-term growth opportunities are less robust...[but] 2026 estimates could prove conservative (and our downgrade could prove overly cautious) if the final LCD is meaningfully less onerous than the draft LCD..."
We all know that curative chemotherapy at one point in your life (for lymphoma at age 20) can raise your risk of cancer later in life. Here's an up-to-date genomic viewpoint of what happens, revealed by ultra-deep sequencing.
See original article in Nature by Pich et al. here.
###
AI CORNER
###
Pich et al. examine how cancer treatment contributes to somatic change in histologically normal tissues over the course of an individual’s life.
Using ultra-deep duplex sequencing of multiple organs from treated cancer patients, the study shows that chemotherapy and other therapies leave distinct, tissue-specific mutational footprints, often rivaling decades of age-related mutation accumulation. Importantly, treatment not only induces mutations but can also shape clonal composition through selection, even without increasing overall mutagenesis.
These findings have direct implications for clinical genomics, particularly for interpreting low-frequency somatic variants in blood and tissue samples from treated patients.
Below, Chat GPT 5.2 reads amd discusses what's new in Johnson et al. 2026, with in put from articles by Siddaway 2025, a news article on Siddaway, and Li 2025.
The second section, below, discusses "Takeaways for Payors."
###
RNA Sequencing Enters Clinical Genomics: From Conceptual Promise to Routine Practice
Introduction: from “adjunct” to clinical substrate
For more than a decade, RNA sequencing (RNA-seq) has been discussed as a powerful complement to DNA-based cancer genomics. Yet until recently, its role remained largely ancillary—most commonly for fusion detection, research applications, or select tumor types. The four papers considered here collectively argue that this phase has ended. Together, Johnson et al., 2026 and the three 2025 studies demonstrate that RNA-seq is no longer experimental or aspirational; rather, it is becoming a practical, scalable, and clinically consequential layer of molecular diagnostics for real-world cancer patients.
What distinguishes this literature from earlier RNA-seq enthusiasm is not technical novelty, but operational maturity: robust performance in FFPE tissue, acceptable failure rates, clinically interpretable outputs, and demonstrated effects on diagnosis and therapy selection.
Johnson et al., 2026: Defining the “Actionable Transcriptome”
Johnson et al. provide the most synthetic and forward-looking contribution of the four studies. Rather than focusing on assay performance alone, they articulate a conceptual framework—the “actionable transcriptome”—that reframes RNA-seq as a unifying diagnostic modality rather than a niche add-on
.
Their central argument is that RNA-seq captures classes of actionability that DNA, immunohistochemistry (IHC), and low-plex assays assess only partially or serially. These include:
oncogene overexpression and tumor-suppressor underexpression independent of copy number,
multigene prognostic and predictive signatures,
immune-related expression states, and
expressed neoantigens and splice variants.
Importantly, Johnson et al. are not proposing RNA-seq as a wholesale replacement for DNA testing. Instead, they position RNA-seq as the first truly multiplexed assay that spans mutation, expression, structure, and functional consequence within a single analyte. This framing explains why RNA-seq is now attractive not only scientifically, but economically and operationally for clinical laboratories.
Siddaway et al., 2025: Real-World Clinical Proof at Scale
If Johnson et al. define the theory, Siddaway et al. provide the proof. In their Nature Medicine study of over 2,300 consecutively tested patients, targeted RNA-seq demonstrated clinical utility in 87% of cases, with actionable or diagnostically relevant findings in nearly two-thirds of tumors
.
Several features of this study are pivotal for clinical adoption:
Specimen realism: ~75% of samples were FFPE, addressing a long-standing concern about RNA stability.
Operational reliability: a failure rate below 5%, comparable to many DNA assays.
Breadth of detection: RNA-seq detected SNVs, indels, fusions, splice alterations, and diagnostically informative negatives.
Clinical consequence: results led to revised diagnoses and altered treatment decisions, including initiation of targeted therapy.
The accompanying Nature Medicine editorial by Siddaway and Hawkins reinforces the central message: RNA-seq can function as a front-line diagnostic assay, reducing tissue consumption, cost, and turnaround time while expanding interpretive depth
.
Li et al., 2025: Bridging DNA to Protein Through Expressed Mutation Detection
Where Siddaway et al. emphasize clinical workflow, Li et al. focus on molecular signal fidelity. Their npj Precision Oncology study evaluates targeted RNA-seq specifically as a tool for detecting expressed somatic mutations, using reference standards and controlled false-positive benchmarking
.
Their findings sharpen an increasingly important clinical distinction:
DNA sequencing identifies potential variants.
RNA sequencing identifies expressed and therefore biologically plausible variants.
Li et al. demonstrate that a substantial fraction of DNA-detected mutations are not transcribed at meaningful levels, while RNA-seq can uncover expressed variants missed by DNA assays due to coverage, tumor purity, or technical limitations. This supports a pragmatic reframing: RNA-seq does not merely add sensitivity; it filters relevance.
Converging Themes Across the Four Studies
Across these papers, several shared conclusions emerge:
1. RNA-seq is clinically viable, not aspirational
All four studies reject the notion that RNA-seq is too fragile, noisy, or slow for routine diagnostics. FFPE compatibility, low failure rates, and standardized pipelines are now established realities.
2. Expression matters as much as mutation
RNA-seq resolves a core limitation of DNA-only profiling: the inability to determine whether a variant is biologically active. Expression-aware interpretation strengthens confidence in therapeutic relevance and deprioritizes silent alterations.
3. Multiplexing drives value
RNA-seq consolidates what would otherwise require multiple assays—fusion testing, expression profiling, splice analysis, and confirmatory variant assessment—into a single workflow.
4. Clinical negatives are informative
Both Johnson et al. and Siddaway et al. emphasize that RNA-seq can generate pertinent negatives—ruling out specific entities or therapeutic targets with confidence, which is itself clinically actionable.
What to Expect Next (2026–2027)
Based on these studies, several near-term developments are likely:
RNA-first testing strategies Clinical laboratories will increasingly deploy RNA-seq as an initial assay, with reflex DNA testing reserved for copy-number changes, mutational burden, or germline assessment.
Expansion beyond oncology panels RNA-based diagnostics will broaden from oncology into immune profiling, resistance monitoring, and therapy-response prediction, particularly for ADCs, bispecifics, and cell therapies.
Regulatory and reimbursement alignment As RNA-seq demonstrates consistent clinical utility, payers and regulators will be forced to engage with expression-based biomarkers as legitimate medical decision tools, not exploratory analytics.
Interpretive standardization Frameworks such as the “actionable transcriptome” will drive standardized reporting vocabularies and decision-support systems, analogous to earlier DNA variant classification schemes.
Conclusion
Taken together, these four papers mark a transition point. RNA sequencing has moved from conceptual bridge to clinical substrate—a technology capable of supporting real-world cancer care at scale. Johnson et al., 2026 provide the intellectual architecture; Siddaway et al. demonstrate feasibility and impact; Li et al. clarify biological relevance; and the accompanying editorial underscores urgency. By 2026–2027, RNA-seq is poised not merely to complement DNA genomics, but to reshape the definition of actionable molecular oncology.
What are the takeaways for payors? (Version 01)
For a reimbursement and payor audience, the central takeaway from this body of work is that RNA sequencing is no longer an exploratory “add-on” technology but a consolidating diagnostic service that can replace multiple sequential or parallel tests while improving clinical decision quality. The studies collectively show that RNA-seq reduces diagnostic uncertainty by determining which genomic alterations are actually expressed and therefore biologically relevant. For payors, this directly addresses a common concern with large DNA panels: that they detect numerous variants of unclear or low clinical consequence, increasing downstream utilization without proportional patient benefit.
A second key takeaway is that RNA-seq supports earlier and more confident treatment selection, which is closely aligned with payor goals of avoiding trial-and-error therapy. Across real-world cohorts, RNA-seq identifies actionable alterations—particularly gene fusions, splice variants, and expression-defined targets—that are either missed or ambiguous with DNA-only testing. Importantly, these findings are not academic: they demonstrably change diagnoses, confirm or refute eligibility for targeted therapies, and guide enrollment into appropriate treatment pathways. From a utilization perspective, this means fewer ineffective therapies, fewer reversals, and better alignment between molecular testing and approved or guideline-supported treatments.
Third, RNA-seq creates value by generating clinically meaningful negative results, which payors often undervalue but rely on implicitly. When RNA-seq shows that a putative DNA alteration is not expressed—or that no actionable transcriptomic drivers are present—it helps rule out costly targeted therapies or immunotherapies that are unlikely to benefit the patient. This ability to produce “pertinent negatives” strengthens medical necessity arguments and supports more defensible coverage decisions, particularly in advanced cancer settings where treatment costs are high and evidence thresholds are scrutinized.
Another important reimbursement-relevant insight is that RNA-seq enables test consolidation and tissue stewardship, both of which have cost implications. By capturing fusion detection, expressed mutation confirmation, splice effects, and multigene signatures in a single assay, RNA-seq can reduce the need for multiple sequential tests such as separate DNA panels, FISH, IHC, and standalone expression assays. This consolidation lowers cumulative testing costs, shortens time to treatment, and reduces the likelihood of repeat biopsies—outcomes that align with payor objectives even when the per-test price of RNA-seq is higher than that of an individual legacy assay.
Finally, these studies signal that RNA-seq is moving toward predictable, standardizable clinical performance, which is critical for coverage policy. Low failure rates in FFPE tissue, consistent turnaround times, and reproducible clinical utility across thousands of patients address historical payor skepticism about RNA-based assays. As the concept of the “actionable transcriptome” matures, RNA-seq results can be framed not as exploratory omics data, but as structured, interpretable medical information that directly informs treatment eligibility, prognosis, and care pathways. For payors, this shifts RNA-seq from a discretionary molecular test to a decision-enabling diagnostic service with a clearer value proposition tied to outcomes, utilization control, and care efficiency.
What are the takeaways for payors? (Version 02)
That's a fairly happy way to present these papers to payers. What about payers that are more skeptical that this is a lot of expensive sequencing and few outcome and benefit studies? Assume payors are more skeptical - which they are - and taking more of a third-party viewpoint, with the "RNA companies" on the left, the payers on the right, and what is likely to be important, what is likely to happen.
That’s a fair pushback—and it’s exactly the stance many payers will take. If we step out of the “author voice” and adopt a third-party, skeptical payer perspective, the same papers read quite differently. What follows reframes the evidence without advocacy, focusing on what payers are likely to accept, question, and eventually demand.
From a skeptical payer viewpoint, these papers do not yet establish RNA sequencing as an outcomes-proven technology; rather, they establish RNA-seq as a technically mature diagnostic platform whose clinical claims are still upstream of hard evidence. The studies convincingly show feasibility, signal detection, and interpretive richness—but they stop short of demonstrating that RNA-guided decisions consistently improve survival, quality of life, or total cost of care relative to existing DNA-based pathways. For payers, that gap matters, because coverage decisions hinge less on molecular elegance than on downstream impact.
A core payer concern is that RNA-seq risks becoming “more information without clearer decisions.” While RNA-seq identifies additional alterations—expressed variants, fusions, splice events—these papers largely report diagnostic yield rather than clinical resolution. From a payer lens, the question is not “did RNA-seq find something?” but “did it replace something, avoid something, or change something in a way that measurably improved outcomes or reduced waste?” The current evidence base is heavy on plausibility and light on comparative effectiveness.
Payers are also likely to challenge the implicit assumption that expression equals actionability. While expressed alterations are biologically appealing, payers will note that regulatory approvals, NCCN categories, and compendia listings remain overwhelmingly DNA-anchored. RNA-defined targets—especially overexpression or underexpression without a corresponding genomic alteration—often lack FDA-labeled companion diagnostics and may lead to off-label therapy use. From a utilization-management standpoint, RNA-seq could therefore increase—not decrease—therapy costs unless carefully constrained by policy.
Another skeptical interpretation is that RNA-seq currently bundles multiple analytical functions that payers prefer to adjudicate separately. DNA sequencing, IHC, FISH, and expression assays each have established coverage precedents and policy guardrails. RNA-seq collapses these into a single test, which is efficient for laboratories but problematic for payers who rely on indication-specific coverage logic. The payer concern is not duplication of information, but loss of control over indication-based reimbursement.
Importantly, payers will also notice that most outcome claims in these papers are proximal, not distal. Changes in diagnosis, identification of targetable alterations, or initiation of targeted therapy are intermediate endpoints. Payers are increasingly wary of these surrogate claims, especially after prior experiences with broad NGS panels where early enthusiasm did not translate into clear population-level benefit. The absence of randomized or quasi-experimental comparisons between RNA-first and DNA-first strategies will be a recurring objection.
That said, skeptical payers are unlikely to reject RNA-seq outright. Instead, the likely near-term outcome is selective containment rather than broad adoption. RNA-seq will be tolerated—and sometimes welcomed—in niches where its value proposition is clearest and risk is lowest: fusion-heavy tumors, cancers of unknown primary, pediatric malignancies, and cases where standard DNA testing has failed. In these contexts, RNA-seq looks less like an added cost and more like a salvage or clarification tool.
What is most likely to matter going forward is how RNA companies reposition their evidence strategy. Payers will increasingly expect studies that show RNA-seq can replace existing tests, not merely augment them; that it shortens diagnostic pathways; or that it reduces ineffective therapy exposure. Budget-impact models, not just diagnostic yield, will become central. Comparative studies against “DNA-only plus standard reflex testing” pathways will carry far more weight than additional catalogs of actionable findings.
In practical terms, the equilibrium between RNA companies and payers will probably settle into a conditional acceptance model. RNA-seq will be covered under defined clinical scenarios, often after prior testing or within institutional programs, while broader first-line use will remain restricted pending stronger outcomes and economic data. Over time, if RNA-seq can demonstrate test substitution, therapy avoidance, or durable response enrichment, payers may begin to view it less as expensive sequencing and more as care pathway optimization.
In short, from a third-party payer perspective, these papers mark RNA-seq as credible but unproven. They lower the technical and operational objections but leave the economic and outcomes questions largely unanswered. What is likely to happen next is not rapid payer enthusiasm, but measured, conditional uptake, with increasing pressure on RNA-based diagnostics to prove that more molecular insight leads to fewer wasted dollars—not just better biology.
Bibliographic Citations
Li D. et al.Augmenting precision medicine via targeted RNA-Seq detection of expressed mutations. npj Precision Oncology. 2025.
…
Siddaway R. et al.Clinical utility of targeted RNA sequencing in cancer molecular diagnostics. Nature Medicine. 2025.
Siddaway R., Hawkins C.The power of RNA sequencing in molecular oncology. Nature Medicine (Research Briefing). 2025.
Johnson F. et al.The actionable transcriptome: a framework for incorporating RNA sequencing into precision oncology. Nature Reviews Clinical Oncology. 2026.
Chat GPT 5.2 read the two essays and had some peppy opinions.
###
AI in Oncology: What’s Actually New in ASCO’s Latest Signal
Two ASCO Post articles published on the same day—How AI Is Ushering in a New Era in Cancer Care and ASCO and AI in Oncology: Rooted in Human-Centered Care—are easy to skim and dismiss as familiar optimism about artificial intelligence. That would be a mistake. Read together, they quietly signal several important shifts in how AI is being framed, governed, and operationalized in oncology—shifts that matter far more than generic claims about speed or scale .
This morning, I asked Chat GPT a question about Medicare Advantage. I noted that many state that Medicare Advantage is cost saving for patients, because its premiums are about the same as Medicare Part B, but you get extra benefits (e.g. vision) and don't have to buy a Medigap plan (about $200/mo).
I suspected that was true, but incomplete. Traditional Medicare A+B, and MediGap, provide nearly "first dollar" coverage all year. (There's a Part B deductible, but it's 'starbucks money" compared to other health costs).
But in Medicare Advantage, you may have office copays ($10-20),specialist copay s($30-40), out of network copays (more, or the whole cost if non-covered). These may reach a $5000 copay cap, adding about $400/month to our total costs of Medicare Advantage (enough to have paid for a Medigap plan). (And any out of network fees might be a different cap.)
###
I asked Chat GPT if this sounded right and to research it, and it "thought" for nine minutes.
I clicked on the "thought for 9 minutes" line, and the result is clipped below.
Don't miss the January 1, 2026 update on newly awarded PLA codes, because AMA seems to use this channel to convey new (or newly emphasized) rules for the PLA code system.
I wrote a blog on November 19 that AMA had put some new PLA text in the 2026 edition of the CPT handbook. I was concerned the updates were vague. E.g. a list of changes that MAY trigger the need for a new PLA code. "MAY???" How about, do or do not require a new code??
There are 17 new codes, 2 revised codes, and some revised text. Some of the 2026 manualized revised text appears now in the PLA code list PDF, but I think there are some items only in one place (manual) or the other (code list PDF).
Tests may be provided by a sole source lab, at the most, there could be multiple locations owned by the same entity. (If you want sequencing at Lab A and bioinformatics at Lab B, you're out.)
PLA codes must not have a physician work component. (E.g. they must all be eliglbe for the Clin Lab fee schedule at CMS - no physician work, which would trigger professional components paid for inpatients and trigger RVU valuation).
An algorithm-only analysis is not eligible for a PLA code.
(I believe this is defined to include digital pathology eg DICOM reads of prior slides).
If a PLA coded tests is elevated to Category I it gets a new code (e.g. 8nnnn). They will time deletion of the PLA code to the first day of the new code.
Like the 2026 6CPT book, that say that certain changes "MAY" require a new code (not that the changes "do" or "do not") which seems like exceptionally cryptic wording to me.
Similarly, an addition or modification of therapeutic applications requires a code change submission, BUT IT MAY OR MAYNOT require a changed code number. "Thanks for clarifying that."
Header: Harvard MBA Theory Collides with Mac & Cheese Industry
###
The Harvard MBA theory is from the late Prof. Clayton Christensen, called "The Innovator's Dilemma" and "Disruptive Innovation." (Wikipedia here and here). Wiki writes, "disruptive innovation is innovation that creates a new market and value network or enters at the bottom of an existing market and eventually displaces established market-leading firms, products, and alliances. In theory, disruptive innovation makes it hard for leading firms to stay at the top of their industry."
The prototypic case was 8" computer drive manufacturers who were undercut and then toppled by pesky little new 5" computer drive manufacturers.
Mac & Cheese industry?
Jesse Newman at WSJ writes a January 1, 2026 article about how Kraft Mac & Cheese is being invaded by upstart far small branders who are now chipping away at significant market shares. Multiple references about how Kraft, a giant incumbent, has been slow and uncertain in its responses.
The WISeR program for AI-mediated autodenials is coming into effect now. For a couple dozen procedures, physicians/hospitals will have to submit pre authorization paperwork to special CMS subcontractors, which will have several days to greenlight or deny the surgery. (Submit to, Zyter, Virtix, Humata, Cohere, Genzeon, Innovaccer.)
CMS has stated that the contractors will simply be submitted existing CMS rules and instructions (LCDs, NCDs). But this won't always work. For example, penile implant codes are among the controlled codes under WISeR: including 54400, 54401, 54405.
CMS says the companies will follow CMS coverage rules under NCD 230.4.
Impotence is a failure of a body part for which the diagnosis, and frequently the treatment, require medical expertise.
Depending on the cause of the condition, treatment may be surgical; e.g., implantation of a penile prosthesis, or nonsurgical; e.g., medical or psychotherapeutic treatment.
So I don't know what rules preauthorization will use, but it's literally impossible for surgeons, hospitals, or patients to know in advance whether a case is "in" or "out" of payment rules that will be applied behind the scenes. Because CMS gives no concrete rules.
(Note also, the NCD is not just about surgery, but covers any kind of medical or psychological intervention).
On December 31, 2025, CMS issued Change Request 14205 (aka Transmittal 13570). The title is,
Implementation of Wasteful and Inappropriate Service Reduction (WISeR) Model Prior Authorization and Medical Review Process and Establishment of New Quarterly Change Request (CR) Process for Possible Future Changes.
The 82-page document covers many implementation aspects for WISeR, the program that will allow AI-facilitated autodenials for a range of clinical interventions.
What caught my attention first was Table 7 (page 28), which applies from January 2026 to December 2031 in a number of large states, including New Jersey, Ohio, Oklahoma, Texas, Arizona, and Washington. (The list of states can be expanded at the discretion of the Center for Innovation, CMMI).
I usually find drug labels via Google, which may flag the FDA website, or a manufacturer website, or other sources.
The CMS OPPS drug passthrough application referred to drug labeling at a website run by National Library of Medicine, DAILY MED. New to me. Here it is:
See an essay at Linked In by the prolific Alex Bendersky on whether it's realistic to expect good results when we dump important health innovations into a chaotic "ecosystem" of healthcare stakeholders.
In a sidebar, I applied the Reed et al. (& Adner) framework to four topics. These were (1) Vahid Asimi's recent article on how-to-IPO in diagnostics (here), (20 Luis Cano's recent article on what's misfit about digital pathology (here), (30 the history of the premature radiopharmaceuticals Bexxar and Zevalin (chat gpt history here.) And fourth ad finally, the stop-and-go history of point of care testing (POCT). Find the four AI essays together here.
Much has been written about pharmacogenetics for 20, maybe 30 years, but basic issues like PGx panel coverage by payors have remained uncertain.
This week the public Pharmacogenomics group at Linked In - which has almost 3000 members - features an article by Rouby et al on the value of community-focused PGx. For my money, the main issue is that it is not an RCT and there can be a lot of issues with assigning "controls" or "comparables" based on things like same age or same number of chronic diseases or other propensity scoes.
In October 2025, Rubrum Advising, in service of Harrison AI, submitted a 34-page petition to the FDA asking for widespread exemptions from review (from premarket notification) of certain medical AI software. Find it here.
In Federal Register December 29, FDA requests public comment on the petition; comment open for 60 days. Find it here.
Sidebar 1: New article by David Horgan et al. on medical AI in Europe - here.
Sidebar 2: In January 2026, FDA announced some change to make AI review as flexible as statutes allow; see StatNews January 6, 2026. See also the original new 26pp FDA document (a final guidance) here. (Prior version was 9/28/2022). Although it's "final," comments are still taken.
Stat writes, "The agency will soften its approach to the regulation of clinical decision support software, which include AI-enabled products that help doctors navigate diagnoses and treatment options. The agency previously considered products that delivered a single recommendation as FDA-regulated medical devices. Now, those products can enter the market without FDA review as long as they fulfill the agency’s other criteria for escaping regulation. -- The new approach appears to open the door to the unregulated use of generative AI products for certain medical tasks, such as summarizing a radiologist’s findings."
It's the holidays, we're traveling, we're using Apple Airtags to track luggage and backpacks. Every two years you change the battery, which takes about 10 seconds and one sentence of instructions.
I asked Chat GPT to make the most comically detailed description possible. Humor only. Below.
See the primary publication by Orenbuch et al. in Nature Genetics (Nov 24) here.
See an interesting earlier publication for backgroud, AlAbdi et al., Nat Comm, 2023, here.
###
I have some concerns that this type of major advance in clinical genomics may collide at high speed and head-on with changes the AMA plans to "Appendix S" (AI) in the AMA CPT, and to a possible future "software registry" (to be called CMAA) placed far outside the normal coding and payment system we use in the clinical laboratory profession and industry. This topic is Agenda 64 at the February AMA CPT Palm Springs meeting (see entry point here.)
###
##
AI Corner
##
Here's a ChatGPT condensed version of the original abstract:
Missense variants pose ongoing challenges for genetic interpretation due to context-dependent effects and poor calibration across the proteome. We developed popEVE, a deep generative model integrating evolutionary and human population data to estimate variant deleteriousness at a proteome-wide scale. popEVE achieves state-of-the-art performance without inflating deleterious variant burden and identifies variants in 442 genes in a severe developmental disorder cohort, including 123 novel candidates. Notably, popEVE can prioritize likely causal variants using child-only exomes, enabling diagnosis without parental sequencing. This work demonstrates a generalizable, evolution-informed framework for rare disease variant interpretation, particularly for singleton cases.
###
Review: PopEVE — A Proteome-Wide, Calibrated Model for Missense Variant Interpretation
1. Context: why missense interpretation remains hard
Missense variants sit at the most difficult intersection of molecular biology and clinical genetics. Unlike loss-of-function variants, their effects are graded, context-dependent, and protein-specific, making them resistant to binary pathogenic/benign classification. Large Mendelian sequencing studies have shown that diagnostic failure is often not due to sequencing gaps but to interpretive bottlenecks, including novel allelic effects, misleading in-silico predictions, and variants whose significance cannot be reliably inferred within a single gene context (Nature Communications).
PopEVE directly targets this bottleneck: not by adding another pathogenicity classifier, but by reframing variant interpretation as a proteome-wide calibration problem (Nature Genetics).
