The past month, I saw a nice white 4 page paper by LabCorp about "Bridging testing gaps by enabling an FDA cleared IVD tumor profiling precision oncology test within your local community setting." Find it here. (See also a 35 page white paper by Labcorp; at bottom of blog).
On a similar theme, in December 2025, CAP Today published a lead article on, "How two labs took on in-house sequencing." Find it here.
I asked Chat GPT 5.2 to contrast and compare the two pieces.
MolDX and Z codes have been around for well over 15 years now, preceding by several years the new AMA CPT genetic coding system of 2012-2013.
How many Z codes are there? We can start to ballpark the number by going to the Dex Diagnostics Exchange Register, registering as a public user, and looking at the two catalogs. These are the TEST catalog and the "LABS & MFGRS" catalog.
Click on Labs & Mfgrs, and you'll see 33 pages at 100/page, and remarks, 3249 items.
###
I don't know to what extend there are separate Commercial and Medicare listing, but this gives you a first approximation of the DEX Z-code library.
Header: OIG Releases Report on 2024 Lab Spending.
What it is.
The PAMA legislation of 2014 requires OIG to issue an annual report on lab spending - in part to show Congress the impact that the new "triennial market price reset" was having on lab spending. As Medicare nerds know, the reset occurred only once (in 2018) based on data from 2016.
My blog September 2025.
CMS released most of this data in September 2025, and I reviewed and discussed it then. Blog. Blog.
Genomeweb Discusses OIG.
Genomeweb/360Dx (Adam Bonislawski) covers the OIG report, and also has a long discussion of the weirder aspects of molecular spending.
What OIG found.
The report has extensive tables and graphics focused on genomic vs conventional lab spending.
The lab report suggests that the increase in genomic tests was due to spending on cancer, infections, and epilepsy.
So for a few years Novitas had about the highest-paid code 81408 rare gene (sic) full length sequencing, circa $500M, and that fell to zero at the same time as 87798, uncontrolled pathogen code, role to circa $500M. Sigh.
Payments outside Novitas hot spots like Texas and Florida was very small, and billing for 87798 and 81419 by known labs (whether Quest-LabCorp type or GeneDx-Ambry type) was about nil.
Get the full report here:
https://oig.hhs.gov/documents/evaluation/11453/OEI-09-25-00330.pdf
On January 26/27, 2026, the big news was that CMS has announced there is virtually no change in rates for Medicare Advantage in CY 2027. They issued a press release, FAQ sheet, and a 160-page policy document.
Read here for Reuters. Read here for HealthcareDive. (HCD also here for comments from administrator Chris Klomp.) Or see picture for the main takeaway:
How to do the policy making processes differ?!?
I asked Chat GPT.
### SUMMARY OF THE TWO DOCUMENTS
###
### LINKS
Fed Reg November 2025 here.
Advanced notice 160pp January 2026 here. Fact sheet here. January 2026 press release here.
###
###
You are absolutely justified in being confused — CMS uses two totally different legal pipelines that both talk about “Part C & D changes for 2027,” but they do different jobs.
Here’s the clean mental model.
| November 28, 2025 document | January 26, 2026 document |
|---|---|
| Federal Register Proposed Rule (137 pages) | Advance Notice (160 pages) |
| Docket: CMS-4212-P | Not a FR rule — payment notice under SSA §1853 |
| Amends regulations (42 CFR Parts 422 & 423) | Updates payment formulas, rates, and risk models |
| Becomes a Final Rule later | Becomes the Rate Announcement (April 2026) |
| Legal authority: APA rulemaking | Legal authority: MA rate-setting statute |
| Changes program rules | Changes how plans get paid |
This is a classic CMS regulation. It changes the rules of the game.
It covers things like:
Part D benefit redesign codification (IRA changes)
Star Ratings methodology changes
Marketing / TPMO oversight
Enrollment rules
SNP policy
MLR / reporting ideas
Network adequacy process changes
It literally amends the CFR.
This is about what plans must do.
Source: The November rule states it revises “Part C and Part D regulations” and includes policy areas like Star Ratings, marketing, enrollment, and drug coverage.
2025 CMS 1128 90FR54894 137p Pr…
👉 Outcome: Final Rule later in 2026.
This one is not a regulation. It does not amend the CFR.
It exists because the Medicare statute requires CMS to annually publish:
MA growth rates
Benchmarks
Risk adjustment model updates (CMS-HCC, RxHCC)
Coding intensity adjustment
Normalization factors
Part D benefit parameters
Payment methodology details
It is explicitly called:
“Advance Notice of Methodological Changes for CY 2027 MA Capitation Rates and Part C and Part D Payment Policies”
2026 CMS 0126 Part C D 2027 Pro…
This is about how CMS calculates plan payments, not beneficiary rules.
👉 Outcome: Becomes the Rate Announcement (April 6, 2026), not a Final Rule.
| Topic | Lives in November Rule? | Lives in January Notice? |
|---|---|---|
| Marketing restrictions | ✔ | ✖ |
| TPMO oversight | ✔ | ✖ |
| SNP eligibility policy | ✔ | ✖ |
| Part D benefit structure (legal codification) | ✔ | ✖ |
| MA benchmark math | ✖ | ✔ |
| Risk adjustment model calibration | ✖ | ✔ |
| Coding pattern adjustment | ✖ | ✔ |
| Star Ratings payment impact | ✖ (policy only) | ✔ |
| Part D risk model updates | ✖ | ✔ |
Medicare Advantage has two separate legal mechanisms:
Made through APA rulemaking → CFR changes
(That’s your November document)
Made through annual rate notice → Actuarial / statutory process
(That’s your January document)
This split goes back to the Medicare Modernization Act (2003) and earlier managed care law. MA payment is treated like insurance rate-setting, not standard rulemaking.
Both documents mention:
Star Ratings
Part D
Risk adjustment
“Part C and D policies”
But:
November = WHAT the program requires
January = HOW CMS calculates money
Example:
November rule: May change Star Ratings measure set.
January notice: Applies Star Ratings to determine bonus percentages in benchmarks.
Different step in pipeline.
| Stage | Document |
|---|---|
| Rules of football | November Proposed Rule |
| How the referee calculates the score | January Advance Notice |
You are not seeing two competing proposed rules.
You are seeing:
1️⃣ Policy rulemaking (regulations)
2025 CMS 1128 90FR54894 137p Pr…
2️⃣ Annual payment methodology notice (actuarial rate process)
2026 CMS 0126 Part C D 2027 Pro…
They run in parallel every year and converge when plans bid.
If you’d like, next I can:
A) Map how a single issue (like Star Ratings) flows through BOTH pipelines
B) Show which parts of the January notice plans actually care about most
C) Explain how Part D IRA changes appear in each document
Your confusion = totally rational. CMS’s architecture is… vintage 1997 managed care law meets modern Medicare.
#######
Biden year 4 Med Adv policymaking had many pages about insuring coverage parity, e.g. validity of LCDs in Part C. Trump Admin went radio silent. No change, I assume?
####
Short answer: No — neither document meaningfully addresses traditional coverage policy (NCDs, benefit categories, Part A/B service definitions, or Part C “same benefit” scope issues).
You’re picking up on a real shift.
This rule is about MA/Part D program operations, not Medicare coverage law. Its focus areas are:
Part D benefit redesign mechanics
Star Ratings methodology
Marketing / TPMO oversight
Enrollment and SNP policy
Reporting, MLR, network adequacy processes
It amends 42 CFR 422 & 423, but does not reopen:
National Coverage Determinations (NCDs)
“Reasonable and necessary” standards
Part A vs Part B benefit definitions
MA “uniformity” or “same benefit as FFS” doctrine in any substantive way
Even the IRA Part D sections are benefit design financing, not coverage scope.
Source: The rule’s purpose is described as revising MA and Part D regulations on Star Ratings, marketing, enrollment, drug benefit structure, etc. — not coverage policy.
This one is even further removed. It is a payment mechanics document:
Benchmarks
Growth rates
Risk adjustment models
Coding intensity adjustment
Part D bid parameters
It explicitly describes itself as updating capitation rates and risk models, not benefits.
No discussion of:
Coverage standards
Clinical benefit categories
Concordance with FFS NCDs
Scope of Part A/B services
You’re right — prior administrations (especially 2021–2024) used MA rulemaking to discuss:
Guardrails around MA “reinterpretation” of coverage
FFS-equivalence expectations
Clinical evidence standards
Prior auth alignment
That language is largely absent here. The focus is:
Program integrity, payment calibration, and administrative simplification
not
Clinical coverage policy
| Topic | November Rule | January Notice |
|---|---|---|
| NCDs | ❌ | ❌ |
| Part A/B benefit definitions | ❌ | ❌ |
| “Reasonable and necessary” | ❌ | ❌ |
| MA must mirror FFS benefits | Only background law, not policy | ❌ |
| Coverage expansions or restrictions | ❌ | ❌ |
These documents operate in the MA/Part D administrative and financial domain, not the Medicare coverage doctrine domain.
Your observation = correct: coverage policy discussion has moved offstage in these CY 2027 materials.
Stephane Budel and DeciBio post a four-page summary of some of their recent research on the moving l ine between tools and clinical diagnostics. Find their post and PDF here.
Writing,
Life sciences research tools (LSRT) companies are pivoting hard toward clinical and translational value pools:
- Illumina: ~60% of consumables now clinical...
- 10x Genomics: launching a CLIA lab...
- [more examples]
Adding, “clinicalization” requires a significant business model transformation. Reimbursement, quality systems, go-to-market, partnerships… it all changes...
We hear about big payors like Optum buying up primary care practices - but how big is that, really?
See remarks at Linked In by Brian Miller - here - and see a new paper in Health Affairs Scholar by Adler et al. here.
Ironically this paper shows that Optum is not a competitive concern in primary care markets:
-Payers own 4% and Optum owns 2.71% of all primary care practices
-Optum max market share 35% in 3 counties
-There are over 3,200 counties
It seems obvious to me that "value based care" and "managed care" won't solve the US outlier position for healthcare costs. But we've got a headline article in Health Affairs that makes the point.
Update: see Peterson Health Technology white paper on "value based contracting playbook" here.
See Health Affairs January 23, 2026:
Value-Based Payment And Managed Care Will Not Solve The Affordability Crisis
Ryan, Rooke-Ley, Berenson (Brown Univ. & Urban Institute)
Among the unsurprising data shown here, Bond and colleagues in 2025 JAMA found that ACOs saved $300/pt after six years. Ryan et al. point out that in the target year, 2019, that was $500M savings compared to $800B spending - a small fraction of 1%. Open access and good reading.
The authors give some airtime to concerns that technology drives healthcare costs. We could go vastly farther in encouraging technologies that save healthcare costs, instead of making them almost impossible to adopt.
___(For one promising example see PaigePredict here. I'm not promoting this particular example, but rather, it stands-in for a whole class of beyond-the-horizon breakthrough efficient technologies that deserve attention, also here.)
The article doesn't have an abstract but here's a 200 word summary [Chat GPT 5.2]:
Policymakers keep backing value-based payment (VBP) as the fix for U.S. health care costs, shifting from fee-for-service to risk contracts like ACOs and Medicare Advantage.
But decades of experience show minimal or no real savings. Even the most optimistic recent ACO studies suggest Medicare savings of only 0.06–0.13%, and those estimates are likely inflated by patient selection and diagnostic upcoding, not true efficiency. Meanwhile, CMS bonus payments and gaming may mean net losses.
Claims that VBP indirectly slowed national spending growth (“spillovers”) don’t hold up. The U.S. slowdown began before the ACA and occurred across OECD countries with very different payment systems. The more plausible driver was a temporary lull in diffusion of expensive medical technologies, not payment reform.
Conceptually, VBP targets the wrong problem. U.S. costs are not mainly due to excess use of routine services; compared to peers, Americans often use fewer hospital days and doctor visits. The real drivers are
Bottom line: VBP outsources cost control to intermediaries while leaving the core pricing and coverage decisions untouched—so it can’t solve long-term cost growth.
Header: More MACs Block Medicare Coverage of Advanced Imaging in Radiology.
##
I've often mentioned that experts interested in the growth of AI in pathology should look to advances in AI in radiology for signposts. See for example a 2022 blog on the topic here. (Also here). This may be even more important now, as AMA is advancing ideas originally thought up in the context of radiology, to apply to both radiology and pathology, which alarms me. 2026 blog here.
The CGS MAC brought out and finalizing, and the NGS MAC has just proposed, an LCD that doesn't pay for any kind of advancd diagnostic algorithms in MRI. Amazingly, the CGS MAC LCD got ZERO comments.
The LCDs are titled, "Automated Detection and Quantification of Brain MRIs."
The use of AI augmented technology does not eliminate the need for appropriate reader training, and the results are not intended to be used independently of physician interpretation. National Government Services does not consider automated detection and quantification of brain MRI imaging to be reasonable and necessary.Investigations have explored the potential of automated quantification technology for the evaluation of ARIA, MS, TBI, epilepsy, brain tumors, and other neurologic conditions. However, progress has been limited by the lack of established measurement standards and insufficient access to large, diverse datasets for training these tools....Without standardized data, it is difficult to determine whether subtle changes reflect true pathology or normal variation. At this time, there is insufficient evidence of clinical utility or validity, and use of this technology is considered investigational and therefore not covered. National Government Services will continue to monitor the evolving research related to these devices.
Continuing a saga that goes back 20 years, FDA goes after some companies not for their in-house LDTs, but for their collection kits shipped through the mails. (See a 2013 warning letter from FDA to 23andMe disallowing a non-approved saliva collection device.) (See a 2025 warning letter from FDA to DRG instruments that a saliva collection device was sold RUO, but with text referring to its "patient samples.")
According to a new item at 360DX, the FDA recently issued warning letters to:
My eye tagged a new article in NEJM about Congress granting coverage privilege to a class of interventions, like those which have FDA breakthrough status. That's Kadakia et al. 2026.
But I'll devote 3/4 of this blog to discussing their citation #2, Kang et al., 2025, on CMS NGS claim denials.
Along the way, I noticed several interesting recent titles which I'll highlight in fine print:
Typically, a PLA code quarterly agenda involves about 20-25 codes.
This quarter, I count 53 codes that are either new or in revision. There are another five or so that are being deleted.
The codes were released around January 13 (sorry I didn't post) and the PLA committee is meeting and voting today, January 20. The 50-some codes will appear as a "consent calendar" or "single vote" at the AMA CPT meeting in Palm Springs, February 5-6, 2026.
PLA page here
https://www.ama-assn.org/practice-management/cpt/cpt-pla-codes
Agenda, 8 page PDF, here:
https://www.ama-assn.org/system/files/february-2026-pla-public-agenda.pdf
Back up here.
AI resources like Chat GPT get better and better at scouring the internet.
Can it find molecular diagnostics funding events? In the last month or two, I had noticed the SPAC for Freenome; the Series B for Mercy BioAnalytics; the Series B for Precede Biosciences.
I asked Chat GPT if it could find those and any others from November 1 to January 13.
Today we ask how large gene-panel lab runs fit to "what's ordered, medically necessary, and billed." Surprise - The OIG was there in 1997-1998.
And, today we ask about pathologist orders to complete a case - such as in this MolDx article. In 2007-2008, OIG was already there, with concepts like the in-office exception and its guardrails.
See the two-part essay below for both topics.
On January 12, 2026, Freenome announced an expansion of its successful collaboration with NVIDIA, bringing artificial intelligence to early cancer detection.
See the press release here. See the Freenome Linked In posting, where you can track real-time comments and feedback.
Freenome was featured in NVIDIA's "Healthcare Diagnostics Industry Brief" a couple months back - here. See the 9 page PDF white paper here. [Highlights from Freenome, Natera, Guardant, Sofia,
###
AI writes in brief:
In Tuesday's business news, January 13, Astrazeneca announces it is acquiring Modella AI.
Here's some background (Chat GPT 5) about Modella:
From time to time I post a HUMOR essay, entirely separate from my day job.
On a podcast, I heard remarks about the 1950s programming language COBOL, and I wondered, could you build a Large Language Model out of COBOL?
_______
* So, I asked Chat GPT to write a two-page business plan for a novel "AI" based entirely on COBOL. And ask for $100M Series A.
And then...
* It's assigned to a young VC associate to write-up - and he's heard his VC Partner likes the idea, so he has to evaluate it cautiously,
And then...
* The company comes back two years later, now ready for a $1 Billion Series B.
____
In the blog, that Series B $1B essay is followed by a "serious" one, where Chat GPT discusses why you don't want to build your new LLM out of COBOL.
____
Find these humor essays here:
https://bqwebpage.blogspot.com/2026/01/ai-humor-investing-in-new-llm-based-on.html
_____
Recent changes in AMA CPT policy
risk unintentionally undermining U.S. leadership in medical artificial
intelligence, particularly in digital pathology and genomics, two areas
where the United States currently holds a global advantage. The issue is not
opposition to AI itself, but a misalignment between CPT’s conceptual framework
and how modern diagnostic AI actually works.
Much of CPT’s thinking about AI
appears to be shaped by radiology use cases, where AI functions as
an adjunct: highlighting suspicious areas, prioritizing worklists,
or prompting a second look by a physician. In that context, CPT’s
principles—resistance to autonomous AI billing, concern about double payment,
and protection of physician RVUs—are internally coherent. However, those same
principles do not translate to computational pathology or genomics, where
AI is not advisory. In these domains, AI is the diagnostic act.
In genomics, billions of sequenced
DNA fragments are algorithmically analyzed to identify mutations that directly
determine therapy selection. There is (usually) no meaningful point at which a
pathologist’s subjective judgment is inserted into the result. AMA would call that autonomous results without physician work - and it's been like that for years. Digital pathology
AI increasingly operates the same way: validated models extract prognostic and
predictive signals from routine histopathology images that no human can
replicate. Treating these outputs as some quirky bird that is “autonomous without physician work” bad misunderstands the whole paradigm of clinical lab tests.
With regard to digital pathology, CPT’s current structure has no
natural home for such services. PLA codes require creation of new biomarkers;
Category I codes prohibit proprietary diagnostics; MAAA explicitly excludes
image-based algorithms (CPT mtg 10/2025); Category III codes are typically unpaid; and the
proposed CMAA category risks becoming a registry rather than a reimbursement
pathway. The result is that “upload image → download report” diagnostic
models—arguably among the most scalable and cost-efficient forms of medical
AI, exactly what the head of CMS wants—are structurally blocked from payment under PLA rules posted 12/31.
The commercial consequences are
already visible. Rather than launching standalone AI diagnostics, companies are
embedding AI invisibly inside existing reimbursed tests, not because the AI
lacks value, but because independent reimbursement is unattainable. Smaller
innovators without large corporate partners are likely to exit or never form.
At every step, these outcomes run counter to
national policy goals. Our system now discourages information-efficient diagnostics,
rewards molecular redundancy, and shifts AI innovation offshore or into
non-clinical domains. The risk is not immediate harm to patients, but long-term
erosion of U.S. leadership in healthcare AI at precisely the moment federal
policy seeks to strengthen it.
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.]
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:
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 a Linked-In article by Joseph Steward here.
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.
_____________________
More and more frequently we're seeing Tempus, Caris, and others bringing together RNA transcriptomes with large-panel or genomic tumor DNA sequencing.
Here's a new review at Nature Review Clinical Oncology that brings you up-to-date. And, we'll add a section, "Takeaways for Payors."
###
AI CORNER
###
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."
###
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. 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:
expressed gene fusions (including unknown partners),
RNA-level confirmation of DNA variants,
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.
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
.
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.
Across these papers, several shared conclusions emerge:
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.
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.
RNA-seq consolidates what would otherwise require multiple assays—fusion testing, expression profiling, splice analysis, and confirmatory variant assessment—into a single workflow.
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.
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.
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.
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.
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.
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.