MolDx held a 2021 CAC on Barrett's molecular testing, published a 2023 non coverage LCD L39256, and have just held a new CAC on the same topic, with new questions for 2025.
First, I asked Chat GPT what it thought about the 2025 questions - alone.
###
Based on the preamble and the 17 questions from the MolDX CAC meeting document, here’s how the tone and framing read:
1. Evidence-Seeking vs. Evidence-Challenging
The questions are structured in a balanced way. They ask both about shortcomings of current EGD surveillance (Q1–4) and about how non-invasive tests might add value (Q5–8). They also probe whether existing molecular tests demonstrate utility beyond current surveillance (Q10–11) and how real-world evidence compares to trials like BEST1–3 (Q15).
That balance suggests the MACs are genuinely canvassing for arguments both for and against current non-coverage.
2. Signs of Skepticism
Several questions highlight limitations and risks:
-
Concerns about false positives/false negatives (Q6, Q8).
-
Concerns about operator error and variable failure rates (Q16).
-
Questions about whether biomarkers are merely correlative vs. causative (Q13).
-
Questions about applicability of UK trial data to the U.S. (Q15).
These lean toward reinforcing a cautious stance.
3. Signs of Openness
Other questions explicitly invite experts to articulate positive use cases:
-
Q5 and Q7 ask how management would change if rule-in or rule-out tests were reliable.
-
Q10 asks what evidence is the “most compelling case” to support their use.
-
Q11 probes whether such tests identify more patients than traditional surveillance.
These provide open doors for panelists to argue for clinical utility.
4. Strategic Reading
Given MolDX’s history, the default position is skepticism (hence the existing non-coverage LCD). However, the framing is not one-sided—questions are written so that if panelists bring strong arguments and data, MACs can justify a shift.
The fact that they are holding a multi-jurisdictional CAC meeting (rather than simply re-affirming non-coverage internally) suggests they are at least willing to re-examine the evidence landscape.
Bottom Line
-
It is not a one-directional “shore up the no” exercise—too many questions invite positive evidence.
-
At the same time, the skeptical undertone (trial generalizability, operator error, biomarker causality) shows they are wary of over-promising.
👉 I would call it too close to call—but leaning toward B (openness to conversion) if stakeholders present convincing data, because otherwise they could have let the negative LCD stand without convening a CAC.
###
Let's step back a step. Yes, the 2025 questions are a mix of more critical and more favorable. But is't the mere fact they have so many intricate questions, a factor in itself? Let's look more broadly. I attached their 2021 Barrett CAC Questions, their 2023 LCD which took input from the CAC meeting, and their new 2025 questions just released. What can we now say about their likely 2025 positioning, for which the 2025 questions are the most recent evidence or clues, but the 2025 questions can be viewed in context of the 2023 non coverage LCD and the 2021 older questions.
###
Great question. Zooming out across cycles (2021 CAC → 2023 LCD → 2025 CAC), here’s the read:
Trajectory & tone shift. The 2021 CAC prompts were relatively “up-stream”—can minimally invasive biomarkers plausibly replace or triage endoscopy; what is the swallow success rate; how often would you repeat testing; how many cases would be missed and with what consequences; what settings might use these assays. That’s exploratory framing, aimed at scoping basic feasibility and early risks.
By 2023, MolDX crystallized a non-coverage LCD but—crucially—wrote a “path to yes” into the policy: tests must demonstrate analytical validity, clinical validity, and clinical utility in the validated population, and show equivalence or superiority to accepted methods for the same intended use; the LCD explicitly says no existing tests yet meet that bar, while pledging to monitor the evidence. This is a “not yet, but tell us when you can meet these conditions” posture—not a philosophical “no.”
The new 2025 CAC questions are more operational and decision-analytic. They drill into: where EGD/biopsy underperforms (indeterminates, sampling error), whether a test could credibly serve rule-out or rule-in roles and how management would change, what accuracy thresholds would be acceptable (including the harms of false results), and whether UK BEST1-3 findings generalize to US practice. They also probe operator error/failure rates and performance across long vs short segment disease and low- vs high-grade dysplasia. That is the language of coverage criteria design (how would we use it, with what cut-points, in whom, and how would it change care), not just abstract skepticism.
Does the sheer intricacy of 2025 questions “mean something”? Yes. In MolDX practice, convening a multi-jurisdiction CAC with granular, threshold-seeking questions is typically about building an administrative record robust enough to support either (1) continuing non-coverage with detailed rationale, or (2) pivoting to limited/conditional coverage with explicit guardrails (population, role as rule-in/rule-out, accuracy minima, operator/process controls, frequency limits, and documented management change). The content of the questions—rule-in/rule-out roles, “would you still do an EGD,” generalizability beyond BEST3, operator failure rates—reads like they’re mapping the exact levers they’d need to write a positive policy if the evidence is there.
So what’s their likely 2025 positioning?
-
Still formally negative today (LCD in effect through March 13, 2025, states non-coverage).
-
Directionally, “cautious-to-open.” Compared to 2021’s feasibility focus, the 2025 prompts now interrogate decision thresholds, management impact, and US-specific implementation risks—exactly what you’d gather to justify a narrow “yes”.
-
Bottom line: Too close to call, but the question set leans toward “seeking a path to yes.” If the panel can credibly specify use-cases (e.g., high-NPV rule-out to avoid unnecessary EGD in defined risk strata; or high-PPV rule-in to accelerate EET/confirmatory EGD) and address operator failure, U.S. validation (not just BEST3), and demonstrable management change, MolDX now has the scaffolding to convert to a limited-coverage policy. Absent that, the record they’re building will be equally serviceable to shore up the “no” under the 2023 criteria.
What would likely flip them to a (narrow) “yes” in 2025–26?
-
US-based validation in the intended-use population showing prespecified accuracy targets aligned to the test’s role (very high NPV for rule-out; strong PPV for rule-in) and stable performance across operators/devices.
-
Clear clinical utility: audited evidence that test-guided pathways reduce unnecessary EGDs or increase detection of treatable dysplasia vs usual care—and that clinicians actually change management when results are ±.
-
Addressing generalizability gaps from BEST1-3 to U.S. practice and payer pathways, plus failure-rate mitigation and long- vs short-segment performance.
If you want, I can convert the 17 items into a one-page “signal map” (which questions imply coverage-framework design vs cautionary hurdles) to brief clients quickly.
Here’s a one-page signal map for the 2025 MolDX Barrett’s / gastric metaplasia CAC questions. It clusters the 17 prompts into two categories:
MolDX 2025 CAC Questions: Signal Map
Signals of Openness (Framework for Limited Coverage)
These questions sketch how MolDX could structure a positive LCD if evidence matures:
-
Alt. to EGD / Management Impact
-
Q4: What are the limitations of EGD?
-
Q5 & Q7: How would reliable “rule-out” or “rule-in” tests change management?
-
Q6 & Q8: What accuracy thresholds would be sufficient? Balance of sensitivity vs specificity.
-
-
Evidence & Utility
-
Q9: Have you used any such test—your experience?
-
Q10: What is the most compelling evidence supporting use?
-
Q11: Have these tests identified more patients than EGD? Reasonable expectation for improved early detection?
-
-
Implementation Parameters
-
Q14: Accuracy in long vs short segment disease.
-
Q15: Relevance of BEST1–3 trial data to U.S. practice.
-
Signals of Skepticism (Hurdles to Overcome)
These highlight risks MolDX could cite to sustain non-coverage:
-
Process Weaknesses
-
Q1–3: Referral gaps and compliance with EGD (highlighting existing care shortfalls not solved by molecular tests).
-
Q12: Limitations in existing data.
-
Q13: Concerns about measuring correlates vs causal markers.
-
-
Operational Risks
-
Q16: Operator error and variable sample failure rates.
-
Q17: Accuracy in detecting low- vs high-grade dysplasia.
-
Overall Signal
-
2021 CAC = exploratory, “what if.”
-
2023 LCD = non-coverage but left an explicit path-to-yes (analytical + clinical validity + clinical utility in the intended population).
-
2025 CAC = intricate, threshold-seeking. Half the questions map to designing a limited coverage policy (use-case, cut-offs, clinical management change). The other half memorialize reasons to continue “no.”
👉 Net read: Too close to call, but the complexity and threshold-focused framing signals MolDX is actively testing a pathway to conditional coverage if new U.S. data can satisfy their criteria.
