When does MolDx change a non-coverage proposed LCD into a coverage final LCD?
Famously, this happened once in the last few years, regarding the PRISM-RA test, which provides some help when choosing advanced medications in arthritis. (However, MolDx remarked in the final that it still found some of the evidence marginal.) (Draft, 9/2022). (Final, 8/2023).
It's an evergreen topic; for example, MolDx just proposed an LCD not covering a risk stratification test in renal cancer, but the comment period is open and the final LCD could be more positive.
On September 2, I published an article about MolDx's decision to postpone finalizing a non-coverage proposed LCD,DL40246, on biomarker risk stratification in DCIS. That particular saga has been going on for a while, and the September 2 article contains a number of links. Test name, DCISionRT, developed by, Prelude Dx.
##
The story continues: See a new posting at Linked In by Dan Forche, CEO of Prelude Dx ("The DCIS Testing Company.") It links to a September 8 Forbes article by Geri Stengel that updates readers on the topic, incuding a MAC contractor advisory meeting in 2024 and a new one pending October 13. See also Stengel at Substack on September 10.
##
I won't attempt to re-summarize everything linked back on September 2, but below Chat GPT summarizes what's new from Forche and Stengel.
##
AI CORNER
##
When Is the Evidence “Good Enough”? The DCISionRT Story Continues
When does MolDX turn a proposed noncoverage LCD into a positive final LCD?
It can happen. A recent example was the PrismRA test, which helps guide selection of advanced therapies in rheumatoid arthritis. MolDX proposed noncoverage in September 2022 but issued a favorable final LCD in August 2023—although the final policy still described portions of the evidence as marginal.
The question is evergreen. MolDX has just proposed noncoverage of a risk-stratification test in renal cancer, for example, and that comment period remains open.
The Back Story in Brief
On September 2, I discussed MolDX’s still-unfinished LCD for biomarker risk stratification in ductal carcinoma in situ, or DCIS. The policy is particularly important for PreludeDx and its DCISionRT test.
MolDX issued proposed LCD DL40246 in July 2025. The draft would deny coverage for biomarker tests intended to estimate DCIS recurrence risk and predict the benefit of radiation therapy.
Ordinarily, the next step would be a final LCD. Instead, in July 2026, the MolDX contractors announced that they were delaying finalization “given the impact of this determination.” They promised additional information in the coming months.
That was unusual. Now we know at least part of what comes next: MolDX has scheduled another Contractor Advisory Committee meeting for October 13.
Two New Articles by Geri Stengel
Geri Stengel has now examined the controversy in two closely related—but distinct—articles.
Her September 8 Forbes article, “Medicare Will Pay for Radiation. Will It Pay to Know Who Needs It?”, tells the story primarily from the patient and clinical perspective.
DCISionRT combines tumor biology with clinical and pathological factors to estimate recurrence risk and potential radiation benefit. The central point is not simply that the test reduces radiation use. In a 2024 clinical-utility study involving 2,007 women at 63 sites, DCISionRT changed physicians’ radiation recommendations in 38% of cases. Some women initially headed toward radiation were advised to avoid it, while some initially thought unlikely to need radiation were advised to receive it.
Thus, the test can move treatment in either direction. The goal is neither more treatment nor less treatment, but better-directed treatment.
Stengel also highlights MolDX’s central objection. Changing a physician’s recommendation is not necessarily the same as proving that the resulting decision improves patient outcomes. MolDX argues that the evidence has not yet shown convincingly enough that biomarker testing adds value beyond conventional clinicopathologic factors and freely available risk tools.
Her September 10 Substack essay, “Who Decides When the Medical Evidence Is Good Enough?”, goes more deeply into that evidence dispute and its business implications.
Stengel begins with a striking fact from the July 2024 MolDX advisory meeting: six of seven breast surgeons and radiation oncologists believed that sufficient evidence supported biomarker testing to help determine which women with DCIS could forgo radiation. Nonetheless, MolDX subsequently proposed noncoverage.
The Substack article also explores several layers of “proof” that are too often blended together: analytical validity, clinical validity, clinical utility, prediction of radiation benefit, and superiority to existing clinical tools. It discusses MolDX’s concern about analyses using Decision Score thresholds of 3.0 and greater than 2.8, as well as PreludeDx’s response that its reported clinical cutoff did not change.
Most importantly, Stengel identifies the commercial bind created when the evidentiary bar remains uncertain. Evidence affects reimbursement; reimbursement affects adoption; adoption affects further evidence generation; and all three affect financing. PreludeDx says it submitted its MolDX technical assessment in 2020. Six years later, the coverage question remains unresolved.
Forche: Why Another Advisory Panel?
On September 21, PreludeDx President and CEO Dan Forche highlighted Stengel’s reporting in a LinkedIn post.
Forche focuses on the disconnect between the July 2024 advisory panel and the subsequent draft LCD. If six of seven physician experts found the evidence sufficient, he asks, why was that strong consensus not reflected in the proposed policy? And why is MolDX convening a second CAC?
The LinkedIn post does not introduce a new clinical study. Rather, it sharpens the procedural and governance questions raised by the two Stengel articles. Who decides what evidence is sufficient? How transparent is the standard? And can a diagnostic developer know the standard early enough to design the studies that will eventually be demanded?
The Next Date to Watch
The next major event is the October 13 CAC meeting.
A second advisory panel does not guarantee that MolDX will reverse the proposed noncoverage policy. Nor does a favorable expert vote bind the contractors. But reconvening a CAC after the draft LCD, public comments, new publications, and the earlier 6–1 vote is plainly significant.
The dispute is no longer about whether DCISionRT has evidence. It clearly does. The dispute is about what kind of evidence MolDX will accept, how much incremental value must be demonstrated over conventional tools, and whether the evidentiary finish line has been defined clearly enough for innovators, physicians, patients—and investors—to see it.
# # #
SIDEBAR:
Advocates Dive Deeper Into the World of Evidence Appraisal
The new advocacy does something unusual: it ventures well beyond familiar arguments about patient access, innovation, and the harms of noncoverage. The Stengel essays and Forche’s response engage directly with the machinery of scientific evidence appraisal.
The dispute is not simply whether DCISionRT has published evidence. It plainly does. The harder questions are how different forms of evidence should be assessed and weighted. How much weight belongs on clinical validity, treatment-decision impact, prediction of radiation benefit, and long-term patient outcomes? Must a test merely add useful biological information, or must it demonstrate superiority to every available clinicopathologic model? How should retrospective analyses of randomized trial specimens be valued? And when an analysis performs differently at cutoffs of 3.0 and 2.8, is that evidence of a moving threshold—or a legitimate effort to understand where predictive information is strongest?
These are questions of scientific judgment, not simple boxes that can be checked mechanically.
The essays also expose the distinction between evidence appraisal and coverage policy. MolDX can reasonably acknowledge that a test is informative while concluding that it has not demonstrated enough incremental value to merit Medicare payment. Conversely, physicians may reasonably believe that evidence capable of changing treatment in both directions has substantial clinical utility, even before a new prospective randomized trial reports long-term outcomes.
Ultimately, “sufficient evidence” is not a self-defining scientific fact. It is a policy judgment produced by selecting endpoints, ranking study designs, weighing uncertainty, choosing comparators, and deciding how much residual risk Medicare will tolerate.
That is why the 6–1 vote at the 2024 advisory panel matters. It does not dictate the correct coverage decision, but it demonstrates that experienced clinicians can examine the evidence and reach a markedly different judgment from MolDX. The question for October 13 is therefore not merely whether more evidence exists. It is how MolDX will weight the evidence already in hand—and whether it can explain clearly where the threshold for “good enough” actually lies.
