MolDx MACs finalize an LCD, proposed last summer, on "Biomarker Testing for Risk Stratification in Metabolic Dysfunction-Associated Steatotic Liver Disease and Metabolic Dysfunction-Associated Steatohepatitis."
See draft DL40187 here. See final L40187 here.
See request letter #1 9/19/2023 here. #2 7/1/2024 here. Posted as MA Connelly and H Imai, respectively; the latter on Siemens letterhead.
# # #
1. Request letter 1
Letter 1 is a Labcorp request, dated September 18, 2023, asking MolDX/Palmetto for a new LCD for NIS4, described as a non-invasive blood test to identify at-risk NASH in patients with features of metabolic syndrome. The letter frames the request as a Medicare Part B diagnostic clinical laboratory test and argues that access would improve treatment decisions and outcomes.
2. Request letter 2
Letter 2 is a Siemens Healthineers request, dated July 1, 2024, asking MolDX for an LCD for proteomics testing for MASH/NASH, specifically the ADVIA Centaur Enhanced Liver Fibrosis / ELF test. Siemens emphasizes that ELF had FDA De Novo authorization in 2021, had been paid by Medicare since launch, had a CPT Category I code, and was at risk of payment interruption after MolDX Z-code/payment-policy changes (from open payment of this type of proteomics code, to, control under MolDx).
The ask is urgent coverage to preserve Medicare access.
3. Main result of the Final LCD
The final LCD creates limited MolDX coverage for molecular or proteomic biomarker testing used to evaluate liver fibrosis in MASLD/MASH, but not blanket coverage for any named test.
The practical result is a category-level coverage pathway: tests can be covered if
- the patient is an adult with MASLD/MASH-compatible liver dysfunction,
- has not been ruled low-risk by primary assessment such as FIB-4,
- lacks recent definitive liver-stiffness imaging,
- the result will guide a documented management decision,
- repeat testing limits are met,
- and the test passes MolDX technical assessment.
The LCD is future effective for services on or after August 10, 2026.
The billing article A60205 lists codes 81479 (of course!), 81599, and specific codes 81517 $176, 0003M $503, 0166U $503, 0344U $792, 0468U $251.
0003M is NASH FibroSure. 81517 is Siemens ELF. 0166U is LiverFAS Fibronostics. 0344U is OWLiver CIMA. 0468U is NASH NEXT LabCorp.
#
#
Draft vs. final LCD
The redline shows that the final LCD made relatively few changes to the evidence narrative or bibliography, but it made several important changes to the coverage language and coverage criteria.
CMS/MolDX characterizes the changes as clarifications based on public comments (see Q&A article on comments received), and the redline supports that: the final version does not reverse the draft’s basic policy, but it makes the policy more precise and operational.
 |
| Example redline (draft-to-final). Click to enlarge. |
First, the final LCD slightly narrows the policy’s framing. The draft described coverage for molecular or proteomic biomarkers for “diagnosis and management” of liver fibrosis; the final describes testing as “aiding the evaluation” of liver fibrosis. That is a softer and more careful coverage formulation: the test is not positioned as independently diagnosing or managing disease, but as contributing to clinical evaluation.
The final also drops “risk stratification” from the title, although risk stratification remains the policy’s functional purpose.
Second, the patient-entry criterion was tightened. The draft allowed an adult with “clinical suspicion or diagnosis” of MASLD or MASH. The final requires clinical evidence of liver dysfunction compatible with MASLD or MASH, based on current professional society guidelines. This is a more concrete medical-necessity gate than mere suspicion, and it pushes the policy toward documented clinical context rather than broad screening use.
Third, the final clarifies the required first-line assessment. The draft required that a primary risk assessment based on non-molecular/non-proteomic laboratory testing did not indicate low risk, using FIB-4 ≥1.3 as an example. The final adds that this primary assessment should be based on integration of clinical evaluation and ordinary laboratory testing, again with FIB-4 ≥1.3 as an example. This makes the LCD less like a simple FIB-4 threshold rule and more like a guideline-based clinical triage requirement.
Fourth, the imaging requirement was sharpened. The draft required liver stiffness measurement by imaging to be “indeterminate or not performed.” The final adds a time frame: LSM by imaging must not have been performed within the prior 12 months, or the results must have been indeterminate. This is an important practical change, because it prevents routine blood-based biomarker coverage when a recent adequate VCTE/MRE-type assessment is already available.
Fifth, the MolDX technical assessment criteria were reorganized and refined. The draft had separate numbered criteria for clinical validity, algorithm validation, technical assessment, and similar-test performance. The final consolidates these under a single MolDX TA requirement. It changes “analytes or profiles” to “novel analytes,” changes reproducibility across study cohorts to comparability across subgroups, and adds an explicit requirement that the test demonstrate equivalent or superior performance to analytes already covered under the policy for the same intended use. The final also removes the draft’s more pointed warning that newer, higher-performing tests could make older tests “no longer compliant.”
The evidence discussion is otherwise largely preserved. The final keeps the same overall rationale: MASLD/MASH is common; fibrosis stage is clinically important; biopsy is impractical as a broad first-line tool; FIB-4 is useful but insufficient as a standalone test; imaging-based LSM is useful but access- and performance-limited; and selected blood-based molecular/proteomic tests may be reasonable as second-line tools after primary triage. The final also keeps the key caution that mentioning specific biomarkers such as ELF, NIS4/NIS2+, or MASEF in the evidence review does not itself imply coverage; coverage depends on satisfying all LCD criteria and completing MolDX TA.
#
The Comments article, A60449, carefully walks through comments received.
Summary of the Comments Article A60449
The comments article reports the public-comment process for DL40187, the proposed MolDX LCD on biomarker testing in MASLD/MASH. The comment period ran from August 28, 2025 to October 12, 2025; the final LCD notice period began June 25, 2026, with the LCD effective August 10, 2026. The title was revised by dropping “Risk Stratification,” becoming simply “MolDX: Biomarker Testing in Metabolic Dysfunction-Associated Steatotic Liver Disease and Metabolic Dysfunction-Associated Steatohepatitis.”
The article contains 10 comments, all broadly supportive of finalizing the LCD. The commenters were AMP, Roche Diagnostics, Labcorp, MetaSight Diagnostics, CAP, ACLA, Siemens Healthineers, AdvaMed, Guardant, and Quest. The repeated theme was: “We support coverage, but please reduce ambiguity or unnecessary access barriers.”
The most important successful industry/commenter request was removal of the draft sentence saying that newer tests with significantly improved performance might render older tests non-compliant. AMP, Roche, Labcorp, ACLA, and Guardant all objected to this concept as unpredictable and potentially disruptive. MolDX removed that sentence, while reserving the right to request further TA documentation if needed.
The second major theme was FIB-4 gatekeeping. Roche and MetaSight argued that FIB-4 should not be treated as a rigid threshold, because clinical findings, metabolic risk, abnormal liver enzymes, or other evidence may justify further testing even when FIB-4 is low. MolDX did not fully accept a broad bypass of FIB-4, but it modified Criterion 2 to state that the primary risk assessment includes clinical evaluation plus non-molecular/non-proteomic laboratory testing. In other words, MolDX softened the “FIB-4 ≥1.3” framing from a simple cutoff into a clinical-risk assessment concept.
The third major theme was the liver stiffness measurement / VCTE condition. Siemens and AdvaMed argued that requiring LSM imaging to be indeterminate or not performed would unnecessarily privilege elastography, which can be unavailable, operator-dependent, affected by BMI and other factors, or delayed. MolDX did not remove the LSM condition. Instead, it revised the syntax to clarify that blood-based testing is precluded mainly when there is a recent definitive LSM result; the final version adds the practical “within the prior 12 months” formulation.
The fourth theme was biopsy and repeat testing. AMP and Guardant objected that the draft could block testing after a biopsy even when biopsy findings were inconclusive, unsuccessful, or possibly false-negative. MolDX revised Criterion 5 to clarify that testing is barred within 12 months after a liver biopsy only when there is definitive histologic interpretation by a pathologist.
The fifth theme was technical assessment and clinical-validity evidence. AMP and Guardant worried that the draft could be read to require each lab to publish its own studies or to require multiple publications/cohorts. MolDX reorganized Criteria 6-9, added the word “novel” for analytes, clarified that prior publications may establish clinical validity for the same test, and revised the language toward comparability/equivalence across subgroups rather than a more rigid reproducibility-across-cohorts formulation.
There were also coding comments. Roche suggested additional ICD-10-CM codes reflecting MASLD clinical contexts, including abnormal liver function studies, type 2 diabetes, mixed hyperlipidemia, and unspecified hyperlipidemia. MolDX thanked Roche and stated that the proposed codes were added to the Billing and Coding Article.
Bottom line: the comments article shows a supportive but tactical comment process. Stakeholders did not oppose the LCD; they wanted it less rigid and less disruptive. MolDX accepted several clarification-oriented changes, especially around Criterion 5, Criterion 9, TA language, ICD-10 coding, and clinical context in Criterion 2. But MolDX preserved the core structure: biomarker testing remains covered only after primary risk assessment, only when recent definitive imaging does not already answer the question, and only after MolDX TA.
# # #
Some MolDx coverage criteria have a random hodge-podge of rules. These are quite a bit more readable and structured - bravo!
Here's how Chat GPT views the new final rules.
#
Yes — this is better than the usual MolDX hodgepodge. It has a coherent clinical sequence:
suspected MASLD/MASH → first-line clinical risk assessment → elastography if available → molecular/proteomic test only if it will change management → no repeated fishing expeditions → test must clear TA/evidence standards.
The improvements I’d suggest are mostly tightening, not rethinking.
Main ways to improve
1. “Clinical evidence of liver dysfunction” may be too narrow.
MASLD/MASH patients may have normal liver enzymes or nonspecific abnormalities. Better wording would be something like: “suspected or established MASLD/MASH based on metabolic risk factors, imaging, laboratory findings, or other clinical evidence, after consideration of competing liver diseases.” The purpose is fibrosis risk stratification, not only “liver dysfunction.”
2. Clarify the FIB-4 gate.
The phrase “e.g., FIB-4 ≥ 1.3” is good, but incomplete. FIB-4 has age issues, indeterminate zones, and false positives in older adults. Better to say the patient must be not low risk under guideline-concordant first-line assessment, using age-adjusted or otherwise guideline-specified thresholds where applicable.
3. Improve the elastography rule.
“LSM has not been performed within 12 months or was indeterminate” is sensible, but should also include cases where LSM was unavailable, technically unsuccessful, unreliable, contraindicated, or discordant with the clinical picture. Otherwise, the policy may push clinicians into artificial documentation games.
4. Define “management decision” more tightly.
This is one of the best parts of the LCD, but it could be strengthened. The test should not merely be “helpful”; it should inform a pending, documented decision, such as hepatology referral, biopsy, treatment eligibility, surveillance intensity, or deferral of invasive testing.
5. Be careful with “monitoring pharmacologic therapy.”
That phrase opens the door to serial testing. The frequency rule limits this, but MolDX should specify that serial monitoring is covered only if the test has evidence for longitudinal interpretation, not merely baseline fibrosis classification.
6. Algorithm rule should say “locked algorithm.”
The current rule says the algorithm must be validated outside the development cohort. Good. But it should also require that the algorithm version, cutoffs, and intended-use categories be locked before validation, and that major algorithm or chemistry changes require bridging or re-review.
7. “Equivalent or superior performance” is underdefined.
Equivalent or superior by what metric? AUROC? sensitivity at fixed specificity? NPV for advanced fibrosis? biopsy endpoint? elastography endpoint? Without defining this, the rule can become arbitrary. It should specify performance against the intended clinical use, not a generic “better test” standard.
8. Last two criteria are partly duplicative.
Both say similar tests must show equivalent or superior performance to already covered tests. They could be consolidated into one rule that applies to similar intended-use claims.
A tightened version
Molecular or proteomic biomarker testing for assessment of liver fibrosis in the setting of suspected or established MASLD or MASH is considered reasonable and necessary when all of the following criteria are met:
The patient is an adult with suspected or established MASLD or MASH based on guideline-concordant clinical evaluation, which may include metabolic risk factors, imaging findings, laboratory abnormalities, or other clinical evidence, after consideration of competing causes of liver disease.
First-line risk assessment using clinical evaluation and non-molecular/non-proteomic laboratory testing, such as FIB-4 or another guideline-supported approach, does not indicate low risk. Age-adjusted or otherwise guideline-specified thresholds should be used where applicable.
Liver stiffness measurement by imaging, such as VCTE or MRE, has not been performed within the prior 12 months, was unavailable or not feasible, was technically unsuccessful or unreliable, produced indeterminate results, or was clinically discordant with the patient’s overall risk assessment.
The test result will be used to directly inform a pending management decision documented in the medical record, such as referral to hepatology, selection or avoidance of liver biopsy, eligibility for pharmacologic therapy, risk-stratified follow-up, or other guideline-concordant management.
Testing is not performed more than once within a 12-month period and is not performed within 12 months after a liver biopsy with definitive histologic interpretation, unless the test has evidence supporting longitudinal use for the specific clinical scenario.
The test has satisfactorily completed a Technical Assessment by the MolDX Molecular Diagnostic Services Program for the claimed intended use and intended-use population.
Clinical validity must be established in peer-reviewed literature for the intended use, intended-use population, and claimed clinical decision point, including appropriate reference standards, prespecified performance metrics, confidence intervals, and evaluation of relevant subgroups.
If the test relies on an algorithm, the algorithm, cutoffs, and reporting categories must be locked before validation, and validation must be performed in a cohort independent of the development cohort. Material changes to chemistry, analytes, algorithm, or cutoffs require appropriate bridging evidence or re-review.
For tests using similar analytes, similar methodology, or the same intended-use claim as an already covered test, the test must demonstrate clinically meaningful comparable or superior performance for the same clinical decision, using direct comparative evidence where feasible or a justified indirect comparison where direct evidence is not available.
The test is not covered for general population screening, for patients already classified as low risk by guideline-concordant first-line assessment, for duplication of an already informative fibrosis assessment, or for serial monitoring unless longitudinal clinical validity and utility have been established.
My bottom line: the LCD is directionally good because it embeds the test in a clinical pathway. The biggest fixes are to reduce ambiguity around MASLD eligibility, elastography exceptions, algorithm versioning, and what “equivalent or superior” actually means.
