Header: CMS proposes X1/X2 modifiers to encode cancer stage on NGS claims—framing it as ICD-10 “clarity,” but potentially creating more complexity than it resolves.
###
My Heart Skipped a Beat: CMS Proposes X1/X2 Modifiers for NCD 90.2 (NGS in Cancer)
In 2017, CMS issued NCD 90.2 for next-generation sequencing (NGS) in cancer. It was not an accidental document. Senior HHS leadership got involved (the Trump 1 administration). The policy was designed to do something specific and modern: create a clean, predictable incentive for FDA approval of comprehensive genomic profiling (CGP) tests.
- The Trump HHS made the lab industry a bargain.
- If an NGS test received FDA approval as a companion diagnostic, it would be covered nationally and immediately—on the day of FDA approval.
- If the label expanded (new gene, new drug), coverage expanded the same day. No MAC-by-MAC delay.
No prolonged local review cycles. A national coverage rule designed to match the pace of precision oncology.
That was the theory.
In practice, implementation has been anything but frictionless. The 14-day rule delays inpatient surgical specimens from being tested promptly. If testing waits for a later oncology visit, even more weeks can pass before genomic results guide therapy. Patients sometimes start urgent but nonspecific chemotherapy during these delays. And while the NCD promises its readers coverage “from the day of FDA approval,” operational reality often looks far worse. [See my blog about erroneous MAC delays in implementing new CMS NCD coverage.]
So one might reasonably ask: if implementation is already uneven, how do we improve it?
At the December 2025 HCPCS meeting (Agenda Item 41), CMS proposed creating two new Level II modifiers to “implement” NCD 90.2:
X1 – NGS for early-stage cancer (Stage I–II)
X2 – NGS for advanced (Stage III–IV), recurrent, relapsed, refractory, or metastatic cancer
CMS’s summary explains the rationale. NCD 90.2 does not cover early-stage cancers, while ICD-10 diagnosis codes do not distinguish stage. The proposed modifiers would “ensure greater clarity,” align submission with coverage requirements, and “maintain data integrity for oversight and claims analysis.”
That seems at first like a coherent administrative theory. But it deserves scrutiny. I'd argue it improves NONE of the real problems, and introduces a half-dozen new ones.
https://www.cms.gov/files/document/b2-2025-public-meeting-agenda-december-17-18-2025.pdf
 |
| click to enlarge |
The decision isn't out yet, but could come out any week, with other non-drug HCPCS narrative decisions (see here, scroll down, right column).
CMS’s Framed Problem: ICD-10 Lacks Stage
CMS’s core claim is that ICD-10 does not reliably encode cancer stage, creating ambiguity in claims submission. That is true. ICD-10 codes generally don't specify Stage I vs Stage IV.
But this has been true across oncology for decades....and:
- Chemotherapy billing does not require a Stage III modifier.
- Radiation therapy billing does not require a Stage IV modifier.
- Surgical oncology claims do not append stage-specific modifiers.
Why does one lab test uniquely in need of a stage-specific HCPCS architecture?
NCD 90.2 has been in effect for eight years. CMS has not tried to assert that ICD-10 stage ambiguity has produced systematic payment error under this NCD. MACs already adjudicate based on diagnosis codes, documentation, coverage articles, and other clinical indicators. The absence of stage coding in ICD-10 is not a new discovery. Deniable claims for Stage 1-2 are already supposed to require a modifier, GZ.
If the problem were widespread miscoding or improper payment, CMS would typically cite the data. The proposal cites none.
Structural and Practical Problems
Even if one accepts CMS’s framing, the proposed solution raises multiple operational and policy concerns.
1. Staging does not map cleanly across all cancers.
Leukemias, lymphomas, and many brain tumors do not follow AJCC Stage I–IV schemas. How are those claims coded? These problems have been elevated to CMS since 2018 with no change.
2. Adjuvant Stage II disease is not simple.
Under MolDx policies, certain Stage II colon and breast cancers qualify for adjuvant chemotherapy decision-making. Clinically, per MolDx, these cases behave as “advanced” for therapeutic purposes. Do they receive X1 (for stage 1 or 2) or X2 (for MolDx-defined "advanced")?
3. Recurrent disease straddles both modifiers.
A recurrent Stage II melanoma may still technically be Stage II but clinically advanced due to "recurrence." X1 or X2? Both? The modifier language introduces interpretive ambiguity rather than eliminating it.
4. Modifier stacking confusion (X1/X2 with GA/GZ).
Most Stage I–II cancers are non-covered under NCD 90.2. Today, such claims would typically involve:
GZ (expected denial, no ABN), or
GA (ABN on file).
If X1 is appended, does that replace GZ?
Are both required?
Is X1 informational only?
The proposal does not clarify interaction with established waiver-of-liability logic.
5. RNA sequencing remains excluded under the NCD.
NCD 90.2 excludes RNA sequencing. Yet RNA-based profiling is increasingly central to oncology practice. X1/X2 do nothing to modernize this scientific mismatch when NGS is used for RNA-SEQ.
6. MRD asymmetry.
NGS-based minimal residual disease (MRD) testing would fall within this modifier framework. Increasingly common digital PCR MRD would not. Same clinical intent, different molecular platform, different coding logic.
7. No demonstrated adjudication crisis.
NCD 90.2 has functioned nationally since 2018. CMS has not shown systemic payment errors attributable to stage ambiguity in claims.
8. Administrative burden without therapeutic benefit.
Every new modifier introduces:
Education requirements
Audit exposure
Software updates
Claim edit revisions
Mistakes and confusion
None of these accelerate patient access to precision therapy.
A Curious Coding Note
A scan of existing HCPCS modifier tables suggests that “X1” and “X2” have appeared in other contexts (e.g., “broad” and “narrow” service distinctions in MIPS). Repurposing modifier codes can generate unintended confusion across systems and historical data.
The Larger Policy Context
NCD 90.2 was designed as a major new pro-innovation lever. FDA approval would trigger immediate national coverage. That principle aligned regulatory science with reimbursement policy.
If implementation delays undermine that principle, the solution lies in:
Addressing MAC lag and other causes of delayed coverage,
Revisiting the 14-day rule’s impact on genomic testing,
Updating the NCD to reflect RNA-based oncology and other staging systems,
Clarifying recurrence and adjuvant logic within coverage text.
Just appending X1 and X2 to claims does not accelerate coverage. It does not modernize the NCD. It does not resolve scientific evolution. Presumably 99% of the claims are X2, anyway.
It encodes cancer stage at the claim line level for one specific technology while leaving the broader oncology billing ecosystem untouched.
That asymmetry and the unintended consequences are what made my heart skip a beat.
If the goal is clarity, CMS should articulate the measurable problem it seeks to solve. If the goal is a capricious choice for data capture, say so. But if the goal is improving patient access to precision oncology, and solving large real problems, this proposal appears to miss the target entirely.
