CMS Final Rules for 2027: Software-Based Laboratory Tests Move Off the CLFS
Executive summary (approximately 100 words)
CMS has finalized a major change in Medicare payment policy for algorithm-only laboratory tests, effective January 2027. Fourteen CPT codes previously paid under the Clinical Laboratory Fee Schedule (CLFS) will instead be treated as other diagnostic tests, paid under the hospital Outpatient Prospective Payment System (OPPS) or Physician Fee Schedule (PFS). Initial payment rates will approximate existing CLFS rates, but conventional Medicare coinsurance will apply, and laboratory-specific rules, including the 14-day date-of-service rule, will no longer govern these services. CMS corrected several code classifications following public comments but retained its fundamental policy. Important questions remain about future pricing, packaging, and regulation.
1. OPPS Final Rule: SaMS Analyses Performed on Laboratory Tests
See CMS-1850-FC, Section X.B.2, esp. Table 85
CMS has largely finalized its proposed policy, with important corrections to the affected CPT codes.
CMS maintains that a stand-alone algorithmic analysis of previously generated laboratory data is not a clinical diagnostic laboratory test (CDLT), even if it has traditionally been coded and paid as one. Instead, CMS considers it an other diagnostic test under Social Security Act §1861(s)(3).
This represents a substantial break from the historical CLFS treatment of these services, including the established AMA CPT coding process and CMS crosswalk/gapfill procedures. CMS rejected laboratory-industry arguments that laboratory-derived data interpretation remains intrinsically part of the laboratory service.
Payment and the 14 affected codes
CMS finalized 14 codes, compared with 10 in the original proposal. For CY 2027, they receive new technology APC assignments with status indicator O1, allowing separate payment without multiple-procedure discounting.
The final APC assignments are:
CPT codes | New technology APC | Standard payment |
|---|---|---|
0510U | 1506 | $450.50 |
0512U, 0513U, 0414U, 0418U, 0220U, 0376U, 1106T, 1107T, 1097T, 1063T, 1064T | 1509 | $750.50 |
81417 | 1505 | $350.50 |
81427 | 1522 | $2,250.50 |
CMS bases the assignments principally on CY 2026 CLFS payment rates. Because new technology APCs have fixed payment bands, the new payments approximate, rather than precisely reproduce, previous CLFS amounts.
Corrections to the proposal
CMS made several important adjustments following public comment that highlighted errors and problems..
Removed 0511U and 81416: CMS accepted that these services include wet-laboratory work and should remain on the CLFS. Deleted code 0208U was also removed.
Retained 0414U and 0220U: CMS rejected objections that these services may entail wet-laboratory components, concluding that their descriptors primarily identify algorithmic analyses.
Added 81417 and 81427: These codes for re-evaluation of previously obtained exome and genome sequences had been omitted from the proposal.
Raised 1063T and 1064T: CMS changed the bladder cancer algorithm codes from proposed APC 1542, paying $350.50, to APC 1509, paying $750.50, based on comparability to other services.
The treatment of codes involving staining or slide digitization when performed remains an interesting definitional issue. CMS insists its policy concerns stand-alone computational services, while retaining several codes whose descriptors contemplate preparatory laboratory activities.
Important consequences beyond price
Beneficiary cost-sharing. These services will enter ordinary OPPS payment rules, including applicable beneficiary coinsurance, rather than remaining under the generally no-copayment CLFS.
The 14-day rule. CMS explicitly confirms that SaMS analyses performed on laboratory tests are no longer governed by the laboratory date-of-service regulation at 42 CFR 414.510. This affects a longstanding issue in separately billing molecular and genomic testing following hospital encounters.
CLIA and FDA. CMS emphasizes that the payment change does not itself revise CLIA requirements. It also distinguishes these laboratory-derived SaMS from the FDA-regulated SaMS devices addressed in the preceding OPPS section: FDA marketing authorization is not currently required for the laboratory-derived category.
Packaging. For 2027, O1 provides separate payment, and new technology APC services are protected from comprehensive APC packaging. That is reassuring in the immediate term. Longer-term treatment is less certain, especially given the OPPS tradition of packaging most ordinary laboratory services, apart from molecular pathology and certain other exceptions.
Pricing new tests. CMS finalizes assignment of future new laboratory SaMS codes to new technology APCs but does not establish a comparably detailed initial valuation process. With these services outside the CLFS, the usual summer crosswalk/gapfill process is no longer the natural entry point for establishing their prices. CMS acknowledges the need for further development of its APC assignment methods.
2. PFS Final Rule: SaMS Laboratory Analyses
See CMS-1848-F, Section (61), esp. Table A-D8
The most significant PFS change from the proposal is CMS's decision to reject contractor pricing.
The proposed rule would have placed the identified algorithmic services under Medicare Administrative Contractor pricing, potentially generating geographic variability and uncertainty.
Commenters objected, pointing out that the underlying services have no obvious jurisdiction-specific cost differences and that the proposal would unnecessarily disrupt nationally established prices.
CMS agreed. The final rule establishes national PFS pricing for the same 14 laboratory SaMS codes, with procedure status indicator A (active).
The conversion from OPPS dollars into RVUs
CMS uses the OPPS new technology APC payment as the starting point. It then calculates the practice expense RVU component by dividing the OPPS payment by the CY 2027 PFS conversion factor.
In broad terms, a service with an OPPS payment of $750.50 is translated into RVUs using the approximately $34 conversion factor, then translated back into dollars when paid under the PFS.
This preserves the approximate payment level while integrating the code into the PFS structure. However, ordinary PFS geographic adjustments and future conversion-factor changes mean that actual payments need not exactly match the OPPS amount in every location or future year.
Other consequences
Like the OPPS rule, the PFS final rule confirms that these services are other diagnostic tests rather than CLFS laboratory services. Consequently, conventional PFS beneficiary cost-sharing will apply, and CMS confirms that the CLFS date-of-service rule will not.
The final rule responds to concerns about whether independent laboratories can bill directly for these services primarily by stating that the codes will be actively priced under the PFS. It does not fully work through the operational implications for laboratories performing computational analyses independently of physician offices or hospital outpatient departments.
The same important question about new test valuation remains. For the first 14 codes, CMS had existing CLFS prices to anchor the OPPS rates, which in turn anchor the PFS RVUs. That historical reference will not necessarily exist for future codes. The final rule does not fully resolve how new services will obtain an initial valuation outside the established CLFS process.
3. OPPS General SaMS Final Rule
See CMS-1850-FC, Section X.B.1, preceding the OPPS laboratory discussion
This broader policy is important because the laboratory changes are an extension of CMS's effort to establish a consistent Medicare payment category for algorithm-based diagnostic software.
From SaaS to SaMS
CMS finalizes the terminology change from Software as a Service (SaaS) to Software as a Medical Service (SaMS).
The category encompasses algorithm-driven services supporting medical decision-making, including analysis of imaging, cardiovascular information, and other previously acquired clinical data.
The original proposal identified 36 SaMS codes, including services already receiving separate payment under clinical APCs or new technology APCs.
Standardization through new technology APCs
CMS generally finalizes reassignment of separately paid SaMS services from clinical APCs to new technology APCs, while maintaining payment levels reasonably close to CY 2026 amounts.
CMS also adopts status indicator O1, with the same payment specifications as status indicator S. Thus, O1 services are separately payable and are not subject to multiple-procedure discounting.
However, the general SaMS policy does not make every algorithm-based service separately payable. CMS preserves existing conditional packaging for certain Q1 services and complete packaging for certain N services. For example, it declines requests to change the conditional packaging status of retinal imaging codes 92227 and 92228.
These distinctions are important when considering how CMS may eventually structure payment for laboratory-derived SaMS.
FDA authorization, coverage, and specific code decisions
For the general SaMS services in Section X.B.1, CMS requires FDA marketing authorization for Medicare payment. This contrasts with the laboratory-derived SaMS category in Section X.B.2, where such authorization is not currently required.
CMS also emphasizes that SaMS classification and APC assignment are payment decisions, not guarantees of Medicare coverage or medical necessity.
Several individual code decisions illustrate the interim approach. CMS added Cardiosense code 1036T at $150.50, burn-analysis code 0972T at $250.50, and EEG-analysis code 1041T at $150.50, resisting some requests for substantially higher initial payment. In other instances, it retained the proposed APC despite industry arguments that existing resource costs justified an increase.
Longer-term SaMS policy remains unsettled
The general SaMS discussion is explicit that 2027 represents an interim step, not a permanent payment framework. CMS received comments advocating dedicated cost centers and revenue codes, better reporting of software licensing and implementation costs, and payment approaches incorporating clinical value.
CMS indicates that it will consider these suggestions in developing possible CY 2028 proposals.
This creates a central uncertainty for both imaging-based and laboratory-based SaMS: whether the eventual permanent policy will retain individual separate payments, introduce more systematic cost-based valuation, adopt different payment categories, or increase packaging.
Overall significance
The immediate effect of the 2027 rules is principally a change in Medicare payment classification, rather than a wholesale repricing of existing tests.
But the longer-term implications are potentially much larger. By removing stand-alone computational analyses from the CLFS, CMS moves these services away from laboratory-specific pricing and billing protections and into the OPPS/PFS structures, with their coinsurance, budget-neutrality, geographic-adjustment, and packaging mechanisms.
The unresolved challenge is particularly striking for new tests: CMS has established where they will be paid, but has not yet fully established how their initial payment amounts will be determined. That may become one of the most consequential questions for algorithm-driven genomics and molecular diagnostics over the next several rulemaking cycles.
The SaMS Paradox: When Physical Work Counts—and When It Doesn't
There is a striking inconsistency in CMS's reasoning, although the agency has left itself a possible legal distinction.
For laboratory-derived SaMS, CMS draws a fairly sharp line: if a CPT code describes only secondary computational analysis of previously generated laboratory data, it is no longer a clinical laboratory test and leaves the CLFS. If the code encompasses actual specimen analysis, as with 0511U or 81416, CMS agrees it is an integrated laboratory service and leaves it on the CLFS.
Yet SaMS generally is not restricted to software-only services.
Consider CPT 92229, which CMS retains as SaMS in OPPS Table 84. The service encompasses retinal imaging and autonomous algorithmic interpretation. The camera, optics, illumination, focusing, and image acquisition are integral to obtaining the diagnostic result. It is hardly a service delivered by an algorithm alone. Nevertheless, CMS assigns the entire service to SaMS, with status indicator O1 and new technology APC 1502.
There are even more explicit examples. CPT 1105T includes actual acquisition of EEG and ECG signals, followed by algorithmic analysis. CMS newly designates it SaMS. Similarly, CPT 0972T includes multispectral imaging, system setup, and image acquisition, yet also qualifies as SaMS.
Thus, CMS is not consistently applying a principle that SaMS must consist exclusively of computational work.
Can the distinction be defended?
CMS could argue that it is answering two different questions.
In the laboratory context, the question is whether a service qualifies under the statutory clinical diagnostic laboratory test benefit category, with its particular payment and regulatory requirements. CMS treats secondary analysis of previously obtained data as an other diagnostic test.
In the broader SaMS context, the question is whether algorithm-driven diagnostic technology belongs in the new SaMS payment category, even when the service includes physical acquisition of the underlying data.
That is a plausible legal distinction. But it is not a consistent functional definition of SaMS.
Indeed, the distinction becomes particularly awkward for pathology-based software that includes preparation or digitization of glass slides. Codes 1063T and 1064T explicitly contemplate H&E staining and slide digitization when performed, yet CMS includes them among the 14 laboratory-derived SaMS services leaving the CLFS.
In other words, the presence of physical work is disqualifying in some laboratory codes, permissible in other laboratory codes, and entirely compatible with SaMS designation in imaging and other diagnostic specialties.
The most charitable interpretation is that CMS is trying to preserve existing integrated laboratory tests while carving out independently identifiable computational services. However, the resulting boundary depends heavily on the individual CPT descriptor, its historical coding conventions, and CMS's interpretation of optional physical components.
That is a precarious foundation for an increasingly consequential payment classification.
And the paradox has practical consequences. Whether physical activity is included in a single CPT code, or separated into two codes describing acquisition and subsequent analysis, may determine whether a service remains on the CLFS, receives conventional coinsurance, escapes the laboratory date-of-service rule, and ultimately becomes subject to OPPS packaging.
The larger lesson may be that CMS has created a payment category called Software as a Medical Service without yet establishing a consistent principle for how much non-software activity that service can contain. The agency acknowledges that its 2027 policy is interim. The unresolved definitional problem may prove considerably harder to resolve than the initial payment assignments.
###
The paradox is not merely semantic. CMS applies opposite reasoning to the presence of physical diagnostic work, depending on whether the service originated in the laboratory or elsewhere.
Consider the contrast:
- CPT 92229 (retinal imaging): Physical image acquisition, involving a camera, optics, illumination, and positioning, is combined with an algorithmic diagnostic interpretation. CMS classifies the entire service as SaMS.
- CPT 0511U (tumor drug-response testing): Physical laboratory activity is combined with computational prediction. CMS concludes that the physical laboratory component prevents the service from being classified as SaMS, retaining it on the CLFS.
Both are integrated diagnostic services involving physical procedures and algorithms. Yet in one case the physical component is compatible with SaMS classification; in the other it is grounds for exclusion.
CMS can invoke the distinct statutory treatment of laboratory tests to explain the administrative result. But that does not resolve the conceptual inconsistency in what constitutes Software as a Medical Service.
The paradox is that physical diagnostic work does not disqualify a service from being SaMS—unless that physical work happens to be laboratory work.
And that makes SaMS less a coherent technological category than a classification shaped by the historical boundaries of Medicare payment systems.
There is no need to force a reconciliation. The inconsistency is itself an important finding about these final rules.
