For years, artificial intelligence has advanced faster than Medicare reimbursement policy. Healthcare organizations have invested in AI-powered diagnostic software, while manufacturers have struggled to explain how products built on algorithms—not physical equipment—fit into a payment system designed decades before modern AI existed.
That may finally be changing.
Within the CY 2027 Hospital Outpatient Prospective Payment System (OPPS) Proposed Rule, CMS introduced one of the most significant AI reimbursement proposals in Medicare history: the creation of a standardized payment pathway for Software as a Medical Service (SaMS).
While much of the industry’s attention has focused on 340B reforms and outpatient payment changes, this proposal could ultimately reshape how hospitals evaluate, purchase, and deploy clinical artificial intelligence.
AI Is No Longer Being Treated Like Traditional Software
CMS acknowledges a problem that hospitals, technology developers, and investors have recognized for years.
Traditional OPPS reimbursement is based upon tangible healthcare resources:
- Equipment
- Supplies
- Personnel
- Facility costs
- Clinical labor
Artificial intelligence creates value differently.
Rather than consuming physical resources, these technologies generate clinical value through proprietary algorithms that analyze patient information and provide diagnostic findings, risk stratification, or treatment recommendations.
That distinction makes conventional APC assignment extremely difficult.
CMS states that its historical payment methodology was never designed to value algorithm-driven clinical decision support and that a new framework is necessary to appropriately reimburse these technologies.
CMS Is Introducing “Software as a Medical Service”
Perhaps the most notable change is actually one of terminology.
Previous CMS rulemaking referred to these technologies as Software as a Service (SaaS).
Beginning with the 2027 Proposed Rule, CMS proposes renaming the category:
Software as a Medical Service (SaMS).
The distinction is intentional.
Outside healthcare, Software as a Service generally refers to cloud-based business software.
CMS believes the new terminology better reflects software functioning as an independent medical technology rather than simply another IT platform.
A New Medicare Payment Category for AI
CMS is proposing several structural changes that collectively create the first dedicated reimbursement pathway for algorithm-driven diagnostic software.
These include:
- Creation of a new Status Indicator O1
- Designation of 36 HCPCS codes as Software as a Medical Service technologies
- Movement of 21 HCPCS codes into New Technology APCs
- Separate payment while CMS develops a permanent valuation methodology
The use of New Technology APCs is particularly important.
Historically, CMS has used these APCs for emerging technologies that lack sufficient utilization data to establish permanent payment rates.
Rather than forcing AI tools into existing payment structures that were never intended for them, CMS is effectively creating a transitional reimbursement environment that allows real-world utilization and cost data to accumulate before establishing long-term payment methodologies.
What Types of AI Are Included?
The proposal extends far beyond simple imaging software.
Examples cited by CMS include:
- AI-assisted retinal image interpretation
- Heart failure detection using echocardiography
- Coronary blood flow estimation from CT angiography
- CT-based fracture risk assessment
- Eye movement analysis for concussion evaluation
- Algorithmic ECG cardiac risk prediction
- Quantitative brain MRI comparisons
- AI-assisted prostate cancer mapping from biopsy imaging
This represents a diverse cross-section of specialties, including cardiology, neurology, orthopedics, ophthalmology, emergency medicine, and oncology.
Oncology May Be One of the Largest Beneficiaries
Although oncology receives relatively little attention within this section of the proposed rule, the implications are enormous.
Modern cancer care increasingly depends upon sophisticated software capable of:
- Automated image segmentation
- Tumor contouring
- Radiomics
- AI-assisted pathology interpretation
- Digital pathology
- Genomic interpretation
- Risk prediction
- Treatment planning optimization
- Decision support
Many of these technologies have struggled to demonstrate a sustainable reimbursement model despite their growing clinical utility.
The proposed SaMS pathway signals that CMS recognizes software itself can constitute a reimbursable clinical resource—not merely an administrative expense.
For oncology programs evaluating investments in AI-assisted contouring, pathology platforms, imaging analytics, radiomics, or predictive treatment software, this represents a meaningful shift in federal payment policy.
CMS Is Also Moving Certain Laboratory Algorithms
One of the proposal’s lesser-discussed provisions may have equally significant consequences.
CMS proposes transferring 10 HCPCS codes involving algorithmic analyses currently reimbursed under the Clinical Laboratory Fee Schedule into the SaMS payment framework.
These include software performing secondary analyses on existing laboratory or genomic data rather than conducting laboratory testing itself.
While the change aligns reimbursement with the software’s clinical function, it introduces an important policy consideration.
Unlike the Clinical Laboratory Fee Schedule, OPPS includes beneficiary cost-sharing and budget-neutrality adjustments.
As a result, some services previously furnished without patient cost-sharing could become subject to Medicare copayments if finalized.
Program Integrity Remains a Major Concern
CMS also acknowledges that reimbursement for AI presents unique compliance challenges.
Unlike traditional medical devices purchased once and depreciated over time, AI software may be licensed through:
- Annual subscriptions
- Enterprise licenses
- Per-study pricing
- Per-click utilization fees
These commercial models create complexities that Medicare has historically not encountered when establishing outpatient payment policy.
CMS specifically requests stakeholder feedback regarding how separate payment should occur and whether reimbursement should be reduced when AI software is billed alongside other procedures.
This Is Only the Beginning
Importantly, CMS repeatedly characterizes this proposal as an interim policy rather than a permanent reimbursement framework.
The agency is actively seeking comments on:
- Additional HCPCS codes that should qualify
- Appropriate payment methodologies
- Whether separate payment should always apply
- Long-term valuation strategies
- Program integrity safeguards
This indicates CMS expects rapid evolution as artificial intelligence continues expanding throughout clinical medicine.
The Bridge Oncology Perspective
The proposed SaMS payment pathway represents far more than another Medicare payment update.
It is the first formal acknowledgment by CMS that clinical algorithms have become healthcare resources deserving independent reimbursement consideration.
For hospitals, this may fundamentally change AI investment strategies. Instead of viewing AI solely as an operational expense, organizations may soon evaluate these technologies as reimbursable clinical assets capable of improving diagnostic accuracy, workflow efficiency, and patient outcomes.
For oncology, the implications are even greater. As artificial intelligence becomes increasingly integrated into imaging, pathology, genomics, treatment planning, and precision medicine, reimbursement policy must evolve alongside clinical innovation.
The CY 2027 OPPS Proposed Rule suggests that evolution has officially begun.
The CY 2027 OPPS Proposed Rule remains a proposal. CMS is accepting public comments before issuing the Final Rule later this year, and payment policies may change before implementation.