Executive Summary 🔗
- Volume is meaningful, but the mix matters. 92 AI/ML-related FDA Clearances in 2025, with image-processing and workflow software product codes dominate. For capital planning, this usually means shorter development cycles and more competitive fields, but also faster iteration and broader distribution channels.
- Coronary CT remains the highest-leverage clinical arena. Plaque quantification, CAC scoring, and CT-FFR sit at the junction of high downstream cost (stents, cath lab utilization, chronic meds) and a radiology-friendly workflow. In 2025, 9 plaque-related tools, 7 CAC-related tools, and several FFR/FFRct/QFR-related tools were cleared.
- Reimbursement is shifting in a way that should increase "investability" for coronary CT stacks. Hospital outpatient payment for CCTA increased materially in 2025 (per ACC reporting), and new 2026 CPT updates include Category I codes tied to CT-derived FFR and CT perfusion.
- PCCPs are no longer theoretical. 8 of 92 2025 clearances are PCCP-authorized, spanning imaging software, cardiology analytics, and consumer device features. That's a signal worth internalizing if you're building a product roadmap that depends on frequent model updates.
- Clearance is a financing catalyst, but the effect is asymmetric. For private companies, the effect is usually visible as a fundraising or partnership event around clearance. For public companies, 510(k) clearances rarely move the stock on their own unless the company is small and the clearance is tightly coupled to near-term revenue.
2025 Cardiology AI/ML Clearances at a Glance 🔗
| Metric | Value |
|---|---|
| Total 2025 AI/ML cardiology clearances | 92 |
| Unique applicants | 69 |
| Unique manufacturers deduplicated | 63 |
| SaMD | 59 |
| Non-SaMD | 33 |
| Panel: Radiology | 69 |
| Panel: Cardiovascular | 19 |
| Panel: Other | 4 |
| 510(k) type: Traditional | 80 |
| 510(k) type: Special | 11 |
| 510(k) type: Abbreviated | 1 |
| Median FDA review time days to clearance | 149 |
| Mean FDA review time days to clearance | 155.2 |
| PCCP-authorized | 8 |
How fast were these cleared? 🔗
Review time is one of the most practical signals for CEOs because it affects cash runway and commercialization timing. The central tendency is closer to "classic 510(k)" than "lightweight software."
| 510(k) type | N | Median days | Mean days |
|---|---|---|---|
| Abbreviated | 1 | 117 | 117 |
| Special | 11 | 28 | 68.2 |
| Traditional | 80 | 153.5 | 167.6 |
A few operating implications:
- Special 510(k) is the speed lever, but it's only a small portion of 2025 clearances (11/92). Many teams talk about "iterating quickly," but the regulatory pathway still determines whether iteration is monthly or annual.
- Traditional 510(k) dominates (80/92). For investors, that usually increases diligence requirements because time-to-revenue is more sensitive to FDA cycle time and evidence generation.
Manufacturer landscape with deduplication 🔗
If you look at applicant names, you’ll see 69 unique applicants. If you “deduplicate” obvious corporate families (for example: GE’s multiple legal entities, Philips’ sub-entities), the list becomes 63 manufacturer groups.
Top manufacturer groups by 2025 clearance count 🔗
| Manufacturer | 2025 Clearances | First AI/ML Clearance | First Year? |
|---|---|---|---|
| Philips | 7 | 2013 | No |
| Siemens Healthineers | 6 | 2014 | No |
| Shanghai United Imaging Healthcare Co., Ltd. | 6 | 2020 | No |
| GE HealthCare | 5 | 2018 | No |
| BrightHeart | 3 | 2024 | No |
| BunkerHill Health | 3 | 2024 | No |
| Canon | 3 | 2019 | No |
| Artrya Limited | 2 | 2025 | Yes |
| DiA Imaging Analysis Ltd. | 2 | 2024 | No |
| TeraRecon, Inc. | 2 | 2020 | No |
A few patterns that show up when you read this like a strategist:
- Incumbent imaging OEMs are shipping AI/ML as a feature, not a standalone product. Siemens, Philips, GE, Canon, and Fujifilm appear repeatedly. That generally compresses standalone “point solution” margins unless the point solution has a unique claim, unique data moat, or a payment tailwind.
- A large share of the long tail is single-clearance companies. That’s typical of an ecosystem where “one killer feature” can support a venture story, but it also means buyers face a vendor management problem. That’s often where platform vendors win.
New entrants signal 🔗
27 of the 63 manufacturer groups appear for the first time (as a group heuristic) in 2025, contributing 28 of the 92 clearances (about 30%). Treat that as a directional signal, not a legal fact, because naming conventions and subsidiaries can hide earlier activity.
What CEOs and investors should take from this:
- There’s still room for new entrants, but the default market structure is “a few platforms plus many niche tools.”
- Category selection is the real moat. In crowded categories, the winning strategy often becomes distribution and workflow integration, not algorithm metrics.
Product codes: what the volume is really about 🔗
The top product codes are concentrated.
| Product code | Count (2025) |
|---|---|
| QIH | 28 |
| IYN | 7 |
| JAK | 6 |
| LNH | 6 |
| QAS | 5 |
| DQK | 4 |
| QKB | 4 |
| QYE | 3 |
| SDJ | 2 |
| QJU | 2 |
| QHA | 2 |
| OWB | 2 |
Two important clarifications for non-regulatory readers:
- A product code is not the clinical indication. It’s an FDA classification bucket. A single code can span a wide range of “features” and modalities.
- In AI/ML, product codes often correlate with a commercialization model:
- Some codes skew toward platform-integrated features (often sold as part of scanner or PACS upgrades).
- Others skew toward workflow tools (triage, notification, prioritization).
- Some skew toward clinical decision support (where reimbursement and clinical evidence become the bottleneck).
What the top codes mean in plain English 🔗
These definitions come from FDA product classification sources.
- QIH is a radiological image processing system category.
- QYE is radiological computer-assisted detection software.
- QAS is radiological computer-assisted triage and notification software.
- QKB is a radiological image processing system category tied to radiation therapy planning. This helps explain why a "cardiology" themed export can still contain radiology-heavy content. In practice, many cardiac AI/ML products live operationally inside radiology infrastructure.
PCCPs in 2025: the update cadence story is real 🔗
8 clearances in 2025 were PCCP-authorized.
| decisiondate | knumber | applicant | devicename | productcode | type |
|---|---|---|---|---|---|
| 2025-09-11 | K250507 | Apple Inc. | Hypertension Notification Feature (HTNF) | QXO | Traditional |
| 2025-04-04 | K242807 | Deski | HeartFocus (V.1.1.1) | QJU | Traditional |
| 2025-07-28 | K250652 | Anumana, Inc. | ECG-AI Low Ejection Fraction (LEF) 12-Lead algorithm (1010) | QYE | Traditional |
| 2025-01-15 | K243065 | Caption Health, Inc. | Cardiac Guidance | QJU | Traditional |
| 2025-05-07 | K243684 | BrightHeart | BrightHeart View Classifier | QIH | Traditional |
| 2025-06-05 | K251456 | BrightHeart | BrightHeart View Classifier | QIH | Special |
| 2025-07-18 | K250902 | HeartFlow, Inc. | HeartFlow Analysis | PJA | Traditional |
| 2025-04-03 | K242551 | Siemens Healthcare GmbH | syngo Dynamics (Version VA41D) | QIH | Traditional |
Why this matters for capital and strategy:
- If your roadmap assumes frequent model updates, PCCP is the regulatory mechanism that can make that roadmap financeable without turning each update into a new fundraise.
- PCCP is also a governance commitment. It forces more maturity in your quality system, monitoring, and change management. In diligence, investors and strategic buyers increasingly treat that as a strength, not overhead.
The highest-leverage clinical wedge: coronary CT (CAC, plaque, CT-FFR) 🔗
Coronary CT is one of the most commercially interesting areas because it can influence downstream utilization while staying close to an imaging workflow. If you want a one-sentence version: it’s where “software that measures” can become “software that changes decisions.”
2025 clearance count for three coronary CT sub-areas 🔗
- Plaque quantification / plaque characterization: 9 cleared devices
- CAC / coronary calcium scoring: 7 cleared devices
- FFR / FFRct / QFR: 3 cleared devices
These counts are directional estimates based on device names and intended use descriptions.
Plaque-related devices (2025) 🔗
| decisiondate | knumber | applicant | devicename | productcode | type |
|---|---|---|---|---|---|
| 2025-03-07 | K242338 | Cleerly, Inc. | Cleerly LABS (v2.0) | QIH | Special |
| 2025-03-27 | K243038 | Artrya Limited | Salix Central | QIH | Traditional |
| 2025-05-14 | K242624 | Shanghai United Imaging Healthcare Co., Ltd. | Medical Image Post-processing Software (uOmnispace.CT) | QIH | Traditional |
| 2025-06-18 | K243672 | GE Medical Systems SCS | CardIQ Suite | JAK | Traditional |
| 2025-07-18 | K250902 | HeartFlow, Inc. | HeartFlow Analysis | PJA | Traditional |
| 2025-08-20 | K251837 | Artrya Limited | Salix Coronary Plaque (V1.0.0) | QIH | Traditional |
| 2025-09-04 | K251656 | Careverse Technology Pte. Ltd. | Careverse CoronaryDoc (Careverse CoronaryDoc) | QIH | Traditional |
| 2025-10-27 | K251027 | Circle Cardiovascular Imaging Inc. | cvi42 Coronary Plaque Software Application | QIH | Traditional |
| 2025-11-28 | K252217 | Canon Medical Informatics, Inc. | CT VScore+ | JAK | Traditional |
CAC-related devices (2025) 🔗
| decisiondate | knumber | applicant | devicename | productcode | type |
|---|---|---|---|---|---|
| 2025-01-27 | K243229 | BunkerHill Health | Bunkerhill AVC | JAK | Traditional |
| 2025-03-27 | K243038 | Artrya Limited | Salix Central | QIH | Traditional |
| 2025-05-14 | K242624 | Shanghai United Imaging Healthcare Co., Ltd. | Medical Image Post-processing Software (uOmnispace.CT) | QIH | Traditional |
| 2025-06-18 | K243672 | GE Medical Systems SCS | CardIQ Suite | JAK | Traditional |
| 2025-08-20 | K251837 | Artrya Limited | Salix Coronary Plaque (V1.0.0) | QIH | Traditional |
| 2025-10-23 | K250288 | TeraRecon, Inc. | TeraRecon Cardiovascular.Calcification.CT | QIH | Traditional |
| 2025-11-28 | K252217 | Canon Medical Informatics, Inc. | CT VScore+ | JAK | Traditional |
FFR-related devices (2025) 🔗
| decisiondate | knumber | applicant | devicename | productcode | type |
|---|---|---|---|---|---|
| 2025-04-04 | K243769 | QFR Solutions bv | QFR (3.0) | QHA | Traditional |
| 2025-07-18 | K250902 | HeartFlow, Inc. | HeartFlow Analysis | PJA | Traditional |
| 2025-10-17 | K251355 | SpectraWAVE, Inc. | X1-FFR | QHA | Traditional |
Why investors care about these three features 🔗
Coronary CTA is a “front door” test. CAC, plaque quantification, and CT-derived functional assessment can do three things that are attractive from an outcomes and cost perspective:
- upgrade the test from “rule-out CAD” to “risk stratification,”
- support more consistent downstream medication decisions, and
- create a measurable narrative around prevention.
If you’ve ever watched a cardiologist scroll through a CTA study while simultaneously asking for a calcium score, you’ve seen the product opportunity. The friction is real, and it’s mostly workflow friction.
Beyond Coronary CT: Other Notable 2025 Trends 🔗
This section is intentionally pragmatic. It’s less about perfect taxonomy and more about “what kinds of products keep showing up.”
1) Heart failure and low ejection fraction screening 🔗
Multiple 2025 clearances involve EF estimation or low EF screening. A capital-relevant point: screening value depends on workflow integration and follow-up pathways, not just AUC.
Representative example:
- Tempus ECG-Low EF (Tempus AI) cleared in 2025.
Related external context: Eko Health raised a large growth round in 2025 to expand its cardio AI platform and commercial footprint (reported publicly).
2) Valvular disease 🔗
Valvular analysis shows up repeatedly because echo remains the most scalable structural heart imaging modality. AI features often focus on chamber quantification, measurement automation, and quality support.
A recurring commercialization truth: echo AI sells when it reduces variability for average users, not when it marginally improves expert performance.
3) ECG AI and consumer-grade screening 🔗
Apple's hypertension notification feature is a reminder that consumer tech companies can and do participate in regulated cardiology features, and they sometimes use PCCP approaches.
From a market lens, consumer screening features can expand awareness and volume, but they can also create “diagnostic spillover” that clinics need to manage.
4) Workflow and triage tools 🔗
Product codes like QAS (triage/notification) point toward an operational reality: hospitals buy tools that change queueing and throughput, even when they're skeptical of "black box diagnosis."
Capital markets: how regulatory milestones show up in financing 🔗
A 510(k) clearance is not a revenue event, but it’s a de-risking event. How much it matters depends on your business model:
- If the clearance unlocks billing, sales can accelerate quickly.
- If the clearance is one feature among many in a platform, it might be a quiet upgrade.
- If the clearance expands indication or population, it can reset TAM in investor narratives.
Below are representative case studies that tie together FDA clearances with financing and related corporate events.
Case study: HeartFlow (coronary CTA analysis, plaque, CT-FFR) 🔗
HeartFlow is useful as a “capital markets + reimbursement + regulatory” example because it sits in a clinically meaningful category and has public-market events in 2025.
Timeline highlights (public sources) 🔗
- March 26, 2025: HeartFlow reported raising $98 million via a convertible note financing (Reuters coverage).
- July 17, 2025: HeartFlow filed for a U.S. IPO (Reuters).
- July 18, 2025: HeartFlow Analysis received FDA clearance (K243178).
- August 2025: HeartFlow's IPO pricing and debut were covered by Reuters, including a reported valuation and proceeds.
Strategy interpretation 🔗
For CEOs: HeartFlow illustrates that when your product is tied to clinical decisions and potential billing, investors can accept longer evidence cycles. But it also shows that capital markets expect a coherent “why now” story that combines:
- reimbursement readiness,
- scalable distribution,
- and a credible evidence plan.
For investors: HeartFlow is a reminder that regulatory clearance can be a timing tool. Companies sometimes line up IPO, reimbursement updates, partnerships, and clearances to build a cohesive narrative.
Case study: Cleerly (coronary plaque analysis) 🔗
Cleerly is a good example of a “software-only” coronary CTA company that leans into clinical and economic value claims.
Timeline highlights (public sources) 🔗
- December 2024: Cleerly announced $106 million in growth financing (public reporting).
- December 2024: Cleerly announced the AMA CPT Editorial Panel approved a Category I CPT code 75577 for plaque analysis from CCTA, effective January 1, 2026 (company announcement).
- March 7, 2025: Cleerly LABS v2.0 received FDA clearance (K243761).
Strategy interpretation 🔗
Cleerly’s approach is a common playbook for “treatment-influencing” imaging AI:
- Regulatory clearance reduces perceived adoption risk.
- Financing supports prospective evidence and commercialization.
- Coding and payment work reduces the “we can’t get paid” objection.
The caution: if you’re building in the same category, you’re competing against a company that has invested in the full stack (regulatory, evidence, payment, narrative). That raises the bar.
Case study: Anumana (ECG-AI screening) 🔗
Anumana provides a useful counterpoint: ECG-AI can be distributed broadly, but monetization depends heavily on clinical pathway integration.
Timeline highlights (public sources) 🔗
- Anumana has received multiple Breakthrough Device Designations for ECG-AI use cases (publicly discussed).
- Anumana announced a Series C financing and strategic collaborations, including Boston Scientific participation and Mayo Clinic Platform collaboration (Business Wire reporting).
- July 28, 2025: ECG-AI Low Ejection Fraction received FDA clearance (K250615).
Strategy interpretation 🔗
For capital: ECG-AI is attractive because distribution can be software-first, but reimbursement and adoption can be tricky. The “unsexy” parts matter:
- how clinicians respond to flagged patients,
- what confirmatory testing is triggered,
- and whether payers see net savings or just increased utilization.
Case study: Eko Health (auscultation + ECG AI platform) 🔗
Eko’s story is a useful example of a platform approach that can combine device distribution with algorithm upgrades.
- May 2025: Eko announced a growth funding round (public reporting).
- August 12, 2025: Eko Duo ECG System received FDA clearance (K251494).
From a CEO lens: the hard part is not building one model. It’s building a system where new models can be deployed, trusted, and paid for.
Predicate network analysis: which devices are the “hubs”? 🔗
Top “predicated by” devices (direct) 🔗
| knumber | year | applicant | devicename | productcode | direct_cited_by | indirect_cited_by |
|---|---|---|---|---|---|---|
| K213998 | 2022 | Circle Cardiovascular Imaging Inc | cvi42 Auto Imaging Software Application | QIH | 6 | 7 |
| K203256 | 2021 | Imbio, LLC | Imbio RV/LV Software | QIH | 6 | 9 |
| DEN190040 | 2020 | Bay Labs, Inc. | Caption Guidance | QJU | 4 | 5 |
| K232699 | 2023 | Anumana, Inc. | Low Ejection Fraction AI-ECG Algorithm | QYE | 4 | 4 |
| K222463 | 2022 | Ultromics Limited | EchoGo Heart Failure | QUO | 4 | 6 |
| K210085 | 2021 | Zebra Medical Vision Ltd. | HealthCCSng | JAK | 4 | 7 |
| K230223 | 2023 | Bunkerhill, Inc | iCAC Device | JAK | 3 | 3 |
| K173780 | 2018 | Bay Labs, Inc. | EchoMD Automated Ejection Fraction Software | LLZ | 3 | 10 |
| K201555 | 2020 | Ultromics Ltd | EchoGo Pro | POK | 3 | 4 |
| K210791 | 2021 | eko.ai Pte. Ltd. d/b/a Us2.ai | Us2.v1 | QIH | 3 | 3 |
Predictions for 2026 🔗
These are not facts. They’re operating hypotheses based on the dataset trendline and public reimbursement and financing signals.
1) Clearance volume likely stays high 🔗
Cardiology AI/ML clearance volume increased from 62 (2024) to 92 (2025). A simple growth extrapolation suggests 2026 could land around 100 to 125 decisions in cardiology AI/ML.
2) Coronary CT stacks should benefit from coding clarity 🔗
With 2025 OPPS increases for CCTA and 2026 CPT updates tied to CT-derived FFR and CT perfusion, expect continued product investment in:
- plaque quantification,
- calcium scoring automation,
- and functional assessment add-ons.
3) PCCP adoption will broaden 🔗
The presence of PCCP-authorized clearances in 2025 suggests more teams will pursue this route, particularly where:
- the business model depends on frequent updates,
- the model is trained continuously, or
- performance drift is a known operational risk.
Don’t get left behind in 2026! 🔗
Regulatory hurdles shouldn't stall your innovation. Let’s get you FDA cleared in 2026. More and more competitors are arriving on the scene and we don’t want you to be left behind.
Already on the market? You need to protect your market share by accelerating your regulatory clearance process and culture. Hire us for one of our done-for-you services so we can teach your team how it is done.
Innolitics guarantees 510(k) clearance if:
- Your algorithm passes our suggested acceptance criteria as we determine during our regulatory strategy and presub service
- And getting FDA clearance is one of the top priorities of your organization
Innolitics guarantees 510(k) submission timeline if :
- Your algorithm is complete and can be containerized or be packaged as an SDK
- And you allow us to project manage the submission and allow us to communicate directly with your engineers
- And getting FDA clearance is one of the top priorities of your organization
