Elucid's FFR-IQ posted 92.2% sensitivity against invasive FFR in a study whose methods have not yet published. Here is what that number means in a real cardiology practice.
The FDA cleared Elucid's FFR-IQ on September 29, 2026. The company announced it on October 21. The full study data land November 1 at the TCT meeting in San Diego.
That six-week gap is the first thing to understand about this clearance.
FFR-IQ is software that estimates fractional flow reserve from a coronary CT angiogram, a process called FFR-CT. Fractional flow reserve measures whether a coronary narrowing restricts blood flow to the point of causing symptoms or risk. The traditional measurement runs a pressure wire through a catheter into the coronary artery. FFR-CT tries to answer the same question from a scan the patient already had.
The headline numbers are strong. Sensitivity 92.2%. Specificity 93.1%. Elucid calls them the highest of any FDA-cleared FFR-CT validation study, a claim the company ties to the FDA 510(k) summary K262015. The full ELUCIDATE study, which generated those numbers, has not published its sample size, patient population, or methods in any trade outlet that covered the clearance. What we know about the study fits in two lines.
This newsletter covers health technology and policy from a health system perspective. When an FDA clearance arrives with performance claims but no published methods, the right response is not skepticism and not enthusiasm. It is precision. What do the numbers mean? What do they not tell us? Who pays, and who can actually use it? This article answers those questions with the evidence in hand and marks clearly where the evidence stops.
1. What the FDA Cleared, and What the Coverage Does Not Say
Three things are clear from the public record.
First, Elucid received FDA 510(k) clearance for FFR-IQ under summary K262015, dated September 29, 2026. 510(k) clearance means FDA found the device substantially equivalent to a predicate device. It is not approval based on clinical evidence of effectiveness. That distinction matters for any software that will influence whether a patient goes to the cath lab.
Second, Elucid is now the third vendor offering commercial FFR-CT in the United States, after HeartFlow in 2016 and KeyaMed in 2022. HeartFlow is the market reference point. Its Q1 2026 revenue was $52.6 million, up 41% year over year, with 2026 guidance of $228 million to $232 million and more than 650,000 patients analyzed worldwide. That is ten years of building a market, regulatory pathway, billing infrastructure, and clinical evidence base. Elucid enters a market that HeartFlow built, which shortens some timelines and creates real competition on performance.
Third, FFR-IQ builds on Elucid's Plaque-IQ, which FDA cleared in October 2024. Plaque-IQ assesses coronary plaque burden and composition from CCTA, and the company says it is the only FDA-cleared CT plaque platform trained and validated on histology. FFR-IQ extends that platform to include a physiologic estimate. The combined product assesses anatomy, physiology, and plaque from a single scan, with what Elucid describes as a new interface and reporting workflow.
What the public record does not give, and what four trade outlets did not provide:
The sample size of the ELUCIDATE study. The number of sites. The patient population. The prevalence of flow-limiting disease in enrolled patients. Any AUC or calibration data. How the software handles image quality variation or patient body habitus. Turnaround time. Pricing. Payer coverage.
Kelly Huang, PhD, Elucid's CEO, said physicians want a more accurate FFR-CT. The claim is that FFR-IQ delivers it. The November 1 presentation at TCT will be the first chance to read the methods behind that claim.
2. Why the Cath Lab Is Where the Money and the Risk Sit
Start with the scale of the problem.
The American Heart Association's 2026 statistics report counts 915,973 cardiovascular deaths in the United States in 2023, including 349,470 from coronary heart disease. The report also counts 130.6 million US adults, or 48.9% of all adults, living with some form of cardiovascular disease in 2021 to 2023. Chest pain is a high-volume referral reason, and every pathway for it ends at the same decision: stress test, CT, or catheterization.
The case against reflexive catheterization is old and large. In a registry analysis presented by Manesh Patel, MD, of Duke University, reported by MDedge, 397,954 stable patients had elective diagnostic angiography at 601 US hospitals from 2004 to 2008. Only 37.5% had obstructive coronary disease. By our arithmetic, about 62.5% of those patients had an invasive procedure and found nothing obstructive. The data are more than fifteen years old and practice has changed, but the size of the problem has not disappeared.
The better modern evidence is the PRECISE trial, summarized by the American College of Cardiology. PRECISE randomized 2,103 adults with stable symptoms suspicious for coronary disease, and no known obstructive disease, to a precision diagnostic strategy or usual testing, with twelve months of follow-up. The precision arm used CCTA with an algorithm to identify patients who needed further testing including FFR-CT.
Catheterization without obstructive disease fell from 10.2% to 2.6%. That is a 75% relative reduction, by our arithmetic. The primary composite outcome, combining unnecessary catheterization with death or nonfatal myocardial infarction, fell from 11.3% to 4.2%. All-cause death was 0.5% versus 0.7%. Nonfatal myocardial infarction was 1.2% versus 0.5%, which ran the other direction.
Two things to be precise about. PRECISE tested a precision diagnostic strategy, not FFR-IQ. The results show the size of the benefit that a better gating tool could deliver, not that this tool delivers it. And the myocardial infarction signal going the other direction in the precision arm is worth noting even if the difference is small.
For any hospital that has tracked its own cath lab referrals, the arithmetic is local. How many patients go invasively and come back clean? That rate is the baseline. A tool that is supposed to reduce it should be measured against that number in your population, not a trial population.
3. The BioIntegrated Approach: One Scan, Three Answers
Earlier FFR-CT products estimate coronary blood flow from vessel geometry. They model fluid dynamics based on the size and shape of the stenosis. Elucid says FFR-IQ does something different.
Per Cardiovascular Business, FFR-IQ incorporates plaque type and volume into the physiologic estimate. The rationale: plaque composition affects vasodilatory capacity and coronary flow even when the degree of stenosis looks similar. A 60% narrowing filled with dense calcified plaque behaves differently from a 60% narrowing with lipid-rich necrotic core. Geometry alone does not capture that difference.
Amir Ahmadi, MD, a cardiologist at Mount Sinai and Elucid's lead scientific advisor, described the logic to Diagnostic Imaging: "Plaque composition, coronary anatomy, and physiology are closely connected." He is a company voice, not an independent reviewer. The design logic is coherent. Whether it improves clinical decisions in a non-trial population is what the ELUCIDATE data will begin to answer.
The practical output: one CCTA now produces a stenosis grade, a functional estimate, and a plaque burden in the same report. That is a more complete picture of a patient's coronary status than anatomy alone.
It also raises staffing questions that no trade outlet addressed.
Who signs the functional estimate? At most hospitals, CCTA is read by a radiologist. FFR and its derivatives have historically lived in cardiology. A combined anatomy-physiology-plaque report sits at the intersection of both specialties. Some institutions have jointly-trained cardiac imagers who handle this already. Many do not.
Who owns the plaque finding? A lipid-rich necrotic core finding on CT is a predictor of risk, not a diagnosis. When it appears in a report, someone has to explain it to the patient, decide whether to act on it, and document the conversation. In a high-volume imaging center, that conversation takes time and infrastructure neither specialty has necessarily built.
Anyone who watched lung screening CT evolve from a single nod-or-no lung finding into a multi-finding report covering lung nodules, incidental cardiac findings, and bone density will recognize the trajectory.
4. Reading 92.2% in a Real Clinic
Sensitivity and specificity describe the test. Clinicians act on predictive values, which shift with the population.
Positive predictive value is the probability that a positive result is a true positive. Negative predictive value is the probability that a negative result is a true negative. Both depend not just on the test's performance but on how common the disease is in the people being tested.
We ran the arithmetic with 92.2% sensitivity and 93.1% specificity across three assumed disease prevalence rates. The ELUCIDATE study's actual prevalence is not in any coverage we read. These figures are our own, chosen to show the range.
At 10% prevalence, a population where the large majority of patients do not have flow-limiting disease: positive predictive value is about 59%. Roughly six in ten positive FFR-IQ results are true positives, and four are false positives who may still proceed to catheterization. Negative predictive value is about 99%.
At 30% prevalence, a higher-risk referral population: positive predictive value rises to about 85%. About 85 of 100 positive results are correct. Negative predictive value is about 97%.
At 50% prevalence, a very high-risk cohort: positive predictive value reaches about 93%. Negative predictive value falls to about 92%.
Here is the same picture in patients. Start with 1,000 patients at 30% prevalence. About 300 have flow-limiting disease. FFR-IQ with 92.2% sensitivity identifies about 277 of them correctly and misses about 23. Among the 700 without flow-limiting disease, FFR-IQ with 93.1% specificity clears about 652 correctly and flags about 48 as positive who do not have the problem. Those 48 patients are the ones who may still go to the cath lab unnecessarily, and the 23 are the ones sent home who were missed.
The lesson: the same software behaves very differently depending on where it is used. A low-risk population presenting to a rural urgent care or a primary care office will see lower positive predictive values. A high-risk referral population at an academic cardiac center will see higher ones. The clinical pathway around the tool has to match. Referral thresholds, downstream protocols, and patient counseling should not be designed for the trial population and applied to a different one.
The ELUCIDATE prevalence is the number that matters most for calibrating this, and it is not yet public.
5. The Real-World Gap, and What Local Validation Costs
Validation studies are the best-case scenario. Real-world performance is the stress test, and for FFR-CT the gap between the two has been measured.
Cardiovascular Business describes a JACC: Cardiovascular Imaging analysis led by Tarun Mittal, MD, MSc, of Imperial College London. The analysis covered more than 2,200 coronary CT angiograms from 12 UK facilities. FFR-CT showed a positive predictive value of 49% and a negative predictive value of 76%. The multicenter trials that supported a 2017 NICE recommendation for FFR-CT had shown 65% and 93%.
The cost comparison is just as pointed. A UK pricing model found that an FFR-CT strategy averaged 2,102 pounds, or about $2,657 per patient, against 1,411 pounds, or about $1,785, for stress imaging. The difference came from higher rates of invasive angiography and revascularization in the FFR-CT arm. The authors concluded that widespread adoption "should be undertaken with caution."
Caveats are real and should be stated. The study may carry referral bias or verification bias, standard concerns in diagnostic imaging analysis. UK pricing and clinical culture differ from the US. The analysis does not name the FFR-CT vendor, and it was not conducted on FFR-IQ, which received clearance this month. Treat it as a warning about the gap between trial and practice, not as a verdict on any product.
What the UK data establish is that the gap between validation-study performance and real-world performance is measurable and large enough to matter. For a tool with 92.2% and 93.1% in its validation study, that history argues for local validation before wide deployment.
What does local validation require?
A baseline rate of catheterization without obstructive disease in your population, tracked before the tool goes live. A tracking system that follows referred patients through to their downstream procedure results. A sample size large enough to detect a meaningful change in that rate. A time horizon of twelve to eighteen months minimum. These are not obstacles to adoption. They are the evidence that adoption is working.
A hospital or health system that builds this infrastructure gets two things: defensible quality data for contracts and credentialing, and an early signal if the tool is performing below expectation in their population.
6. Reimbursement, Access, and the Patient Who Cannot Get a CT
Clearance and reimbursement are separate events.
CPT 75580, the code for noninvasive FFR-CT assessment, became a Category I code on January 1, 2024, replacing four Category III codes. AuntMinnie reported in January 2024 that the American College of Cardiology, the American College of Radiology, and the Society of Cardiovascular Computed Tomography jointly petitioned the AMA for the single Category I code, and that CMS raised hospital payment by approximately 7% for 2024. A HeartFlow payer document places CPT 75580 in APC 5724 with a calendar year 2024 national average hospital outpatient rate of $997.22. That is a company document and a two-year-old figure; confirm current rates with CMS payment files before building a budget.
The open billing questions for FFR-IQ are unanswered in every source we read. How will Elucid price the software? How will it be billed, and will the existing CPT code apply to a combined anatomy, physiology, and plaque report? How do commercial and Medicaid payers treat the combined product? Those answers belong to your revenue cycle team, your Medicare contractor, and your payer contracts, and they need to be in hand before the first order.
Workflow adds another variable. A 2017 TCTMD report described FFR-CT at that time as calculated off-site, with turnaround measured in hours to days. Hasan Jilaihawi, MD, of NYU Langone, called it a possible gatekeeper for the cath lab. None of the FFR-IQ coverage addresses turnaround, on-site versus cloud processing, or how fast results return to the ordering clinician. For a chest pain pathway, turnaround time determines whether the result changes today's clinical decision or reaches the chart after the patient has already been referred.
For patients at federally qualified health centers, the access question starts before the scan.
HRSA's 2025 Uniform Data System counts 32,746,392 health center patients. Of those, 47.92% were on Medicaid or CHIP and 17.21% were uninsured. Among patients with reported income, 67.73% were at or below the federal poverty guideline. Hypertension, the sentinel condition for cardiovascular risk, was reported in 5,573,807 health center patients in 2025, with blood pressure controlled in 69.08% of those measured. That is roughly 1.7 million health center patients with uncontrolled hypertension, all of whom are at elevated risk for coronary disease.
Most FQHCs do not own CT scanners. They refer. Referral to CCTA, and from there to FFR-CT interpretation, runs through a chain of steps that each carry wait times, coordination costs, and the possibility that results never return to the primary care team. The question for an FQHC executive is not whether to buy FFR-IQ software. It is whether the CT referral pathway is functional enough that a patient who needs the test can get it and have the result acted on.
Five questions define whether that pathway works. Where is the CCTA done, and how long is the wait? Who reads it, and does the report include a functional estimate? How does the result return to the health center? What happens to a patient with a borderline or positive result? And what does an uninsured patient pay out of pocket?
A health center that can answer all five is positioned to benefit from the improvements FFR-CT may offer. One that cannot has found the real project.
Connecting the Dots: What Comes Next for FFR-IQ
The clearance is real progress. A CT scan that answers coronary anatomy, physiology, and plaque composition in one pass is more information from an exam the patient was having anyway. The early performance numbers, if they hold in independent data, are genuinely strong.
The proof comes in steps.
The November 1 ELUCIDATE presentation at TCT will show what the numbers actually mean: sample size, patient population, prevalence of flow-limiting disease, and methodology. Read those carefully before committing to a deployment timeline.
After that, the path that matters is independent real-world data in populations that look like the patients a given health system or FQHC actually serves. The UK analysis of more than 2,200 scans across 12 sites is not a verdict on FFR-IQ. It is a reminder that trial populations and real-world populations are not the same, and that the gap between them should be measured, not assumed away.
For radiologists and cardiologists, the workflow question is immediate. Who interprets the combined anatomy, physiology, and plaque report? Who documents the plaque finding and counsels the patient? Those staffing agreements need to exist before the software does.
For health systems, set a baseline now. What is your current rate of catheterization without obstructive coronary disease? Track it prospectively. Any tool that promises to reduce that rate should be held to that metric in your facility.
For FQHCs and community health centers, the question is access, not software. Map the CT referral pathway. Find where patients fall out between the order and the result returning to the primary care team. Fix that pathway first.
This connects directly to the work Oatmeal Health is doing in lung cancer screening. LungAI is our AI tool for lung CT analysis in communities that screening guidelines have historically missed. The FDA clearance picture there is different: LungAI is investigational and not FDA cleared. But the operating questions are the same. A clearance or a study result does not help a patient unless the scan happens, the result returns, and someone acts on it. Building that infrastructure in communities that have been underserved by imaging is the work. FFR-IQ's clearance is a step forward for coronary CT. The step that matters for most patients is whether the referral chain that gets them to the scan and back exists in the first place.
Jonathan Govette is Co-Founder and CEO of Oatmeal Health, an AI lung cancer diagnostics company catching cancers earlier in underserved communities. He writes about radiology, AI diagnostics, and health policy for health system leaders, FQHCs, and imaging teams.
Key References
Cardiovascular Business, Dave Fornell, October 21, 2026: "FDA clears Elucid's new AI-enabled FFR-CT software." Source for the clearance, ELUCIDATE figures, third-vendor statement, Plaque-IQ's October 2024 clearance, TCT November 1 date, and plaque-physiology design rationale. https://cardiovascularbusiness.com/topics/artificial-intelligence/fda-clears-elucids-new-ai-enabled-ffr-ct-software
DAIC (Diagnostic and Interventional Cardiology), October 23, 2026: Elucid FFR-IQ announcement coverage. Source for FDA 510(k) summary K262015, the footnote on the "highest" claim, and the "majority of unnecessary catheterizations" language. https://www.dicardiology.com/node/85568
AuntMinnie, October 21, 2026: "Elucid gets FDA clearance for plaque-informed FFR-CT software." Source for the combined platform description. https://www.auntminnie.com/clinical-news/ct/news/15838110/elucid-gets-fda-clearance-for-plaqueinformed-ffrct-software
HMP Global Learning Network, Cath Lab Digest, October 21, 2026: "FDA Clears Elucid's FFR-IQ Noninvasive FFR-CT Assessment." Source for Kelly Huang CEO quote and ELUCIDATE TCT presentation details. https://www.hmpgloballearningnetwork.com/site/cathlab/press-release/fda-clears-elucids-ffr-iq-noninvasive-ffr-ct-assessment
Diagnostic Imaging, Jeff Hall, October 2026: FFR-IQ coverage and Amir Ahmadi, MD, comments. https://www.diagnosticimaging.com/view/fda-ai-enabled-ffr-iq-software-estimating-fractional-flow-reserve-from-ccta
American College of Cardiology: PRECISE trial summary. Randomized trial, 2,103 patients, 10.2% vs. 2.6% unnecessary catheterization. https://www.acc.org/Latest-in-Cardiology/Clinical-Trials/2023/11/20/14/54/precise
American Heart Association: 2026 Heart Disease and Stroke Statistics, Circulation, January 21, 2026. Source for 915,973 cardiovascular deaths, 349,470 from coronary heart disease, 130.6 million adults with cardiovascular disease. https://professional.heart.org
MDedge, Hospitalist News, Bruce Jancin, March 1, 2009: Manesh Patel registry analysis. 397,954 patients, 601 hospitals, 37.5% obstructive disease. https://www.mdedge.com/content/many-elective-angiographies-are-deemed-unnecessary
Cardiovascular Business, Michael Walter: JACC Cardiovascular Imaging analysis, Tarun Mittal, MD, MSc. 2,200+ CCTAs, 12 UK sites, PPV 49%, NPV 76% vs. 65%/93% in NICE trial data. 2,102 vs. 1,411 GBP per patient. https://cardiovascularbusiness.com/topics/cardiac-imaging/researchers-question-ffr-cts-value-compared-other-cardiac-imaging-options
AuntMinnie, January 24, 2024: CPT 75580 Category I code effective January 1, 2024. https://www.auntminnie.com/clinical-news/ct/article/15662727/new-cpt-code-issued-for-ffrct
HeartFlow, Category I transition document for health care providers: OPPS APC 5724, $997.22 national average hospital outpatient rate, CY2024. https://www.heartflow.com/wp-content/uploads/2025/02/cat-i-transition-for-hcps.pdf
Cardiovascular Business: HeartFlow Q1 2026 financial results. $52.6M revenue, 41% growth, 650,000+ patients analyzed worldwide. https://cardiovascularbusiness.com/topics/healthcare-management/healthcare-economics/heartflow-reports-significant-revenue-growth-ccta-sees-wider-adoption
TCTMD, Yael L. Maxwell, July 9, 2017: FFR-CT workflow, off-site calculation, hours-to-days turnaround, ~$1,500 analysis cost, Hasan Jilaihawi, MD comment. https://www.tctmd.com/news/faster-site-ffrct-may-be-horizon-through-innovation-and-collaboration
HSRA Health Center Program Uniform Data System, national data, 2025: 32,746,392 health center patients; 5,573,807 with hypertension (28.68%); Medicaid/CHIP 47.92%; uninsured 17.21%; 67.73% at or below FPG; BP control 69.08%. https://data.hrsa.gov/topics/healthcenters/uds/overview/national
Our arithmetic (labeled in text): PPV/NPV table computed from 92.2% sensitivity and 93.1% specificity at assumed 10%, 30%, and 50% prevalence. 75% relative reduction in unnecessary catheterization (10.2% to 2.6% in PRECISE). 62.5% of angiography patients without obstructive disease (Patel registry). Dollar conversions from GBP at approximate exchange rate.