What happened
The Scottish Health Technologies Group (SHTG) published an overview of AI-estimated CT fractional flow reserve (CT-FFR) in May 2026. The software computes FFR, a measure of blood flow past a coronary narrowing, from a standard CT angiogram. In some patients, it could replace an invasive pressure-wire test. SHTG reviewed three meta-analyses. In one, covering 17 studies and 3,255 patients, pooled sensitivity was 0.86 and specificity 0.87 per patient. The tools in that analysis included cFFR and DEEPVESSEL-FFR. In an indirect comparison, AI tools were as specific as tools based on computational fluid dynamics but less sensitive. SHTG concluded that AI-estimated CT-FFR agrees well with invasive FFR and could spare some patients invasive testing. It found substantial uncertainty because studies defined AI differently, used different thresholds and were mostly retrospective. No study looked at later tests, health outcomes, patient experience or costs. No study tested whether CT-FFR adds value to CT angiography alone.
Sources
-
Assessment report, by authors with no tie to the system.
Scottish Health Technologies Group, Healthcare Improvement Scotland. Innovative Medical Technology Overview: Artificial intelligence (AI) estimated computed tomography fractional flow reserve (CT-FFR). May 2026, Key messages.
AI-estimated CT-FFR showed good agreement with results from invasive fractional flow reserve (FFR) measurement for assessing functionally significant coronary artery stenosis. The use of non-invasive AI-estimated CT-FFR could avoid the need for invasive FFR in some patients.
Details
AI system
- Reported configuration
-
CFFR versions 1.4 to 3.2 in one meta-analysis
- Workflow
-
The software computed FFR from standard CT coronary angiography images. The studies measured invasive FFR in the same patients as the reference test.
Follow-up
- Conditions for replication
-
CT images are of high quality, since some studies excluded poor-quality images and so inflated accuracy. Patients are adults with stable chest pain and suspected coronary artery disease.
Entry details
- Sector
-
Healthcare.
- Purpose
-
Siemens Healthineers, Keya Medical and other makers built machine-learning CT-FFR software to calculate fractional flow reserve from standard CT coronary angiography images.
30 more questions not yet checked
model, provider, role, configuration, inputs and outputs, autonomy, human oversight, AI contribution, beneficiaries, reach, distribution and access, duration, every measurement question, every money question, ownership and control, dependencies, cross-border connections, geopolitical implications, replication attempts, availability, institutional decision, grounds, action and result, success criterion, every evidence question, observed or intended, alternative assessments, related entries, other identifiers, category mappings.
Cite this record
Peter H. Vartanian. (2026). AI-estimated CT-FFR agrees well with invasive heart tests, SHTG finds (AIBD-1126). Artificial Intelligence Benefits Database, edition of 1 November 2026. https://petervartanian.xyz/aibd/records/AIBD-1126.html (archived at https://doi.org/10.7910/DVN/AADSII, version 7.0)