CT-derived fractional flow reserve

The physiological question

FFR estimates the pressure consequence of epicardial disease during hyperaemia

Fractional flow reserve is conceptually the maximal flow distal to a stenosis divided by the maximal flow that same artery could supply without the stenosis. Under low, stable hyperaemic microvascular resistance, the invasive pressure ratio Pd/Pa estimates that flow ratio.

Anatomical stenosis is an imperfect predictor of flow limitation. Lesion length, vessel size, supplied myocardial mass, serial lesions, diffuse disease, and microvascular conditions all matter. CT-FFR adds physiology to an already diagnostic CTCA.

No second scan

CT-FFR computes a virtual hyperaemic pressure field from CTCA

Software segments the aorta and coronary tree, constructs a three-dimensional anatomical model, estimates boundary conditions and microvascular response, and applies computational fluid dynamics or validated machine-learning methods. The output provides values along epicardial vessels.

• No additional iodinated contrast, radiation, or pharmacological stress is required beyond the source CTCA.
• The analysis may be off-site or on-site depending on the validated platform.
• Different algorithms are not automatically interchangeable; follow platform instructions and local validation.
• A coloured vessel model is an analysis product, not a substitute for reviewing the source CTCA.

Garbage in, physiology out

Source CTCA requirements

1. Diagnostic enhancement: consistent luminal contrast with adequate contrast-to-noise.
2. Motion control: minimal cardiac, respiratory, body, transition, and misregistration artifact.
3. Nitroglycerin: use sublingual nitrate unless contraindicated; small vessels and spasm degrade segmentation and specificity.
4. Complete coverage: aorta and the full relevant coronary tree without truncation.
5. Thin reconstruction: appropriate slice thickness/increment and a validated kernel; preserve the best phase.
6. Manual QC: confirm vessel model against source axial/curved MPR, especially around calcium, bifurcations, stents, and motion.

The 2026 SCCT/SCAI consensus recommends CT-FFR only from high-quality CCTA, with heart-rate control when required and nitrate administration.

Use it where it changes a decision

Best use: anatomically intermediate disease

Current 2026 consensus recommends CT-FFR assessment for 50–90% diameter stenosis when the result could change downstream investigation or management beyond CTCA alone.

• Moderate/intermediate stenosis whose functional significance is uncertain.
• Selected severe-appearing lesions when blooming or lesion characteristics make anatomy–physiology integration useful and immediate invasive evaluation is not already clear.
• Multivessel or serial disease where the physiologically dominant lesion may differ from the visually tightest lesion.
• Diffuse disease where the pressure pattern can distinguish focal step-down from gradual whole-vessel decline.
• Selected proximal 40–49% stenosis with concerning morphology when functional information would alter care, consistent with CAD-RADS context.

Do not send every CTCA automatically. A normal study, mild distal plaque, or clearly actionable high-risk anatomy often gains little from CT-FFR.

Know the boundaries

Situations where CT-FFR is unreliable, unvalidated, or low value

Location changes the number

Measure 2 cm distal to the stenosis

The 2026 consensus standardises three complementary values:

• Post-lesional CT-FFR: measured approximately 2 cm distal to the stenotic plaque; include this in every CT-FFR report. It has the strongest evidence for lesion-specific diagnosis and risk.
• Proximal value: immediately before the lesion; used with the post-lesional value to calculate ΔCT-FFR.
• Distal vessel nadir: lowest value at the distal tip; reflects cumulative vessel disease, size/mass relationship, and possible artifact—not automatically one focal lesion.

Read the curve, not one number

Focal step-down versus diffuse decline

Two related frameworks

Use continuous risk; state which cut-points you apply

The traditional binary clinical threshold is ≤0.80, but values near that boundary are uncertain and risk is continuous. Two current frameworks appear in practice:

The newer consensus stresses the continuous value: lower CT-FFR generally means greater probability of flow limitation and adverse outcome. The “grey zone” is not solved by rounding. Integrate lesion site, supplied territory, symptoms, anatomy, ΔCT-FFR, and confidence.

Translesional gradient

ΔCT-FFR localises the pressure loss

ΔCT-FFR is the value immediately proximal to a lesion minus the post-lesional value. A larger drop supports focal lesion-specific significance.

Example: proximal 0.94, 2 cm distal 0.78 → Δ = 0.16. The 2026 consensus suggests considering a Δ around 0.12 with CT-FFR ≤0.80 as a high-probability focal pattern, while acknowledging that prospective validation of the exact cut-point is still developing.

Delta is particularly helpful for grey-zone values and when a proximal lesion has a preserved post-lesion number but the distal vessel eventually falls below 0.80.

Whole-tree physiology

The tightest-looking lesion may not be the dominant physiological lesion

• Serial lesions: pressure loss across one lesion changes flow across the other; simple independent interpretation can mislead.
• Diffuse disease: gradual decline may produce a low distal value without a revascularisable focal step.
• Multivessel disease: compare lesion-specific values, territories, symptoms, and anatomical feasibility rather than ranking only distal nadirs.
• Small distal branches: values normally decline and segmentation is more vulnerable; do not plan PCI from the terminal tip alone.
• Heart-team cases: complex left-main, multivessel, CTO, and surgical disease require multidisciplinary integration beyond a colour map.

Special situations

What the model may not capture cleanly

When analysis fails

Common rejection and false-result mechanisms

• Cardiac motion, slab misregistration, respiratory motion, or poor ECG phase.
• Low/inhomogeneous coronary enhancement or excessive image noise.
• Severe blooming, stent metal, or beam hardening obscuring vessel boundaries.
• Truncated distal vessels or incomplete z-axis coverage.
• Incorrect centreline/lumen segmentation at bifurcations, calcified plaque, or veins.
• Small distal vessels, absent nitrate, or anatomical variants outside algorithm validation.

If broader phase data were acquired, a cleaner reconstruction may permit resubmission. Never manipulate the model to force a plausible answer; record non-diagnostic CT-FFR and use an alternative pathway.

A complete physiology handoff

Minimum CT-FFR report

1. State platform/method and whether the analysis was diagnostic.
2. Repeat the anatomical lesion location and stenosis range from CTCA.
3. Report CT-FFR approximately 2 cm distal to each relevant lesion.
4. Describe proximal value and ΔCT-FFR when informative, particularly in grey-zone results.
5. Describe focal step-down, gradual diffuse decline, serial lesions, and distal nadir separately.
6. State concordance/discordance with anatomy and any segmentation limitation.
7. Apply CAD-RADS I+, I−, or I± only under its defined thresholds and context.

Read the pattern

Worked CT-FFR cases

Case A · 55% LAD; post-lesion 0.86, distal tip 0.77

Low probability that the focal LAD lesion is flow limiting. The gradual distal decline may reflect diffuse disease or small-vessel/model effects. Do not label the lesion I+ from the terminal nadir.

Case B · 60% RCA; 0.94 before and 0.78 after

Grey-zone post-lesional value with Δ 0.16. This focal drop raises the probability of lesion-specific significance; integrate symptoms, location, territory, and confidence rather than using 0.80 as a switch.

Case C · 45% proximal LAD with HRP; post-lesion 0.72

Anatomy and physiology are discordant in severity but concordant in concern. Reinspect the source/model, confirm measurement and territory, then report lesion-specific I+ if valid.

Case D · Severe left-main narrowing; CT-FFR requested

Do not use CT-FFR to decide left-main significance under the 2026 consensus. Communicate the high-risk anatomy and pursue the appropriate invasive/anatomic pathway.

Case E · No visible stenosis; distal LCx 0.75

Search for missed plaque, motion, segmentation error, small-calibre vessel, and absent nitrate. An isolated low distal value without anatomical correlate is not proof of a focal culprit.

Active recall

Self-test: answer before opening

Printable quick revision

One-page CT-FFR card

• CT-FFR: virtual hyperaemic epicardial pressure physiology; no extra scan/stress drug.
• CAD-RADS I: ≤0.75 positive · 0.76–0.80 borderline · >0.80 negative.
• 2026 probability: <0.70 high · 0.70–0.80 intermediate · >0.80 low.
• Delta: proximal − post-lesional; ~0.12 with ≤0.80 supports focal significance, but remains contextual.
• Distal low alone: may mean diffuse disease, small vessel, or artifact—not one focal lesion.
• Avoid: non-diagnostic CTCA, LM significance, myocardial bridge, anomalous artery; platform caution for stent/graft/CTO.
• QC: nitrate, motion-free phase, contrast, complete tree, manual segmentation review.
• Final rule: colour map + curve + source CTCA + clinical question.

Provenance

Sources and scope

This chapter follows Dr Bhavana Reddy’s Webinar 4 CT-FFR teaching and discussion. Automated captions contain transcription errors; key practice points were updated to the June 2026 SCCT/SCAI/ACC-endorsed consensus.

1. 2026 SCCT/SCAI expert consensus on CT-FFR — patient selection, quality, 2-cm measurement, focal/diffuse patterns, probability bands, ΔCT-FFR, pitfalls, and reporting.
2. CAD-RADS 2.0 open consensus — I modifier thresholds and lesion-specific interpretation.
3. SCCT scientific document index — current official CT-FFR and CCTA guidance.

Educational material only. CT-FFR software, accepted indications, and regulatory status vary by region and platform. Results must be integrated with diagnostic CTCA, symptoms, and local coronary-care pathways. External references require internet; this chapter works offline.