← All 13 topics

Topic 11 · Structured reporting

CAD-RADS 2.0

Convert a complete coronary interpretation into one standard string: maximal stenosis first, overall plaque burden second, then only the modifiers that are actually supported.

Webinar 4 · Dr Zeeshan LakhaniCAD-RADS 2.0Reporting drillsPrint-friendly
Start reading

The grammar

Stenosis / plaque burden / modifiers

CAD-RADS is a patient-level communication system for coronary atherosclerotic disease on CTCA. It does not replace the vessel-by-vessel findings; it compresses them into a standard summary linked to clinical context.

1 · StenosisWorst evaluable atherosclerotic diameter stenosis: 0, 1, 2, 3, 4A, 4B, 5—or N.
2 · PlaqueWhole-tree burden: P1 mild, P2 moderate, P3 severe, P4 extensive.
3 · ModifiersN, S, G, HRP, I, E when applicable and supported.

Example: CAD-RADS 3/P2/HRP/I− means a moderate maximal stenosis, moderate overall plaque burden, high-risk plaque morphology, and a performed CT ischaemia test that is negative for lesion-specific ischaemia.

Never code before interpreting

First complete anatomy, segmental plaque/stenosis, quality, stents/grafts, and functional analysis. The code is the final summary—not a shortcut around the study.

The first field

Patient category follows the worst atherosclerotic stenosis

CategoryMaximal diameter stenosisMeaning
CAD-RADS 00%; no plaqueNo coronary atherosclerosis on CTCA.
CAD-RADS 11–24%, or plaque without visible narrowingMinimal nonobstructive CAD.
CAD-RADS 225–49%Mild nonobstructive CAD.
CAD-RADS 350–69%Moderate stenosis.
CAD-RADS 4A70–99% in one or two vesselsSevere stenosis.
CAD-RADS 4BLeft main ≥50% or three-vessel ≥70% diseaseHigh-risk severe anatomical pattern.
CAD-RADS 5100%Total/subtotal occlusion.
CAD-RADS NSignificant disease cannot be excludedNon-diagnostic study/segment under the N rules.

Apply the scale to vessels/branches ≥1.5 mm when image quality is sufficient. The category is based on atherosclerotic stenosis; non-atherosclerotic narrowing belongs under modifier E and descriptive reporting.

The high-grade fork

4A, 4B, and 5 are not interchangeable

  • 4A: one or two vessels with 70–99% stenosis. A severe proximal LAD lesion is 4A unless the 4B definition is met.
  • 4B: left main ≥50%, or obstructive three-vessel disease with ≥70% in all three major epicardial territories.
  • 5: a total/subtotal occlusion, whether acute or chronic; describe chronicity clues and viability/territory separately.

Count major vessels, not every branch as an independent “vessel.” A 70–99% lesion in LAD plus another in D1 is not automatically two-vessel disease. The body of the report must show the anatomy behind the patient-level category.

The second field

P1–P4: grade the entire coronary plaque burden

P gradeBurdenCACSISVisual guide
P1Mild1–100≤21–2 vessels with mild plaque.
P2Moderate101–3003–41–2 vessels with moderate plaque, or 3 with mild plaque.
P3Severe301–9995–73 vessels with moderate plaque, or 1 with severe plaque.
P4Extensive>1000≥82–3 vessels with severe plaque.

No P0 is required. CAD-RADS 0 already means no plaque and no stenosis. In CAD-RADS 1–5, append the selected P grade after the stenosis category: for example, CAD-RADS 2/P3.

CAC, SIS, visual assessment, or validated quantitative plaque volume may be used according to local practice. State or internally standardise the method so follow-up and audits are meaningful.

When tools disagree

CAC counts calcium; CTCA sees total plaque

The P methods need not give the same category. A patient with CAC 20 but diffuse noncalcified plaque can have mild calcified burden yet substantial total CTCA plaque. Conversely, a densely calcified limited segment can produce a high CAC with fewer involved segments.

  • Use total quantitative plaque burden when technically valid and available.
  • Otherwise integrate CTCA visual burden/SIS rather than allowing CAC to erase noncalcified plaque.
  • Document the method and choose the grade that best represents overall plaque according to institutional policy.
  • Do not average incompatible grades into a fictitious number.

The optional fields

Use modifiers for defined facts—not decoration

ModifierMeaningTrigger
NNon-diagnosticAt least one clinically relevant segment >1.5 mm cannot be confidently interpreted.
SStentAt least one coronary stent is present.
GGraftAt least one coronary bypass graft is present.
HRPHigh-risk plaqueAt least two HRP features within a plaque.
ICT ischaemia assessmentCT-FFR or stress CT perfusion performed: I+, I−, or I±.
EExceptionClinically important non-atherosclerotic coronary abnormality.

Recommended syntax is stenosis, then P grade, then modifiers separated by slashes. Keep the impression readable: the code never replaces explanatory prose.

The most tested rule

N is a category or a modifier depending on what is already proven

Evaluable segmentsNon-evaluable segment(s)Correct form
No stenosis, minimal, or mild only (<50%)Could hide ≥50% diseaseCAD-RADS N/Px if burden can be assessed; N replaces the numeric stenosis category.
At least one proven ≥50% stenosisAdditional segment non-evaluableCAD-RADS 3–5/Px/N; retain the known numeric category and append N.

Example: mild LCx plaque plus motion-obscured proximal RCA = CAD-RADS N/P1, not CAD-RADS 2/N. But moderate LAD stenosis plus motion-obscured RCA = CAD-RADS 3/P2/N.

State the exact non-diagnostic segment and cause. “Limited study” is insufficient because downstream action depends on the territory and whether obstructive disease remains possible.

Prior revascularisation

S and G identify the altered coronary map

Stent modifier S

Use S when any coronary stent is present. Grade in-stent restenosis using the same stenosis bands when the lumen is evaluable. If the stent is non-evaluable and no other ≥50% stenosis is seen, use N/Px/S.

Graft modifier G

Use G when bypass grafts are present. Evaluate origin, body, distal anastomosis, runoff, and native vessels. A severe native lesion successfully bypassed by a fully patent graft is not used to inflate the CAD-RADS category; a graft occlusion is CAD-RADS 5/Px/G.

Map before grading

In CABG, the clinically relevant circulation includes grafts, anastomoses, native targets, competitive flow, and ungrafted territories. A numeric category without the graft map is unsafe.

Morphology modifier

HRP requires at least two features in the same plaque

The four features are positive remodelling, low attenuation (<30 HU focus with careful technique), napkin-ring sign, and spotty calcification. Append HRP only when at least two are present.

  • One feature: describe it in the body; do not use HRP.
  • HRP can accompany nonobstructive disease—for example, CAD-RADS 2/P2/HRP.
  • HRP increases concern but does not independently prove culprit status or ischaemia.
  • Do not call a plaque “vulnerable” as a promise of future rupture; CAD-RADS 2.0 uses “high-risk plaque.”

Functional modifier

I+, I−, and I± require a CT ischaemia test

ModifierCT-FFRStress CT perfusion
I+Lesion-specific value ≤0.75 in a vessel suitable for PCI, interpreted with the pressure curve.Reversible defect or peri-infarct ischaemia.
I−Lesion-specific value >0.80.No reversible ischaemia; a fixed infarct alone is I− and should be described.
0.76–0.80, requiring location, symptoms, ΔCT-FFR, and context.Borderline/equivocal perfusion result.

Do not assign I from a resting first-pass hypoperfusion impression on routine CTCA. Modifier I specifically indicates CT-FFR or stress myocardial CT perfusion analysis.

Outside atherosclerosis

E flags non-atherosclerotic coronary abnormalities

Examples include dissection, anomalous origin/course, aneurysm/pseudoaneurysm, vasculitis, fistula, extrinsic compression, and arteriovenous malformation. E indicates that the atherosclerotic stenosis code does not capture the whole problem.

Example: no atherosclerotic plaque, but an anomalous RCA with interarterial/intramural course: CAD-RADS 0/E, followed by a complete anomaly description. The E modifier is not a substitute for that description or for urgent communication when required.

Same code, different context

Stable and acute chest pain tables guide different actions

The stenosis and P categories are the same, but clinical recommendations differ. In acute chest pain, troponin, ECG, haemodynamic state, culprit morphology, and alternative diagnoses determine urgency. CAD-RADS is not an ACS rule by itself.

  • CAD-RADS 0 in an appropriate acute-pain CT pathway can support exclusion of coronary atherosclerotic cause, but still assess noncoronary emergencies and clinical data.
  • Nonobstructive plaque with HRP may be important in an acute syndrome even without ≥50% narrowing.
  • CAD-RADS 4B/5 or a suspected culprit requires timely specialist communication—not merely a code in a routine report.

Codes inform; patients decide

Keep recommendations contextual

CAD-RADS tables provide management considerations, but the final recommendation must reflect the indication, symptoms, acute versus stable setting, comorbidities, anatomy, physiology, local resources, and clinician–patient decision-making.

  • 0: consider non-atherosclerotic causes of symptoms.
  • 1–2: preventive risk-factor assessment; no routine functional test solely for mild stenosis.
  • 3: consider functional assessment if it will change management; optimise prevention/anti-anginal care.
  • 4A: functional assessment or invasive angiography depending on location, symptoms, and context.
  • 4B: invasive evaluation is generally recommended.
  • 5: invasive/functional/viability assessment according to acute/chronic presentation and territory.
  • N: alternative/additional evaluation may be needed because obstructive disease is not excluded.

Final assembly

A seven-step CAD-RADS workflow

  1. Confirm indication and acute/stable context.
  2. Finish segmental interpretation; list non-evaluable segments.
  3. Assign the worst atherosclerotic stenosis category.
  4. Determine P1–P4 from the institution’s chosen whole-tree burden method.
  5. Apply N logic before other modifiers.
  6. Add S, G, HRP, I result, and E only when their definitions are met.
  7. Write a plain-language impression and actionable recommendation beside the code.

Order: CAD-RADS [stenosis or N]/P[1–4]/[modifiers]. Avoid undocumented custom abbreviations.

Code it yourself

Worked CAD-RADS cases

Case A · Plaque without narrowing; SIS 1

CAD-RADS 1/P1. Visible plaque excludes category 0 even when the lumen is normal.

Case B · Maximum 40% stenosis; diffuse plaque in six segments

CAD-RADS 2/P3. Mild worst stenosis and severe total burden are deliberately allowed to diverge.

Case C · 60% LAD; calcium-obscured RCA

CAD-RADS 3/Px/N. A proven ≥50% lesion preserves numeric category; N flags the additional uncertainty.

Case D · Mild evaluable disease; proximal RCA non-diagnostic

CAD-RADS N/Px. N replaces category 1/2 because significant RCA disease cannot be excluded.

Case E · 80% LAD in-stent restenosis; P3

CAD-RADS 4A/P3/S. Grade the evaluable stent lumen like native stenosis and append S.

Case F · Patent LIMA–LAD bypasses occluded proximal LAD; mild RCA/LCx

CAD-RADS 2/Px/G if graft/anastomosis/runoff are patent. The successfully bypassed native LAD occlusion does not set the category.

Case G · 55% lesion; CT-FFR 0.73 distal; two HRP features

CAD-RADS 3/Px/HRP/I+. State lesion location, CT-FFR sampling position, and symptoms/context in prose.

Case H · No plaque; coronary fistula

CAD-RADS 0/E. Then describe donor, course, drainage, aneurysm, and shunt effects completely.

Common errors

What breaks the system

  • Calling plaque-free arteries CAD-RADS 1 instead of 0—or visible plaque CAD-RADS 0.
  • Using the P grade as if it were the stenosis grade.
  • Assigning HRP for one feature.
  • Adding I without CT-FFR or stress CT perfusion.
  • Using CAD-RADS 4B for any two severe lesions or any left-main plaque.
  • Writing CAD-RADS 2/N when N should replace the number.
  • Letting a bypassed native occlusion inflate an otherwise patent graft circulation.
  • Reporting only the code and omitting vessel-level disease, artifacts, and recommendation.

Active recall

Self-test: answer before opening

1. What are the three CAD-RADS 2.0 fields?

Worst stenosis category, whole-tree plaque burden P1–P4, then applicable modifiers.

2. What distinguishes 4A from 4B?

4A is 70–99% in one or two vessels; 4B is left main ≥50% or three-vessel obstructive disease with ≥70% stenosis.

3. What are P1–P4 CAC ranges?

P1 1–100; P2 101–300; P3 301–999; P4 >1000.

4. What are P1–P4 SIS ranges?

P1 ≤2; P2 3–4; P3 5–7; P4 ≥8 involved segments.

5. When does N replace the numeric category?

When at least one important segment is non-diagnostic and evaluable segments show no stenosis ≥50%, so obstructive disease cannot be excluded.

6. When is N appended to a number?

When a ≥50% stenosis is already proven in an evaluable segment but another important segment remains non-diagnostic.

7. What does the HRP modifier require?

At least two high-risk plaque features within a plaque.

8. What CT-FFR ranges define I+, I±, and I−?

I+ ≤0.75; I± 0.76–0.80; I− >0.80, using a lesion-specific value and full context.

9. How is a severe native lesion treated if a patent graft bypasses it?

It does not set the CAD-RADS category when the graft, anastomosis, and runoff are fully patent; grade the clinically functioning grafted circulation and append G.

10. What belongs under E?

Clinically important non-atherosclerotic coronary abnormalities such as anomaly, dissection, aneurysm, vasculitis, fistula, or extrinsic compression.

Printable quick revision

One-page CAD-RADS 2.0 card

Stenosis0 · 1 · 2 · 3 · 4A · 4B · 5 · N.
BurdenP1 mild · P2 moderate · P3 severe · P4 extensive.
ModifiersN · S · G · HRP · I · E.
  • 0: no plaque/no stenosis. 1: 1–24 or plaque/no narrowing. 2: 25–49. 3: 50–69.
  • 4A: 70–99% in 1–2 vessels. 4B: LM ≥50 or 3-vessel ≥70. 5: occlusion.
  • P by CAC: 1–100 · 101–300 · 301–999 · >1000.
  • P by SIS: ≤2 · 3–4 · 5–7 · ≥8.
  • N: no known ≥50 → N/Px; known ≥50 → numeric/Px/N.
  • HRP: ≥2 features. I: only after CT-FFR or stress CTP.
  • S/G: presence of stent/graft; still describe patency and restenosis.
  • E: anomaly/dissection/aneurysm/fistula/other non-atherosclerotic disease.
  • Final rule: code + plain-language vessel findings + clinical recommendation.

Provenance

Sources and scope

This chapter follows Dr Zeeshan Lakhani’s Webinar 4 CAD-RADS 2.0 teaching. Automated captions contain transcription errors; the coding rules and thresholds were checked against the multisociety consensus and 2026 SCCT interpretation update.

  1. CAD-RADS 2.0 multisociety expert consensus — definitive categories, P grades, modifiers, and management tables.
  2. CAD-RADS 2.0 open consensus version — N logic, plaque methods, I thresholds, and worked examples.
  3. SCCT Interpretation and Reporting of CCTA: 2026 Update — current integration of CAD-RADS 2.0 into structured CCTA reporting.

Educational material only. Management tables require clinical context and local pathways; urgent findings require direct communication. CAD-RADS is a summary of a complete CTCA interpretation, not a standalone diagnosis. External references require internet; this chapter works offline.

Back to top ↑