ACR CAD-RADS Calculator

CAD-RADS Categories Quick Reference

CAD-RADS 2.0 assigns one of seven assessment categories — including two subcategories of category 4 — based on the most severe stenosis identified:

Worked Clinical Example

A 58-year-old man with hypertension, hyperlipidemia, and a 25 pack-year smoking history presents to the emergency department with 45 minutes of substernal chest pressure that began at rest and partially resolved with sublingual nitroglycerin. Initial troponin-I is 0.08 ng/mL (mildly elevated, 1–3× the upper reference limit). The ECG shows non-specific ST flattening in V4–V6 without frank ST elevation or T-wave inversion. The emergency physician computes a HEART Score of 5 — history moderately suspicious, ECG non-diagnostic for ischemia, age 45–64, three risk factors, troponin 1–3× normal — placing the patient in the intermediate-risk band where CCTA is an evidence-supported disposition pathway.

CCTA is acquired on a 320-row detector scanner with prospective ECG gating, heart rate brought to 58 bpm with oral metoprolol, and sublingual nitroglycerin administered prior to contrast injection. The Agatston calcium score is 142 (mild burden), effective dose 2.1 mSv, and image quality is diagnostic in all 18 segments (no CAD-RADS M modifier needed). On the contrast-enhanced series, the proximal left anterior descending (LAD) artery demonstrates a focal 60% diameter stenosis with positive remodeling (remodeling index 1.18), a hypodense plaque core measuring 22 HU (low-attenuation component), and a thin hyperattenuating rim — the classic napkin-ring sign of high-risk plaque. The left circumflex has 30% non-obstructive plaque in the mid-vessel, and the right coronary artery has minimal disease without HRP features. The radiologist formally applies CAD-RADS with the HRP modifier to the LAD lesion.

Walking the inputs into this calculator: Modality → CCTA; Highest-grade stenosis → 3 (50–69%, moderate); HRP → checked (positive remodeling, low-attenuation core, napkin-ring sign all present); Non-obstructive plaque → Yes (LCx and RCA have plaque but no higher-grade stenosis). The result panel returns CAD-RADS 3 + HRP in rose, with the HRP-driven escalation shifting the recommendation from functional testing (rec-moderate) toward earlier invasive angiography consideration (rec-high). The driving class line reads CAD-RADS 3 · HRP+ (3), reflecting the underlying anatomic category alongside the HRP modifier overlay.

Cardiology is consulted the same admission. Invasive coronary angiography confirms a 65% LAD stenosis with a measurement of FFR 0.74 (positive, ≤0.80). The patient undergoes single-vessel PCI with a drug-eluting stent and is discharged on dual antiplatelet therapy for 12 months, high-intensity statin, and aggressive risk-factor modification (smoking cessation program, BP control). A repeat CCTA is planned at 12 months for in-stent patency and assessment of any new lesions in the LCx plaque that has not yet crossed the 50% threshold. Teaching point: stenosis percent alone drives only part of the CAD-RADS story; HRP features identify the plaque that is most likely to become symptomatic despite a moderate luminal narrowing, and CAD-RADS 2.0 codifies this as a one-tier management upgrade.

Methodology and Validation

CAD-RADS 1.0 was published in 2016 by Cury, Abbara, Achenbach, and colleagues through a multi-society expert consensus process led by the American College of Radiology with representation from the American College of Cardiology, the Society of Cardiovascular Computed Tomography, and industry collaborators including Siemens. It was not derived from a single derivation dataset; rather, it synthesized the prevailing CCTA literature into a six-category reporting schema (0 through 5, with the M modifier for non-diagnostic studies) intended to standardize communication between radiology report authors and the referring cardiology or emergency-medicine teams. The framework was deliberately modeled on the existing ACR BI-RADS, Lung-RADS, and LI-RADS templates to ease adoption within radiology departments already accustomed to category-based reporting.

CAD-RADS 2.0 (Cury, Leipsic, Abbara, et al., JACC: Cardiovascular Imaging, 2022) extended the original framework along three clinically meaningful axes. First, it formalized the split of category 4 into 4A (severe stenosis 70–99% in one or two vessels) and 4B (severe stenosis in the left main or three-vessel disease), bringing the reporting granularity closer to heart-team decision-making for CABG versus PCI. Second, it codified the four modifiers (M, HRP, S, G) as explicit overlays — HRP in particular now carries a one-tier recommendation upgrade when present, even on a low category. Third, CAD-RADS 2.0 integrated CT-derived fractional flow reserve (CT-FFR) and stress CT myocardial perfusion into the anatomic grading, allowing a CAD-RADS 3 lesion to be functionally re-staged using on-site or cloud-based FFR computation. The category-to-surveillance-interval mapping shown in the FAQ above is also drawn directly from the 2.0 document.

Validation of CAD-RADS as a clinical decision tool comes from the prospective randomized trials that established CCTA itself as a viable chest-pain pathway. The PROMISE trial (Douglas et al., NEJM 2015) randomized 10,003 symptomatic outpatients to CCTA versus functional testing and demonstrated non-inferiority for clinical outcomes at 2 years, providing the evidence base for using CCTA — and therefore CAD-RADS — as a first-line anatomic workup. ROMICAT-II (Hoffmann et al., NEJM 2012) established that CCTA-based ED rule-out shortens length of stay compared with standard evaluation without increasing missed ACS. For CAD-RADS 2.0 specifically, the CONFIDENCE registry and subsequent CT-FFR outcome studies have provided category-specific prognostic data, showing that CAD-RADS 0 carries a near-zero 1-year MACE rate while CAD-RADS 4B/5 carry 5–10% one-year event rates even on optimal medical therapy. Sources for these trials are listed under Clinical sources & references below.

Per-category diagnostic accuracy: against invasive coronary angiography as the reference standard, modern CCTA shows per-patient sensitivity of 95–99% and specificity of 64–90% for ≥50% stenosis, with the highest accuracy in CAD-RADS 0 (ruled-out disease) and the lowest in CAD-RADS 3 where functional correlation is most often needed. CAD-RADS 1 and 2 are highly concordant with ICA — non-obstructive plaque is reliably identified — but CCTA tends to overestimate stenosis in heavily calcified vessels (CAD-RADS 4A on CCTA may be only CAD-RADS 3 on ICA) and underestimate in low-attenuation plaque where blooming artifact is minimal but the reader's eye trends toward normal-appearing lumens. These systematic biases are why the CAD-RADS 2.0 framework pairs the anatomic grade with HRP and CT-FFR modifiers that recover some of the lost signal at the moderate-stenosis band, and why the calculator surfaces HRP-driven escalation as a first-class output rather than a footnote.

How to Use This Calculator

The CAD-RADS calculator follows the ACR 2.0 stepwise assignment. First, choose the imaging modality (CCTA is the default; ICA is also valid since CAD-RADS is increasingly applied at invasive angiography). Next, enter the maximum stenosis grade identified — across all vessels. The calculator maps the highest stenosis to a category automatically:

1. Modality — Toggle CCTA or ICA. CCTA is the primary modality for which CAD-RADS was developed; ICA is increasingly reported using the same framework.
2. Maximum stenosis — Pick the highest single stenosis grade across all assessed vessels. If multiple lesions span multiple grades, only the highest drives the category.
3. Modifiers — Check any modifier overlay that applies: M for non-diagnostic, HRP for high-risk plaque features, S for stent, G for graft. Modifiers alter management without changing the stenosis-based category.
4. Non-obstructive plaque — If stenosis is 0% but plaque is present, toggle plaque Yes to elevate CAD-RADS 0 → 1 per Cury et al. (2016).
5. Management — The result panel returns the CAD-RADS category, recClass color band, and a recommendation matching ACC/AHA chest-pain and chronic CAD guidelines.

LM = left main coronary artery. 3-vessel = three major epicardial arteries (LAD, LCx, RCA) each with ≥70% stenosis. The 4B carve-out should be applied at the vessel-pattern level — not per-lesion.

Modality-Specific Guidance

CAD-RADS recognizes two modality contexts with different technical considerations:

• Coronary CT angiography (CCTA) — the primary modality for CAD-RADS. Gated cardiac CT performed with IV iodinated contrast; breath-hold 6–10 seconds; ideal heart rate 50–65 bpm (β-blocker prep if needed). Effective dose typically 1–5 mSv with modern scanners. Calcium blooming is the dominant limitation — at very high Agatston scores (>400), lumen assessment becomes unreliable and CAD-RADS M is often the appropriate assignment. Stress CT perfusion (CT-FFR) is increasingly reported alongside the anatomic grade.
• Invasive coronary angiography (ICA) — the historic gold-standard for stenosis grading. CAD-RADS is increasingly applied at ICA for consistency with CCTA reports, especially when both modalities coexist in a single workup. For intermediate lesions (CAD-RADS 3 equivalent), ICA enables immediate fractional flow reserve (FFR) or instantaneous wave-free ratio (iFR) measurement rather than deferring to a non-invasive stress test.

Modifiers in CAD-RADS 2.0

CAD-RADS 2.0 (2022) introduced four modifiers that overlay on the numeric category and adjust management without changing the stenosis grade:

• M (Non-diagnostic) — Used whenever the study is uninterpretable for any segment due to motion artifact, calcification blooming, beam hardening, or arrhythmia. The result label reports CAD-RADS M (not a numeric value) and management becomes repeat CCTA with HR control, dual-source scanning, or a calcium-score-guided modified protocol.
• HRP (High-risk plaque) — High-risk plaque features include positive remodeling (remodeling index >1.1), low-attenuation plaque (<30 HU), napkin-ring sign (low-attenuation core with hyperattenuating rim), and spotty calcification. HRP shifts management one tier higher — for example, CAD-RADS 2 with HRP warranting functional testing rather than risk-factor modification alone.
• S (Stent) — Indicates a stented segment with assessment of in-stent restenosis. Reported alongside the stenosis grade for the native or non-stented segments. CCTA in-stent evaluation is reliable for stent diameters ≥3 mm.
• G (Graft) — Indicates one or more coronary artery bypass grafts (post-CABG). Graft patency assessment is appended; LIMA-LAD grafts have the highest long-term patency. Stenosis grading in the graft plus native vessels combined drives the final category.

Patient Management Recommendations

CAD-RADS management pairs stenosis severity with ACC/AHA guideline-directed therapy:

• CAD-RADS 0 — No plaque or stenosis identified. If symptomatic, consider non-coronary etiology. Continue guideline-directed primary prevention (statin if ASCVD ≥7.5% 10-year risk, BP control, smoking cessation, glycemic control).
• CAD-RADS 1 — Minimal stenosis (1–24%) or non-obstructive plaque. Reassurance. Optimize risk factors; statin per ASCVD. No further cardiac testing for chest-pain rule-out unless high clinical suspicion.
• CAD-RADS 2 — Mild stenosis (25–49%). Non-obstructive CAD. Aggressive risk-factor modification (high-intensity statin, BP control, smoking cessation, weight management). Re-evaluation if symptoms change.
• CAD-RADS 3 — Moderate stenosis (50–69%). Consider functional testing (stress CT perfusion, stress MRI, PET, or invasive FFR) to assess hemodynamic significance. Cardiology referral appropriate.
• CAD-RADS 4A — Severe stenosis (70–99%) in 1–2 vessels. High-intensity statin + dual antiplatelet therapy per guideline. Consider invasive coronary angiography (ICA) for definitive evaluation and revascularization candidacy.
• CAD-RADS 4B — Severe stenosis (70–99%) in left main or 3-vessel disease. Invasive coronary angiography recommended; heart-team discussion for CABG vs PCI. Guideline-directed medical therapy in parallel.
• CAD-RADS 5 — Total coronary occlusion (100%). Invasive coronary angiography indicated; revascularization (PCI or CABG) per lesion complexity, viability, and SYNTAX score. Urgent cardiology referral.

Limitations and Considerations

• Prior imaging comparison. Stability over 2+ years can downgrade a moderate stenosis; a new lesion is more suspicious. This calculator does not incorporate prior CCTA data.
• Calcium blooming. Heavily calcified lesions (Agatston ≥400) often overestimate stenosis severity and trigger CAD-RADS M (non-diagnostic) rather than numeric assignment.
• Plaque vs. stenosis. Atherosclerotic plaque burden (volume, HRP features) predicts MACE independent of stenosis percent; do not equate low CAD-RADS with low event risk in patients with diffuse non-obstructive plaque.
• Non-coronary findings. CCTA also detects non-coronary findings (pericardial, valvular, pulmonary, aortic) that this calculator does not address.
• Not a substitute. This tool is for educational reference; final assignment requires full imaging review, vessel-by-vessel analysis, plaque morphology, and clinical judgment.
• Agatston-score threshold. When the total calcium score exceeds ~400, calcification blooming broadly distorts luminal assessment; in this range the appropriate CAD-RADS output is often M with a repeat protocol (high-pitch scan, β-blocker preparation above the usual 65 bpm target, or shift to a non-CTA modality such as stress MRI) rather than a numeric stenosis estimate.
• HRP derivation and interobserver variability. High-risk plaque feature identification (positive remodeling, low-attenuation plaque <30 HU, napkin-ring sign, spotty calcification) carries meaningful interreader variability — published kappa values range from 0.45 to 0.70 depending on plaque type and reader experience, with napkin-ring sign being the most subjective of the four features.
• Serial CCTA interpretation. Interval change versus new lesion carries different prognostic weight. Progression from CAD-RADS 1 to CAD-RADS 3 over 18 months is a distinct clinical event from a new category-3 lesion in a previously normal study, but the calculator outputs are identical. Read the index report alongside this tool rather than relying on the category alone.
• Pediatric, young adult, and post-CABG subpopulations. Patients under 40 rarely meet indication for CCTA, and CAD-RADS categories are not validated in this group. Post-CABG patients require the G modifier — but graft atherosclerosis progresses faster than native CAD, and the modifier alone does not capture graft-specific lesion morphology warranting separate reading.
• Women-specific considerations. Smaller coronary caliber, higher resting heart-rate variability, and differences in plaque distribution (more non-calcified plaque, often proximal LAD involvement) alter CAD-RADS interpretation. Symptom–lesion discordance is more common, so an anatomic CAD-RADS 1 or 2 in a symptomatic woman does not reliably exclude microvascular or functional disease.
• Stent ≤2.5 mm diameter caveat. The S modifier assumes CCTA in-stent evaluation. Blooming effects become dominant in stents with inner diameters at or below ~2.5 mm, and in-stent restenosis assessment in this size range frequently degrades to non-diagnostic despite good image quality in the rest of the study.
• Anchoring bias risk. If an outside radiology report exists, the calculator's recommendation should not be used as confirmation of that prior read; the calculator is a category-mapping tool, not a second interpretation. Re-review at the treating institution when the prior report is the only documented source.