Cardiac CT in Stroke: A Non-Invasive Alternative to TEE

Transesophageal echocardiography (TEE) has long been the gold standard for detecting cardioembolic sources in stroke. However, it is invasive, requires sedation, and is often deferred or unavailable in the acute setting. An emerging body of evidence supports cardiac CT (cCT) as a practical, high-yield alternative that can be performed during the initial stroke workup—identifying thrombi and structural abnormalities that alter management in a significant proportion of patients.

Why Cardiac CT?

Traditional transthoracic echocardiography (TTE) has limited sensitivity for detecting left atrial appendage (LAA) thrombi and other high-risk cardioembolic sources. The Cardiac CT vs TTE study (2022) demonstrated that ECG-gated cardiac CT detected high-risk embolic sources in 11.4% vs 4.9% with TTE (OR 5.60), with thrombus detection rates of 7.1% vs 0.6%. In cryptogenic strokes alone, cardiac CT identified a cause in 6.3% of cases—often changing management to anticoagulation(1).

🔹 Bottom Line: Cardiac CT in Stroke Workup

  • Extend your CTA: Adding cardiac coverage to routine stroke CTA adds minimal time (<5 sec) and no extra contrast, while detecting thrombus in 8–17% of patients.
  • Superior to TTE: Cardiac CT detects high-risk sources in 11% vs 5% with TTE, and thrombi in 7% vs <1%.
  • Spectral CT: Dual-energy reconstructions improve thrombus detection by 20% over conventional CT.
  • Changes management: 21% of ESUS patients had anticoagulation newly indicated based on cardiac CT findings.
  • Still need TEE for: Infective endocarditis, PFO with bubble study, complex congenital disease, prosthetic valve assessment.

ENCLOSE: Nongated CTA as a Practical Solution

The ENCLOSE trial (2023) evaluated admission nongated cardiac CTA—simply extending the standard stroke CTA scan to include the heart—and found remarkable results. Cardiac thrombi were detected in 12% of patients (9% LAA, 4% LV). Notably, LAA thrombus was found in 6% of patients without atrial fibrillation and 7% not on anticoagulation(2).

An expert panel changed stroke etiology to cardioembolic in 68% of cases after reviewing cardiac CTA, compared to 0% before imaging (p<0.001). The additional scan time was minimal (<5 seconds), with no extra contrast required.

DAYLIGHT: Extended CTA During Stroke Code

DAYLIGHT randomized stroke code patients to extended CTA (≥6 cm below carina to include the LAA) versus standard head-and-neck CTA. Cardioaortic thrombus detection increased from 1.7% to 8.8% (OR 5.70, p=0.002), yielding a number needed to screen of just 14. There was no delay in time to CTA completion, and 50% of patients with detected thrombus had their anticoagulation management changed(3).

CaCTUSES: ESUS Evaluation

The CaCTUSES study (2022) specifically evaluated cardiac CT in embolic stroke of undetermined source (ESUS). Among 129 ESUS patients undergoing CCTA, intra-cardiac thrombus was detected in 17% (10% LV, 6% LAA, 1% LA). Anticoagulation was newly indicated in 21% based on CCTA findings alone. These thrombi were not visible on TTE but clearly identified on CT(4).

Spectral CT: Enhanced Detection

Spectral (dual-energy) cardiac CT further improves diagnostic yield. The Spectral Cardiac CT study (2023) found that 20% of thrombi missed on conventional CT were detected using spectral reconstructions. Monoenergetic images at 55 keV (monoE55) achieved 95% detection rates, with iodine density maps reaching 85%. This technique also improved visualization of minor embolic sources such as PFO and mitral annular calcification(5).

Beyond Thrombi: Structural Findings

The Cardiac CT in LVO Stroke study (2025) examined patients without AF or visible thrombus and found that structural abnormalities on cCT were strongly predictive of cardioembolic etiology(6):

  • Left ventricular dilatation: Adjusted OR 32.4
  • Interatrial shunt: Adjusted OR 30.1
  • Valve implants: Adjusted OR 24.5
  • Aortic arch atheroma ≥grade II: Adjusted OR 6.9
  • Old myocardial infarction: Adjusted OR 6.3

A combined imaging model achieved an AUC of 0.83 for predicting cardioembolism, demonstrating that cardiac CT provides valuable information even when no thrombus is present.

When Cardiac CT Is Not Enough: Indications for TEE

Despite its advantages, cardiac CT cannot replace TEE in all scenarios. The following indications still require transesophageal echocardiography:

Infective Endocarditis

TEE remains essential for evaluating suspected infective endocarditis. The high spatial resolution and real-time imaging allow detection of small vegetations, abscess formation, and valve perforation. CT lacks the temporal resolution to assess vegetation mobility and may miss small lesions (<3mm). The Modified Duke Criteria incorporate TEE findings as a major criterion.

Patent Foramen Ovale (PFO) and Right-to-Left Shunts

While spectral CT can visualize anatomic PFO, functional assessment requires TEE with agitated saline (bubble study). The Valsalva maneuver during bubble contrast injection demonstrates right-to-left shunting—critical for determining PFO significance and closure candidacy. CT cannot replicate this dynamic assessment.

Atrial Septal Defects and Complex Congenital Heart Disease

Detailed characterization of ASD size, location (secundum vs. primum vs. sinus venosus), rim adequacy for closure, and associated anomalous pulmonary venous return requires TEE. This is particularly important when planning percutaneous or surgical closure.

Prosthetic Valve Dysfunction

While CT can detect valve thrombus and pannus, TEE provides superior assessment of prosthetic valve regurgitation, dehiscence, and paravalvular leak. Real-time color Doppler imaging is irreplaceable for flow dynamics.

Left Atrial Appendage Closure Device Assessment

Post-LAAC device surveillance for residual leak requires TEE with color Doppler. CT can assess device position but cannot evaluate flow dynamics.

Indication Cardiac CT TEE Required Rationale
LAA/LV thrombus ✓ Preferred Only if CT inconclusive Higher sensitivity, performed acutely
Aortic arch atheroma ✓ Preferred Not needed Better visualization of entire aorta
Cardiac structure (LV, valves) ✓ Adequate If functional assessment needed CT identifies structural abnormalities
Infective endocarditis ✗ Insufficient ✓ Required Need real-time, high-resolution imaging
PFO with shunt assessment ✗ Insufficient ✓ Required Bubble contrast with Valsalva needed
ASD sizing for closure ~ Limited ✓ Required Rim assessment, flow quantification
Prosthetic valve regurgitation ✗ Insufficient ✓ Required Doppler assessment of flow essential
Trial Year Population Technique Thrombus Detection Key Finding
Cardiac CT vs TTE 2022 Acute ischemic stroke ECG-gated cCT 7.1% vs 0.6% (TTE) OR 5.60 for high-risk sources
CaCTUSES 2022 ESUS CCTA 17% 21% newly anticoagulated
ENCLOSE 2023 AIS/TIA Nongated CTA 12% 68% reclassified as cardioembolic
Spectral Cardiac CT 2023 LVO stroke Spectral/dual-energy 95% (monoE55) 20% thrombi missed by conventional CT
DAYLIGHT 2025 Stroke code Extended nongated CTA 8.8% vs 1.7% NNS = 14; no delay
Cardiac CT in LVO 2025 LVO without AF/thrombus ECG-gated cCT N/A (structural) AUC 0.83 for cardioembolism

Practical Implementation

Integrating cardiac CT into acute stroke protocols requires minimal adjustment. The DAYLIGHT and ENCLOSE studies demonstrate that simply extending the CTA field to include the heart adds <5 seconds of scan time with no additional contrast. For centers with dual-energy CT capability, spectral reconstructions can be generated from existing acquisitions to improve thrombus detection.

ECG-gated acquisitions provide superior image quality but require more time and specialized protocols. Nongated CTA is sufficient for thrombus detection in most cases and is more practical for acute stroke workflows.

References

  1. Cardiac CT vs TTE. Circulation. 2022.
  2. ENCLOSE Investigators. Stroke. 2023.
  3. DAYLIGHT Investigators. Neurology. 2025.
  4. CaCTUSES Investigators. JAMA Neurol. 2022.
  5. Spectral Cardiac CT in Acute Stroke. Radiology. 2023.
  6. Cardiac CT in LVO Stroke. Stroke. 2025.