(2026) Kirschfink | Cardiac PCCT reliably rules out LAA thrombus in acute ischemic stroke

Daniel Stromer

Cardiac PCCT integrated directly into the acute stroke CT workflow achieved 96% diagnostic accuracy, 90% sensitivity, 97% specificity, and a particularly strong 99% negative predictive value for left atrial appendage (LAA) thrombus compared with TEE.

The cardiac acquisition was added immediately after stroke CTA with only an ~1.5-s scan duration and an 8-s delay after CTA, without measurable delay to acute revascularization treatment. Prospective ECG triggering was targeted at 35% of the R–R interval, although successful gating was achieved in only about 60% of patients.

Diagnostic performance was numerically better with ECG-gating (AUC 0.99) than without gating (AUC 0.89), supporting ECG synchronization as an important technical optimization. The study nevertheless positions cardiac PCD-CT primarily as a strong rule-out tool for LAA thrombus, complementing rather than replacing echocardiography.

Access publication

Figure 2. Cardiac PCCT imaging of the three false positive cases (a–c). Left atrial appendage thrombi were found in PCCT imaging (arrows) but not in transesophageal echocardiography.

Protocol

Detail	Value
Scanner	NAEOTOM Alpha
Application	Acute ischemic stroke / TIA with added cardiac CT for LAA thrombus assessment
Study design	Retrospective single-center diagnostic accuracy study
Reference standard	Transesophageal echocardiography (TEE)
Stroke CT workflow	Non-contrast brain CT → CTA vertex to main pulmonary artery → cardiac CT aortic arch to diaphragm → optional brain CT perfusion
Cardiac scan timing	Automatically triggered 8 s after end of CTA
Cardiac acquisition time	~1.5 s
Tube voltage	120 kV
Collimation	144 × 0.4 mm
Pitch	3.2
Rotation time	0.25 s
ECG-gating	Prospective ECG-triggering targeted at 35% R–R interval
ECG-gating success	~60%; not achieved in 39.6% because of ECG lead/signal problems
Bandwidth quality level	200
Slice thickness	0.9 mm
Reconstruction phase	40% R–R interval
CTA contrast agent	Nonionic low-osmolality contrast, 300 mg iodine/mL
Contrast volume	100 mL total
Additional cardiac contrast	None; cardiac acquisition used the same CTA contrast bolus
CTA triggering	Manual trigger after contrast arrival in the internal carotid arteries
Thrombus definition	Homogeneous, well-defined LAA filling defect with ≤100 HU and/or filling defect/ascending-aorta HU ratio ≤0.19
ROI measurements	LAA, left atrium, ascending aorta, descending aorta; 100 mm² ROIs
Median cardiac CTDIvol	8.9 mGy
Median cardiac DLP	186 mGy·cm
Median total examination DLP	3013 mGy·cm
Primary endpoint	Diagnostic accuracy for LAA thrombus compared with TEE

Key Results

Parameter	Main finding
Study population	101 patients with acute ischemic stroke or TIA undergoing both cardiac PCD-CT and TEE
LAA thrombus prevalence by PCD-CT	12/101 patients (11.9%)
Diagnostic accuracy	96.0%
Sensitivity	90.0%
Specificity	96.7%
Positive predictive value	75%
Negative predictive value	98.9%
Positive likelihood ratio	27.3
Interobserver agreement – PCD-CT	Excellent, κ = 0.91
Interobserver agreement – TEE	Substantial, κ = 0.63
ECG-gated performance	AUC 0.99
Non-ECG-gated performance	AUC 0.89
≤3-day CT-to-TEE subgroup	Accuracy 98.7%, specificity 100%, NPV 98.5%
Workflow impact	Adding cardiac PCD-CT caused no measurable delay in initiating acute revascularization therapy
Main implication	Cardiac PCD-CT is particularly strong for excluding LAA thrombus and may help select patients who most need subsequent TEE
Limitation	It should complement rather than replace echocardiography, because other potential cardiac stroke sources may be missed

Source: Kirschfink A, Elhalal M, Seidel S, Kammer S, Hasan D, Frick M, Marx N, Schulz JB, Wiesmann M, Altiok E, Ridwan H, Reich A, Pinho J. Cardiac photon-counting detector computed tomography in acute ischemic stroke: diagnostic accuracy for detection of left atrial appendage thrombus. Therapeutic Advances in Neurological Disorders. 2026;19:17562864261466073. doi:10.1177/17562864261466073.