Comparative Accuracy of TCD, TTE, TEE, and Cardiac CT in Detecting Right-to-Left Shunt in Embolic Stroke of...

Identifying a right-to-left shunt (RLS), commonly due to a patent foramen ovale (PFO), is essential in evaluating embolic stroke of undetermined source (ESUS). While several imaging modalities are used for PFO detection, direct comparisons in a single cohort are limited. This study evaluates and compares the diagnostic performance of 4 major modalities: transcranial Doppler (TCD), transthoracic echocardiography (TTE), transesophageal echocardiography (TEE), and cardiac CT in patients with ESUS.

We retrospectively identified patients with ESUS referred to our neurosonology laboratory from July 2023 to May 2025, who underwent at least 3 of the 4 diagnostic modalities. A confirmed PFO was defined by concordant positive findings on at least 2 tests. Diagnostic performance metrics (sensitivity, specificity, positive predictive value, negative predictive value [NPV], accuracy, and area under the curve [AUC]) were calculated, with pairwise comparisons assessed using the DeLong test.

Of 469 screened patients, 130 met inclusion criteria. TCD showed the highest sensitivity (100%), NPV (100%), and an accuracy of 87.6%, although specificity was moderate at 75.0%. After excluding patients with mild PFO or extracardiac shunts, specificity and accuracy improved to 81% and 90.8%, respectively. TTE achieved a sensitivity of 83.3% and specificity of 93.3%, with an overall accuracy of 88.1%. TEE had a sensitivity of 95.1%, specificity of 87.5%, and the highest overall accuracy at 91.3%. Cardiac CT had perfect specificity (100%) but the lowest sensitivity (45.8%), limiting its standalone diagnostic utility. AUC comparisons revealed superior diagnostic performance for TCD, TTE, and TEE over cardiac CT (p values <0.05), but no significant differences among TCD, TTE, and TEE.

TCD is a highly sensitive, noninvasive first-line tool for detecting RLS in ESUS, particularly in acute or resource-limited settings. Confirmatory TTE, TEE, or CT can be tailored to context, with a tiered approach optimizing PFO detection and prevention strategies. Results should be interpreted cautiously, given the referral and selection bias. Larger prospective multicenter studies are needed to validate this approach and improve outcomes.

This study provides Class IV evidence that (1) TCD and TEE were highly sensitive in detecting PFO; (2) TTE and cardiac CT were highly specific in detecting PFO; (3) TCD, TTE, and TEE had superior diagnostic performance over cardiac CT (as defined by AUC comparison analysis); and (4) there were no significant differences in overall diagnostic performance among TCD, TTE, and TEE.