FOCUSED TRANSTHORACIC ECHOCARDIOGRAPHY IN STROKE: A FEASIBILITY STUDY

Abstract

Introduction Transthoracic echocardiography (TTE) use in the diagnostic work-up of suspected cardio-embolic stroke has historically had a low diagnostic yield. The goal of this project was to evaluate the practicality of an abbreviated, thus more cost-effective, approach to TTE using an easily portable ultrasound probe in ischemic stroke. Methods In this cross-sectional study, we evaluated patients undergoing echocardiography for evidence of possible cardioembolic stroke, examined with both standard (Philips© EPIQ 7 Device) and focused (Philips© Lumify Device) imaging approaches. The focused protocol had a smaller number of imaging sequences and used a handheld ultrasound device. Results 58 paired standard and focused TTE’s were used for agreement calculation and 121 TTE’s were used for frequency data calculation. The mean time for image acquisition for focused and standard TTE was 7 minutes and 37 minutes (p<0.0001), respectively. Substantial agreement by kappa analysis was noted between the focused and standard TTE for left atrial enlargement (>34 ml/m²), severe left ventricular dysfunction (ejection fraction <30%), and presence of atrial septal aneurysm. Moderate agreement was noted for aortic valve calcification and presence of wall motion abnormality. A logistic regression model was constructed using Focused TTE reports to assess whether BMI could help predict the use of uncertain language in the report. There was a statistically significant association between BMI and use of the words “cannot exclude” for wall motion abnormality, atrial septal aneurysm, ventricular mass, atrial mass, ventricular mass, and atrial septal defect. Using a minimum sensitivity threshold of 70%, the only cardioembolic sources which could be readily identified at BMIs higher than 30 included wall motion abnormality, atrial septal aneurysm, and atrial mass. Conclusion For basic imaging findings, including left ventricular dimensions, left atrial size, and left ventricular function, there was moderate to substantial agreement between the focused TTE and standard TTE. In contrast, with the caveat that statistical requirements of non-inferiority were not met, there were multiple indications that the focused TTE using the Philips© Lumify device did not readily identify major cardioembolic sources compared to the EPIQ 7. In clinical practice, our results suggest handheld ultrasound is a poor screening modality in ischemic stroke patients, but that abnormalities identified are reproducible on more standard imaging (high specificity). More research is needed prior to making recommendations about handheld imaging in terms of guiding clinical decision-making in real time for ischemic stroke patients.