Real-time validation of transcranial Doppler criteria in assessing recanalization during intra-arterial procedures for acute ischemic stroke: an international, multicenter study.

BACKGROUND AND
PURPOSE: We sought to evaluate the diagnostic accuracy of ultrasound criteria for recanalization during real-time transcranial Doppler monitoring of intra-arterial reperfusion procedures in acute ischemic stroke patients in an international, multicenter study.
METHODS: Consecutive acute ischemic stroke patients with proximal intracranial occlusions underwent intra-arterial reperfusion procedures with simultaneous real-time transcranial Doppler monitoring at 3 tertiary-care stroke centers. Residual flow signals at the site of angiographically confirmed occlusions were monitored at a constant transtemporal insonation angle using a standard head-frame. Recanalization was assessed simultaneously by digital subtraction angiography and ultrasound using thrombolysis in myocardial infarction and thrombolysis in brain ischemia (TIBI) criteria, respectively. Independent readers blinded to digital subtraction angiography performed validation of TIBI flow grades. The interrater reliability for assessment of TIBI grades was investigated.
RESULTS: We evaluated time-linked real-time digital subtraction angiography transcranial Doppler images from 96 diagnostic digital subtraction angiography runs during intra-arterial reperfusion procedures in 62 acute ischemic stroke patients (mean age, 59 ± 17 years; 58% men; median baseline National Institutes of Health Stroke Scale score, 18 [interquartile range 12-21]; median time from symptom onset to intra-arterial procedure initiation, 240 minutes [interquartile range 163-308]). The interrater reliability for evaluation of TIBI grades and assessment of recanalization was good (Cohen κ: 0.838 and 0.874, respectively; P<0.001). Compared with angiography, transcranial Doppler had the following accuracy parameters for detection of complete recanalization (TIBI 4 and 5 versus thrombolysis in myocardial infarction 3, flow grades): sensitivity, 88% (95% confidence interval, 72%-96%); specificity, 89% (79%-95%); positive predictive value, 81% (65%-91%); negative predictive value, 93% (84%-98%); and overall accuracy 89% (80%-94%).
CONCLUSIONS: At laboratories with high-interrater reliability, TIBI criteria can accurately predict brain recanalization in real time as compared with thrombolysis in myocardial infarction angiographic scores.