Abstract WP164: Validation of Noninvasive Methods of Intracranial Stenosis Measurement Against Conventional Angiography

Abstract

Introduction: Stenosis caused by intracranial atherosclerotic plaques has long driven treatment algorithms. Conventional digital subtraction angiography (DSA) is considered the gold standard for measuring such stenosis, but fewer patients are evaluated with this modality as utilization of endovascular treatment has decreased. Vessel wall MRI (VWMRI) offers improved characterization by allowing inspection of the vessel wall and disease itself rather than narrowing of the lumen. This study compares measurements of stenosis in VWMRI against those made on conventional angiography or older noninvasive imaging techniques. Methods: VWMRI studies performed for patients with plaques that had caused a stroke were retrospectively analyzed. Stenoses were measured according to the WASID technique on DSA, post-contrast T1 DANTE, non-contrast time of flight (TOF), post-contrast TOF, and CTA images. Intraclass correlation coefficient (ICC), weighted kappa, and Cohen’s kappa were calculated as appropriate to assess agreement among stenosis measurements on these techniques. Results: Studies from 54 patients met inclusion criteria. Mean stenosis as measured by post-contrast T1 DANTE was 53.8% (±23.2%). 13 (24.1%) had stenosis ≥70%. ICC analysis of agreement between stenosis measurements on DSA demonstrated highest agreement with post-contrast T1 DANTE (ρ=0.927) and CTA (ρ=0.935), with excellent agreement also seen with post-contrast (ρ=0.915) and non-contrast TOF (ρ=0.8110). When separating measurements into ≥70% or <70%, agreement was highest for CTA (κ=1.0) and post-contrast T1 DANTE (κ=0.857), and post-contrast TOF (κ=0.857), while non-contrast TOF agreement was much lower (κ=0.400). Classifying stenosis ordinally by quartiles, greatest agreement with DSA as identified by weighted kappa analysis was seen with post-contrast T1 DANTE (wκ=0.331), followed by post-contrast TOF (wκ=0.268), CTA (wκ=0.253), and non-contrast TOF (wκ=0.222). Conclusion: Contrast-enhanced noninvasive angiographic techniques had greatest agreement with DSA when measuring plaque stenosis, with best agreement seen with post-contrast T1 DANTE and CTA. Non-contrast TOF agreement was inferior. The majority (75.9%) of symptomatic lesions had stenosis <70%.