Stroke: Vascular and Interventional Neurology (Jan 2023)

Imaging Indicators for Parenchymal Hemorrhage After Mechanical Thrombectomy in Acute Stroke

  • Ryutaro Kimura,
  • Sotaro Shoda,
  • Tomonari Saito,
  • Kentaro Suzuki,
  • Akihito Kutsuna,
  • Takuya Kanamaru,
  • Takehiro Katano,
  • Toru Nakagami,
  • Shinichiro Numao,
  • Satoshi Suda,
  • Yasuhiro Nishiyama,
  • Kazumi Kimura

DOI
https://doi.org/10.1161/SVIN.122.000499
Journal volume & issue
Vol. 3, no. 1

Abstract

Read online

Background In patients with acute ischemic stroke with large vessel occlusion, the presence of hyperintense lesions on fluid‐attenuated inversion recovery (FLAIR‐positive) before mechanical thrombectomy (MT) and that of hyperdense lesions on noncontrast computed tomography (CT‐positive) after MT are associated with parenchymal hemorrhage, but the significance of these findings is unclear. Methods Patients with acute stroke with large vessel occlusion underwent magnetic resonance imaging, including diffusion‐weighted imaging and FLAIR, before MT. Noncontrast CT was obtained within 60 minutes after MT. The occurrence of parenchymal hemorrhage type 2 (PH2) was assessed on CT within 7 days after MT. FLAIR and CT‐positive were defined as a case with hyperintense and hyperdense lesions immediately before and after MT, respectively. Clinical and imaging factors associated with PH2 were evaluated by multivariate regression analysis. Results Enrolled were 412 patients (median age, 76 years; men, 58.3%; median National Institutes of Health Stroke Scale score, 16). The site of occlusion was the internal carotid artery (n=122, 29.6%), M1 (n=180, 43.7%), and M2 (n=96, 23.3%). FLAIR‐positive, CT‐positive, and PH2 were found in 149 (36.2%), 223 (54.1%), and 34 patients (8.3%), respectively. PH2 was significantly more frequent in CT‐positive than CT‐negative patients (14.3% versus 1.1%; P<0.001), but not in FLAIR‐positive than FLAIR‐negative patients (10.7% versus 6.8%; P=0.193). Regardless of FLAIR status, PH2 was significantly more frequent in CT‐positive than CT‐negative patients (12.8% versus 0.8% [P<0.001] for FLAIR‐negative, and 16.7% versus 1.7% [P=0.003] for FLAIR‐positive). Multivariate regression analysis demonstrated that CT‐positive was the only independent factor associated with PH2 (odds ratio, 12.699 [95% CI, 2.964–54.416]; P<0.001). Sensitivity, specificity, positive predictive value, and negative predictive values of CT‐positive for predicting PH2 occurrence were 94.1%, 49.5%, 14.3%, and 98.9%, respectively. Conclusion In patients with acute ischemic stroke with large vessel occlusion, CT‐positive immediately after MT was strongly predictive of PH2, but no such relationship was found for FLAIR‐positive before MT.

Keywords