E-007 Immediate post-thrombectomy dual-energy CT virtual non-contrast imaging for final infarct prediction. (23rd July 2022)

• Record Type: Journal Article
• Authors: Kim, C; Moghimi, S; Rohr, A; Settecase, F
• Is Part Of: Journal of neurointerventional surgery. Volume 14(2022)Supplement 1
• Page Start: A78
• Page End: A78
• Publication Date: 2022-07-23
• Subjects: Nervous system -- Surgery -- Periodicals; Cerebrovascular disease -- Surgery -- Periodicals; 617.48
• DOI: 10.1136/neurintsurg-2022-SNIS.118
• Languages: English
• ISSNs: 1759-8478

Abstract

Abstract : Purpose: Early detection of infarct volume immediately after mechanical thrombectomy may aid in prognostication and guiding post-treatment clinical management decisions. Iodine-staining of infarcted tissue may mask low-attenuation changes and virtual non-contrast (VNC) reconstruction of post-procedure dual-energy CT (DECT) may improve visualization of infarcted brain. In mTICI 3 reperfused patients, we rated infarct extent using ASPECTS on immediate post-MT mixed energy DECT and VNC, compared to follow-up single energy noncontrast CT ASPECTS . We hypothesize that VNC ASPECTS would more closely approximate follow-up CT ASPECTS than mixed energy DECT that has not been post-processed. Materials & Methods: Institutional review board approval was obtained. Patients who underwent MT for treatment of anterior circulation ischemic stroke between January and September 2019 were consecutively reviewed. 28 patients (aged 68 ± 17 years, 16 female) with a mTICI score of 3 and without hemorrhagic transformation were included. Dual-energy CT imaging was obtained immediately post-MT and post-processed with vendor software (syngo.via, Siemens, Erlangen, Germany) to produce VNC images. ASPECTS was rated by two independent blinded assessors for the immediate post-MT DECT and VNC as well as for the follow-up conventional single energy CT at 24–48 hours. A Wilcoxon rank sum test was applied to compare the ASPECTS across DECT, VNC and follow-up CT studies. Cohen's kappa was measured to assess inter-rater agreement. Results: ASPECTS for the DECT (rater A; median = 9 (IQR 8–9), rater B; median = 8.5 (IQR 6.75–9)) were significantly different from the follow-up at 24–48 hours for both raters (p<0.001 for both). There was no statistical difference, however, between the VNC (rater A; median = 6.5 (IQR 6–8.25), rater B; median = 6 (IQR 4–6)) and the follow-up CT (rater A; median = 7.5 (IQR 6–8), rater B; median = 6 (IQR 5–7)) for both raters (rater A; p=0.985, rater B; p=0.058). Inter-rater agreement for VNC was low (κ = -0.029) and follow-up CT (κ =-0.031) and DECT ((κ =0.235) . The caudate and lentiform nuclei were areas most commonly affected on the follow-up CT and demonstrated the greatest discrepancy between the VNC and the DECT. For rater A, VNC was 88.9% and 83.3% sensitive for caudate and lentiform infarcts, respectively, versus 22.2% and 27.8% for DECT. Similarly, for rater B, VNC was 88.2% and 90.9% sensitive for caudate and lentiform infarcts, respectively, versus 20% and 22.7% for DECT. Conclusions: Immediate post-thrombectomy DECT with VNC post-processing may more accurately visualize infarcted tissue and aid in prognostication and guiding treatment decisions despite suffering from low inter-rater agreement. Further larger scale studies are warranted. Disclosures: C. Kim: None. S. Moghimi: None. A. Rohr: None. F. Settecase: 1; C; Stryker Neurovascular, Microvention. 2; C; Stryker Neurovascular. 6; C; Medtronic.