Hypoperfusion intensity ratio for refinement of elderly patient selection for endovascular thrombectomy.

2021 
Background Patients ≥80-year-old presenting with large-vessel occlusion treated with endovascular thrombectomy (EVT) have worst outcomes than younger individuals. Improved patient selection in this age range is warranted. We investigated the hypoperfusion-intensity-ratio (HIR) and its associations with baseline parameters and clinical outcomes in a cohort ≥80-year-old to assess whether it could an option in improving their selection for EVT. Methods We performed retrospective analysis of consecutive patients treated with EVT at our center between 2015 and 2019. Inclusion criteria were age ≥80-year-old, any baseline modified Rankin Scale (mRS), and anterior circulation occlusion. Demographic information, baseline characteristics, clinical data, and radiological imaging parameters were collected. HIR was dichotomized into favorable and unfavorable based on median value of the cohort. Good outcome was defined as mRS ≤2 at 90-days. Results We included 82 patients. HIR was significantly correlated with baseline ischemic core volume, NIHSS, and time-of-onset to groin puncture. Good outcome was achieved in 18.3% and mortality occurred in 34.1%. In patients with baseline mRS ≤2, the rate of good outcome was significantly higher in favorable vs unfavorable HIR (52.6% vs 20%, P=0.02). In shift-analysis, unfavorable HIR was significantly associated with downshift to mRS ≥3 (P=0.02). Regression analysis found lower baseline mRS (P=0.009), higher ASPECTS (P=0.02), complete recanalization (P=0.04), and lower HIR (P=0.02) to be associated with increased rate of good outcome. Hierarchical regression showed HIR to independently predict good outcome. Conclusions In our cohort, HIR was correlated with baseline parameters and predicted clinical outcomes. Future studies should investigate perfusion parameters such as HIR to improve the selection of elderly patients for EVT.
    • Correction
    • Source
    • Cite
    • Save
    • Machine Reading By IdeaReader
    19
    References
    0
    Citations
    NaN
    KQI
    []