This study enrolled a total of 125 AIS patients. After excluding some patients with incomplete imaging data (n = 9), posterior circulation occlusion (n = 12), missing follow-ups (n = 8), no substantial tissue reperfusion as seen with mTICI score < 2b (n = 15), not the large vessel vascular occlusion(n = 2), seventy-nine patients (Fig. 1) were finally included in our analysis (Table 1). Twenty-two patients (27.85%) received intravenous thrombolysis before mechanical thrombectomy. No significant differences were observed between favorable and unfavorable outcome group in age, gender, TIA, IV thrombolysis, risk factors, baseline INR, reperfusion score, TOAST, occlusive location and hemorrhagic transformation (Table 1). There was a statistically significant difference in mortality between the two groups(p = 0.003).
Table 1
Comparisons of baseline and clinical characteristics between favorable and unfavorable outcome group.
Factor | All patients N = 79 | Favorable mRS ≤ 2 N = 37 | Unfavorable mRS༞2 N = 42 | Χ2/t/U | P value |
Age, mean ± SD | 62.96 ± 11.33 | 61.56 ± 10.16 | 64.19 ± 12.32 | 1.027 | 0.301 |
Females/Males, n | 34/45 | 12/25 | 22/20 | 3.193 | 0.074 |
Hypertension, n (%) | 55(69.62) | 28(75.67) | 27(64.29) | 1.207 | 0.272 |
Diabetes, n (%) | 19(24.05) | 8(21.62) | 11(26.19) | 0.225 | 0.635 |
Ischemic heart disease, n (%) | 10(12.66) | 4(10.81) | 6(14.29) | 0.215 | 0.643 |
Atrial fibrillation, n (%) | 27(34.17) | 10(27.03) | 17(40.48) | 1.582 | 0.209 |
TIA, n (%) | 5(6.33) | 3(8.11) | 2(4.76) | 0.412 | 0.521 |
Alcohol intake, n (%) | 21(26.58) | 8(21.62) | 13(30.95) | 0.878 | 0.349 |
History of smoking, n (%) | 27(34.18) | 13(35.14) | 14(33.33) | 0.028 | 0.866 |
INR, mean ± SD | 1.07 ± 0.19 | 1.06 ± 0.18 | 1.08 ± 0.20 | 0.012 | 0.913 |
IV thrombolysis, n (%) | 22(27.85) | 11(29.73) | 11(26.19) | 0.123 | 0.726 |
Admission NIHSS, Median (IQR) | 11(8–14) | 11(7–14) | 11(10–14) | -1.623 | 0.105 |
OTI, Median (IQR) | 7.7(4.75-11.00) | 7.43(4.32–11.23) | 7.88(5.95–10.83) | -0.757 | 0.449 |
OTR, Median (IQR) | 12.53(8.37–17.5) | 11.78(8.5-15.82) | 11.75(9-16.55) | -0.418 | 0.676 |
Reperfusion grade, n (%) | | 0.177 | 0.764 |
2b | 34(43.04) | 15(40.54) | 19(45.24) | | |
3 | 45(56.96) | 22(59.46) | 23(54.76) | | |
Location of occlusion, n (%) | | 0.944 | 0.624 |
MCA | 59(74.68) | 27(72.97) | 32(76.19) | | |
ICA | 12(15.19) | 7(18.92) | 5(11.90) | | |
T-type | 8(10.13) | 3(8.11) | 5(11.90) | | |
TOAST, n (%) | | 3.516 | 0.172 |
large-artery atherosclerosis | 53(67.09) | 28(75.68) | 25(59.52) | | |
cardioembolism | 23(29.11) | 7(18.92) | 16(38.10) | | |
other determined etiology | 2(2.53) | 1(2.70) | 1(2.38) | | |
undetermined etiology | 1(1.27) | 1(2.70) | 0(0) | | |
Hemorrhagic transformation, n (%) | 16(20.25) | 7(18.92) | 9(21.43) | 0.077 | 0.782 |
sICH,n(%) | 4(5.06) | 1(2.70) | 3(7.14) | 0.142 | 0.677 |
Death, n (%) | 9(11.39) | 0(0) | 9(21.43) | 8.948 | 0.003 |
IQR, Interquartile range; mRS, modified Rankin Scale; NIHSS, National Institutes of Health Stroke Scale; INR = international normalized ratio; TIA = transient ischemic attack; OTI = onset-to-image time; OTR = onset-to-reperfusion time; MCA, Middle cerebral artery; ICA, Internal carotid artery; sICH = symptomatic intracerebral hemorrhage. |
The imaging characteristics were compared between recanalized AIS patients with favorable outcome and with unfavorable outcome (Table 2). The lesion volumes of rCBF < 30%, Tmax > 10s and > 12s in the unfavorable group were significantly higher than those in the favorable group (Fig. 2). Specifically, the median volume of ischemic core was 4.79 mL (IQR 2.41–15.90) versus 13.98 mL (IQR 4.07–31.33) for favorable outcome group and unfavorable group (p = 0.011), respectively. The median volume of Tmax > 10s was 28.27 mL (IQR 9.72–69.8) versus 54.12 mL (IQR 30.20-73.11) for favorable and unfavorable group, respectively (p = 0.032). The median volume of Tmax > 12s was 9.60 mL (IQR 4.69–44.13) versus 33.66 mL (IQR 18.98–54.66) for favorable and unfavorable group, respectively (p = 0.008). In addition, a higher median HIR was seen in the unfavorable group compared to the favorable outcome group (p = 0.002, 0.30 (IQR 0.20–0.40) vs 0.40 (IQR 0.3–0.4), Fig. 2). Spearman correlation analysis (Fig. 3) found that volume of rCBF < 30% (ischemic core) was significantly and positively correlated with the volume of Tmax > 12s (r = 0.454).
Table 2
Comparisons of imaging characteristics between recanalized AIS patients with favorable outcome and with unfavorable outcome.
Imaging parameter | All Patients N = 79 | Favorable mRS ≤ 2 N = 37 | Unfavorable mRS༞2 N = 42 | U | z | P value |
Volume of rCBF < 30%, median(IQR) | 8.36(3.12–27.17) | 4.79(2.41–15.9) | 13.98(4.07–31.33) | 521.0 | -2.556 | 0.011 |
Mismatch Volume,median(IQR) | 121.345(60.02-150.35) | 93.14(42.49-126.59) | 127.18(72.74-164.31) | 673.0 | -1.022 | 0.307 |
MismatchRatio, median(IQR) | 10(5.6–24) | 14(8.3–22) | 8.7(5.1–25) | 380.5 | 0.604 | 0.546 |
Volume of Tmax > 4s, median(IQR) | 189.395(114.14-258.27) | 157.53(89.18-208.28) | 221.84(166.46-268.71) | 687.0 | -0.884 | 0.377 |
Volume of Tmax > 6s, median(IQR) | 130.625(73.94-178.01) | 105.98(47.47-167.19) | 151.87(120.15-185.17) | 639.0 | -1.356 | 0.175 |
Volume of Tmax > 8s, median(IQR) | 73.15(34.96-103.26) | 52.29(19.29-101.57) | 87.12(61.55-103.26) | 586.0 | -1.877 | 0.061 |
Volume of Tmax > 10s, median(IQR) | 47.91(20.7-73.11) | 28.27(9.72–69.8) | 54.12(30.2-73.11) | 558.5 | -2.147 | 0.032 |
Volume of Tmax > 12s, median(IQR) | 31.375(7.4-54.34) | 9.6(4.69–44.13) | 33.66(18.98–54.66) | 507.5 | -2.648 | 0.008 |
HIR, median(IQR) | 0.3(0.2–0.4) | 0.3(0.2–0.4) | 0.4(0.3–0.4) | 408 | -3.026 | 0.002 |
IQR: Interquartile range; rCBF, relative cerebral blood flow; Tmax, time-to-maximum (Tmax); HIR, the volumetric ratio of the Tmax > 10s/Tmax > 6s. |
The pretreatment predictors or factors could be associated with poor/good outcomes even in the successfully recanalized AIS patients, and univariate logistic regression (Table 3) showed an association between volume of Tmax > 12s (critically hypoperfused region) and favorable outcome, but only had a marginal significance (OR, 0.986; 95% CI, 0.971-1, p = 0.055). Similar associations occurred for the volume of rCBF < 30% (OR, 0.975; 95% CI, 0.948–1.003, p = 0.083), volume of Tmax > 8s (OR, 0.993; 95% CI, 0.984–1.001, p = 0.086), volume of Tmax > 10s (OR, 0.990; 95% CI, 0.980–1.001, p = 0.073), admission NIHSS (OR, 0.905; 95% CI, 0.815–1.044, p = 0.059) ,but remained insignificant. On the other hand, there was a statistically significant association of HIR (OR, 0.008; 95% CI, 0.001–0.254, p = 0.006) with a favorable outcome (Table 3).
Table 3
Association of factors with favorable outcome (mRS 0–2) in the univariable analysis.
Factor | Odds ratio | 95% CI | P value |
Age | 0.979 | 0.941–1.019 | 0.304 |
Female sex | 0.436 | 0.174–1.019 | 0.076 |
Hypertension | 1.728 | 0.648–4.610 | 0.274 |
Diabetes | 0.777 | 0.274–2.204 | 0.636 |
Ischemic heart disease | 0.727 | 0.188–2.807 | 0.664 |
Atrial fibrillation | 0.545 | 0.210–1.411 | 0.211 |
TIA | 1.818 | 0.287–11.536 | 0.526 |
Alcohol | 0.615 | 0.222–1.707 | 0.351 |
History of smoking | 1.083 | 0.427–2.749 | 0.866 |
INR | 0.674 | 0.064–6.804 | 0.726 |
IV thrombolysis | 1.192 | 0.445–3.193 | 0.726 |
Location of occlusion | | | |
MCA | | | 0.628 |
ICA | 1.406 | 0.307–6.431 | 0.660 |
T-type | 2.333 | 0.463–14.613 | 0.365 |
TOAST | | | |
large-artery atherosclerosis | | | 0.182 |
cardioembolism | 1.160 | 0.069–19.519 | 0.918 |
other determined etiology | 0.438 | 0.500-8.036 | 0.578 |
Admission NIHSS | 0.905 | 0.815–1.004 | 0.059 |
OTI | 0.964 | 0.872–1.067 | 0.481 |
OTR | 0.981 | 0.939–1.025 | 0.394 |
Volume of rCBF < 30% | 0.975 | 0.948–1.003 | 0.083 |
Mismatch Volume | 0.998 | 0.992–1.003 | 0.419 |
Mismatch Ratio | 0.989 | 0.967–1.011 | 0.320 |
Volume of Tmax > 4s | 0.998 | 0.994–1.003 | 0.449 |
Volume of Tmax > 6s | 0.997 | 0.992–1.002 | 0.243 |
Volume of Tmax > 8s | 0.993 | 0.984–1.001 | 0.086 |
Volume of Tmax > 10s | 0.990 | 0.980–1.001 | 0.073 |
Volume of Tmax > 12s | 0.986 | 0.971-1 | 0.055 |
HIR | 0.008 | 0.001–0.254 | 0.006 |
Abbreviations: 95% CI = 95% confidence interval; ASPECTS = Alberta Stroke Program Early Computed Tomography Score; NIHSS = National Institutes of Health Stroke Scale; OTI = onset-to-imaging time; INR = international normalized ratio; TIA = transient ischemic attack. Six variables in italic were significant at the p < 0.1 level and entered to the multivariable-integrated model for predicting poor/good outcomes. T-type = TOAST type. |
Based on the results of univariate analysis, HIR was selected as a significant imaging predictor to further assess its performance for predicting outcome in the recanalized patients. A receiver operating characteristic curve (ROC) analysis demonstrated an area under the ROC of 0.701; with a sensitivity of 62.9% and a specificity of 69.2% (Fig. 4). The optimum cut-off value of HIR for differentiating favorable or unfavorable outcome, was calculated using the Youden index. The HIR cut-off was determined (HIR ≥ 0.491) to predict an unfavorable functional outcome even when the patients were successfully recanalized (mTICI score = 2b or 3).
Furthermore, seven variables, which had significance at P < 0.1 level in the univariable logistic regression, were selected to build up an integrated multivariate regression model for predicting unfavorable outcome. The gender, admission NIHSS, HIR, volume of Tmax > 8s, of Tmax > 10s, of Tmax > 12s and volume of rCBF < 30% were then integrated, and the performance of the integrated prediction model revealed a ROC-AUC of 0.782, with a model sensitivity of 82.1% and specificity of 62.9% (Fig. 5).