The growth rate of early DWI lesions is highly variable and associated with penumbral salvage and clinical outcomes following endovascular reperfusion

Abstract

Background: The degree of variability in the rate of early diffusion-weighted imaging expansion in acute stroke has not been well characterized.

Aim: We hypothesized that patients with slowly expanding diffusion-weighted imaging lesions would have more penumbral salvage and better clinical outcomes following endovascular reperfusion than patients with rapidly expanding diffusion-weighted imaging lesions.

Methods: In the first part of this substudy of DEFUSE 2, growth curves were constructed for patients with >90% reperfusion and <10% reperfusion. Next, the initial growth rate was determined in all patients with a clearly established time of symptom onset, assuming a lesion volume of 0 ml just prior to symptom onset. Patients who achieved reperfusion (>50% reduction in perfusion-weighted imaging after endovascular therapy) were categorized into tertiles according to their initial diffusion-weighted imaging growth rates. For each tertile, penumbral salvage [comparison of final volume to the volume of perfusion-weighted imaging (Tmax > 6 s)/diffusion-weighted imaging mismatch prior to endovascular therapy], favorable clinical response (National Institutes of Health Stroke Scale improvement of ≥8 points or 0-1 at 30 days), and good functional outcome (90-day modified Rankin score of ≤2) were calculated. A multivariate model assessed whether infarct growth rates were an independent predictor of clinical outcomes.

Results: Sixty-five patients were eligible for this study; the median initial growth rate was 3·1 ml/h (interquartile range 0·7-10·7). Target mismatch patients (n = 42) had initial growth rates that were significantly slower than the growth rates in malignant profile (n = 9 patients, P < 0·001). In patients who achieved reperfusion (n = 38), slower early diffusion-weighted imaging growth rates were associated with better clinical outcomes (P < 0·05) and a trend toward more penumbral salvage (n = 31, P = 0·103). A multivariate model demonstrated that initial diffusion-weighted imaging growth rate was an independent predictor of achieving a 90-day modified Rankin score of ≤2.

Conclusions: The growth rate of early diffusion-weighted imaging lesions in acute stroke patients is highly variable; malignant profile patients have higher growth rates than patients with target mismatch. A slower rate of early diffusion-weighted imaging growth is associated with a greater degree of penumbral salvage and improved clinical outcomes following endovascular reperfusion.

Keywords: MRI; ischemic stroke; lesions; radiology; reperfusion; stroke.

Figure 1

Predicted growth curves for patients with <10% reperfusion vs. >90% reperfusion. Kolmogorov-Smirnov test shows a significant difference between the two distributions (p<0.001).

Figure 2

Initial DWI growth rate for the 65 patients with known time of stroke onset. Graphed based on the assumption of infarct volume of 0 ml just prior to symptom onset and linear growth (based on initial findings in Part I of this study). Overall, median initial growth rate was 3.1 ml/hr, with a range from 0 ml/hr to 117 ml/hr.

Figure 3

Initial DWI growth rates and clinical and radiographic outcomes. Thirty-eight patients with successful reperfusion divided into three statistical tertiles to evaluate for association of initial growth rate and patient outcomes. Tertile 1 represents the slowest initial growth rate tertile (median initial growth rate 0.7 ml/hr). Tertile 2 is the middle growth tertile (median initial growth rate 3.2 ml/hr). Tertile 3 is the fastest growth tertile (median initial growth rate 18.3 ml/hr). Cochran-Armitage trend test shows p = 0.010 for Favorable Clinical Response. Cochran-Armitage trend test shows p = 0.004 for Good Functional Outcome. Mann-Whitney U Test comparing 1st and 2nd tertiles to 3rd tertile shows p = 0.057 for percent penumbral salvage.