Abstract Background and aims We investigated whether radiological lesion severity, quantified by voxel-level intensity on apparent diffusion coefficient (ADC) maps, provides prognostic information beyond total lesion volume. Methods We included patients from the MR CLEAN-NO IV randomized trial with available follow-up MRI 12–120 hours after randomization. Ischemic lesions were manually segmented. Radiological severity was characterized by (1) intensity-weighted lesion volume, inversely weighting voxel volume by ADC intensity; (2) volume of voxels with ADC 470; and (3) mean ADC intensity of the lesion. The primary outcome was the 90-day modified Rankin Scale (mRS). Associations were assessed using ordinal regression adjusted for prognostic covariates. Model discrimination was evaluated using optimism-corrected C-statistics. Results Among 539 participants, 127 had follow-up MRI (median 26 hours; IQR 22–38). Median lesion volume was 8 mL (IQR 3–32), and median volume with ADC470 was 2 mL (IQR 1–12). The median intensity-weighted lesion volume was 2 mL (IQR 0–6). Lesion volume was associated with outcome (aOR per 10 mL:1.10; 95% CI=1.01–1.21). Mean lesion intensity was also associated with functional outcomes (aOR 0.84; 95%CI=0.77 –0.91) after adjusting for total lesion volume, randomization to MRI time, and predefined prognostic factors. The predictive performance of models considering only the total lesion volume was modest (C-index=0.69, 95%CI=0.62-0.76) and was not significantly improved with inclusion of average lesion intensity (C-index=0.72, 95%CI=0.65-0.79) or the intensity weighted lesion volume (C-index=0.71, 95%CI=0.64-0.77). Conclusions ADC-derived radiological lesion severity captures biologically relevant information beyond infarct size but provided only marginal gains in outcome prediction. Conflict of interest Dr Majoie reports grants from European Commission to other; grants from Health Evaluation Netherlands to other; grants from Stryker Corporation to other; stock holdings in Nicolab; grants from Dutch Heart Foundation to other; and grants from Boehringer Ingelheim to other. Dr Roos reports stock holdings in Nicolab. The other authors report no conflicts. All other authors have nothing to disclose.
Cavalcante et al. (Fri,) studied this question.