Alejandro Rodríguez-Vázquez, MD
Malignant stroke is a devastating condition that often leads to death or severe impairment. In large middle cerebral artery (MCA) infarcts, decompressive hemicraniectomy (DH) practiced within the first 48 hours of symptom onset improves both survival rates and functional outcomes to a moderate but remarkable extent. Infarct volume is one of the most important elements that define a stroke as malignant and, therefore, is often used as a main landmark in making clinical decisions. However, there are controversies around the infarct volume threshold that would allow to make the best prognosis estimation.
In this study, Casolla et al. tried to determine optimal infarct volume using magnetic resonance imaging (MRI) to predict a catastrophic outcome 1 year after DH, defined as a modified Rankin scale (mRS) score 5 or death. They studied 173 patients who underwent DH using b1000 diffusion-weighted image (DWI) and apparent diffusion coefficient (ADC) maps on MRI performed in admission, 24-36 hours after treatment or earlier in case of clinical worsening. In patients who received 2 or more MRI before DH, the closest imaging to intervention was analyzed. Of those 173 patients, 42 had a catastrophic outcome (34 died, and 8 mRS 5) in a year. The optimal threshold of infarct volume to predict this outcome was 211 mL in b1000 DWI, with a sensitivity of 59.5% (955 CI, 43.3-74.4) and a specificity of 61.8% (95% CI, 52.9-70.2). On ADC maps, the optimal threshold was 181 mL, with a sensitivity of 57.1% (95% CI, 41.0-72.3) and a specificity of 62.5% (95%, 53.5-70.9). The area under the curve for both sequences’ optimal volumes as predictors for catastrophic outcome was 0.64. An infarct volume of 274 mL ob b1000 DWI and 244 mL in ADC maps had a specificity of 90% for death or mRS 5, but there was not a volume able to predict a catastrophic outcome with a specificity higher than 90%.
With the aforementioned number of patients of 173, this is one of, if not the largest, prospective cohorts of consecutive large MCA strokes with systematized MRI analysis using semiautomatic calculations and, hence, allowing a reliable standardized measure of volume. However, patients who had not received DH were excluded, so there could be a potential selection bias, especially in the largest infarct volumes.
Following the study results, although infarct volume measured both in b1000 DWI and ADC highly correlates with a catastrophic outcome, it does not predict by itself death or bedriddeness, and thus they should not be used exclusively to deny DH in potential candidates. There is more than meets the eye in these often complex patients, so a multidisciplinary approach is mandatory. However, combining neuroimaging with clinical and analytic results in future studies could help us clinicians to establish a consensus regarding critical life-or-death dilemmas such as DH.
