Back

Early prediction of space-occupying hemispheric infarction: Development and validation of the Malignant Edema Risk Assessment (MERA) score - Results from the PREDICT-MMI study

Dabbagh, A.; Welle, F.; Mielke, J.; Wawrzyniak, M.; Scherlach, C.; Hoffmann, K.-T.; Pelz, J. O.

2025-11-23 neurology
10.1101/2025.11.21.25340775 medRxiv
Show abstract

BackgroundSpace-occupying hemispheric infarction is a severe complication after acute ischemic stroke (AIS) due to large-vessel occlusion (LVO). As decompressive hemicraniectomy improves outcome, early identification of patients at risk is clinically crucial. We developed and validated the Malignant Edema Risk Assessment (MERA) score to predict the individual risk of space-occupying infarction after LVO. MethodsFour models, each an extension of the previous one, were defined a priori: Net water uptake (estimated from the first non-contrast enhanced CT (NCCT); Model 0) + volume of intracranial cerebrospinal fluid (Model 1) + clinical data (age, localization of LVO, NIHSS at admission, successful recanalization; Model 2) + laboratory parameters, that reflected the patients fluid status (Model 3). The best performing model was identified based on a retrospective derivation cohort composed of patients with AIS due to LVO of the anterior circulation and first NCCT within 24 hours after symptom onset, while patients for the validation cohort were prospectively enrolled between September 2023 and February 2025. Results155 patients with AIS and LVO were included in the derivation cohort, of whom 84 (54.2%) had a space-occupying hemispheric infarction. Model 2 performed best (MERA score [≥] 66.6%: sensitivity 75.0%, specificity 86.3%, area under the curve (AUC) of the receiver operated curve (ROC) 0.853 (0.801 - 0.918)). 156 patients with AIS (28 (17.9%) patients with a space-occupying hemispheric infarction) were enrolled in the validation study. Validating the MERA score resulted in a sensitivity of 85.7%, a specificity of 85.9%, a positive predictive value of 57.1%, and a negative predictive value of 96.5%. The AUC of the ROC amounted to 0.944 (0.909 - 0.980). ConclusionsThe MERA score allows an early and accurate identification of patients at high risk developing a space-occupying hemispheric infarction after AIS due to LVO. Clinical Trial Registration: https://www.drks.de. Unique identifier: DRKS00033266. Clinical Perspective What Is New?O_LIThis study introduces the MERA score, a simple and semi-automatic tool that integrates early CT-based and basic clinical variables to estimate the individual risk of malignant brain edema in patients with large-vessel occlusion stroke. C_LI What Are the Clinical Implications?O_LIBecause all input data are available before patients reach the stroke unit or neurointensive care unit, the MERA score enables early and practical risk stratification. C_LIO_LIAn estimated cutoff of 66.6% allows clinicians to classify patients into higher-and lower-risk categories for the development of malignant brain edema using a straightforward and quickly applicable method. C_LI O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=200 SRC="FIGDIR/small/25340775v1_ufig1.gif" ALT="Figure 1"> View larger version (99K): org.highwire.dtl.DTLVardef@1bf8a41org.highwire.dtl.DTLVardef@dd841org.highwire.dtl.DTLVardef@f09995org.highwire.dtl.DTLVardef@1f336eb_HPS_FORMAT_FIGEXP M_FIG C_FIG

Matching journals

The top 5 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.