Rehemorrhage of brainstem cavernous malformations: a benchmark approach to individualized risk and severity assessment

Abstract

OBJECTIVE Brainstem cavernous malformations (BSCMs) represent a unique subgroup of cavernous malformations with more hemorrhagic presentation and technical challenges. This study aimed to provide individualized assessment of the rehemorrhage clustering risk of BSCMs after the first symptomatic hemorrhage and to identify patients at higher risk of neurological deterioration after new hemorrhage, which would help in clinical decision-making. METHODS A total of 123 consecutive BSCM patients with symptomatic hemorrhage were identified between 2015 and 2022, with untreated follow-up > 12 months or subsequent hemorrhage during the untreated follow-up. Nomograms were proposed to individualize the assessment of subsequent hemorrhage risk and neurological status (determined by the modified Rankin Scale [mRS] score) after future hemorrhage. The least absolute shrinkage and selector operation (LASSO) regression was used for feature screening. The calibration curve and concordance index (C-index) were used to assess the internal calibration and discrimination performance of the nomograms. Cross-validation was further performed to validate the accuracy of the nomograms. RESULTS Prior hemorrhage times (adjusted OR [aOR] 6.78 per ictus increase) and Zabramski type I or V (OR 11.04) were associated with rehemorrhage within 1 year. A lower mRS score after previous hemorrhage (aOR 0.38 for a shift to a higher mRS score), Zabramski type I or V (OR 3.41), medulla or midbrain location (aOR 2.77), and multiple cerebral cavernous malformations (aOR 11.76) were associated with worsened neurological status at subsequent hemorrhage. The nomograms showed good accuracy and discrimination, with a C-index of 0.80 for predicting subsequent hemorrhage within 1 year and 0.71 for predicting neurological status after subsequent hemorrhage, which were maintained in cross-validation. CONCLUSIONS An individualized approach to risk and severity assessment of BSCM rehemorrhage was feasible with clinical and imaging features.