Assessment and validation of proposed classification tools for brainstem cavernous malformations.

OBJECTIVE

Treatment indications for patients with brainstem cavernous malformations (BSCMs) remain difficult and controversial. Some authors have tried to establish classification tools to identify eligible candidates for surgery. Authors of this study aimed to validate the performance and replicability of two proposed BSCM grading systems, the Lawton-Garcia (LG) and the Dammann-Sure (DS) systems.

METHODS

For this cross-sectional study, a database was screened for patients with BSCM treated surgically between 2003 and 2019 in the authors' department. Complete clinical records, preoperative contrast-enhanced MRI, and a postoperative follow-up ≥ 6 months were mandatory for study inclusion. The modified Rankin Scale (mRS) score was determined to quantify neurological function and outcome. Three observers independently determined the LG and the DS score for each patient.

RESULTS

A total of 67 patients met selection criteria. Univariate and multivariate analyses identified multiple bleedings (p = 0.02, OR 5.59), lesion diameter (> 20 mm, p = 0.007, OR 5.43), and patient age (> 50 years, p = 0.019, OR 4.26) as predictors of an unfavorable postoperative functional outcome. Both the LG (AUC = 0.72, p = 0.01) and the DS (AUC = 0.78, p < 0.01) scores were robust tools to estimate patient outcome. Subgroup analyses confirmed this observation for both grading systems (LG: p = 0.005, OR 6; DS: p = 0.026, OR 4.5), but the combined use of the two scales enhanced the test performance significantly (p = 0.001, OR 22.5).

CONCLUSIONS

Currently available classification systems are appropriate tools to estimate the neurological outcome after BSCM surgery. Future studies are needed to design an advanced scoring system, incorporating items from the LG and the DS score systems.