The Spetzler-Martin grading system is a poor predictor of outcomes of patients with cerebellar arteriovenous malformations (AVMs), research indicates.
The researchers report that their supplementary grading system, previously shown to improve on the predictive accuracy of the Spetzler-Martin grading system, is the more accurate grading system in patients specifically with cerebellar AVMs.
"Key components of the Spetzler-Martin scale such as venous drainage and eloquence are distorted by cerebellar anatomy in ways that the components of the supplementary scale are not, resulting in greater predictive accuracy with the supplementary scale," they believe.
The supplementary scale is based on just three items: patient age, bleeding, and compactness. These are "not susceptible to anatomic differences between the cerebellum and cerebrum," says the team.
In the current study of 60 patients with cerebellar AVMs and 401 with cerebral AVMs, cerebellar AVMs were more likely than cerebral ones to drain deep (57 vs 41%), but were less likely to be in an eloquent area (30 vs 61%). And they were significantly more likely to be ruptured at presentation (73 vs 49%).
The increased bleeding of cerebellar AVMs warrants aggressive management, say study author Michael Lawton (University of California at San Francisco, USA) and team, who advise surgery for all patients with ruptured AVMs.
"We also recommend treatment in most patients with unruptured cerebellar AVMs, selecting patients with low- and intermediate-grade AVMs for surgery and high-grade AVMs for radiosurgery," they write in Neurosurgery.
All patients in the study underwent surgery. After an average of 1.1 years follow up, 74% of patients had good outcomes (modified Rankin Scale [mRS] 0-2), 16% had poor outcomes (mRS 3-4), and 11% had died. Patients with tonsillar and tentorial AVMs had better outcomes than those with vermian, suboccipital, and petrosal AVMs.
The ability of the supplementary grading system to predict which patients would have good or poor outcomes was identical for those with cerebellar and cerebral AVMs, with areas under the receiver operating characteristic (ROC) curve of 0.70 and 0.70 (where 1.0 is perfect discrimination and 0.5 no better than chance).
By contrast, the Spetzler-Martin scale was more accurate for cerebral than cerebellar AVMs, with corresponding values of 0.62 and 0.56. Of note, the supplementary system was actually more accurate than the Spetzler-Martin scale for cerebral AVMs.
"Areas under the ROC curve exceeded thresholds for clinical utility with the supplementary scale but not with the Spetzler-Martin scale," says the team.
Licensed from medwireNews with permission from Springer Healthcare Ltd. ©Springer Healthcare Ltd. All rights reserved. Neither of these parties endorse or recommend any commercial products, services, or equipment.