BACKGROUND: Cerebral cavernous malformations (CCM) may lead to repetitive intracerebral hemorrhage. In selected cases, a surgical resection is indicated. OBJECTIVE: To identify magnetic resonance imaging (MRI) features of CCM that correlate with the difficulty of dissection and postoperative outcome. METHODS: This study prospectively analyzed pre- and postoperative MRI features, intraoperative findings (surgical questionnaire), and postoperative outcome of 41 patients with eloquent CCM. Based on the results of the surgeon's questionnaire and postoperative MRI findings, all surgical procedures were dichotomized in a "difficult" (group A) or "not difficult" (group B) lesion dissection. Based on the correlation of preoperative MRI features with groups A and B, a 3-tiered classification was established and tested for sensitivity and specificity. RESULTS: In 22 patients, dissection of the lesion was rated difficult. This was significantly correlated with amount of postoperative diffusion restriction on MRI ( P = .001) and postoperative outcome ( P = .05). Various preoperative MRI features were tested for correlation and combined in a 3-tiered classification. Receiver operating characteristics revealed excellent and good results for predicting difficulty of dissection for the different classification types. CONCLUSION: We provide a meticulous analysis and new classification of preoperative MRI features that seem to be involved in the microsurgical resection of CCM.
BACKGROUND:Cerebral cavernous malformations (CCM) may lead to repetitive intracerebral hemorrhage. In selected cases, a surgical resection is indicated. OBJECTIVE: To identify magnetic resonance imaging (MRI) features of CCM that correlate with the difficulty of dissection and postoperative outcome. METHODS: This study prospectively analyzed pre- and postoperative MRI features, intraoperative findings (surgical questionnaire), and postoperative outcome of 41 patients with eloquent CCM. Based on the results of the surgeon's questionnaire and postoperative MRI findings, all surgical procedures were dichotomized in a "difficult" (group A) or "not difficult" (group B) lesion dissection. Based on the correlation of preoperative MRI features with groups A and B, a 3-tiered classification was established and tested for sensitivity and specificity. RESULTS: In 22 patients, dissection of the lesion was rated difficult. This was significantly correlated with amount of postoperative diffusion restriction on MRI ( P = .001) and postoperative outcome ( P = .05). Various preoperative MRI features were tested for correlation and combined in a 3-tiered classification. Receiver operating characteristics revealed excellent and good results for predicting difficulty of dissection for the different classification types. CONCLUSION: We provide a meticulous analysis and new classification of preoperative MRI features that seem to be involved in the microsurgical resection of CCM.