Identifying locally invasive brain metastases to improve control and survival

AbstractBACKGROUNDImaging and pathological investigations of patients with brain metastases are identifying more local invasion than previously appreciated. Recurrence rates are 15–20% after focal treatment, and targeted treatment of the invasive areas may help improve local control.METHODProspective image guided study using matched MRI and biopsy samples at the brain-metastasis interface. Retrospective case note review of operative approach and survival.RESULTSSamples were obtained during image guided neurosurgical resection of 26 metastases. Pathological analysis of 71/118 samples showed tumour cells present beyond the apparent brain-metastasis interface whereas 20/26 tumours were reported to show a clear macroscopic plane by the operating surgeon indicating significant discordance (kappa 0.02). There was no association of an invasive growth pattern with the type of primary cancer (Fisher’s exact test p=0.94) nor subsequent survival (HR 95% CI 0.42–1.28, p=0.28). It was not possible to identify a locally invasive phenotype using preoperative conventional, diffusion or perfusion MRI. In a retrospective review of 370 previously operated cases, gross total resection versus piecemeal resection was associated with an improved overall survival which was lost after 12 months, which may be due to the competing risk of progressive extracranial disease.CONCLUSIONThere is a biological basis for extended local treatments such as supramarginal resection or cavity boost radiosurgery in patients with brain metastases. Identifying patients who will benefit from such treatments remains problematic.