Author(s): Zakaria R.; Jenkinson M.D.; Chen Y.J.; Wang S.; Chawla S.; Mohan S.; Hughes D.M.; Berghoff A.S.; Preusser M.; Poptani H.
Source: British Journal of Neurosurgery; 2018; vol. 32 (no. 1); p. 79
Publication Date: 2018
Publication Type(s): Conference Abstract
Abstract:Objectives: Brain metastases (BM) are a common, serious pathology but selecting patients for treatment and trials is difficult due to the heterogeneous nature of the disease. Since diffusion weighted MRI (DWI) is a promising biomarker in cancer imaging, we investigated whether the addition of a DWI parameter to existing clinical models improves survival modelling in BM. Design: Retrospective multicentre study. Subjects: Patients undergoing surgical resection of a brain metastasis and having preoperative diffusion weighted MRI brain. Methods: Combined retrospective analysis of 223 adult patients from three different centres with BM treated from diagnosis to death. All patients underwent preoperative DWI followed by neuro-surgical resection then adjuvant therapy. Results: Median overall survival (OS) was 9.6 months (95% CI 7.5-11.7). Adjuvant whole brain radiotherapy, younger age, improved performance status and favourable status of the extracranial disease were all associated with longer survival. Tumour ADC did not predict primary cancer of origin but when added to conventional models such as Recursive Partitioning Analysis and Graded Prognostic Assessment improved the prediction of overall survival, with higher tumour ADC being favourable. Conclusions: Combining biological information obtained at the time of diagnosis with known clinical factors is a logical means of improving survival prediction in BMs and further work is needed to incorporate existing and future imaging data into such models.