Validation of the Medical Research Council and a newly developed prognostic index in patients with malignant glioma: how useful are prognostic indices in routine clinical practice?

摘要

Although different prognostic indices for malignant gliomas have been developed, their validity outside of clinical trials has not been widely tested. The aim of this study was to determine whether the Medical Research Council (MRC) brain tumour prognostic index was able to stratify patients for survival managed in routine practice, and secondly to compare the results with our newly developed prognostic score which included tumour grade and only 3 prognostic groups. The MRC and the new prognostic index were calculated for a group of 119 adult patients with malignant glioma managed by surgical resection/biopsy and post-operative radiotherapy. For the MRC and new score, 6 and 3 prognostic groups were defined, respectively. For all patients median survival was 11 (2-66) months. The overall survival rate at 12 and 24 months were 43% and 18%, respectively. The MRC median and two-year survival rates were 14 months and 26% for a score of 1-10, 14 months and 27% for a score of 11-15, 13 months and 22% for a score of 16-20, 8 months and 10% for a score of 21-25, 8 months and 0% for those scoring 26-33. There was only one patient in the 34-38 group. For the new prognostic index, median and two-year survival rates were respectively 16 and 26%; 12 and 23%; 8 and 7% for the good, intermediate and poor prognostic groups. Both indices were significant factors for survival in univariate analysis (MRC index, p = 0.0089, new index p = 0.0002), but not in multivariate analysis. Both the MRC and our newly devised prognostic score were able to separate patients into good and poor prognostic groups, which may aid in treatment decisions, although there was less differentiation between the MRC groups especially over the first year. Both scores use routinely available factors. However, inclusion of tumour grade in the new score may be an advantage over the MRC index.