Treatment of Glioblastoma Multiforme at Progression or Recurrence: A Service Evaluation

Abstract AIMS After initial treatment, glioblastoma multiforme (GBM) invariably reoccurs or progresses. When this happens, there is no established standard of care and treatment choice is made on a case-by-case basis. In our service, there is a multidisciplinary approach to progression, with oncologists and surgeons cooperating closely. This means that our cohort is ideal for studying treatment at progression. METHOD We performed a retrospective cohort service evaluation on patients diagnosed with isocitrate dehydrogenase wildtype GBM over a 4-year, 5-month period. We focused on patients who had received Stupp regimen treatment. Analysis was though a cox-proportional hazard model with multiple imputation analysis for missing data when appropriate. RESULTS The variable most strongly correlated with the treatment chosen at progression was patient performance status (Kruskal Wallis chi-squared = 22.804, p=0.00014). Some variables, such as age at progression or initial surgical management, were unexpectedly not correlated with the chosen treatment. Of the possible treatments at progression, only surgical management approached borderline significance for increasing survival relative to best supportive care (p=0.06). Performance status had at most a moderate, non-significant impact upon survival post-progression. The MGMT methylation status of the tumour emerged as a powerful predictor of survival post-progression (p=0.01). CONCLUSION A range of variables are correlated with which treatment is chosen at progression, but they do not always relate to the post-progression survival. Of all treatments, we can only demonstrate surgical management as having an impact upon survival, suggesting a more moderate effect of medical therapies relative to best supportive care.