34 Extent of resection in glioblastoma: Incorporating clinical and molecular data

Abstract

Background: The benefits of increasing extent of resection (EOR) for both overall survival and progression-free survival (PFS) in glioblastoma has been well documented. However, models predicting surgical outcomes have failed to incorporate a patient’s IDH status, a known prognostic factor. We isolate the impact of IDH on surgical outcomes. We determine the effect modification of increasing EOR and decreasing residual tumor volume (RTV) on IDH status. Methods: We performed a retrospective cohort study of 98 patients with glioblastoma who had undergone either biopsy or surgical resection. Tumor volumes were determined by volumetric analysis. Univariable and multivariable Cox PH Regression models were built using overall survival and PFS as endpoints. Results: Increasing EOR and decreasing RTV were both associated with prolonged overall survival and PFS. When IDH status was added to multivariable models, the model utilizing RTV provided a slightly better fit compared to EOR. An interaction term between RTV and IDH status was characterized, such that at low RTVs the prognosis of an IDH mutant is significantly better than that of an IDH wild-type, an effect that is less important as RTV increases. The significance of this term was confirmed by improved fit upon insertion into multivariable models. Conclusion: Minimizing RTV and increasing EOR are important prognostic factors for both IDH wild-type and IDH mutant glioblastoma. The protective benefit of the IDH mutation at lower RTVs suggests these patients are the best candidates for aggressive surgical resection.