Comparison of two dose calculation methods of intra-arterial carboplatin in the treatment of recurrent glioblastoma multiforme.
e13503 Background: Glioblastoma is a deadly brain cancer, and as part of the natural evolution, its relapses will mostly occur 32 to 36 weeks after the initial diagnosis. The standard of care for recurrent glioblastomas is not well defined and multiple options exist. Intra-arterial chemotherapy as one option is currently practiced at the Centre hospitalier universitaire de Sherbrooke, the only center doing this procedure in Canada. The aim of this study is to evaluate the impact of using the area under the curve (AUC) or body surface area (BSA) dosing formula to guide intra-arterial carboplatin administration. It was designed to explore the relation of two dose calculation methods on efficacy and toxicity, determined by progression-free survival (PFS) and hematologic toxicity, respectively. Methods: A retrospective observational study was conducted, which includes adult in-patients who received intra-arterial carboplatin for recurrent glioblastoma from June 1st, 2000 to December 31st, 2017. It examined the associations of the method used to calculate the doses, in mg or mg/mL x min-1, with clinical outcomes. Results: One hundred ninety patients were included in the analysis. For the primary endpoint, an exposure greater than 6 of AUC was found potentially harmful in terms of efficacy compared to AUC under 6 (HR 0.529, p = 0.0044). In terms of toxicity, a higher AUC is highlighted to correlate with a slightly higher incidence of thrombocytopenia and neutropenia (OR 1.051 p = 0.0271; OR 1.072 p = 0.049). No significant association between toxicity and dose calculated with BSA could be established. Conclusions: AUC greater than 6 is detrimental in terms of effectiveness according to progression free survival. Moreover, AUC could be used to predict neutropenia and thrombocytopenia. However no significant predictions were found using BSA method. We recommend the use of AUC method in intra-arterial carboplatin instead of BSA method since the latter was not statistically related to toxicity. A prospective trial comparing the two dosage methods would be needed.