Tumor treating electric fields (TTFields) are an established treatment for glioblastoma patients. But the anatomical and physical characteristics of the brain and tumor contributing to treatment efficacy are unknown. We contoured gross tumor volume (GTV) using ScanIP and measured tumor size according to the RANO criteria on 5 patients with recurrent glioblastomas who benefited from TTFields (cohort 1) and 5 who did not (cohort 2). Tumor surface area and geometric centroid distance (GCD) from the bilateral ventricles were computed. Wilcoxon rank sum test was used to compare these physical parameters between the two cohorts. The results showed that the respective median GTV was 11.6cm3 and 38.1cm3 (P=0.0591), while the respective median GCD was 5.0cm and 5.3cm (P=0.6761), in cohort 1 and 2. The tumor size had a median of 8.2cm2 in cohort 1, as compared to a median of 53.9cm2 in cohort 2 (P=0.0591). The surface area had a median of 56.2cm2 in cohort 1, as compared to a median of 214.0cm2 in cohort 2 (P=0.4034). After removing an outlier from cohort 1 and another from cohort 2, the respective median GTV was 9.7cm3 and 41.3cm3 (P=0.1003), and the respective median surface area was 46.5cm2 and 236.0cm2 (P=0.0304). The respective median surface area/GCD was 13.3cm and 38.9cm (P=0.0304) and the respective median GTV/GCD was 2.7cm2 and 8.5cm2 (P=0.0304). The data suggest that the ratio of surface area/GCD and GTV/GCD, which are proportional to the tissue capacitance, may be an important parameter for optimizing TTFields efficacy in the treatment of glioblastoma.