Quantification of the Impact of Slice Thickness on Volumetric Analysis of Brain Tumors

The partial volume effect influences the results of volumetrics performed on MRI sequences with different slice thicknesses. Although a standardized brain tumor imaging protocol with high-quality 1 mm slices is recommended, segmentation is still commonly performed on thicker slices, especially in retrospective studies. We aimed to quantify the impact of different slice thicknesses on volumetric measurements of brain tumors, using treatment-naive glioblastoma as a representative example.

Volumes were measured on T1-weighted contrast-enhanced (CE) sequences using semiautomatic segmentation across three predefined slice thicknesses: 1 mm ("thin"; reference measurement), 3 mm ("intermediate"), and 5 mm ("thick"). Data from 46 patients with the availability of all three slice thicknesses were statistically analyzed, including repeated-measures ANOVA with post hoc comparisons, Bland-Altman analysis, correlations, and linear regression.

The tumor volumes measured increased with slice thickness, with mean values of 36.4 cm³ ("thin"), 38.6 cm³ ("intermediate"), and 41.0 cm³ ("thick"). Repeated measures ANOVA showed significant differences between groups (p < 0.01). The mean deviation of the tumor volumes on "thick" relative to "thin" slices was 16% (±10%). Bland-Altman analysis and linear regression confirmed systematically higher measured volumes on "thick" slices. Tumor size on "thin" slices correlated negatively with the percentage deviation on "thick" slices (r = -0.56; p < 0.01).

Volumetric measurements obtained from "thick" slices showed systematically higher tumor volumes and greater variability relative to the 1 mm reference. Volumetric data obtained from different slice thicknesses should not be combined, or concepts for adjustment should be considered to maintain methodological consistency and reliability in volumetric studies. We advocate the use of high-quality 3D 1 mm imaging for volumetric analysis whenever possible.