gms | German Medical Science

Objective: For surgery of WHO°II astrocytomas, postoperative extent of resection (EOR) has consistently been shown to correlate with overall survival as well as with progression- and malignance-free intervals. However, times when to determine the amount of residual tumor to define EOR have been chosen very heterogeneously in past studies limiting the comparability of results. Particularly in non contrast-enhancing tumors, surgical trauma and formation of edema are sources of bias for the interpretation of early postoperative imaging. This study sought to evaluate differences of early vs. Late resection control imaging to determine the most appropriate time for the assessment of EOR.

Method: Cases with primary resections for WHO°II astrocytomas and simultaneous early and late postoperative imaging were retrospectively identified since 2004. Besides adjusting for histology, previous radiation therapy lead to exclusion from the study to account for confounders of image interpretation. 29 cases matched inclusion criteria. Volumetric analysis and calculation of EOR was performed on FLAIR sequences (Fluid-Attenuated Inversion Recovery) early (<72h), as well as late postoperatively (median 3,6months, range 1,9-7,5m). Statistical analysis of volumetric values and correlation to patient progression-free survival (PFS) was performed.

Results: Median postoperative EOR significantly increased by 23% from 67% early, to 90% late postoperatively (p<0,01, Wilcoxon matched pairs test). Moreover, correlation analysis confirmed this difference to be systematic (Spearman r=0,71, p<0,0001). Evidence of surgical trauma (ischemia, edema) was found on 66% (19/29) of early postoperative MRI diffusion-weighted-imaging (DWI). In survival analysis, only late EOR was significantly associated with PFS (HR=0,23, p<0,05), early EOR was not a significant confounder of survival.

Conclusions: After resection of WHO°II astrocytomas, EOR depends strongly upon the time when it is defined after surgery. EOR was systematically underestimated early after surgery and only late EOR was a significant prognosticator for PFS in this series. Post-surgical trauma can mimic FLAIR hyperintensity that bias early postoperative image interpretation and might account for differences in EOR calculation. Results from this analysis strongly advocate the use of late instead of early postoperative imaging for the evaluation of residual tumor and definition of EOR in non contrast-enhancing astrocytomas.