To evaluate the spatial relationship between peritumoral edema and recurrence pattern in patients with glioblastoma (GBM).

Forty-eight primary GBM patients received three-dimensional conformal radiotherapy that did not intentionally include peritumoral edema within the clinical target volume between July 2000 and June 2001. All 48 patients have subsequently recurred, and their original treatment planning parameters were used for this study. New theoretical radiation treatment plans were created for the same 48 patients, based on Radiation Therapy Oncology Group (RTOG) target delineation guidelines that specify inclusion of peritumoral edema. Target volume and recurrent tumor coverage, as well as percent volume of normal brain irradiated, were assessed for both methods of target delineation using dose-volume histograms.

A comparison between the location of recurrent tumor and peritumoral edema volumes from all 48 cases failed to show correlation by linear regression modeling (r(2) = 0.0007; p = 0.3). For patients with edema >75 cm(3), the percent volume of brain irradiated to 46 Gy was significantly greater in treatment plans that intentionally included peritumoral edema compared with those that did not (38% vs. 31%; p = 0.003). The pattern of failure was identical between the two sets of plans (40 central, 3 in-field, 3 marginal, and 2 distant recurrence).

Clinical target volume delineation based on a 2-cm margin rather than on peritumoral edema did not seem to alter the central pattern of failure for patients with GBM. For patients with peritumoral edema >75 cm(3), using a constant 2-cm margin resulted in a smaller median percent volume of brain being irradiated to 30 Gy, 46 Gy, and 50 Gy compared with corresponding theoretical RTOG plans that deliberately included peritumoral edema.