PURPOSE: The objective of this study was to perform the dosimetry and evaluate the dose-response relationships in newly diagnosed patients with malignant brain tumors treated by direct injections of (131)I-labeled 81C6 monoclonal antibody (MAb) into surgically created resection cavities (SCRCs). METHODS AND MATERIALS: Absorbed doses to the 2-cm-thick shell as measured from the margins of the resection cavity interface were estimated for 42 patients with primary brain tumors. MR images were used to assess the enhanced-rim volume as a function of time after radiolabeled MAb therapy. Biopsy samples were obtained from 15 patients and 1 autopsy. RESULTS: The average absorbed dose [range] to the 2-cm shell region was 32 [3-59] Gy. For the endpoint of minimal time to MR contrast enhancement, the optimal absorbed dose and initial dose-rate were 43 +/- 16 Gy and 0. 41 +/- 0.10 Gy/h, respectively. There was a correlation between the absorbed dose and dose rate to the shell region and biopsy outcome (tumor recurrence, radionecrosis, and tumor recurrence and/or radionecrosis). In this Phase I study, the maximum tolerated dose (MTD) was 120 mCi. At this MTD, the estimated average absorbed dose and initial dose rate to the 2-cm shell were 41 [9-89] Gy and 0.51 [0.24-1.13] Gy/h, respectively. These values are in agreement with the optimal values based on the time to MR lesion rim enhancement. CONCLUSIONS: The average absorbed dose to the 2-cm shell region varied considerably and mainly depended on cavity volume. In future clinical trials, the administered activity of (131)I-labeled 81C6 MAb may be adjusted based on cavity volume in order to deliver the optimal absorbed dose of 43 Gy rather than giving a fixed administered activity.