gms | German Medical Science

Objective: PET using radiolabelled amino acids has become a constant tool for preoperative diagnostics and monitoring treatment response in patients with brain metastases. Interestingly, tumor margins in FET-PET and MRI are rarely completely congruent. We investigated which imaging shows more accurately the extension of viable tumor tissue.

Methods: We included patients ≥18 years old with metastatic tumor disease who received MRI (3T MR scanner, Achieva 3T; Philips Medical Systems) and FET-PET (SIEMENS Biograph 64 PET/CT scanner) imaging who were eligible for tumor resection of ≥1 brain metastases. Navigation trajectories and tumor areas were created on BrainLab iPlan 3.0 cranial planning software (BrainLAB AG) targeting contrast enhancing PET negative tumor areas and PET positive areas without contrast enhancement. During surgery we performed navigated biopsies of the respective tumor areas based on trajectories or we switched intraoperatively to freehand biopsies and took screenshots of the targets. Volumes of interest (VOI) were later of the tumor voxel with the highest FET uptake and of the biopsy targets. Tumor-to-background standardized uptake ratios (TBRs) relating maximal counts in the tumor VOI to the maximal counts in a background VOI derived from a cortical region in the contralateral (non tumor) hemisphere, were calculated by a method described by Popperl et al. previously and used in other studies.

Results: 13 patients (7 females/6 males, median age 58, [range 46-73]) were included. 4 patients were diagnosed with malignant melanoma, 4 patients with NSCLC and 5 patients with different primary. In 3 cases of PET positive and 3 cases of PET negative biopsies we were not able to detect viable tumor cells. Median TBRmax value was 3,07 [range 1,72-7,49], TBRmax in PET positive VOI without contrast enhancement was 2,22 [1,21-3,77] and in the PET negative VOI 1,13 [range 0,23-1,99]. Malignant melanoma metastases in general appeared to have higher TBR then the rest but non-significantly. There was no significant difference in the TBR values based on Karnofsky index, localization, tumor primary, prior radiochemotherapy or application of steroids.

Conclusion: We were not able to define the tumor margins based on FET-PET and MRI targeted biopsies reliably in all cases and did not find any statistical differences of TBR in brain metastases based on different histological type or localizations. Further investigations based on MIB-1 immunohistochemistry could provide evidence which imaging is more accurate.

Figure 1 [Fig. 1]