BACKGROUND: Conventional interpretation of methionine PET (Met-PET) scans in suspected brain tumors uses the ratio of the tracer uptake within the lesion to the corresponding contralateral area. The precise location at which the region of interest used to calculate the reference value is placed is vital, because local variations in methionine uptake may significantly alter the calculated ratio. Identifying a precise mirror region is complicated by the distorting effect of the tumor and the need for manual realignment of the image. METHOD: Patients with low-grade primary brain tumors or benignlesions were identified on the basis of a tissue diagnosis or surveillance neuroimaging that excluded a high-grade tumor. These conditions were selected because they primarily involve a single hemisphere, with methionine uptake in the unaffected hemisphere being essentially normal. A total of 180 Met-PET scans performed during 2003 2005 were identified from the database at the Max Planck Institute for Neurological Research, coded, and anonymized for analysis. Scans demonstrating midline lesions, significant mass effect, or evidence of substantial previous surgery were then excluded. A methionine template was prepared using data from patients who had undergone both FDG and Met-PET scans within eight weeks, with normalization to a previously developed FDG template. Methionine scans were coregistered to the template, after masking of any tumor, and the diseased hemispheres stripped. Mean uptake maps for each hemisphere were calculated on a voxel-by-voxel basis and merged to create the normal methionine uptake map. Scans unsuitable for inclusion into the normal map were reanalyzed using the contralateral hemisphere and the normal uptake map for reference values, allowing the methods to be compared. RESULTS: Good correlation was found between uptake ratios using reference values calculated by both methods. Reference values could be reliably calculated in tumors that were previously problematic to analyze, such as those that cross the midline. Coregistration of the normal map was impaired in some cases by loss of the normal architecture, but valid reference values were obtained despite this. CONCLUSION: Use of a normal uptake may facilitate calculation of PET uptake ratios in brain tumors. Further research is required to evaluate the correlation with histological findings and the accuracy of image coregistration in the presence of distorting tumors.