This study evaluated various quantitative criteria for analysis of breast imaging with PET using the radiolabeled glucose analog 18F-fluorodeoxyglucose (FDG).
Methods: In a prospective study, 73 patients with abnormal mammography or palpable breast masses scheduled for biopsy were investigated with PET. A total of 97 breast tumors were evaluated by histology, including 46 benign and 51 malignant tumors. Using a whole-body PET scanner, attenuation-corrected images were acquired between 40 and 60 min after tracer injection. For Patlak analysis, dynamic data acquisition was obtained in 24 patients. To differentiate between benign and malignant breast tumors, receiver operating characteristic curves were calculated using incrementally increasing threshold values for tumor/ nontumor ratios based on average and maximum activity values per region of interest, standardized uptake values (corrected for partial volume effect, normalized to blood glucose, partial volume effect and blood glucose, using the lean body mass as well as the body surface area) and calculating the FDG influx rate (K) assessed by Patlak analysis.
Results: Quantification of FDG uptake in breast tumors provided objective criteria for differentiation between benign and malignant tissue with similar diagnostic accuracy as compared with visual analysis. Applying correction for partial volume effect and normalization by blood glucose yielded the highest diagnostic accuracy.
Conclusions: These quantitative methods provided accurate evaluation of PET data for differentiating benign from malignant breast tumors. Quantitative assessment is recommended to complement visual image interpretation with the potential benefit of reduced interobserver variability.