Résumé
PURPOSE: This study aimed to investigate the relationship between the SUVmax of primary breast cancer lesions and the molecular subtypes based on the recommendations of the St. Gallen consensus meeting 2013.METHODS: Clinical records of patients who underwent F-18 FDG PET/CT for initial staging of invasive ductal carcinoma (IDC) of SUVmax was correlated with the molecular subtypes defined by the St. Gallen Consensus Meeting 2013, i.e., luminal A-like (LA), luminal B-like HER2 negative (LBHER2−), luminal Blike HER2 positive (LBHER2+), HER2 positive (HER2+), and triple negative (TN), and with the clinicohistopathologic characteristics.RESULTS: The molecular subtype was LA in 38 patients, LBHER2− in 72, LBHER2+ in 21, HER2+ in 30, and TN in 22. The mean SUVmax in the LA, LBHER2−, LBHER2+, HER2+, and TN groups were 4.5 ± 2.3, 7.2 ± 4.9, 7.2 ± 4.3, 10.2 ± 5.5, and 8.8 ± 7.1, respectively. Although SUVmax differed significantly among these subtypes (p < 0.001), the values showed a wide overlap. Optimal cut-off SUVmax to differentiate LA from LBHER2−, LBHER2+, HER2+ and TN were 5.9, 5.8, 7.5, and 10.2 respectively, with area under curve (AUC) of 0.648, 0.709, 0.833, and 0.697 respectively. The cut-off value of 5.9 yielded the highest accuracy for differentiation between the LA and non-LA subtypes, with sensitivity, specificity, and AUC of 79.4 %, 57.9 %, and 0.704 respectively.CONCLUSION: The SUVmax showed a significant correlation with the molecular subtype. Although SUVmax measurements could be used along with immunohistochemical analysis for differentiating between molecular subtypes, its application to individual patients may be limited due to the wide overlaps in SUVmax.