Prognostic values of quantitative and morphological parameters of dbPET in patients with luminal-type breast cancer: A pilot study

550 Objectives: To investigate prognostic values of quantitative and morphological parameters of dedicated breast PET (dbPET) in patients with luminal-type breast cancer in comparison with whole body PET/CT (WB-PET/CT), MRI and clinicopathological parameters using progression-free survival (PFS) as the main outcome. Methods: This retrospective study included a total of 29 mass-forming luminal-type invasive breast cancers in 29 females (age 32-80 years), who underwent 18F-fluorodeoxyglucose (FDG) dbPET, FDG WB-PET/CT and dynamic contrast enhanced MRI at diagnosis. For index primary tumors, SUVmax and morphological appearance (tumor shape, margin, internal pattern and rim uptake) on dbPET, those on WB-PET/CT, and tumor size and morphological appearance (tumor shape, margin, internal pattern and rim enhancement) on MRI were evaluated and used as imaging variables. Morphological tumor appearance on PET and MRI was subjectively determined by independent four readers (2 breast radiologists and 2 nuclear medicine physicians). We correlated imaging variables as well as clinicopathological parameters (age, menopausal status, clinical stage, presence of nodal metastasis, distant metastasis, and multicenter/multifocal lesions, histological grade, subtype, expression levels of estrogen receptor, progesterone receptor, HER2, and Ki67) to PFS using Kaplan-Meier plots and the log-rank test. Unadjusted hazard rations (HRs) were calculated from a univariate Cox proportional hazards model. Results: With a median follow up of 6.9 years (range, 1.4 to 9.1 years), disease progression events occurred in seven patients (recurrence after surgery in 5 cases, progression after chemotherapy in 2 cases). Patients with higher tumor SUVmax on dbPET (>15.1 vs. ≤15.1, log-rank p = 0.034; HR 7.13, 95% confidence interval [CI] 1.22 to 134.66, p = 0.027) and those with distant metastasis (positive vs. negative, log-rank p = 0.008; HR 6.97, 95% CI 0.98 to 33.08, p = 0.052) had a significantly shorter PFS. The other imaging and clinicopathological variables, including tumor SUVmax on WB-PET/CT (>7.0 vs. ≤7.0, log-rank p = 0.070; HR 4.05, 95% CI 0.87 to 28.33, p = 0.075) and morphological parameters, were not significantly associated with PFS in this study. Conclusions: Although sample size was limited, our study demonstrated that SUVmax of primary tumor on dbPET and presence of distant metastasis at diagnosis were predictive of PFS. FDG dbPET could be helpful to predict prognosis in patients with luminal-type breast cancer. $$graphic_AF83AD62-8E48-4D36-AEA4-120C57E04E49$$