Prognostic values of dedicated breast PET parameters in patients with breast cancer: Analysis by tumor size

Objectives: This retrospective study was to investigate prognostic potential of morphological and quantitative parameters of dedicated breast PET (dbPET), whole-body PET/CT (WB-PET/CT), breast MRI and clinicopathological parameters using progression-free survival (PFS) and disease-free survival (DSS) in breast cancer patients with stratified analysis by tumor size.

Methods: We retrospectively reviewed a total of 111 female patients (mean age, 57.2±13.6 years; range, 33-89 years) with known or suspicious breast cancer, who underwent 18F-fluorodeoxyglucose (FDG) dbPET, FDG WB-PET/CT, and dynamic contrast-enhanced MRI at our institution between June 2010 and February 2014. DbPET images were acquired using a ring-shaped prototype dbPET system. For primary tumors, SUVmax and morphological features of tumor uptake (mass-like, nonmasss-like and focus with their subcategory classifications) on dbPET, SUVmax on WB-PET/CT, and tumor size on MRI were evaluated and used as imaging parameters. Optimal cutoff values for SUVmax and size were determined using the Youden index. Two board certified nuclear medicine physicians independently determined morphological appearance on PET. All disagreements were resolved by a third reader. Patients were divided into two groups by tumor size on MRI (≤2cm or >2cm) that were named small tumor group and large tumor group, respectively. We correlated imaging parameters and clinicopathological parameters (age, menopausal status, clinical TNM staging, histological grade, subtype, expression levels of estrogen receptor and progesterone receptor, HER2 and Ki67) to PFS and DSS using Kaplan-Meier analysis, the log-rank test and multivariate Cox regression model in the whole subjects as well as in each subgroup.

Results: The median length of follow-up was 86.8 months (range, 8.9 to 106.8 months) for all patients, disease progression events occurred in 22 patients and 16 patients died with the cancer progression. PFS and DSS in all patients were significantly shorter in patients with higher tumor SUVmax on dbPET (≥15.1) or WB-PET (≥11.6), larger tumor size on MRI (≥34mm), the existence of lymph node metastasis or distant metastasis, cStage III or higher and non-luminal type (all, p<0.05). Multivariate analysis showed that tumor SUVmax on dbPET and WB-PET, and clinical stage were independent predictors of PFS and DSS (all, p<0.05). Morphological dbPET parameters had no prognostic significance. In stratified analysis by tumor size, univariate analysis demonstrated that, in large tumor group (n=70), higher tumor SUVmax on dbPET, existence of nodal or distant metastasis were significantly associated with both shorter PFS and shorter DSS, while SUVmax on WB PET/CT and non-luminal type were significantly associated only with shorter DSS. In small tumor group (n=40), higher tumor SUVmax on dbPET and the existence of distant metastasis were still significant predictors for shorter PFS and DSS, however the others, including SUVmax on WB PET/CT, were not. Multivariate analysis was not performed in the stratified analysis due to sample size insufficiency. Conclusions: The study suggests that FDG accumulation level in the primary tumor is a significant prognostic indicator for PFS and DSS in patients with breast cancer, and that dbPET may be helpful in assessing FDG accumulation in small tumors (≤T1) that can be difficult to evaluate with conventional WB PET.