Metabolic Tumor Burden Assessed by Dual Time Point [18F]FDG PET/CT in Locally Advanced Breast Cancer: Relation with Tumor Biology.

Abstract

The aim of the study was to investigate the influence of dual time point 2-deoxy-2-[18F]fluoro-D-glucose ([18F]FDG) positron emission tomography/x-ray computed tomography (PET/CT) on the standard uptake value (SUV) and volume-based metabolic variables of breast lesions and their relation with biological characteristics and molecular phenotypes. Retrospective analysis including 67 patients with locally advanced breast cancer (LABC). All patients underwent a dual time point [18F]FDG PET/CT, 1h (PET-1) and 3h (PET-2) after [18F]FDG administration. Tumors were segmented following a three-dimensional methodology. Semiquantitative metabolic variables (SUVmax, SUVmean, and SUVpeak) and volume-based variables (metabolic tumor volume, MTV, and total lesion glycolysis, TLG) were obtained. Biologic prognostic parameters, such as the hormone receptors status, p53, HER2 expression, proliferation rate (Ki-67), and grading were obtained. Molecular phenotypes and risk-classification [low: luminal A, intermediate: luminal B HER2 (-) or luminal B HER2 (+), and high: HER2 pure or triple negative] were established. Relations between clinical and biological variables with the metabolic parameters were studied. The relevance of each metabolic variable in the prediction of phenotype risk was assessed using a multivariate analysis. SUV-based variables and TLG obtained in the PET-1 and PET-2 showed high and significant correlations between them. MTV and SUV variables (SUVmax, SUVmean, and SUVpeak) where only marginally correlated. Significant differences were found between mean SUV variables and TLG obtained in PET-1 and PET-2. High and significant associations were found between metabolic variables obtained in PET-1 and their homonymous in PET-2. Based on that, only relations of PET-1 variables with biological tumor characteristics were explored. SUV variables showed associations with hormone receptors status (p<0.001 and p=0.001 for estrogen and progesterone receptor, respectively) and risk-classification according to phenotype (SUVmax, p=0.003; SUVmean, p=0.004; SUVpeak, p=0.003). As to volume-based variables, only TLG showed association with hormone receptors status (estrogen, p<0.001; progesterone, p=0.031), risk-classification (p=0.007), and grade (p=0.036). Hormone receptor negative tumors, high-grade tumors, and high-risk phenotypes showed higher TLG values. No association was found between the metabolic variables and Ki-67, HER2, or p53 expression. Statistical differences were found between mean SUV-based variables and TLG obtained in the dual time point PET/CT. Most of PET-derived parameters showed high association with molecular factors of breast cancer. However, dual time point PET/CT did not offer any added value to the single PET acquisition with respect to the relations with biological variables, based on PET-1 SUV, and volume-based variables were predictors of those obtained in PET-2.