Intrinsic susceptibility contrast (R2*) in the evaluation of tumour oxygenation at baseline and in response to neoadjuvant chemotherapy in breast cancer

Introduction: The ability to image tumour hypoxia and evaluate oxygenation changes in response to treatment using R2* is a powerful yet underexplored tool in breast cancer in humans [1]. This study evaluates the relationship between baseline histology and dynamic contrast enhanced (DCE) and dynamic susceptibility enhanced (DSC) MRI parameters with R2* in breast cancer. The role of R2* as an imaging biomarker of response to neoadjuvant chemotherapy (NAC) is also explored. Methods: 37 patients with solid, well defined, primary invasive ductal breast adenocarcinomas were imaged with a spoiled multi-gradient echo T2*-weighted MRI sequence (TE 5-75ms, TR100s, flip angle (α) 40 ̊, 8mm slice thickness, FOV 260mm, 256 matrix). T1-weighted DCE-MRI sequences (TE 4.7ms, TR 11ms, α 35 ̊, 256 matrix) and DSC-MRI sequences (TE 20ms, TR 30ms, α 40 ̊, 128 matrix) were also performed using 0.1mmol/kg and 0.2 mmol/kg body weight of Gd-DTPA respectively. R2* values were calculated using a least-squares fitting routine on ln[signal] plotted against TE. DCE-MRI images were analysed with Tofts’ pharmacokinetic model [2] and a modified Fritz-Hansen assumed arterial input function [3] using specialist MRIW software (Institute of Cancer Research, London) [4]. DSC parameters were calculated from a fitted Γ−variate function using MRIW [4]. Whole tumour ROI parametric values were acquired: R2*, K, ve, kep, IAUGC60, rBV, rBF and the MTT of the fitted curve. Relationships between baseline R2* and tumour characteristics (size, grade, ER/PR/HER2 status) and DCE and DSC-MRI parameters were explored using Spearman’s rank correlation for continuous variables and the Mann-Whitney U test for discrete variables. Baseline R2* and changes in R2* with treatment were also correlated with final pathological response using paired t-testing. R2* was compared with DCE and DSC kinetic parameters as a predictor of response using ROC (receiver operating characteristic curve) analyses.