Absolute Breast 99mTc-sestamibi Quantification with Molecular Breast Imaging in Patients Undergoing Neoadjuvant Chemotherapy

Introduction: Recently published techniques to absolutely quantify ^99mTc concentrations in normal and malignant breast tissue with dual-headed planar Molecular Breast Imaging (MBI) systems have been proposed and validated in Monte Carlo simulations. In this work, we present the initial clinical implementation of these techniques and compare calculated concentration values to image- and whole-body-based surrogate estimates.

Methods: Sixty-four patients with locally advanced breast cancer underwent ^99mTc-sestamibi MBI and diagnostic mammography prior to undergoing neoadjuvant therapy. A breast radiologist contoured tumor, normal tissue, and entire breast regions of interest (ROIs) on all acquired craniocaudal (CC) and mediolateral oblique (MLO) MBI views and measured tumor volumes (V_t) on mammography. These contours were used to calculate the ^99mTc activity in tumors (A_t), activity concentrations in tumor (C_t) and normal breast tissue (C_n) (See Supporting Document), the relative concentration ratio (C_r = C_t/C_n), and the ratio of average tumor ROI counts to average background ROI count (TBR) in the standard photopeak images. Average whole-body tissue concentration (C_wb) at time of imaging was calculated as ratio of decay-corrected administered activity and patient weight. We provide summary statistics for V_t, A_t, C_t, C_n, C_wb, C_r, and TBR. To assess robustness, we report Bland-Altman analysis for C_r and TBR calculated between different views and detectors. Furthermore, we report correlations between C_wb and C_n breast concentrations and between count-based (TBR) and concentration-based (C_r) tumor uptake ratios.

Results: The median (interquartile range, IQR) tumor volume was V_t = 7.1 (21.1) cm³. The median (IQR) uptake across all images (N=256) were: A_t = 2.2 (3.7) mCi, C_t = 276 (323) nCi/cm³, C_n = 29 (15) nCi/cm³, C_wb = 91 (29) nCi/g, C_r = 11.2 (11.7), and TBR = 2.3 (1.2). In both views, the mean ± standard deviation (SD) difference between detectors was lower for C_r (1.8% ± 2.8%) compared to TBR 9.7% (6.7%), i.e., Cr is more robust as a measure of tumor uptake. However, the mean (SD) discrepancy in calculated values between CC and MLO views increased from 6.4% (5.5%) in TBR to 13.2% (12.3%) in C_r. No correlation was noted between TBR and C_r suggesting image-based metrics cannot be used as surrogate for activity concentration ratios. Additionally, no correlation was observed between C_wb and C_n. Interestingly, whole-body concentration overestimated normal tissue concentration, on average, by a factor of 3.5 (range 1.4-11.6), highlighting 1) the limited availability of ^99mTc-sestamibi for breast tissue uptake after its rapid uptake in heart, liver, and even muscular tissue, and 2) the appropriateness of normalizing malignant tissue uptake by the corresponding benign organ tissue uptake instead of simply assuming a uniform whole-body uptake.

Conclusions: Initial calculations of tumor and normal tissue ^99mTc-sestamibi uptake in 64 patients have demonstrated the feasibility of quantifying absolute activity uptake in tumor and normal breast tissue with a commercial planar MBI system and absolute activity concentrations by incorporating anatomical tumor volume information. Absolute uptake measurements were more consistent than image-based metrics within a given MBI view, but additional work is still necessary to resolve sources of variability between MBI views. These initial findings highlight the importance of consistent measurement conditions for longitudinal comparisons (e.g., changes in TBR or C_r during or after neoadjuvant therapy). Evaluation of these quantitative uptake metrics (both absolute and relative changes over treatment cycle) as early predictors of response to neoadjuvant therapy is currently underway.

• Download figure
• Open in new tab
• Download powerpoint