Backscatter coefficient estimation from human thyroid in vivo

Currently, fine-needle aspiration (FNA) biopsy is the gold standard for the diagnosis of thyroid cancer. Therefore, there is a need to develop non-invasive tools that aid in the identification of malignant tissues in this gland. Quantitative ultrasound imaging using backscatter coefficients (BSCs) has shown potential to characterize thyroid tissues in rodent models ex vivo. As an initial step towards the goal of thyroid cancer diagnosis on a clinical setting, in this study the feasibility of BSC-based imaging in humans in vivo is evaluated. Radiofrequency data was collected using a scanner equipped with a 9-MHz linear probe from ten patients with no clinical records of thyroid disease and no thyroid nodules visible through ultrasonic examination. Backscatter coefficients were estimated using the reference phantom method and considering analysis regions of 4.5 mm by 4.5 mm. Attenuation compensation was performed considering the presence of sub-cutaneous fat, muscle, and thyroid tissues. Images were constructed by calculating the mean BSC within the analysis bandwidth spanning from 3 to 8 MHz The average value of the backscatter coefficients in normal thyroids was 0.056±0.037 1/(sr.cm) within the analysis bandwidth. The coefficient of variation of the mean BSC values of all the thyroid samples was less than 2dB. These results suggest that consistent imaging of BSC-derived parameters from human thyroids in vivo is possible and may play a role in thyroid tissue characterization.