MON-571 The Performance Of An Advanced Ultrasound Technique, Quantitative Ultrasound, Compared To Conventional Ultrasound In The Evaluation Of Thyroid Nodules

Abstract

INTRODUCTION: Thyroid cancer is the most common endocrine malignancy. While thyroid nodules (TN) can been seen in up to 50% of the adult population, only about 5 percent of the nodules are malignant. Currently, management of TN involves risk stratification prior to biopsy using B-mode grey-scale ultrasound (US) characteristics including echogenicity, micro- and macrocalcifications, and margins. A less-subjective method of TN evaluation, shear wave elastography, has been evaluated but both the grey-scale classification and elastography are subject to operator and machine variability. This study assesses the use of quantitative ultrasound (QUS) to differentiate between malignant and benign thyroid nodules. B-mode appearance based on the radiofrequency (RF) signal envelope and discards much of the information in the RF signal. QUS uses the normalized power spectrum and extracts discarded RF signal information to obtain parameters that reflect scatterer size, concentration and relative acoustic impedance. The advantage of using QUS is that the parameter values reflect the microarchitecture (<120 µm) of the tissue and are theoretically operator and machine independent. RESULTS: In this preliminary cohort, US data from 52 TN (with 57 data sets) were collected using a GE Logiq E9 US system and QUS estimates were obtained. 8 TN were diagnosed as malignant and 44 TN were benign as determined by cytology, molecular testing or surgical pathology. The B-mode images were reviewed by an experienced ultrasonologist and TI-RADS (Russ et al) scores were calculated. A combination of QUS parameters (EAC, I0, µ) produced an AUC value of 0.96. The AUC by using the TI-RADS scoring criteria alone was 0.94 and when combined with QUS parameters (EAC, µ) it was 0.97. When the grey-scale US images were reviewed by a senior fellow in training, the TI-RADS AUC was lower. This highlights the subjective variability and effect of operator experience in risk stratification of TN using conventional B-mode characteristics and the importance of investigating advanced US techniques that are operator and machine independent for better cancer detection and improved identification of benign TN to avoid unnecessary biopsy and surgery.