Comparison study of displacement estimation methods for microwave ablation procedures using electrode displacement elastography

Abstract

Minimally invasive procedures such as microwave ablation (MWA) are growing in popularity as a substitute for surgical resection of hepatocellular carcinomas. Ultrasound electrode displacement elastography (EDE) has demonstrated the potential to provide a feasible non-ionizing approach for monitoring the region of thermal necrosis. We present results on a comparison of the contrast, signal-to-noise (SNR) and contrast-to-noise (CNR) ratios for two displacement estimation algorithms; namely the Quality-Guided Displacement Tracking (QGDT) and the Coupled Subsample Displacement Estimation (CSDE) methods. Twenty patients with EDE strain images estimated from the same radiofrequency frame pair using both algorithms are presented. CSDE depicts well delineated ablated regions, while QGDT in many instances depicts noisy boundaries. Quantitatively, the SNR, contrast, and CNR obtained using CSDE were higher than those obtained with QGDT. The axial SNR, contrast, and CNR for QGDT were found to be 34.6 ± 11.9 dB, 7.8 ± 4.2 dB, and 2.7 ± 6.4 dB, while the CSDE presented with values of 56.9 ± 20.4 dB, 22.5 ± 12.3, and 3.8 ± 2.5 dB, respectively. This study indicates that EDE using CSDE provides improved strain images however, both approaches produced reproducible results.