Monitoring of thermal-induced changes in liver stiffness during controlled hyperthermia and microwave ablation in an ex-vivo bovine model with point shear wave elastography

ABSTRACT Purpose In point shear wave elastography (pSWE), an acoustic radiation force impulse (ARFI) is used to generate shear­ waves in liver tissue. Shear wave velocity (SWV) reflects the underlying tissue stiffness, which varies during the heating process. In order to investigate how liver stiffness changes during heating, we designed this two phases study. The first phase was conducted in controlled hyperthermia and aimed to (1) quantify the temperature dependence of liver tissue stiffness as measured with SWV (2) identify the threshold value of SWV related to coagulation (3) assess the reproducibility of this measurement and (4) evaluate the irreversibility of stiffness changes. The second phase was conducted during microwave (MW) ablation to (1) confirm the SWV threshold for the stiffness of coagulated tissue during MW ablation and (2) determine the reliability of pSWE to delineate the boundaries of ablation zone. Materials and Methods In phase 1 of the study 5 cuboidal samples of ex-vivo bovine liver were uniformly heated to target temperatures ranging 40-100°C and then cooled down until 50°C. B-mode ultrasound (US) imaging and pSWE were acquired simultaneously (Virtual TouchTM Tissue Quantification, Siemens Healthcare) and SWV, expressed in m/s, was measured in a fixed region of interest (ROI, < 1cm2) at set temperatures (2.5-5°C intervals). Tissue temperature was recorded by two thermocouples positioned at 1.5 cm from the ROI. The threshold value of SWV at 60°C (avg60), the temperature for immediate tissue coagulation, was identified. In phase 2, 11 samples were ablated by a commercial MW ablation system (HS AMICA®, H.S. Hospital Service SpA). B ­mode US imaging and SWV were acquired simultaneously. MW ablation was performed at 60 W until avg60 + 0.5 m/sec was reached in a ROI that was placed 1.5 cm radially from the antenna feed point. Afterwards, SWV was measured in several ROIs at established distances ranging from 10 mm to 40 mm from the antenna feed, acquiring 10 SWVs for each ROI. Finally, the specimens were cut along the antenna to obtain gross ­pathology of the coagulation necrosis. A SWV contour map was created, superimposing the pathology picture of the liver necrosis with the B­ mode US images of ROIs’ positions. To each point at pathology the following values were assigned: 0=outside, 0.5= at the border, and 1=within necrosis. The correlation of mean values of SWV with location at gross-pathology was evaluated. Results In phase 1 experiments a steep transition in liver stiffness was observed at 63.0 ± 2.4°C with an average SWV value of 3.54 ± 0.68 m/s (baseline value: 1.41 ± 0.17 m/s at room temperature). A decrease in SWV was observed up to 42°C (­0.31 ± 0.07 m/s with respect to baseline), a gradual increase was detected up to 55-60°C, and the described transition followed above 60°C. Avg60, defined as mean SWV at 60°C, was of 2.5 m/sec. This pattern was observed in all experiments, with very similar SWV/temperature curves. The changes in liver stiffness proved to be irreversible, as average SWV values of 4.40 ± 0.41 m/s were measured in the cool-down cycles. In phase 2, in 8 of the experiments, interrupted when SWV of 3 m/s was measured in the studied ROI, the ROI was at the inner side of the necrotic area border at pathology (accuracy 89%). SWV values measured in ROI valued at pathology as outside (0), border (0.5), and within (1) necrosis were dispersed. No correlation between SWV values for outside (0), border (0.5), and within (1) necrosis could be identified. Conclusion SWV is useful to monitor thermal changes in tissue in the setting of controlled hyperthermia. pSWE can provide a velocity threshold predictive of the presence of coagulation necrosis during MW ablation in ex­ vivo liver model. The complexity of ablation process in tissues, and possibly the shrinkage occurring after ablation, makes pSWE not able to reliably capture changes of stiffness within, at the border, and outside the necrotic zone in this experimental model