Differential Imaging of Liver Tumors before and after Microwave Ablation with Electrode Displacement Elastography

Abstract

Liver cancer is a leading cause of cancer-related deaths; however, primary treatment options such as surgical resection and liver transplant may not be viable for many patients. Minimally invasive image-guided microwave ablation (MWA) provides a locally effective treatment option for these patients with an impact comparable to that of surgery for both cancer-specific and overall survival. MWA efficacy is correlated with accurate image guidance; however, conventional modalities such as B-mode ultrasound and computed tomography have limitations. Alternatively, ultrasound elastography has been used to demarcate post-ablation zones, yet has limitations for pre-ablation visualization because of variability in strain contrast between cancer types. This study attempted to characterize both pre-ablation tumors and post-ablation zones using electrode displacement elastography (EDE) for 13 patients with hepatocellular carcinoma or liver metastasis. Typically, MWA ablation margins of 0.5-1.0 cm are desired, which are strongly correlated with treatment efficacy. Our results revealed an average estimated ablation margin inner quartile range of 0.54-1.21 cm with a median value of 0.84 cm. These treatment margins lie within or above the targeted ablative margin, indicating the potential to use EDE for differentiating index tumors and ablated zones during clinical ablations. We also obtained a high correlation between corresponding segmented cross-sectional areas from contrast-enhanced computed tomography, the current clinical gold standard, when compared with EDE strain images, with r² values of 0.97 and 0.98 for pre- and post-ablation regions.

Keywords: Elastography; Liver cancer treatment; Microwave ablation; Segmentation; Treatment monitoring.

Figure 1 –

Flowchart describing EDE estimation process.

Figure 2 –

Flowchart describing registration and estimated ablative margin calculation.

Figure 3 -

An example of differentiating pre and post ablation regions from a patient with hepatocellular carcinoma (HCC). Subfigures (a) and (b) show the clinical B-mode images of pre and post ablation respectively, (c) and (d) show the CT slices complementing the planes shown in (a) and (b), (e) and (f) show the B-mode images from the frames which elastography images were estimated, while (g) and (h) show results of EDE with improved lesion visualization. Blue and red curves in (g) and (h) represent the pre-ablation tumor and post-ablation zone segmentations respectively from strain tensor images.

Figure 4 -

An example of differentiating pre and post ablation regions from a patient with HCC. Subfigures (a) and (b) show the clinical B-mode images of pre and post ablation respectively, (c) and (d) show the CT slices complementing the planes shown in (a) and (b), (e) and (f) show the B-mode images from the frames which elastography images were estimated, while (g) and (h) show results of EDE with improved lesion visualization. Blue and red curves in (g) and (h) represent the pre-ablation tumor and post-ablation zone segmentations respectively from strain tensor images.

Figure 5 -

An example of differentiating pre and post ablation regions from a patient with colon-adenocarcinoma. Subfigures (a) and (b) show clinical B-mode images of pre and post ablation respectively, (c) and (d) show the CT slices complementing the planes shown in (a) and (b), (e) and (f) show the B-mode images from the frames which elastography images were estimated, while (g) and (h) show results of EDE with improved lesion visualization. Blue and red curves in (g) and (h) represent the pre-ablation tumor and post-ablation zone segmentations respectively from strain tensor images.

Figure 6 -

An example of differentiating pre and post ablation regions from a patient with a carcinoid tumor. Subfigures (a) and (b) show the clinical B-mode images of pre and post ablation respectively, (c) and (d) show the CT slices complementing the planes shown in (a) and (b), (e) and (f) show the B-mode images from the frames which elastography images were estimated, while (g) and (h) show results of EDE with improved lesion visualization. Blue and red curves in (g) and (h) represent the pre-ablation tumor and post-ablation zone segmentations respectively from strain tensor images.

Figure 7 –

An example of differentiating pre and post ablation regions from a patient with HCC. Subfigures (a) and (b) show the clinical B-mode images of pre and post ablation respectively, (c) shows the pre-procedural MR image of the target tumor and (d) shows the CT slice complementing the planes shown in (a) and (b), (e) and (f) show the B-mode images from the frames which elastography images were estimated, while (g) and (h) show results of EDE with improved lesion visualization. Blue and red curves in (g) and (h) represent the pre-ablation tumor and post-ablation zone segmentations respectively from strain tensor images.

Figure 8–

Distributions of EAMs for 13 patients with hepatocellular carcinoma or liver metastases shown in blue or red, respectively.

Figure 9 –

Comparison of EDE segmented areas against the clinical segmented areas from respective MR or CT images. (a) represents pre-ablation tumors and (b) represents the post-ablation zones where the red asterisks show data points and dashed line represents the 45° correlation line.

Figure 10 -

An example of differentiating pre and post ablation regions from a patient with a squamous lung tumor metastasized in the liver. Subfigures (a) and (b) show the B-mode images for pre and post ablation respectively, while (c) and (d) show axial strain tensor imaging results of electrode displacement elastography with improved lesion visualization.