Lagrangian Deformation Tracking for Microwave Ablation Zones

Minimally invasive treatments, namely microwave ablation (MWA), show promise in extending liver cancer patients survival particularly when surgical resection and transplantation are not viable. However, for MWA, treatment success requires an imaging modality for visualizing ablative zones and margins. Ablative margins ideally contain 5–10 mm of normal tissue surrounding the tumor to reduce recurrence, while retaining as much normal functioning liver tissue as possible. Conventionally, this role is filled by contrast enhanced computed tomography (CECT) (clinical gold standard) however some shortcomings exist. An alternative approach, electrode displacement elastography, shows potential for visualizing ablation zones, yet motion artifacts from physiological motion degrade consistent delineation. We present a Lagrangian deformation tracking (LDT) algorithm to track tissue movement thereby reducing decorrelation noise and motion artifacts. Results are shown on a single spherical inclusion phantom scanned freehand with both qualitative and quantitative comparisons. In this paper, we demonstrate greater consistency among visualized LDT strain tensor images showing a standard deviation in inclusion radius of only 0.4 mm while conventional Eulerian tracking resulted in 1.5 mm over 0.5 seconds of tracking the 1.25 cm radius inclusion. Phantom results are validated using in vivo data on human subjects where consistency of LDT strain tensor tracking is demonstrated by the improved Dice and Pearson's correlation coefficient values obtained with respect to CECT, where the Eulerian approach performs worse when compared to LDT. Additionally, as Lagrangian tracking accumulates displacements over time, LDT increased visualization success rate from 75% seen with Eulerian tracking to 100% visualization. LDT therefore provides higher confidence that clinicians can determine ablative margins for defining ablation efficacy.