Original Contribution
Visualization of HIFU-Induced Lesion Boundaries by Axial-Shear Strain Elastography: A Feasibility Study

https://doi.org/10.1016/j.ultrasmedbio.2010.11.009Get rights and content

Abstract

In this paper, we report on a study that investigated the feasibility of reliably visualizing high-intensity focused ultrasound (HIFU) lesion boundaries using axial-shear strain elastograms (ASSE). The HIFU-induced lesion cases used in the present work were selected from data acquired in a previous study. The samples consisted of excised canine livers with thermal lesions produced by a magnetic resonance–compatible HIFU system (GE Medical System, Milwaukee, WI, USA) and were cast in a gelatin block for the elastographic experiment. Both single and multiple HIFU-lesion samples were investigated. For each of the single-lesion samples, the lesion boundaries were determined independently from the axial strain elastogram (ASE) and ASSE at various iso-intensity contour thresholds (from –2 dB to –6 dB), and the area of the enclosed lesion was computed. For samples with multiple lesions, the corresponding ASSE was analyzed for identifying any unique axial-shear strain zones of interest. We further performed finite element modeling (FEM) of simple two-inclusion cases to verify whether the in vitro ASSE obtained were reasonable. The results show that the estimation of the lesion area using ASSE is less sensitive to iso-intensity threshold selection, making this method more robust compared with the ASE-based method. For multiple lesion cases, it was shown that ASSE enables high-contrast visualization of a “thin” untreated region in between multiple fully-treated HIFU-lesions. This contrast visualization was also noticed in the FEM predictions. In summary, the results demonstrate that it is feasible to reliably visualize HIFU lesion boundaries using ASSE. (E-mail: [email protected])

Introduction

High-intensity focused ultrasound (HIFU) has become a promising noninvasive technique to thermally ablate and destroy volume of tissue lying deep under the skin surface (Lele, 1967, Hill, 1995). The HIFU technique is synonymously used for focused ultrasound surgery (FUS), which was first investigated and described for neurosurgical applications about five decades ago (Lynn and Putnam, 1944, Fry et al., 1950). In the last few decades, the HIFU technique has found several applications that include, among others, ablation of tumors in prostate (Sanghvi et al., 1999, Foster et al., 1993, Gelet et al., 1993) and other organs (Coleman et al. 1985; Vallancien et al. 1993; Vaezy et al., 2001, Hynynen et al., 2001). The success of the HIFU procedure required that the progress of the treatment be accurately visualized. This was a major challenge until the late 1970s because there was no reliable imaging modality for the purpose. In the last two decades, magnetic resonance imaging (MRI)–based methods as well as ultrasound (US)–based methods have been investigated for monitoring the HIFU treatment.

MRI methods used for monitoring and guiding HIFU therapy have used real-time tracking of the resulting temperature elevation in the target region (Cline et al. 1994; Hynynen et al. 1996) to visualize the thermal damage (referred to as a HIFU lesion). Although MRI has proven to be quite capable of imaging HIFU lesions, it has several drawbacks and limitations. Most notably, MRI is expensive, not portable and may not be suitable for some populations such as those with pacemakers, pregnant women, children or large patients.

The desire to have a unified system for HIFU treatment and inexpensive real-time lesion visualization has fueled research on several ultrasound-based methodologies. However, the extent of the HIFU lesion is difficult to quantify with current B-mode imaging (sonography) techniques because of lack of contrast between lesion and normal tissue boundaries, as well as shadows that limit the entire view of the lesion. Bush et al. (1993) reported that there was a significant increase in the speed of sound and attenuation associated with HIFU lesion, whereas the backscatter coefficient did not change significantly.

Several methods have been proposed that attempt to estimate the temperature increase during HIFU exposure by detecting change in the speed of sound (Sehgal et al., 1986, Seip and Ebbini, 1995; Miller et al. 2002; Arthur et al., 2003, Pernot et al., 2004, Anand et al., 2007). The major limitation of this approach is that the data relating temperature and speed of sound are limited. Further, a nonlinear speed of sound profile is reported at 50° (Bamber and Hill, 1979, Bloch et al., 1998), limiting the temperature estimation accurately at these levels. Moreover, tissue composition varies from patient to patient and this may lead to different temperature-dependent speed of sound profiles (Miller et al. 2002). This interpatient variability would also limit the estimation of accurate temperature measurements. There have been other methods proposed that use the change in attenuation to visualize HIFU lesion (Ribault et al., 1998, Bevan and Sherar, 2001a, Bevan and Sherar, 2001b). These have been shown to be successful to an extent. Recently, a method for HIFU lesion localization has been proposed that is based on tracking the change in backscattered radiofrequency (RF) signals (Zheng and Vaezy 2010). Results from experiments performed on ex vivo chicken breast were reported. In their method, the decay of RF signal post HIFU exposure is tracked temporally on a pixel-by-pixel basis for the entire image. Later, an image is formed by mapping the decay rate at each pixel. The contrast caused by differences in decay rate between the HIFU lesion and the surrounding tissue was used to locate the lesion. Although the initial results are promising, the applicability of this method in vivo, where target motion may challenge pixel-by-pixel tracking for decay rate estimation, has yet to be demonstrated.

Ultrasound elastography was introduced by Ophir et al. (1991) as a technique to image the stiffness variation in soft tissues. The technique involves acquiring US (RF/envelope) signals from an imaging plane before and after the application of a small (∼1%) quasistatic compression. Typically, the pre- and postcompression frames are processed to generate images of local strain, commonly known as elastograms. When the elastogram depicts axial strain values, it is referred to as an axial strain elastogram (ASE) (Ophir et al. 1999).

It is known that protein denaturation causes changes in stiffness of the HIFU lesion compared with the surrounding soft tissue (Fasano et al., 1983, Consigny et al., 1989, Sapin-de Brosses et al., 2010). Therefore, axial strain elastography was proposed as a possible technique to visualize HIFU lesions (Stafford et al. 1998). The feasibility of using ASE to detect HIFU lesions was demonstrated in an ex vivo animal model (Stafford et al., 1998, Kallel et al., 1999) as well as in ex vivo human prostate (Souchon et al. 2003). Righetti et al. (1999) performed elastographic experiments with controlled compression on canine liver in vitro. They characterized the appearance of HIFU lesions on ASE and showed that lesion size measured from ASE was virtually unbiased and correlated well with that from the pathology image, when the lesion boundary on ASE was set at a particular iso-intensity strain contrast level of –2.5dB. Although the HIFU lesion had appreciable contrast with the surrounding soft tissue, a challenge was to define the boundary of the lesion. Righetti et al. (1999) showed that the lesion size depended on the choice iso-intensity strain contrast level and that –2.5 dB was appropriate for their case. However, for in vivo applications and applications where the HIFU lesion in a heterogeneous tissue is of interest, defining lesion boundary in a consistent and reliable manner becomes important.

Recently, axial shear strain elastography has been introduced as a method to visualize the boundaries of breast lesions (Thitaikumar et al. 2007). In this technique, the axial-shear strain experienced by the tissue element caused by quasistatic compression (as in elastography) is imaged and referred to as axial-shear strain elastograms (ASSE). The axial-shear strain is estimated using eqn (1).ɛaxialshear=(vx),where v is the displacement along the direction of compression (axial) and x is the lateral direction. We have shown previously that axial-shear strains are generated at the boundaries of a firmly bonded inclusion, thus producing contrast to visualize the boundary directly. We hypothesize that this method may also be useful in reliable HIFU lesion boundary visualization.

Section snippets

Original in vitro data acquisition and data reprocessing

The HIFU-induced lesions used in the present work were from data acquired in a previous study (Righetti et al. 1999). The samples consisted of excised canine livers with thermal lesions produced by a prototype magnetic resonance (MR)–compatible HIFU system (GE Medical System, Milwaukee, WI, USA). After thermal exposure, the samples were cast in a gelatin block and placed in the compression apparatus. The elastographic experiments were conducted using a linear array scanner (Diasonics Spectra

Single HIFU lesion case

A total of five single lesion cases were used in this work. The HIFU-induced lesion area was estimated from the two elastograms (ASE and ASSE) separately, based on the lesion boundary determined by the iso-intensity contour threshold values ranging from –2 dB to –6 dB.

Figure 2 shows the sonogram, ASE and ASSE of an example case of HIFU lesion. The lesion boundaries corresponding to the two extreme iso-intensity contour levels used (–6 dB and –2 dB) are also shown. The difference between the

Single HIFU lesion case

In this paper, we investigated the ASSE of in vitro samples of canine liver containing HIFU-induced lesion. Unlike ASE, where a stiff lesion is visualized as low strain region, ASSE does not visualize the stiff lesion directly. Rather, the finite axial-shear strain values at the lesion boundary contrasts the boundary from zero axial-shear strain inside the lesion. The ASSE of the single HIFU lesion case (Fig. 2c) appears similar to the ASSE of the firmly bonded inclusion case published in our

Conclusions

We have demonstrated that it is feasible to reliably visualize HIFU lesion boundaries using ASSE. Further, we have shown that the estimation of the lesion area using ASSE is less sensitive to iso-intensity threshold selection, making this method more robust compared with the ASE-based method. We have also shown that ASSE enables high-contrast visualization of the presence of a thin untreated tissue layer subtended between multiple fully treated HIFU-lesions.

Acknowledgments

Data used in this study were acquired previously for projects supported by NIH Program Project grants P01-CA64597 and P01-EB02105-13. The current work was supported by NIH grant R21-CA135580. The authors would like to thank the authors of the previous study (Righettti et al. 1999), who acquired and archived the data that is used in the present study.

References (39)

  • N.T. Sanghvi et al.

    Noninvasive surgery of prostate tissue by high intensity focused ultrasound: An updated report

    Eur J Ultrasound

    (1999)
  • C.M. Sehgal et al.

    Measurement and use of acoustic nonlinearity and sound speed to estimate composition of excised livers

    Ultrasound Med Biol

    (1986)
  • R. Souchon et al.

    Visualisation of HIFU lesions using elastography of the human prostate in vivo: Preliminary results

    Ultrasound Med Biol

    (2003)
  • R.J. Stafford et al.

    Elastographic imaging of thermal lesions in soft-tissue: A preliminary study in-vitro

    Ultrasound Med Biol

    (1998)
  • A.K. Thittai et al.

    Axial-shear strain distributions in an elliptical inclusion model: Experimental validation and in vivo examples with implications to breast tumor classification

    Ultrasound Med Biol

    (2010)
  • X. Zheng et al.

    An acoustic backscatter-based method for localization of lesions induced by high-intensity focused ultrasound

    Ultras Med Biol

    (2010)
  • A. Anand et al.

    Three-dimensional spatial and temporal temperature imaging in gel phantoms using backscattered ultrasound

    IEEE Trans Ultrason Ferroelectr Freq Control

    (2007)
  • R.M. Arthur et al.

    Noninvasive temperature estimation based on the energy of backscattered ultrasound

    Med Phys

    (2003)
  • S.H.B. Bloch et al.

    Measurements of sound speed in excised tissue over temperatures expected under high-intensity focused ultrasound conditions

    J Acoust Soc Am

    (1998)
  • Cited by (17)

    • Imaging of Thermal Effects during High-Intensity Ultrasound Treatment in Liver by Passive Elastography: A Preliminary Feasibility in Vitro Study

      2020, Ultrasound in Medicine and Biology
      Citation Excerpt :

      The liver is an interesting organ to use for this feasibility study, because it is already well studied. It has been characterized in terms of elasticity through axial-shear-strain elastography (Thittai et al. 2010), shear-wave elastography (Sapin-de Brosses et al. 2010; Arnal et al. 2011) and MRI (Righetti et al., 2001). However, the most important point is that the in vivo feasibility for passive elastography imaging of the liver has already been reported by Gallot et al. (2011).

    • Real-Time Monitoring of High-Intensity Focused Ultrasound Treatment Using Axial Strain and Axial-Shear Strain Elastograms

      2014, Ultrasound in Medicine and Biology
      Citation Excerpt :

      Thittai et al. (2011) suggested that ASSEs may become a useful modality in such situations to visualize thin untreated region under certain conditions. The primary condition is that the untreated “thin” region that forms a boundary between two adjacent lesions should be at a non-normal orientation with respect to the axis of compression (Thittai et al. 2007, 2011). We further hypothesized that once this untreated (soft) region is treated, the high-contrast axial-shear zone between the lesions will vanish.

    • Ex vivo characterization of canine liver tissue viscoelasticity after high-intensity focused ultrasound ablation

      2014, Ultrasound in Medicine and Biology
      Citation Excerpt :

      Over the past decade, high-intensity focused ultrasound (HIFU) has been found to be a promising, non-invasive therapeutic method (Li et al. 2004; Vallancien et al. 1992; Wu et al. 1999, 2005a, 2005b). An emerging application for ultrasound-based elasticity imaging techniques is the assessment and monitoring of change in tissue mechanical properties during HIFU (Arnal et al. 2011; Basauri and Lele 1962; Chenot et al. 2010; Fry et al. 1954; Kallel et al. 1999b; Righetti et al. 1999; Souchon et al. 2005; Thittai et al. 2011; Warwick and Pond 1968; Wu et al. 2001). Harmonic motion imaging for focused ultrasound (HMIFU) is a radiation force-based HIFU treatment monitoring technique, the feasibility of which has been studied in silico, in vitro, ex vivo and in vivo (Hou et al. 2010, 2011, 2012a, 2012b; Maleke and Konofagou 2008, 2009).

    View all citing articles on Scopus
    View full text