Evaluation of visualization of the prostate gland in vibro-elastography images

Abstract

In this paper, vibro-elastography (VE), an ultrasound-based method that creates images of tissue viscoelasticity contrast, is evaluated as an imaging modality to visualize and segment the prostate. We report a clinical study to characterize the visibility of the prostate in VE images and the ability to detect the boundary of the gland. Measures for contrast, edge strength characterized by gradient and statistical intensity change at the edge, and the continuity of the edges are proposed and computed for VE and B-mode ultrasound images. Furthermore, using MRI as the gold standard, we compare the error in the computation of the volume of the gland from VE and B-mode images. The results demonstrate that VE images are superior to B-mode images in terms of contrast, with an approximately six fold improvement in contrast-to-noise ratio, and in terms of edge strength, with an approximately two fold improvement in the gradient in the direction normal to the edge. The computed volumes show that the VE images provide an accurate 3D visualization of the prostate with volume errors that are slightly lower than errors computed based on B-mode images. The total gland volume error is 8.8 ± 2.5% for VE vs. MRI and 10.3 ± 4.6% for B-mode vs. MRI, and the total gland volume difference is −4.6 ± 11.1% for VE vs. MRI and −4.1 ± 17.1% for B-mode vs. MRI, averaged over nine patients and three observers. Our results show that viscoelastic mapping of the prostate region using VE images can play an important role in improving the anatomic visualization of the prostate and has the potential of becoming an integral component of interventional procedures such as brachytherapy.

Introduction

Prostate cancer is the most prevalent type of cancer among men and is projected to affect 24,600 men in Canada (Canadian Cancer Society, 2010) and 217,730 in the United States (National Cancer Institute, 2010) in 2010. It is the most numerous cancer diagnosed in European men (382,000 cases in 2008) (Ferlay et al., 2010). Common treatment options include brachytherapy and radical prostatectomy. Low dose rate (LDR) prostate brachytherapy is generally used for early stage, intra-capsular prostate cancer and has rapidly gained acceptance due to its highly successful clinical results (Morris et al., 2009). In this treatment, 40-150 small radioactive seeds (Iodine-125 or Palladium-103) are inserted through the perineum and permanently implanted into the prostate and periprostatic tissue. In high dose rate (HDR) brachytherapy, temporary catheters are placed inside the prostate which allow the placement of high dose rate sources delivering the radiation treatment over a number of fractions, typically over a few days. Radical prostatectomy is a surgical option in which the prostate is removed either by laparoscopic or open surgery. The surgical margin within which the prostate is removed depends on the stage of the disease.

Effective treatment of prostate cancer, regardless of the treatment method used, requires accurate visualization of the gland and its surrounding region. Accurate delineation of the prostate and appropriate visualization of the prostatic region has the potential to reduce some of the possible side-effects of the current treatment methods. These complications include urinary incontinence, impotence, and damage to the rectum and urethra (Thompson et al., 2007).

Ultrasound is the most commonly used modality for imaging of the prostate. This is due to its availability, safety and ease of use. However, ultrasound B-mode images do not always delineate the prostate reliably. As a result, prostate boundary extraction becomes a highly subjective process (Smith et al., 2007). This is observed specifically at the base of the gland, where the prostate merges with the bladder neck, and the apex, where it blends into the pelvic floor muscles. It has been shown that user segmentation variability is large in these areas (Choi et al., 2009, Tong et al., 1998).

Many attempts have been made to improve the visibility of the prostate in ultrasound B-mode images. These vary from the processing of the images (Sahba et al., 2005, Pathak et al., 2000), to the use of additional information from other modalities such as MR (Daanen et al., 2006). Another recently developed option is the use of ultrasound elastography.

Elastography (Ophir et al., 1991, Ophir et al., 1996) is a promising technique for imaging soft tissues and relies upon measuring tissue strain in response to a mechanical excitation. Indeed, when compressed by an external mechanical exciter, e.g., by the inward motion of the ultrasound transducer in the image axial direction, softer tissue will compress more than stiffer tissue, and therefore experience larger strain, which can be measured and displayed by processing the ultrasound echo data. Alternatively, tissue vibration induced by the exciter can be measured with Doppler ultrasound (Lerner et al., 1988), with larger vibrations corresponding to softer tissue. In transient elastography, the propagation of a shear wave is imaged with parallel receive ultrasound (Tanter et al., 2002). When the excitation is dynamic, the shear modulus can be estimated as a complex function of frequency, thus providing information on the viscoelastic properties of tissue. In the last few years various clinical applications of elastography have been reported in the literature. These include, but are not limited to, breast lesions (Garra et al., 1997, Sinkus et al., 2000, Kadour and Noble, 2009, Li et al., 2009), liver fibrosis (Castéra et al., 2005, Huwart et al., 2008, Yin et al., 2007), vascular vulnerable plaque (Schaar et al., 2003), elastic properties of skeletal muscle (Dresner et al., 2001), thyroid gland tumors (Lyshchik et al., 2005), and assessment of thermal tissue ablation (Wu et al., 2001) and for the detection of prostate cancer (Cochlin et al., 2002, Souchon et al., 2003, Miyagawa et al., 2009, Zhang et al., 2008, Pallwein et al., 2007, Gravas et al., 2009, Kamoi et al., 2008, Fleming et al., 2009).

Elastography has been shown to be promising in improving the visibility of the prostate gland and the cancer within it. In Egorov et al. (2006) the stress pattern on the rectal wall is directly measured with the use of a transrectal probe equipped with a pressure sensor array. Temporal and spatial changes in the stress pattern provide information for calculating prostate features such as size, shape and hardness. Phantom and in vivo results show that such a method has the potential to replace digital rectal examination (DRE). The usefulness of elastography was evaluated on 311 patients in Miyagawa et al. (2009). They showed that the sensitivity of elastography and elastography + TRUS imaging in detecting cancer (confirmed by biopsy) is higher than that of DRE or TRUS only. A higher prostate-specific antigen (PSA) level and smaller prostate volume are reported to increase the sensitivity of elastography and elastography + TRUS. However, the high frequency of false-positive elastography results and difficulty in the detection of cancer in the peripheral zone are two main problems reported in their work.

Vibro-elastography is a dynamic ultrasound elastography method (Turgay et al., 2006) which models viscoelastic properties of tissue. The approach is illustrated in Fig. 1. The technique relies on the continuous real time acquisition of unprocessed ultrasound echo data as a time series of ‘radio-frequency’ (RF) data images, while, simultaneously, tissue is externally vibrated with a broad-band mechanical excitation. A time series of tissue displacements or strain images are computed from consecutive RF data images. The tissue displacement as a function of time at a given spatial location can be viewed as the output of a linear system whose input is the motion of the exciter as a function of time. Therefore a frequency response or Transfer Function (TF) that relates the tissue motion at any spatial location with the exciter motion as a reference can be computed in the frequency domain. Alternatively, if the exciter motion is not measured, a tissue region, typically in the focal area of the ultrasound beam, can be selected as the reference. For display, the change in the transfer functions from one spatial location to another can be computed, for example as the L₂-norm of the difference between the transfer functions over the vibration frequency range.

In Salcudean et al. (2006) we introduced prostate ultrasound vibro-elastography. The signal processing used to obtain the transfer function images used in this paper are presented in detail in Salcudean et al. (2006), together with phantom images and initial in vivo prostate imaging data from three patients. We showed qualitatively that this method has the potential to improve the visibility of the prostate. In this article, our goal is to evaluate quantitatively the ability of vibro-elastography to visualize the prostate in comparison to the commonly used B-mode imaging.

Toward this goal, two groups of evaluation measures are proposed: image-based and volume-based measures. We first evaluate the quality of the images, by computing the standard measure of ‘contrast-to-noise ratio’ (CNR) in VE and B-mode images. Since the CNR measure does not effectively assess edge quality, we also analyze the quality of edges based on edge strength and edge continuity. For edge strength, we use a gradient-based edge filter and a statistical edge detector. For edge continuity, we compute the similarity of adjacent edge points using a correlation-based measure. We compare the edge evaluation results in VE vs. B-mode images. Then, in order to assess whether the delineated prostate boundary in VE images is indeed the prostate, we use volume-based measures to compare the overall shape and size of the gland as seen in VE and B-mode ultrasound images, with MR images as the gold standard.

The quality assessment criteria used in this paper are based on both standard measures (percentage volume difference, percentage volume error, CNR, gradient filter results) and on new measures, such as changes in image statistics at edges and edge continuity. These new measures had to be developed to deal with the particular situation of prostate segmentation, which requires the identification of a thin capsule within a background of relatively uniform echogenicity.

A preliminary version of the results presented in this paper, however with limited details and analyses, and fewer patients, has appeared in Salcudean et al., 2009, Mahdavi et al., 2009.

The paper is organized as follows: In Section 2 we summarize the patient data acquisition and the vibro-elastography imaging method used in this study. In Section 2.1, our data acquisition and the resulting images are described. The proposed evaluation measures of this paper are presented in Section 2.2. The results of applying these measures to the collected MRI, B-mode ultrasound and VE data are described in Section 3. Finally, conclusions and a discussion of our results are found in Section 4, which also provides avenues for future research.