Dynamic frame pairing in real-time freehand elastography
21 May 2014-IEEE Transactions on Ultrasonics Ferroelectrics and Frequency Control (IEEE)-Vol. 61, Iss: 6, pp 979-985
TL;DR: A one-prediction-one- correction method that dynamically selects pre- and post- compression frames to form an elastogram, based on the applied axial strain level is described.
Abstract: Quasi-static ultrasound (US) elastography is now a well-established technique that involves acquiring US (RF/ envelope) signals from an imaging plane before and after a small quasi-static compression to form axial strain elastograms (ASE). The image quality of the ASEs is a function of the applied axial strain. This relationship was extensively investigated and formalized in terms of strain filter in the literature. Most of the work in elastography formed elastograms by choosing pre- and post-compression frames separated by a desired compression strain. Although this approach is feasible in simulations and in vitro/in vivo experiments that involve controlled compression, it has been a challenge to do this during freehand compression in real time. In this work, we describe a one-prediction-one- correction method that dynamically selects pre- and post- compression frames to form an elastogram, based on the applied axial strain level. We validate the method using controlled compression experiments on phantoms and compare the performance of the dynamic frame pairing method against successive-frame pairing method in terms of the contrast-to-noise ratio (CNRe). Further, we demonstrate the advantages of the new method with the help of freehand acquired data from phantom experiments and in vivo breast data. The results demonstrate that the frame-pairing identified by the dynamic method matched the frame pairing that was designed to yield an applied axial strain of ˜1%. The CNRe obtained by the traditional approach varied from as low as ˜5 to as high as ˜25, depending on the choice of skip number and compression rate. However, the dynamic frame pairing method provided elastograms with a CNRe that was consistently around ˜20, irrespective of the compression rate. The results from analysis of 22 in vivo breast data demonstrated that the dynamic pairing method generated elastograms such that the frame-average axial strain (FAAS) of each frame in the cineloop is consistently ˜1% (0.011 ± 0.001).
Citations
More filters
TL;DR: This work names the proposed method tOtal Variation Regularization and WINDow-based time delay estimation (OVERWIND) and shows that it is robust to signal decorrelation and generates sharp displacement and strain maps for simulated, experimental phantom and in-vivo data.
Abstract: A major challenge of free-hand palpation ultrasound elastography (USE) is estimating the displacement of RF samples between pre- and post-compressed RF data. The problem of displacement estimation is ill-posed since the displacement of one sample by itself cannot be uniquely calculated. To resolve this problem, two categories of methods have emerged. The first category assumes that the displacement of samples within a small window surrounding the reference sample is constant. The second class imposes smoothness regularization and optimizes an energy function. Herein, we propose a novel method that combines both approaches, and as such, is more robust to noise. The second contribution of this work is the introduction of the L1 norm as the regularization term in our cost function, which is often referred to as the total variation (TV) regularization. Compared to previous work that used the L2 norm regularization, optimization of the new cost function is more challenging. However, the advantages of using the L1 norm are twofold. First, it leads to substantial improvement in the sharpness of displacement estimates. Second, to optimize the cost function with the L1 norm regularization, we use an iterative method that further increases the robustness. We name our proposed method tOtal Variation Regularization and WINDow-based time delay estimation (OVERWIND) and show that it is robust to signal decorrelation and generates sharp displacement and strain maps for simulated, experimental phantom and in-vivo data. In particular, OVERWIND improves strain contrast-to-noise ratio (CNR) by 27.26%, 144.05%, and 49.90% on average in simulation, phantom, and in-vivo data, respectively, compared to our recent Global Ultrasound Elastography (GLUE) method.
44 citations
TL;DR: This study attempts to provide a systematic investigation of a 2-D cross-correlation-based USI method in a theoretical framework and finds a compromise between optimal kernel sizes and estimation accuracy of various strain components was required in complex kinematic scenarios.
Abstract: Estimation of tissue motion in the lateral direction remains a major challenge in 2-D ultrasound strain imaging (USI). Although various methodologies have been proposed to improve the accuracy of estimation of in-plane displacements and strains, the fundamental limitations of 2-D USI and how to choose optimal algorithmic parameters in various tissue deformation paradigms to retrieve the full strain tensor of acceptable accuracy are scattered throughout the literature. Thus, this study attempts to provide a systematic investigation of a 2-D cross-correlation-based USI method in a theoretical framework. Our previously developed cross-correlation-based USI method was revisited, and additional estimation strategies were incorporated to improve in-plane displacement and strain estimation. The performance of the presented method using different matching kernel sizes (axial: from 1λ to 14λ, where λ = wavelength; lateral: from 1 to 13 pitches) and two data formats (radiofrequency and envelope) in various kinematic scenarios (normal, shear or hybrid deformation) was investigated using Field II simulations, in which coherent plane wave compounding with 64 steered angles was realized. For radiofrequency-based USI, smaller axial and larger lateral kernel sizes were preferred in scenarios with normal strains, whereas larger kernel sizes along the shearing direction and smaller ones orthogonal to the shearing direction were more suitable in scenarios with shear strains. For envelope-based USI, in contrast, the kernel size requirement was relatively relaxed. A compromise between optimal kernel sizes and estimation accuracy of various strain components was required in complex kinematic scenarios. These practical strategies for accurate motion estimation using 2-D cross-correlation-based USI were further tested in a tissue-mimicking phantom under quasi-static compression and in a preliminary in vivo examination of a normal human median nerve at the wrist during active finger motion.
24 citations
TL;DR: A novel technique to estimate tissue displacement in quasi-static elastography is introduced and two approaches are proposed to get third dimension, which substantially outperforms NCC using simulation, phantom and in-vivo experiments.
24 citations
Cites background from "Dynamic frame pairing in real-time ..."
...Low cost and ease to availability are two advantages of free-hand palpation ultrasound elastography [28, 29]....
[...]
TL;DR: The results indicate that the automatic frame selection method described here may provide an objective way to select a representative frame while saving time for the radiologist.
Abstract: This study was aimed at developing a method for automatically selecting a few representative frames from several hundred axial-shear strain elastogram frames typically obtained during freehand compression elastography of the breast in vivo. This may also alleviate some inter-observer variations that arise at least partly because of differences in selection of representative frames from a cine loop for evaluation and feature extraction. In addition to the correlation coefficient and frame-average axial strain that have been previously used as quality indicators for axial strain elastograms, we incorporated the angle of compression, which has unique effects on axial-shear strain elastogram interpretation. These identified quality factors were computed for every frame in the elastographic cine loop. The algorithm identifies the section having N contiguous frames (N = 10) that possess the highest cumulative quality scores from the cine loop as the one containing representative frames. Data for total of 40 biopsy-proven malignant or benign breast lesions in vivo were part of this study. The performance of the automated algorithm was evaluated by comparing its selection against that by trained radiologists. The observer- identified frame that consisted of a sonogram, axial strain elastogram and axial-shear strain elastogram was compared with the respective images in the frames of the algorithm-identified section using cross-correlation as a similarity measure. It was observed that there was, on average (∼standard deviation), 82.2% (∼2.2%), 83.4% (∼3.8%) and 78.4% (∼3.6%) correlation between corresponding images of the observer-selected and algorithm-selected frames, respectively. The results indicate that the automatic frame selection method described here may provide an objective way to select a representative frame while saving time for the radiologist. Furthermore, the frame quality metric described and used here can be displayed in real time as feedback to guide elastographic data acquisition and for training purposes.
7 citations
Cites background or methods from "Dynamic frame pairing in real-time ..."
...Xia and Thittai (2015) found that deviation from uni-axial compression can potentially lead to artifacts in ASSE and used a threshold of 10 in their analysis on different in vivo data sets....
[...]
...Thus, they emphasized the importance of including angle of compression as a quality indicator in freehand compression elastography for correct interpretation of ASSE (Xia and Thittai 2015)....
[...]
...Real-Time elastic imaging software developed by Xia and Thittai (Xia et al. 2014) was used for displacement tracking and strain estimation....
[...]
...Angle of compression was estimated from the axial displacement map using the method proposed by Xia and Thittai (2015)....
[...]
...Xia et al. (2014) proposed a ‘‘one prediction–one correction’’ method that selects pre-compression frames dynamically, such that the pairing frames have a desired axial compression strain level between them (typically set to 1%)....
[...]
TL;DR: This study utilizes ASSE from tissue mimicking phantom experiments and freehand - acquired in vivo breast lesion data to analyze the effect of segmentation threshold on ASSE feature value, specifically, the "fill-in" feature that was introduced recently.
5 citations
Cites methods from "Dynamic frame pairing in real-time ..."
...We recently described an approach to dynamically pair RF frames such that the frameaverage axial strain is around 1% [27]....
[...]
References
More filters
TL;DR: Initial results of several phantom and excised animal tissue experiments are reported which demonstrate the ability of this technique to quantitatively image strain and elastic modulus distributions with good resolution, sensitivity and with diminished speckle.
3,636 citations
"Dynamic frame pairing in real-time ..." refers methods in this paper
...U (Us) elastography is now well established as a technique to image the stiffness variation in soft tissues [1]....
[...]
TL;DR: Elastography has the potential to be useful in the evaluation of areas of shadowing on the sonogram and also may be helpful in the distinction of benign from malignant masses.
Abstract: PURPOSE: To determine the appearance of various breast lesions on elastograms and to explore the potential of elastography in the diagnosis of breast lesions. MATERIALS AND METHODS: A total of 46 breast lesions were examined with elastography. Patients underwent biopsy or aspiration of all lesions, revealing 15 fibroadenomas, 12 carcinomas, six fibrocystic nodules, and 13 other lesions. The elastogram was generated from radio-frequency data collected with use of a 5-MHz linear-array transducer. The elastogram and corresponding sonogram were evaluated by a single observer for lesion visualization, relative brightness, and margin definition and regularity. The sizes of the lesions at each imaging examination and at biopsy were recorded and compared. RESULTS: Softer tissues such as fat appear as bright areas on elastograms. Firm tissues, including parenchyma, cancers, and other masses, appear darker. The cancers were statistically significantly darker than fibroadenomas (P < .005) and substantially larger on...
980 citations
"Dynamic frame pairing in real-time ..." refers methods in this paper
...Pre- and post-compression frames separated by a fixed number (referred to as skip frame number in this paper) are generally used to generate an elastogram in quasistatic elastography [17]....
[...]
01 Mar 1999
TL;DR: The strain filter formalism and its utility in understanding the noise performance of the elastographic process is given, as well as its use for various image improvements.
Abstract: The basic principles of using sonographic techniques for imaging the elastic properties of tissues are described, with particular emphasis on elastography. After some preliminaries that describe some basic tissue stiffness measurements and some contrast transfer limitations of strain images are presented, four types of elastograms are described, which include axial strain, lateral strain, modulus and Poisson's ratio elastograms. The strain filter formalism and its utility in understanding the noise performance of the elastographic process is then given, as well as its use for various image improvements. After discussing some main classes of elastographic artefacts, the paper concludes with recent results of tissue elastography in vitro and in vivo.
837 citations
"Dynamic frame pairing in real-time ..." refers background or methods in this paper
...The factors determining the statistical image quality of asE are well established in literature [2]....
[...]
...When an elastogram depicts the axial strain values, it is referred to as an axial strain elastogram (asE) [2]....
[...]
TL;DR: In this article, a device and procedure for measuring elastic properties of gelatin for elasticity imaging (elastography) was described. And the measured compression forces were comparable to results obtained from finite element analysis when linear elastic media are assumed.
Abstract: Acoustic and mechanical properties are reported for gelatin materials used to construct tissue-like phantoms for elasticity imaging (elastography). A device and procedure for measuring elastic properties are described. The measured compression forces were comparable to results obtained from finite element analysis when linear elastic media are assumed. Also measured were the stress relaxation, temporal stability, and melting point of the materials. Aldehyde concentration was used to increase the stiffness of the gelatin by controlling the amount of collagen cross-linking. A broad range of tissue-like elastic properties was achieved with these materials, although gels continued to stiffen for several weeks. The precision for elastic modulus measurements ranged from less than 0.1% for 100 kPa samples to 8.9% for soft (<10 kPa), sticky samples.
511 citations
TL;DR: The preliminary data suggest that the strain image sequences for various breast pathologies are unique, and that a comparison of the lesion area measured in B-mode vs. strain images appears to be a sensitive criterion for separating invasive ductal carcinoma from cyst and fibroadenoma.
Abstract: Previous experience with laboratory fixtures and off-line processing of elasticity data showed that problems occurring in data acquisition often resulted in poor elasticity image quality. A system for real-time estimation and display of tissue elastic properties using a clinical ultrasonic imaging system has been developed. A brief description of that system and the initial clinical tests of that system are reported. Experience with real-time freehand elasticity imaging shows that images with high contrast-to-noise ratios are consistently obtained. Images of breast lesions were acquired with freehand palpation using standard linear-array ultrasound (US) transducers. Results in volunteer patients show that high-quality elasticity images are easily obtained from in vivo breast studies. The key element to successful scanning is real-time visual feedback of B-mode and strain images that guide the patient positioning and compression direction. Results show that individual images of axial strain in tissues can be quite misleading, and that a "movie loop" of side-by-side B-mode and strain images provides significantly more information. Our preliminary data suggest that the strain image sequences for various breast pathologies are unique. For example, strain images of fibroadenomas lose contrast with increasing precompression, but those of invasive ductal carcinoma have high negative contrast (dark relative to "normal" tissue) for a wide range of precompression. In addition, a comparison of the lesion area measured in B-mode vs. strain images, for a representative image from the sequence, appears to be a sensitive criterion for separating invasive ductal carcinoma from cyst and fibroadenoma.
410 citations
"Dynamic frame pairing in real-time ..." refers background or methods in this paper
...2% seemed to produce good quality elastograms for freehand acquired in vivo breast data [16]– [18]....
[...]
...For example, applying a maximum of 3% axial compression with a typical 1-Hz compression-relaxation cycle [16], [18] the inter-frame compression may be as small as 0....
[...]
...Typically, the operator is recommended to have a compression-relaxation cycle of about 1 Hz [16], [18]....
[...]
...We set the desired Faas to be ~1% based on the reported findings that compression strain around this value produces good quality elastograms for freehand-acquired in vivo breast data [16]–[18]....
[...]