Analysis of linear, area and volume distortion in 3D ultrasound imaging

Calibration for position tracking of swept motor 3-D ultrasound

Tracking the position and orientation of a 3-D ultrasound transducer has many clinical applications. Tracking requires calibration to find the transformation between the tracking sensor and the ultrasound coordinates. Typically the set of image slice data are scan converted to a Cartesian volume using assumed motor geometry and a single transformation to the sensor. We propose, instead, the calibration of individual slices using a 2-D calibration technique. A best fit to a subset of slices is performed to decrease data collection time compared with that for calibration of all slices, and to reduce the influence of random errors in individual calibrations. We compare our technique with four scan conversion-based techniques: 2-D N-wire on the center slice, N-wire using a 3-D volume, N-wire using a 3-D volume including the edge points and a new closed-form planar method using a 3-D volume. The proposed multi-slice technique produced the smallest point reconstruction error (0.82 mm using a tracked stylus).

Preoperative assessment of extrathyroidal extension of papillary thyroid carcinoma: comparison of 2- and 3-dimensional sonography

OBJECTIVES

The purpose of this study was to evaluate the diagnostic performance and interobserver agreement of 2-dimensional (2D) and 3-dimensional (3D) sonography for evaluating extrathyroidal extension of papillary thyroid carcinoma.

METHODS

A total of 79 papillary thyroid carcinomas in 79 patients who underwent both 2D and 3D sonography for preoperative staging of papillary thyroid carcinoma were included. When the lesion was abutting on the thyroid capsule on 2D sonography, 3D sonography was performed. Three radiologists reviewed 3 data sets: 2D sonography, 3D sonography, and a combined set of both for tumor staging. The diagnostic performance, including sensitivity, specificity, positive predictive value, negative predictive value, and accuracy, for extrathyroidal extension was analyzed. Interobserver agreement of the 3 radiologists was assessed with κ values.

RESULTS

The overall accuracy rates for 2D sonography, 3D sonography, and the combined set in predicting extrathyroidal extension were 60.8%, 66.2%, and 67.9%, respectively. The accuracy of the combined set was significantly higher than that of 2D sonography (P = .016). The interobserver agreement of the 3 reviewers was fair (κ = 0.33) for 2D sonography and moderate for 3D sonography (κ = 0.46) and the combined set (κ = 0.49).

CONCLUSIONS

Adding 3D sonography to 2D sonography could improve the accuracy and interobserver agreement for predicting extrathyroidal extension of papillary thyroid carcinoma.

Differentiating benign from malignant thyroid nodules: comparison of 2- and 3- dimensional sonography

OBJECTIVES

To compare the diagnostic performance of radiologists and to determine interobserver and intraobserver variability with regard to differentiation of benign and malignant thyroid nodules using prospectively obtained 2-dimensional (2D) and 3-dimensional (3D) sonograms.

METHODS

This study had Institutional Review Board approval, and the requirement for patient informed consent was waived. Conventional 2D and 3D sonograms were obtained from 82 patients (age range, 20-77 years; mean age, 51 years) with 91 thyroid nodules (15 cancers, 13 indeterminate, and 63 benign lesions) before diagnostic fine-needle aspiration. Three radiologists reviewed stored 2D and 3D images for internal content, shape, margin, echogenicity, echo texture, and the presence of calcification and estimated the level of suspicion as to the probability of malignancy according to known sonographic criteria. The diagnostic performance of 2D images was compared with that of 3D images.

RESULTS

For all readers, interpretation using 3D images was more sensitive and specific than that using 2D images for diagnosis of malignant thyroid nodules, with the exception of specificity for reader 1. However, differences were not statistically significant (P > .05). Area under the receiver operating characteristic curve values were 0.83 for 2D images and 0.92 for 3D images for reader 1; 0.78 for 2D images and 0.89 for 3D images for reader 2; and 0.89 for 2D images and 0.93 for 3D images for reader 3. Interobserver agreement between the 3 radiologists for differentiation of benign and malignant thyroid nodules was better for 3D images (κ = 0.49) than for 2D images (κ = 0.15). Intraobserver variability for nodule descriptions and assessments using 3D and 2D images was fair to moderate.

CONCLUSION

The performance of radiologists and interobserver and intraobserver agreement for characterization of thyroid nodules were better when 3D sonograms were used than when 2D sonograms were used.

Anniversary paper: evolution of ultrasound physics and the role of medical physicists and the AAPM and its journal in that evolution

Ultrasound has been the greatest imaging modality worldwide for many years by equipment purchase value and by number of machines and examinations. It is becoming increasingly the front end imaging modality; serving often as an extension of the physician's fingers. We believe that at the other extreme, high-end systems will continue to compete with all other imaging modalities in imaging departments to be the method of choice for various applications, particularly where safety and cost are paramount. Therapeutic ultrasound, in addition to the physiotherapy practiced for many decades, is just coming into its own as a major tool in the long progression to less invasive interventional treatment. The physics of medical ultrasound has evolved over many fronts throughout its history. For this reason, a topical review, rather than a primarily chronological one is presented. A brief review of medical ultrasound imaging and therapy is presented, with an emphasis on the contributions of medical physicists, the American Association of Physicists in Medicine (AAPM) and its publications, particularly its journal Medical Physics. The AAPM and Medical Physics have contributed substantially to training of physicists and engineers, medical practitioners, technologists, and the public.

Stereotactic mammography imaging combined with 3D US imaging for image guided breast biopsy

Stereotactic X-ray mammography (SM) and ultrasound (US) guidance are both commonly used for breast biopsy. While SM provides three-dimensional (3D) targeting information and US provides real-time guidance, both have limitations. SM is a long and uncomfortable procedure and the US guided procedure is inherently two dimensional (2D), requiring a skilled physician for both safety and accuracy. The authors developed a 3D US-guided biopsy system to be integrated with, and to supplement SM imaging. Their goal is to be able to biopsy a larger percentage of suspicious masses using US, by clarifying ambiguous structures with SM imaging. Features from SM and US guided biopsy were combined, including breast stabilization, a confined needle trajectory, and dual modality imaging. The 3D US guided biopsy system uses a 7.5 MHz breast probe and is mounted on an upright SM machine for preprocedural imaging. Intraprocedural targeting and guidance was achieved with real-time 2D and near real-time 3D US imaging. Postbiopsy 3D US imaging allowed for confirmation that the needle was penetrating the target. The authors evaluated 3D US-guided biopsy accuracy of their system using test phantoms. To use mammographic imaging information, they registered the SM and 3D US coordinate systems. The 3D positions of targets identified in the SM images were determined with a target localization error (TLE) of 0.49 mm. The z component (x-ray tube to image) of the TLE dominated with a TLEz of 0.47 mm. The SM system was then registered to 3D US, with a fiducial registration error (FRE) and target registration error (TRE) of 0.82 and 0.92 mm, respectively. Analysis of the FRE and TRE components showed that these errors were dominated by inaccuracies in the z component with a FREz of 0.76 mm and a TREz of 0.85 mm. A stereotactic mammography and 3D US guided breast biopsy system should include breast compression for stability and safety and dual modality imaging for target localization. The system will provide preprocedural x-ray mammography information in the form of SM imaging along with real-time US imaging for needle guidance to a target. 3D US imaging will also be available for targeting, guidance, and biopsy verification immediately postbiopsy.

Fast prostate segmentation in 3D TRUS images based on continuity constraint using an autoregressive model

In this article a new slice-based 3D prostate segmentation method based on a continuity constraint, implemented as an autoregressive (AR) model is described. In order to decrease the propagated segmentation error produced by the slice-based 3D segmentation method, a continuity constraint was imposed in the prostate segmentation algorithm. A 3D ultrasound image was segmented using the slice-based segmentation method. Then, a cross-sectional profile of the resulting contours was obtained by intersecting the 2D segmented contours with a coronal plane passing through the midpoint of the manually identified rotational axis, which is considered to be the approximate center of the prostate. On the coronal cross-sectional plane, these intersections form a set of radial lines directed from the center of the prostate. The lengths of these radial lines were smoothed using an AR model. Slice-based 3D segmentations were performed in the clockwise and in the anticlockwise directions, where clockwise and anticlockwise are defined with respect to the propagation directions on the coronal view. This resulted in two different segmentations for each 2D slice. For each pair of unmatched segments, in which the distance between the contour generated clockwise and that generated anticlockwise was greater than 4 mm, a method was used to select the optimal contour. Experiments performed using 3D prostate ultrasound images of nine patients demonstrated that the proposed method produced accurate 3D prostate boundaries without manual editing. The average distance between the proposed method and manual segmentation was 1.29 mm. The average intraobserver coefficient of variation (i.e., the standard deviation divided by the average volume) of the boundaries segmented by the proposed method was 1.6%. The average segmentation time of a 352 x 379 x 704 image on a Pentium IV 2.8 GHz PC was 10 s.

Detection of bladder tumors with 3-dimensional sonography and virtual sonographic cystoscopy

OBJECTIVE

Bladder tumors are among the most common types of malignant neoplasms of the urinary tract. The purpose of this study was to evaluate the potential value of 3-dimensional (3D) sonography and sonographic cystoscopy in detection of bladder tumors.

METHODS

Thirty-one patients with suspected or known bladder tumors were included this study. All patients underwent 3D sonography and conventional cystoscopy within 15 days. The number, size, location, and morphologic features of the lesions were evaluated on gray scale, 3D virtual, and multiplanar reconstruction images obtained from the patients. The results of 3D sonographic cystoscopy were compared with the findings from conventional cystoscopy, which was considered the reference standard.

RESULTS

Twenty-eight (90.3%) of 31 3D virtual sonographic cystoscopic studies had good or excellent image quality. Conventional cystoscopy revealed 47 lesions in 22 of 28 patients; 3D sonographic virtual cystoscopy showed 41 (87.2%) of 47 lesions. Three-dimensional virtual sonography alone had sensitivity of 96.2%, specificity of 70.6%, a positive predictive value of 93.9%, and a negative predictive value of 80% for tumor detection. The combination of gray scale sonography, multiplanar reconstruction, and 3D virtual sonography had sensitivity of 96.4%, specificity of 88.8%, a positive predictive value of 97.6%, and a negative predictive value of 84.2% for tumor detection.

CONCLUSIONS

Three-dimensional sonography is a promising alternative noninvasive technique for use in detection of bladder tumors, their localization, and perivesical spreading. The location, size, and morphologic features of the tumors shown on 3D sonography agreed well with the findings of conventional cystoscopy.

Spectral clustering for TRUS images

Background

Identifying the location and the volume of the prostate is important for ultrasound-guided prostate brachytherapy. Prostate volume is also important for prostate cancer diagnosis. Manual outlining of the prostate border is able to determine the prostate volume accurately, however, it is time consuming and tedious. Therefore, a number of investigations have been devoted to designing algorithms that are suitable for segmenting the prostate boundary in ultrasound images. The most popular method is the deformable model (snakes), a method that involves designing an energy function and then optimizing this function. The snakes algorithm usually requires either an initial contour or some points on the prostate boundary to be estimated close enough to the original boundary which is considered a drawback to this powerful method.

Methods

The proposed spectral clustering segmentation algorithm is built on a totally different foundation that doesn't involve any function design or optimization. It also doesn't need any contour or any points on the boundary to be estimated. The proposed algorithm depends mainly on graph theory techniques.

Results

Spectral clustering is used in this paper for both prostate gland segmentation from the background and internal gland segmentation. The obtained segmented images were compared to the expert radiologist segmented images. The proposed algorithm obtained excellent gland segmentation results with 93% average overlap areas. It is also able to internally segment the gland where the segmentation showed consistency with the cancerous regions identified by the expert radiologist.

Conclusion

The proposed spectral clustering segmentation algorithm obtained fast excellent estimates that can give rough prostate volume and location as well as internal gland segmentation without any user interaction.

A comparative study of 2D and 3D ultrasonography for evaluation of solid breast masses

OBJECTIVE

To compare image quality and diagnostic accuracy of 2D with 3D ultrasonography in solid breast masses.

METHODS AND MATERIAL

To rate image quality, two radiologists compared lesion contrast and characterization of 507 solid breast masses in 2D and 3D ultrasonography and then graded the 3D imaging in 3-point scale. To characterize the masses, the same radiologists rated the examination for clarity of margin, posterior acoustic feature, and clustered microcalcifications within a mass. In addition, the masses were assigned BI-RADS categories as proposed by the American College of Radiology, criteria using just ultrasonographic features. In the 202 pathologically confirmed cases, sensitivity, specificity, positive predictive value, negative predictive value, and false negative rate for diagnosis of breast cancer in 2D and 3D ultrasonography were assessed. Image quality and diagnostic accuracy were further evaluated according to the size of the masses.

RESULTS

Two observers rated 3D imaging superior to 2D imaging in terms of lesion contrast and characterization of the masses. Especially, superiority of 3D ultrasonography in terms of image quality was increasing in more than 10 mm sized masses. However, diagnostic accuracy including sensitivity, specificity, positive predictive value, negative predictive value, and false negative rate for diagnosis of breast cancer of 3D imaging was not different from 2D imaging.

CONCLUSION

In spite of superior image quality on 3D ultrasonography, it does not provide additional benefits to diagnostic accuracy for diagnosis of breast cancer.

Prostate segmentation algorithm using dyadic wavelet transform and discrete dynamic contour

Knowing the location and the volume of the prostate is important for ultrasound-guided prostate brachytherapy, a commonly used prostate cancer treatment method. The prostate boundary must be segmented before a dose plan can be obtained. However, manual segmentation is arduous and time consuming. This paper introduces a semi-automatic segmentation algorithm based on the dyadic wavelet transform (DWT) and the discrete dynamic contour (DDC). A spline interpolation method is used to determine the initial contour based on four user-defined initial points. The DDC model then refines the initial contour based on the approximate coefficients and the wavelet coefficients generated using the DWT. The DDC model is executed under two settings. The coefficients used in these two settings are derived using smoothing functions with different sizes. A selection rule is used to choose the best contour based on the contours produced in these two settings. The accuracy of the final contour produced by the proposed algorithm is evaluated by comparing it with the manual contour outlined by an expert observer. A total of 114 2D TRUS images taken for six different patients scheduled for brachytherapy were segmented using the proposed algorithm. The average difference between the contour segmented using the proposed algorithm and the manually outlined contour is less than 3 pixels.

The development and evaluation of a three-dimensional ultrasound-guided breast biopsy apparatus

We have designed a prototype three-dimensional ultrasound guidance (3D USB) apparatus to improve the breast biopsy procedure. Features from stereotactic mammography and free-hand US-guided biopsy have been combined with 3D US imaging. This breast biopsy apparatus accurately guides a needle into position for the sampling of target tissue. We have evaluated this apparatus in three stages. First, by testing the placement accuracy of a needle in a tissue mimicking phantom. Second, with tissue mimicking phantoms that had embedded lesions for biopsy. Finally, by comparison to free-hand US-guided biopsy, using chicken breast phantoms. The first two stages of evaluation quantified the mechanical biases in the 3D USB apparatus. Compensating for these, a 96% success rate in targeting 3.2 mm "lesions" in chicken breast phantoms was achieved when using the 3D USB apparatus. The expert radiologists performing biopsies with free-hand US guidance achieved a 94.5% success rate. This has proven an equivalence between our apparatus, operated by non-experts, and free-hand biopsy performed by expert radiologists, for 3.2 mm lesions in vitro, with a 95% confidence.

Three-dimensional ultrasound imaging

Two-dimensional viewing of three-dimensional anatomy by conventional ultrasound limits our ability to quantify and visualize a number of diseases and is partly responsible for the reported variability in diagnosis. Over the past two decades, many investigators have addressed this limitation by developing three-dimensional imaging techniques, including three-dimensional ultrasound imaging. In this paper we describe the development of a number of three-dimensional ultrasound imaging systems that make use of B mode, color Doppler, and power Doppler. In these systems, the conventional ultrasound transducer is scanned mechanically or by a freehand technique. The ultrasound images are digitized and then reconstructed into a three-dimensional volume, which can be viewed and manipulated interactively by the diagnostician with a variety of image-rendering techniques. These developments as well as future trends are discussed with regard to their applications and limitations.

Three-dimensional ultrasound imaging

Ultrasound is an inexpensive and widely used imaging modality for the diagnosis and staging of a number of diseases. In the past two decades, it has benefited from major advances in technology and has become an indispensable imaging modality, due to its flexibility and non-invasive character. In the last decade, research investigators and commercial companies have further advanced ultrasound imaging with the development of 3D ultrasound. This new imaging approach is rapidly achieving widespread use with numerous applications. The major reason for the increase in the use of 3D ultrasound is related to the limitations of 2D viewing of 3D anatomy, using conventional ultrasound. This occurs because: (a) Conventional ultrasound images are 2D, yet the anatomy is 3D, hence the diagnostician must integrate multiple images in his mind. This practice is inefficient, and may lead to variability and incorrect diagnoses. (b) The 2D ultrasound image represents a thin plane at some arbitrary angle in the body. It is difficult to localize the image plane and reproduce it at a later time for follow-up studies. In this review article we describe how 3D ultrasound imaging overcomes these limitations. Specifically, we describe the developments of a number of 3D ultrasound imaging systems using mechanical, free-hand and 2D array scanning techniques. Reconstruction and viewing methods of the 3D images are described with specific examples. Since 3D ultrasound is used to quantify the volume of organs and pathology, the sources of errors in the reconstruction techniques as well as formulae relating design specification to geometric errors are provided. Finally, methods to measure organ volume from the 3D ultrasound images and sources of errors are described.

Variability and accuracy of measurements of prostate brachytherapy seed position in vitro using three-dimensional ultrasound: an intra- and inter-observer study

This paper is a step in investigating whether three-dimensional (3D) ultrasound can be used intraoperatively to replace Computed Tomography (CT) for localization of brachytherapy seeds. In order to quantify the accuracy and variability of seed localization without introducing effects due to tissues, we first report our results with test phantoms. An inter- and intra-observer study was performed to assess the variability of 2 3D ultrasound scan acquisition methods: Tilt 3D scanning and pull-back 3D scanning. Seven observers measured the positions of gold seed markers in an agar phantom twice in each of the three orthogonal image planes. An analysis of variance (ANOVA) was performed to determine the intra- and inter-observer standard errors of measurement (SEM) and the minimum detectable changes in marker position (deltap). Average intra- and inter-observer SEMs for the tilt scan 3D image were 0.36 and 0.40 mm, respectively. Measurements of the pull-back scan 3D image yielded average intra- and inter-observer SEM of 0.46 and 0.49 mm, respectively. A paired difference analysis showed that the lower SEM for the tilt 3D scan image were statistically significant at a significance level of alpha= 0.05. The accuracy of the US measurements was tested by determining marker coordinates from CT images of the phantom in a stereotactic head frame. CT coordinates were matched to the ultrasound (US) coordinates by means of an affine transform. Average matching errors in x, y, and z were 0.02, 0.10, and -0.02 mm, respectively.

Analysis of geometrical distortion and statistical variance in length, area, and volume in a linearly scanned 3-D ultrasound image

A linearly scanned three-dimensional (3-D) ultrasound imaging system is considered. The transducer array is initially oriented along the x axis and aimed in the y direction. After being tilted by an angle theta about the x axis, and then swiveled by an angle phi about the y axis, it is translated in the z direction, in steps of size d, to acquire a series of parallel two-dimendional (2-D) images. From these, the 3-D image is reconstructed, using the nominal values of the parameters (phi, theta, d). Thus, any systematic or random errors in these, relative to their actual values (phi0, theta0, d0), will respectively cause distortions or variances in length, area, and volume in the reconstructed 3-D image, relative to the 3-D object. Here, we analyze these effects. Compact linear approximations are derived for the relative distortions as functions of the parameter errors, and hence, for the relative variances as functions of the parameter variances. Also, exact matrix formulas for the relative distortions are derived for arbitrary values of (phi, theta, d) and (phi0, theta0, d0). These were numerically compared to the linear approximations and to measurements from simulated 3-D images of a cubical object and real 3-D images of a wire phantom. In every case tested, the theory was confirmed within experimental error (0.5%).

Clinical utility of three-dimensional US

Three-dimensional (3D) ultrasonography (US) is rapidly gaining popularity as it moves out of the research environment and into the clinical setting. This modality offers several distinct advantages over conventional US, including 3D image reconstruction with a single pass of the US beam, virtually unlimited viewing perspectives; accurate assessment of long-term effects of treatment; and more accurate, repeatable evaluation of anatomic structures and disease entities. In obstetric imaging, 3D US provides a novel perspective on the fetal anatomy, makes anomalies easier to recognize, facilitates maternal-fetal bonding, and helps families better understand fetal abnormalities. Three-dimensional pelvic US allows volume data sets to be acquired with both transvaginal and transabdominal probes. Viewing multiple 3D power Doppler US images in a fast cine loop has proved useful in angiographic applications. Three-dimensional prostate US can help make accurate volume assessments for dosimetry planning or for estimating prostate-specific antigen levels. In breast imaging, 3D US has the capacity to demonstrate lesion margins and topography, thereby helping differentiate benign from malignant masses. Three-dimensional US can also help determine the need for biopsy and help facilitate needle localization and guidance during biopsy. With recent advances in computer technology and display techniques, 3D US will likely play an increasingly important role in medicine.