Synthetic Aperture Imagingin Breast Ultrasound:

A Preliminary Clinical Study

Won Hwa Kim, MD, Jung Min Chang, MD, PhD, Choye Kim, BS, Jongho Park, MS,Yangmo Yoo, PhD, Woo Kyung Moon, MD, PhD, Nariya Cho, MD, PhD, Byung Ihn Choi, MD, PhD

Ac

Fr(WSeJ.Bi28Ce(2e-

ªdo

Rationale and Objectives: To compare the image quality between conventional and synthetic aperture (SA) imaging in breast ultrasound

(US).

Materials and Methods: Twenty-four patients with 31 breast lesions were included in our study. The US data were processed with SAalgorithm. For quantitative analysis, contrast-to-noise ratio (CNR) and signal-to-noise ratio (SNR) were calculated. For qualitative analysis,

conventional and SA images were reviewed by three radiologists and diagnostic preference (conspicuity, margin sharpness, and contrast)

was assessed. The radiologists also determined whether artifacts were present. Parameters were analyzed using a paired t-test, Wilcoxonsigned-rank test, and chi-square test.

Results: The mean CNRs were higher in SA images compared with conventional images (mean, 2.56 versus 2.28, P = .004). The mean

SNRs were higher in SA images compared with conventional images (31.62 versus 25.26, P < .0001). SA images were considered as being‘‘better’’ or ‘‘much better’’ in 16–23 (51.6–74.2%) lesions of total 31 lesions for conspicuity, in 17–24 (45.2%–77.4%) for margin sharpness,

and in 13–23 (41.9%–74.2%) for contrast. Significant preferences in SA imageswere demonstrated (conspicuity,P < .05 for all radiologists;

margin sharpness and contrast in two radiologists). Refraction and speckle artifactswere less frequently observed in SA images (refraction,

P < .05 for all radiologists; speckle,P < .05 for two radiologists), whereas reflection artifactsweremore frequent in SA images (P < .05 in tworadiologists).

Conclusion: SA imaging provides better image quality than conventional imaging in patients with focal breast lesions in breast US.

Key Words: Ultrasound; synthetic aperture imaging; breast lesions.

ªAUR, 2012

INTRODUCTION image quality and contribute to the inconsistent and inaccu-

Ultrasonography (US) of the breast plays a critical role

in the diagnosis and management of breast diseases.

US with a high-frequency transducer is essential

for noninvasive diagnosis of screening-detected or palpable

masses, and it allows radiologists to differentiate between

benign or malignant masses using descriptive parameters (1).

The limitations of US, however, are that it is operator-

dependent and has considerable interobserver variability in

terms of radiologists’ descriptions and assessments of breast

lesions (2). Because inherent artifacts, such as speckle, acoustic

shadowing, and acoustic enhancement, can compromise

ad Radiol 2012; 19:923–929

om the Department of Radiology, Seoul National University Hospital.H.K., J.M.C., W.K.M., N.C., B.I.C.), 28, Yongon-dong, Chongno-gu,oul, 100-744, Seoul, Korea; Departments of Electronic Engineering (C.K.,P.) and Electronic Engineering and Interdisciplinary Program of Integratedotechnology (Y.Y.), Sogang University, Seoul, Korea. Received February, 2012; accepted April 2, 2012. Supported by the Converging Researchnter Program through the Ministry of Education, Science and Technology010K001113 and 2010K001110). Address correspondence to: J.M.C.mail: imchangjm@gmail.com

AUR, 2012i:10.1016/j.acra.2012.04.005

rate interpretation of breast images, several US techniques

had been developed and are commercially used to overcome

these problems (3–5). Nevertheless, conventional US still

has its own set of problems (eg, fixed transmit focus and

limitation of penetration) that is responsible for

compromising image quality.

Synthetic aperture (SA) imaging is a US algorithm technique

that can potentially provide higher image quality than conven-

tional imaging by using the superposition of the acoustic fields

The SA technique was originally developed from synthetic

aperture radar for geological and sonar applications, but it has

been modified for use in medical imaging for several decades

(6,7). The major difference between SA and conventional

imaging is that SA imaging simultaneously produces the full

range of images at each transmission, whereas a single image

line is built sequentially to form the full range of an image in

conventional imaging. In SA imaging, the simultaneously

obtained full range of the images yields low-resolution images,

which are synthesized to create a high-resolution image

(Fig 1) (8,9). During this process, the SA imaging system

calculates the transmit focus in all points of an image with

information from each emission, which provides more

uniform resolution beyond the fixed focus depth.

923

Figure 1. Schematic illustrations of conven-

tional and synthetic aperture (SA) imaging. Inconventional imaging, an image is formed by

sequentially building up each image line

achieved at each transmission. In SAimaging, a full line of low-resolution image is

produced from each transmission, and the

low-resolution imagesare synthesizedcoher-

ently to provide a high-resolution image.

KIM ET AL Academic Radiology, Vol 19, No 8, August 2012

Although previous theoretical and experimental in vitro

studies have generally shown that the image quality of SA

imaging is superior to conventional imaging, little research

has been conducted on the effects of SA imaging in clinical

settings (8–10). One in vivo SA imaging study used

a convex array transducer in the abdomen; however, the

study population was small and limited to healthy volunteers

(6). To the best of our knowledge, SA imaging for the appli-

cation of breast US has never been investigated.

Therefore, we conducted this study to compare the image

quality between conventional and SA imaging in breast US on

the basis of quantitative and qualitative analyses.

MATERIALS AND METHODS

Patients

This study was approved by our institutional review board,

and informed consent was obtained from all patients. Between

April 2011 and May 2011, US imaging was performed on 24

patients (with a total of 31 breast lesions) whowere to undergo

surgery or US-guided biopsy (mean age 46 years; age range

27–68 years). Nine of the 31 lesions were biopsied with

a 14-gauge automated needle, and 11-gauge vacuum-

assisted biopsies were performed for two lesions. The remain-

ing 20 lesions were surgically excised. Histological diagnoses

of the lesions were made by a pathologist with 20 years of

experience in breast pathology. There were 15 benign masses

(4 fibroadenomas, 3 adenoses, 1 intraductal papilloma, and 7

fibrocystic diseases) and 16 malignant masses (3 cases of ductal

carcinoma in situ and 13 cases of invasive carcinoma).The

mean sizes of the benign and malignant lesions were 1.25

cm (0.3–2 cm) and 2.03 cm (0.5–5.4 cm), respectively.

US Unit and Imaging Acquisition

The equipment used in this study was a commercial US

scanner (SonixTouch, Ultrasonix Medical Co., Richmond,

924

BC, Canada) that was equipped with a research package using

a 7.5-MHz linear array transducer and a SonixDAQ parallel

data acquisition system (because no SA system is currently

available). For conventional imaging, a 64-element

transmit-and-receive aperture was used, and the transmit

focus point was adjusted depending on the location of the

breast lesions. Similar to conventional imaging, a 64-

element transmit-and-receive aperture was used in SA

imaging, and 32 intermediate image lines from 32

emission-and-reception events were synthesized to produce

a high-resolution image. Images in both methods were

obtained on the same plane of the lesion by two experienced

radiologists. Except for focus depth, the other parameters (eg,

central frequency, dynamic range) were identical among

patients and techniques.

Quantitative Image Analysis

For quantitative image analysis, pre-beamformed raw channel

data acquired from the US machine were transferred to

a personal computer and processed offline with in-house soft-

ware running in MATLAB (Mathworks, Natick, MA).Two

parameters were used for an objective analysis: contrast-to-

noise ratio (CNR) and signal-to-noise ratio (SNR). To obtain

a sample value of the CNR, rectangular regions of interest

(ROIs) with a size of approximately 5 mm2 were placed in

the lesion and adjacent breast parenchyma. The CNR was

calculated by the following equation:

CNR ¼ jm1� m2jffiffiffiffiffiffiffiffiffiffiffiffiffiffiffis21 þ s2

2

pwhere m1 and m2 are the mean gray intensities within a rect-

angular ROI drawn within the lesion and in the back-

ground, respectively, and s1 and s2 are the variances

(standard deviation) of the mean gray intensities within

the rectangular ROI in the lesion and the background,

Academic Radiology, Vol 19, No 8, August 2012 SYNTHETIC APERTURE IMAGING IN BREAST ULTRASOUND

respectively. Similarly, the SNR is given by the following

equation:

SNRdB ¼ 10 log

�Psignal

Pnoise

�;

where Psignal and Pnoise are the powers of the signal and the

noise, respectively.

Figure 2. The box plots show the distribution of contrast-to-noise

ratios (CNRs) and signal-to-noise ratios (SNRs) in conventional andsynthetic aperture imaging. (a) CNR values were significantly higher

in synthetic aperture (SA) imaging (P = .0004). (b) SNR values were

significantly higher in SA imaging (P < .0001). Box values are from

lower to upper quartiles, and the central lines represent the medianvalues. The whiskers extend from minimal to maximal values, and

the dots represent outliers.

Qualitative Image Analysis

The processed images were saved in a bitmap format, and all

demographic and technical information was removed from

the images. Conventional and SA images were arranged by

pairs, but the left-to-right layout of the paired images was

random. Three experienced breast imaging radiologists

(W.K.M, N.C., J.M.C, with 6–13 years of breast radiology

experience) independently reviewed and scored the subjective

preference in terms of lesion conspicuity,margin sharpness, and

lesion contrast. Lesion conspicuity was defined as the general

visibility or clarity of the lesion compared with the adjacent

structures. Margin sharpness was defined as the capacity to

properly visualize the lesion border. Contrast was defined as

the subjective difference in the echo intensities between the

lesion and the background. All radiologists were asked to select

the image that showed better quality and note the degree of

their preference (ie, equal, better, or much better). A 5-point

scale score was assigned for each paired dataset from -2 to +2,

where a score of -2 indicated that the conventional image was

much better than the SA image, a score of -1 indicated that

the conventional image was better than the SA image, a score

of 0 indicated that there was no difference between the two

images, a score of +1 indicated that the SA image was better

than the conventional image, and a score of +2 indicated that

the SA image was much better than the conventional image.

In terms of artifacts, the radiologists were asked to identify if

any artifacts (ie, reverberation, beam thickness, refraction,

posterior shadowing or enhancement, speckle or reflection)

were present in the images.

Statistical Analysis

Apaired t-testwas used to compare thevaluesofCNRandSNR

between conventional and SA imaging. A Wilcoxon signed-

rank test was used for qualitative analysis of subjective preference

score, and a chi-square test was used for comparison of the

frequency of the artifacts. Interobserver agreement between

the radiologists was calculated by weighted k statistics. k values

ranging from 0 to 0.20, 0.21–0.40, 0.41–0.60, 0.61–0.80, and

0.81–1.00 were considered to represent slight agreement, fair

agreement, moderate agreement, substantial agreement, and

almost perfect agreement, respectively. Statistical analyses were

performed with the use of SPSS (version 19.0 software; SPSS,

Chicago, IL) andMedCalc (Ver. 11.6.1. Mariakerke, Belgium).

For all statistical analyses, a two-tailed P< .05 was considered to

indicate a statistically significant difference.

RESULTS

Quantitative Image Analysis

The mean values of CNR were significantly higher in the SA

images compared with the conventional images (mean � SD

= 2.56 � 0.74 versus 2.28 � 0.81, P = .004; Fig 2a). The

mean CNR increase in the SA images was 12%. In addition,

the mean values of SNR were significantly higher in the SA

images compared with the conventional images (mean �SD = 31.62 � 3.75 vs. 25.26 � 6.75, P < .0001; Fig 2b).

The mean SNR increase in the SA images was 25%.

Qualitative Image Analysis

The subjective preference scores for SA and conventional

imaging given by the radiologists are shown in Table 1 and

Figure 3. For lesion conspicuity, the three radiologists

925

TABLE 1. Subjective Preference Scores of Synthetic Aperture Imaging Compared with Conventional Imaging

Parameters

Conventional Images

Much Better

(Score = �2)

Conventional

Images Better

(Score = �1) Same (Score = 0)

SA Images

Better

(Score = +1)

SA Images

Much Better

(Score = +2)

Average

Score P Valuey

Conspicuity

Radiologist 1 0 (0) 6 (19.4) 9 (29) 13 (41.9) 3 (9.7) 0.42 .019*

Radiologist 2 0 (0) 4 (12.9) 9 (29) 18 (58.1) 0 (0) 0.45 .003*

Radiologist 3 0 (0) 5 (16.1) 3 (9.7) 12 (38.7) 11 (35.5) 0.94 <.0001*

Margin sharpness

Radiologist 1 0 (0) 6 (19.4) 8 (25.8) 15 (48.4) 2 (6.5) 0.42 .016*

Radiologist 2 0 (0) 6 (19.4) 11 (35.5) 14 (45.2) 0 (0) 0.26 .074

Radiologist 3 0 (0) 6 (19.4) 1 (3.2) 16 (51.6) 8 (25.8) 0.84 <.0001*

Contrast

Radiologist 1 0 (0) 6 (19.4) 9 (29) 13 (41.9) 3 (9.7) 0.42 .019*

Radiologist 2 0 (0) 6 (19.4) 12 (38.7) 13 (41.9) 0 (0) 0.23 .074

Radiologist 3 0 (0) 3 (9.7) 5 (16.1) 16 (51.6) 7 (22.6) 0.87 <.0001*

Data are presented as the mean values, and the numbers in parentheses are percentages.

*P < .05, two-tailed.yWilcoxon signed-rank test.

Figure 3. The bar graphs show the frequency distribution of preference scores of synthetic aperture imaging compared with conventional

imaging for (a) lesion conspicuity, (b) margin sharpness, and (c) contrast.

KIM ET AL Academic Radiology, Vol 19, No 8, August 2012

926

Figure 4. A 45-year-old woman with inva-sive ductal carcinoma in her left upper breast.

(a) Conventional ultrasound imaging demon-

strates an ill-defined hypoechoic mass

(arrow). (b) Synthetic aperture imaging showsimprovements inmargin sharpness and lesion

contrast of the mass (arrow) and connective

tissue in the subcutaneous fat (arrowhead).

For lesion conspicuity and margin sharpness,two of the three radiologists scored (b) as

‘‘better,’’ and the other radiologist scored (b)as ‘‘much better.’’ For the contrast measure-ment, all radiologists scored (b) as ‘‘better.’’

The contrast-to-noise ratio (CNR) and

signal-to-noise ratio (SNR) were higher in

synthetic aperture imaging (CNR, 2.69 vs.2.47; SNR, 36.46 vs. 34.93).

TABLE 2. Artifacts of Synthetic Aperture Imaging versusConventional Imaging

Artifacts

Conventional

(n = 31)

Synthetic

Aperture (n = 31) P Value

Refraction

Radiologist 1 12 7 .025*

Radiologist 2 12 5 .020*

Radiologist 3 19 7 .001*

Speckle

Radiologist 1 14 5 .029*

Radiologist 2 10 3 .052

Radiologist 3 15 7 .033*

Reflection

Radiologist 1 2 14 .003*

Radiologist 2 4 24 <.0001*

Radiologist 3 5 12 .071

Posterior shadowing

Radiologist 1 3 3 1.000

Radiologist 2 14 6 .083

Radiologist 3 2 0 .317

Posterior enhancement

Radiologist 1 1 0 .317

Radiologist 2 5 3 .157

Radiologist 3 0 0 1.000

Numbers are frequencies of artifacts in the 31 image pairs.

*P < .05, two-tailed.

Academic Radiology, Vol 19, No 8, August 2012 SYNTHETIC APERTURE IMAGING IN BREAST ULTRASOUND

indicated a score of ‘‘better’’ or ‘‘much better’’ for SA images in

16 (51.6%), 18 (58.1%), and 23 (74.2%) of the 31 total image

pairs, which was compared with 6 (19.4%), 4 (12.9%), and 5

(16.1%) for the conventional image pairs (Fig 4). The mean

scores of lesion conspicuity in the SA images were 0.42,

0.45, and 0.94 for the three radiologists, and all three radiol-

ogists rated significantly different scores between SA and

conventional images (P = .019, P = .003, P < .0001).

For margin sharpness, the three radiologists indicated

a score of ‘‘better’’ or ‘‘much better’’ for SA images in 17

(54.9%), 14 (45.2%), and 24 (77.4%) of the image pairs, which

was compared with 6 (19.4%), 6 (19.4%), and 6 (19.4%) for

the conventional image pairs. The mean average scores of

SA images for margin sharpness was 0.42, 0.26, and 0.84 in

the three radiologists, and two out of the three radiologists

rated significantly different scores between SA and conven-

tional images (P = .016, P < .0001).

For the measure of contrast, the three radiologists indicated

a score of ‘‘better’’ or ‘‘much better’’ for SA images in 16

(51.6%), 13 (41.9%), and 23 (74.2%) of the image pairs, which

was compared with 6 (19.4%), 6 (19.4%), and 3 (9.7%) for the

conventional image pairs. The mean scores of the SA images

for contrast were 0.42, 0.23, and 0.87 for the three radiolo-

gists, and two of the three radiologists rated significantly

different scores between SA and conventional images

(P = .019, P < .0001).

Refractions, which are caused by differences in sound

speeds between tissues and are usually seen as shadows around

the edges of the lesion or Cooper’s ligament, were less

frequently observed in the SA images compared with the

conventional images, and the differences were significant for

all three radiologists (conventional versus SA: 38.7%, 38.7%,

and 61.3% versus 22.6%, 16.1%, and 22.6%, P < .05 for all

three radiologists; Table 2, Fig 5). Speckles, which give noisy

and granular artifacts, were also less frequently found in the SA

images compared with the conventional images (conventional

versus SA: 45.2%, 32.3%, and 48.4% versus 16.1%, 9.7%, and

22.6%, P < .05 in two radiologists and P = .052 in one radi-

ologist). In contrast, reflections, which were seen as strong

echo signals around the tissue interface (eg, connective tissue

within fat lobules), were more frequently seen in the SA

images (conventional versus SA: 6.5%, 12.9%, and 16.1%

versus 45.2%, 77.4%, and 38.7%, P < .05 in two radiologists,

and P = .071 in one radiologist).

In overall, there were moderate interobserver agreements

between the radiologists in the grading of subjective prefer-

ence; kappa scores are summarized in Table 3.

DISCUSSION

In the present study, we assessed quantitative and qualitative

parameters for a new SA imaging technique for the

927

Figure 5. A 68-year-old woman with invasive

ducal carcinoma in her left upper outer breast.(a) Conventional ultrasound imaging demon-

strates a hypoechoic round mass (arrow),

which is difficult to distinguish from thesurrounding breast parenchyma. The refrac-

tion artifacts (arrowhead) are severe. (b) In

synthetic aperture imaging, the mass and the

surrounding normal tissues are more conspic-uous and there is a sharper margin and higher

contrast of the mass. In addition, a minimal

degree of refraction artifact is visible. For

lesion conspicuity, margin sharpness andcontrast, two of the three radiologists scored

(b) as ‘‘much better,’’ and the other radiologist

scored (b) as ‘‘better.’’ The contrast-to-noise

ratio (CNR) and signal-to-noise ratio (SNR)were higher in synthetic aperture imaging

(CNR, 2.53 vs. 2.19; SNR, 34.57 vs. 23.98).

TABLE 3. Levels of Agreement for Interobserver Variability forConventional and Synthetic Aperture Images

Parameters

Radiologists

1 and 2

Radiologists

1, 2, and 3

Radiologists

2 and 3

Conspicuity 0.59 � 0.13 0.65 � 0.1 0.56 � 0.08

Margin sharpness 0.46 � 0.15 0.44 � 0.01 0.44 � 0.1

Contrast 0.68 � 0.07 0.57 � 0.11 0.44 � 0.09

Artifact 0.32 � 0.06 0.50 � 0.06 0.27 � 0.06

Data are presented as the mean � standard deviation. A kappa

value of 0 to 0.20 was considered to be slight agreement, 0.21–

0.40 was considered fair agreement, 0.41–0.60 was considered

moderate agreement 0.61–0.80 was considered substantial agree-

ment, and 0.81–1.00 was considered almost perfect agreement.

KIM ET AL Academic Radiology, Vol 19, No 8, August 2012

visualization of breast lesions, and we compared the SA tech-

nique with conventional US. We noted higher values for

CNR and SNR in the SA images, and we found a higher

preference in terms of lesion conspicuity, margin sharpness,

and contrast in the SA images. In addition, artifacts (eg,

speckle and refraction artifacts) were less frequently observed

in the SA imaging. We found that the SA images were higher

quality than the conventional images, which was consistent

with previous in vitro and in vivo studies (6,10,11). Our

findings, together with previous theoretical and

experimental reports, suggest that SA imaging can offer

a promising alternative to conventional techniques in

medical US.

Conventional US imaging systems have several technical

constraints because of the use of a fixed transmit focus,

and high-resolution images are only achieved in the region

of the focus depth. Although the conventional system may

transmit with multiple transmit focusing along the scan

line to further improve resolution, the frame rate is conse-

quently reduced by the number of transmit focus. In addi-

tion, the attempts of current systems to increase the

number of active elements in the beamformer and increase

928

the central frequency of the transducer have the potential

to lower the frame rate to maintain the same level of image

quality.

In contrast with conventional imaging, the SA imaging

system is capable of calculating the transmit focus for all

points in the image (ie, dynamic focus in both transmit

and receive); thus, high resolution is obtained throughout

the image. This advantage in SA imaging can enhance lesion

visualization in deeper locations in the tissue. Several

previous studies have also shown an improvement in pene-

tration depth with the use of SA (6,9). In addition, a high

frame rate is still maintained in SA imaging because only

a sparse set of emissions is needed to create a full image

(12). SA imaging results in uniform improvements in image

resolution at all points of the image, which might also

provide better image quality in recently introduced auto-

mated whole-breast US systems. Indeed, automated US

using conventional imaging might suffer from the predeter-

mined focus, which would obscure the objects beyond the

focus depth.

Even with high-resolution images in breast US imaging,

artifacts (eg, shadowing and refractions caused by normal

anatomic structures) might interfere with fine visualization

of target lesions (13). In our study, we found that the

frequency of refraction and speckle artifacts was lower in the

SA imaging compared with the conventional imaging. The

reason might be caused by the fact that SA imaging averages

the various artifact patterns by synthesizing the low-

resolution images; however, reflection artifacts from specular

reflectors were accentuated in the SA imaging, which agreed

with a previous reports by Pederson et al (6,14). Interestingly,

most of the reflection artifacts in our images were found just

above the rib.

The present study had several limitations. First, this study

consisted of a small number of breast masses and did not

include normal breast tissues and benign breast cysts. Further

studies with a larger number of patients are needed to

Academic Radiology, Vol 19, No 8, August 2012 SYNTHETIC APERTURE IMAGING IN BREAST ULTRASOUND

generalize the feasibility of SA imaging in clinical applications.

Second, we did not evaluate the effect of image quality by

motion and penetration depth. In several previous reports,

SA imaging has shown a substantial amount of susceptibility

to patient motion (6,15). In this study, we could not assess

the effect of motion on image quality of SA imaging

because the equipment we used was not real-time ultrasound.

In regard to penetration depth in SA imaging, we also

observed a tendency toward improved penetration, which

resulted in a more uniformed resolution, even in deeper

tissues beyond the focal lesion; however, this was not proved

in this study. Future studies examining the penetration depth

in SA imaging should be investigated in various clinical

settings. Third, image acquisition was performed using rela-

tively low-frequency (7.5-MHz) probe that is related to

higher rate of artifact than usual images because this study

was a preclinical study and the commercial real-time equip-

ment with high-frequency probe for SA imaging is not

currently available. For this reason, scanning using high-

frequency probe was preceded before the scanning using

low-frequency probe for research purpose that allowed us to

detect small (< 5 mm) lesions that may not be plausible with

low-frequency probe. Therefore, the effect of SA imaging

on detection rate especially for small lesions could not be

investigated. Further research on implementation of SA

imaging using more relevant commercial equipment that

provides real-time scanning and a high-frequency probe is

needed.

In conclusion, SA imaging provides better image quality

than conventional imaging in breast US. Although our study

was limited to a small number of patients, further clinical

investigations should be performed to validate the medical

applications of SA imaging.

REFERENCES

1. Stavros AT, Thickman D, Rapp CL, et al. Solid breast nodules: use of

sonography to distinguish between benign and malignant lesions. Radi-

ology 1995; 196:123.

2. Baker JA, Kornguth P, Soo MS, et al. Sonography of solid breast lesions:

observer variability of lesion description and assessment. AJR Am J

Roentgenol 1999; 172:1621.

3. Szopinski KT, Pajk AM, Wysocki M, et al. Tissue harmonic imaging: utility

in breast sonography. J Ultrasound Med 2003; 22:479.

4. Mesurolle B, Helou T, El-Khoury M, et al. Tissue harmonic imaging,

frequency compound imaging, and conventional imaging: use and benefit

in breast sonography. J Ultrasound Med 2007; 26:1041.

5. Huber S, Wagner M, Medl M, et al. Real-time spatial compound imaging in

breast ultrasound. Ultrasound Med Biol 2002; 28:155–163.

6. Pedersen MH, Gammelmark KL, Jensen JA. In-vivo evaluation of convex

array synthetic aperture imaging. Ultrasound Med Biol 2007; 33:37–47.

7. O’Donnell M. Coded excitation system for improving the penetration of

real-time phased-array imaging systems. IEEE Trans Ultrason Ferroelec-

trics Frequency Control 1992; 39:341–351.

8. Jensen JA, Nikolov SI, Gammelmark KL, et al. Synthetic aperture ultra-

sound imaging. Ultrasonics 2006; 44:e5–e15.

9. Bae MH, Jeong MK. A study of synthetic-aperture imaging with virtual

source elements in B-mode ultrasound imaging systems. IEEE Trans Ul-

trason Ferroelectrics Frequency Control 2000; 47:1510–1519.

10. Gammelmark KL, Jensen JA. Multielement synthetic transmit aperture

imaging using temporal encoding. IEEE Trans Med Imaging 2003; 22:

552–563.

11. Kim K, Hwang J, Jeong J, et al. An efficient motion estimation and

compensation method for ultrasound synthetic aperture imaging. Ultrason

Imaging 2002; 24:81–99.

12. Nikolov S, Jensen JA. Synthetic aperture tissue and flow ultrasound

imaging. In: Synthetic aperture tissue and flow ultrasound imaging.

Denmark: Technical University of Denmark, Department of Electrical

Engineering, Biomedical Engineering; 2001.

13. Baker JA, Soo MS, Rosen EL. Artifacts and pitfalls in sonographic imaging

of the breast. AJR Am J Roentgenol 2001; 176:1261.

14. Pedersen MH, Misaridis TX, Jensen JA. Clinical evaluation of chirp-coded

excitation in medical ultrasound. Ultrasound Med Biol 2003; 29:895–905.

15. Trahey GE, Nock LF. Synthetic receive aperture imaging with phase

correction for motion and for tissue inhomogeneities. II. Effects of and

correction for motion. IEEE Trans Ultrason Ferroelectrics Frequency

Control 1992; 39:496–501.

929