(Radiographics. 2000;20:1127-1135.)
© RSNA, 2000
IMAGING & THERAPEUTIC TECHNOLOGY |
Comparison of Tissue Harmonic Imaging with Conventional US in Abdominal Disease1
Sabina Choudhry, MD,
Brian Gorman, MD,
J. William Charboneau, MD,
Donald J. Tradup, RDMS,
Rebecca J. Beck, RDMS,
James M. Kofler, MS and
Debra S. Groth
1 From the Department of Diagnostic Radiology, Mayo Clinic and Mayo Foundation, 200 First St SW, Rochester, MN 55905. Recipient of a Magna cum Laude award and an Excellence in Design award at the 1998 RSNA scientific assembly. Received June 29, 1999; revision requested July 26; final revision received February 16, 2000; accepted March 10. Supported in part by an institutional grant from the Mayo Research Fund. Address correspondence to B.G.
 |
Abstract
|
|---|
Harmonic waves are generated from nonlinear distortion of an acoustic signal as an ultrasound wave insonates tissues in the body. These beams are integer multiples of a fundamental transmitted frequency. Potential advantages of harmonic imaging include improved axial resolution due to higher frequencies and better lateral resolution due to narrower beams. Decreased noise from side lobes improves signal-to-noise ratios and reduces artifacts. Deleterious effects of the body wall are also reduced. The authors prospectively studied ultrasonographic (US) findings in 100 adult patients with 202 abdominal lesions by comparing harmonic US images with conventional US images. The results were subjected to statistical analysis. Harmonic imaging was superior to conventional US in regard to lesion visibility and diagnostic confidence. Harmonic imaging was particularly useful for depicting cystic lesions and those containing echogenic tissues such as fat, calcium, or air. In patients with a body mass index of 30 or more, harmonic imaging was clearly better for lesion visibility and confidence of diagnosis. The authors recommend routine use of harmonic imaging for abdominal US studies in all adult patients.
Index Terms: Abdomen, US, 70.1298, 70.12989 • Ultrasound (US), harmonic study, 70.12989
|
Introduction
|
|---|
Tissue harmonic ultrasonography (US) is based on the phenomenon of nonlinear distortion of an acoustic signal as it travels through the body (1,2) (Fig 1). Imaging begins with insonation of tissue with ultrasound waves of a specific transmitted frequency. Harmonic waves are generated within the tissue and build up with depth to a point of maximal intensity before they decrease because of attenuation (1–4). In comparison, conventional ultrasound waves are generated at the surface of the transducer and progressively decrease in intensity as they traverse the body (Fig 2). Harmonic wave frequencies are higher integer multiples of the transmitted frequency, much like the overtones of a musical note. Current technology uses only the second harmonic (twice the transmitted frequency) for imaging. The processed image is formed with use of the harmonic-frequency bandwidth in the received signal after the transmitted frequency spectrum is filtered out (5–7) (Fig 3).
View larger version (17K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 1. Time sequence illustrates the generation of harmonic frequencies. Time step 1 contains a wave of single frequency. As the wave travels into the tissue (time step 2), it becomes distorted. Additional frequency components are created that are integer multiples of the initial frequency. These components are called "harmonic frequencies." As the wave continues to travel, it becomes highly distorted (time step 3) and very rich in harmonic frequencies. Note that the harmonic frequencies are created and accumulate as the wave travels through the tissue. Although many harmonic frequencies are produced with nonlinear wave propagation, the amplitudes of the higher harmonics are extremely small. Therefore, current technology uses only the second harmonic (2f), which is twice the nominal transmitted frequency.
|
|
View larger version (9K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 3. Frequency spectrum of the transmitted and received waves. The fundamental wave (f) is generated at the transducer surface and attenuates linearly as it is transmitted through the body. The harmonic wave (2f) is generated as the fundamental wave travels through the body. The harmonic wave increases exponentially in intensity before attenuating within the deeper tissues. Harmonic imaging uses only the harmonic frequency in the echo signal received by filtering out the transmitted frequency spectrum in the signal.
|
|
Potential advantages of harmonic imaging include improved axial resolution due to shorter wavelength, better lateral resolution due to improved focusing with higher frequencies, and less artifact than with conventional US. Reduced artifact in harmonic imaging is due to the relatively small amplitude of the harmonic waves, which reduces detection of echoes from multiple scattering events. Additionally, side lobes are less likely to occur and degrade the image. Furthermore, harmonic beams are produced beyond the body wall, thereby reducing the defocusing effect of the body wall. This is a theoretic advantage in imaging of obese patients (1,2).
The purpose of this study was to evaluate the potential benefits of tissue harmonic imaging compared with conventional US in clinical practice. We evaluated cystic, solid, or echogenic lesions containing highly reflective tissues (such as fat, calcium, or air) and diffuse abnormalities in echo pattern of the organ involved. Harmonic and conventional US images were compared by two radiologists on the basis of lesion visibility and confidence of diagnosis and associations with age, sex, body mass index (BMI), organ of involvement, characteristics of the lesions, and technical difficulty of the examination.
|
Methods
|
|---|
Subjects
From September 1997 through April 1998 at our institution, 100 adult patients (57 men and 43 women; age range, 30–87 years; mean age, 62.3 years) underwent abdominal US examinations at which lesions were observed by two designated sonographers (D.J.T., R.J.B.). The sonographers were US technologists certified by the American Registry of Diagnostic Medical Sonographers, each with about 10 years of scanning experience. Informed consent was obtained from all patients.
US Studies
Imaging was performed with use of a transducer with a single-vector format (model 2V4) and a commercially available US scanner (Sequoia 512; Acuson, Mountain View, Calif). The sonographer obtained similar views of abdominal lesions with both harmonic imaging and conventional US. Scanning parameters were optimized independently for each method, and all images were obtained with use of the same focal zone. The sonographers used an external video with cine playback mode to obtain identical images in two standard orthogonal planes. Images were obtained with the two methods in random sequence to facilitate their masking for the observers, and they were displayed side by side on film by using a single laser printer (model 2180; Eastman Kodak, Rochester, NY).
Harmonic images were acquired at a transmitting frequency of 2.0 MHz and a receiving harmonic bandwidth of 4.0 MHz. Conventional US images were obtained at a frequency of 3.5 MHz, which is a frequency used commonly at abdominal imaging in adults. The harmonic and conventional US modes were switched by means of a toggle switch on the scanner control panel.
Lesions
In the 100 patients, 202 abdominal lesions were evaluated. The diagnosis for each lesion was confirmed by means of pathologic correlation or comparison with pertinent imaging studies or results at long-term follow-up of more than 12 months. Of the 202 lesions, 76 (38%) were in the liver, 67 (33%) in the kidneys, 30 (15%) in the biliary tract including the gallbladder, 12 (6%) in the blood vessels, and 17 (8%) in other abdominal organs, including the pancreas and spleen. Lesions were classified by the radiologists into four groups based on their US characteristics regardless of the organ of involvement: cystic, solid, or echogenic lesions containing highly reflective tissues such as fat, calcium, or air; and diffuse abnormalities in echo pattern of the organs involved.
|
Review and Analysis of Harmonic and Conventional US Images
|
|---|
A mask was placed over the technical parameters of the paired images, and they were analyzed independently by two radiologists (B.G., J.W.C.). They evaluated each image for lesion visibility and confidence of diagnosis.
Lesion visibility was graded on a five-point scale: -2, conventional US markedly better than harmonic imaging; -1, conventional slightly better than harmonic; 0, equal visibility; 1, harmonic slightly better than conventional; 2, harmonic markedly better than conventional. The rating scale was chosen so positive values indicated harmonic imaging was preferred over conventional US.
Confidence of diagnosis was similarly graded on a five-point scale: -2, confidence markedly higher with conventional US; -1, confidence slightly better with conventional US; 0, confidence equal with both methods; 1, confidence slightly higher with harmonic imaging; 2, confidence markedly higher with harmonic imaging.
The ratings for lesion visibility and confidence of diagnosis were then averaged for both radiologists and classified according to whether harmonic imaging was better than conventional US. Associations between lesion visibility and confidence of diagnosis with age, sex, BMI, organ of involvement, characteristics of the lesions, and technical difficulty of the examination were assessed with Wilcoxon rank sum tests. Ratings for each radiologist and comparisons between radiologists were evaluated with the Wilcoxon signed rank test or the Wilcoxon matched pair signed rank test (7).
|
Results
|
|---|
Lesion Visibility
Both radiologists rated harmonic imaging better than conventional US for lesion visibility (P < .001, Wilcoxon signed rank test). Mean lesion visibility scores were 0.53 (SD, 0.96) for radiologist 1 and 0.55 (SD, 0.70) for radiologist 2. Median lesion visibility scores were 1.00 for both radiologists (P = .78, Wilcoxon signed rank test). The mean difference between the radiologists was -0.01 (SD, 0.97). The median difference was 0.00. This difference in lesion visibility ratings between the radiologists was not statistically significant (P = .78, Wilcoxon signed rank test).
As there was no significant difference between the radiologists, lesion visibility ratings were averaged across the two observers for subsequent analyses. Both radiologists rated lesion visibility as better with harmonic US for 126 (62%) lesions, equal with both techniques for 52 (26%), and better with conventional US for 24 (12%). They rated visibility of a small polyp on the anterior wall of the gallbladder as better with harmonic US (Fig 4).
View larger version (128K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 4a. Gallbladder polyp. Small polyp on the anterior surface of the gallbladder is seen on the harmonic image (a) with marked reduction of noise in the gallbladder lumen compared with that on the conventional US image (b).
|
|
View larger version (129K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 4b. Gallbladder polyp. Small polyp on the anterior surface of the gallbladder is seen on the harmonic image (a) with marked reduction of noise in the gallbladder lumen compared with that on the conventional US image (b).
|
|
The organ of involvement was marginally associated with lesion visibility (P = .07), indicating that harmonic US may not perform as well in lesions in the liver as compared with other organs. The underlying characteristic of the lesion was significantly associated with lesion visibility (P = .029) (Table 1). Harmonic images depicted cystic lesions (Fig 5) and lesions with highly reflective tissues (Fig 6) more clearly than did conventional US scans. Differences in lesion visibility between harmonic and conventional US images were less marked for solid masses (Fig 7) and diffuse abnormalities in the echotexture of solid organs (Fig 8).
View larger version (121K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 5a. Renal cyst. (a) Because of marked clearing of noise, the harmonic image demonstrates septation and a mural nodule within the cyst. (b) Conventional US image shows artifact from noise.
|
|
View larger version (112K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 5b. Renal cyst. (a) Because of marked clearing of noise, the harmonic image demonstrates septation and a mural nodule within the cyst. (b) Conventional US image shows artifact from noise.
|
|
View larger version (114K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 6a. Cholelithiasis. Multiple small gallstones in the dependent portion of the gallbladder are depicted more clearly on the harmonic image (a) than on the conventional US image (b).
|
|
View larger version (102K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 6b. Cholelithiasis. Multiple small gallstones in the dependent portion of the gallbladder are depicted more clearly on the harmonic image (a) than on the conventional US image (b).
|
|
The radiologists found that results with harmonic imaging were better than those with conventional US as the BMI of the patient increased. Patients were considered obese when their BMI was more than 30 (8). Median BMI in patients with lesions scored as "harmonic better than conventional" was 28.3 and in the remaining patients was 25.3 (P = .017). In obese patients, lesion visibility was rated "harmonic better than conventional" for 43 (73%) lesions, "harmonic the same as conventional" for 13 (22%), and "harmonic inferior to conventional" for three (5%) (P = .017) (Table 2). An exophytic mass arising from the lower pole of the left kidney in an obese patient was more visible on the harmonic image (Fig 9). There were no statistically significant associations between lesion visibility and the age or sex of the patients.
View larger version (126K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 9a. Renal carcinoma in an obese patient. Exophytic solid mass in the lower pole of the left kidney is seen on the harmonic image (a) but not on the conventional US image (b), which demonstrates the benefits of harmonic imaging in obese patients.
|
|
View larger version (139K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 9b. Renal carcinoma in an obese patient. Exophytic solid mass in the lower pole of the left kidney is seen on the harmonic image (a) but not on the conventional US image (b), which demonstrates the benefits of harmonic imaging in obese patients.
|
|
Confidence of Diagnosis
Both radiologists were more confident of their diagnosis with harmonic images (P < .001, Wilcoxon signed rank test). Mean scores for the confidence of diagnosis were 0.18 (SD, 0.56) for radiologist 1 and 0.24 (SD, 0.51) for radiologist 2. Median scores were 0.00 for both radiologists. The mean difference between the radiologists was 0.06 (SD, 0.61). The median difference was 0.00. Differences in confidence of diagnosis between the two radiologists were not statistically significant (P = .19, Wilcoxon signed rank test).
Confidence of diagnosis ratings were averaged for the two radiologists, and confidence was higher with harmonic imaging for 64 (32%) lesions, equal for 127 (63%), and higher with conventional US for 11 (5%). In a patient with hepatic metastases, confidence of diagnosis was improved with the harmonic image (Fig 10). In another patient with obstructive jaundice, the harmonic image clearly depicted the obstructing lesion (Fig 11).
View larger version (105K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 10a. Hepatic metastases. In the liver, multiple isoechoic masses with peripheral hypoechoic halos are depicted more clearly on the harmonic image (a) than on the conventional image (b). The clarity in a improved diagnostic confidence.
|
|
View larger version (112K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 10b. Hepatic metastases. In the liver, multiple isoechoic masses with peripheral hypoechoic halos are depicted more clearly on the harmonic image (a) than on the conventional image (b). The clarity in a improved diagnostic confidence.
|
|
View larger version (117K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 11a. Choledocholithiasis with liver metastases. Both the harmonic (a) and conventional (b) US images show a dilated common duct and multiple liver metastases. The obstructing stone in the common duct was detected on only a, and diagnostic confidence was improved.
|
|
View larger version (119K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 11b. Choledocholithiasis with liver metastases. Both the harmonic (a) and conventional (b) US images show a dilated common duct and multiple liver metastases. The obstructing stone in the common duct was detected on only a, and diagnostic confidence was improved.
|
|
For both radiologists, confidence of diagnosis increased with harmonic imaging as the BMI increased. Median BMI in patients with lesions rated as "harmonic better than conventional" was 29.2 and in the remaining patients was 26.2 (P < .001). Table 3 reports the effect of obesity (BMI, 
30) on the confidence of diagnosis with either method: Confidence was higher with harmonic imaging compared with conventional US for 27 (46%) lesions, equal for 29 (49%), and higher with conventional US than with harmonic imaging for three (5%) (P = .002).
When ratings for diagnostic confidence were averaged for all lesions for each patient, BMI was still strongly associated with a higher confidence with harmonic imaging. Median BMI for patients with lesions scored as "harmonic better than conventional" was 29.3 and for the remaining patients was 25.8 (P = .003). There were no statistically significant associations between confidence of diagnosis and age, sex, US characteristics of the lesion, or technical difficulty of the examination.
|
Discussion
|
|---|
Results showed the superiority of harmonic imaging over conventional US for lesion visibility, with harmonic imaging being preferred over conventional US for 62% of lesions. The lesions were clearer and better defined with harmonic imaging, which thereby improved the detection of subtle lesions. The use of higher effective frequencies for imaging yielded an improvement in axial resolution and thereby improved visualization of smaller objects along the axis of the beam. Higher frequency ultrasound waves produced narrower beams as a result of better focusing of the waves, which improved lateral resolution. An improvement in lateral resolution provides better visualization of smaller lesions that exist side by side within the field of view. In conventional US, optimal resolution was achieved by using a higher frequency. This, at times, compromised adequate penetration, which necessitated use of a lower frequency. Harmonic imaging provided better resolution by using a higher frequency for imaging without compromising depth penetration.
Harmonic waves were generated within the body as tissues were insonated by ultrasound waves of specific transmitted frequencies. These waves were produced from nonlinear distortion of the transmitted frequencies as they passed through body tissues. Harmonic imaging theoretically improved signal-to-noise ratios by reducing noise from side lobe artifact in the near field and echo detection from multiple scattering events (9). This reduced noise was most likely responsible for the superiority of harmonic imaging over conventional US in the visualization of cystic lesions and in the improved confidence of diagnosis for most cystic lesions. Harmonic imaging helped differentiate cysts from hypoechoic solid masses. There was better clarity of the contents of cysts, which helped differentiate artifact from true echogenic structures within cysts, such as debris, hemorrhage, and septations.
Harmonic imaging was superior to conventional US in the visualization of lesions containing highly reflective tissues such as fat, calcium, and air. Better definition of the posterior acoustic shadows in calcifications and lesions containing air also contributed to improved lesion visibility.
The two radiologists believed that diagnostic confidence was equal with both methods in most (63%) lesions. They were more confident of their diagnosis with harmonic imaging over conventional US, however, in about 32% of lesions. In contrast, conventional US was preferred over harmonic imaging in only 5% of lesions. The radiologists preferred harmonic imaging over conventional US six times as often when a preference for either method was expressed.
Harmonic imaging was especially helpful in obese patients because of reduction in the deleterious effects of the body wall. The intensity of harmonic waves generated depends on the nonlinearity coefficient (B/A coefficient) of the tissue insonated, which is a measure of the capability of a tissue to support a nonlinear wave. Of all body tissues, those containing a high proportion of fat had the highest nonlinearity coefficients, which increased the intensity of harmonic waves generated, other conditions being equal. This increased intensity was one of the factors that improved lesion visibility in obese patients, whose body walls typically contain a high proportion of fat. Generation of harmonic waves within the tissues of the body provided ultrasound beams with higher effective frequency, which allowed imaging without significant attenuation or distortion of these waves in the body wall, thereby providing better depth penetration. Harmonic imaging improved lesion visibility and diagnostic confidence in obese patients. Improvement in lesion visibility correlated strongly with increasing BMI, which demonstrates that harmonic imaging is a useful method for imaging of such patients.
|
Limitations
|
|---|
Harmonic imaging was better for demonstrating most abdominal lesions, but at times it did not have any additional benefit over conventional US for demonstrating a lesion, improving diagnostic confidence, or altering clinical management. Such cases included the minority of lesions that were better demonstrated with conventional US. As shown in Table 1, differences in lesion visibility based on US characteristics were less marked for solid masses (Figs 12, 13) and diffuse abnormalities in echotexture (Fig 8). For patients with diffuse fatty infiltration in the liver, penetration of the ultrasound beam was better on conventional US images. These observations were not analyzed statistically because of the small sample size.
View larger version (101K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 12a. Right renal mass in an obese patient. Solid mass in the lower pole of the right kidney is seen equally well on the harmonic (a) and conventional (b) US images. Clinical management was not altered because a did not contribute information in addition to that in b.
|
|
View larger version (101K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 12b. Right renal mass in an obese patient. Solid mass in the lower pole of the right kidney is seen equally well on the harmonic (a) and conventional (b) US images. Clinical management was not altered because a did not contribute information in addition to that in b.
|
|
View larger version (99K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 13a. Liver abscess. The harmonic (a) and conventional (b) US images depicted the partially cystic abscess equally well. No additional information was contributed in b that altered the diagnosis.
|
|
View larger version (98K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 13b. Liver abscess. The harmonic (a) and conventional (b) US images depicted the partially cystic abscess equally well. No additional information was contributed in b that altered the diagnosis.
|
|
|
Conclusions
|
|---|
With use of the equipment in our study, we found significantly improved image quality with harmonic imaging. This improvement was most apparent in obese patients and in lesions that were cystic or contained highly reflective tissues such as calcium, gas, or fat. At our institution, therefore, we use harmonic imaging as the initial method for all abdominal US examinations.
|
Footnotes
|
|---|
Abbreviation: BMI = body mass index
|
References
|
|---|
-
Ward B, Baker AC, Humphrey VF. Nonlinear propagation applied to the improvement of resolution in diagnostic medical ultrasound. J Acoust Soc Am 1997; 101:143154.
-
Starritt HC, Duck FA, Hawkins AJ, Humphrey VF. The development of harmonic distortion in pulsed finite-amplitude ultrasound passing through liver. Phys Med Biol 1986; 31:14011409.
-
Starritt HC, Perkins MA, Duck FA, Humphrey VF. Evidence for ultrasonic finite-amplitude distortion in muscle using medical equipment. J Acoust Soc Am 1985; 77:302306.
-
Muir TG. Nonlinear effects in acoustic imaging. Acoust Imag 1980; 9:93109.
-
Ragavendra N, Chen H, Powers JE, et al. Harmonic imaging of porcine intraovarian arteries using sonographic contrast medium: initial findings. Ultrasound Obstet Gynecol 1997; 9:266270.
-
Wu JY, Shung KK. Nonlinear energy exchange among harmonic modes and its applications to nonlinear imaging. J Acoust Soc Am 1990; 88:28522858.
-
Siegel S. Nonparametric statistics for the behavioral sciences New York, NY: McGraw-Hill, 1956; 6875.
-
National Institutes of Health. First federal obesity clinical guidelines released; National Institutes of Health, National Heart, Lung, and Blood Institute, June 17, 1998. Available at: http://www.nhlbi.nih.gov/nhlbi/news/obere14f.htm. Accessed June 25, 1999..
-
Shapiro RS, Wagreich J, Parsons RB, Stancato-Pasik A, Yeh HC, Lao R. Tissue harmonic imaging sonography: evaluation of image quality compared with conventional sonography. AJR Am J Roentgenol 1998; 171:12031206.
Top
Abstract
Introduction
