| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Department of Radiology, Mayo Clinic, 200 First St., S.W., Rochester, MN
55905.
2 Present address: Department of Radiology, Veterinary Hospital, 3800 Spruce
St., University of Pennsylvania, Philadelphia, PA 19104.
3 Department of Physiology and Biophysics, Mayo Clinic, Rochester, MN
55905.
4 Department of Oncology, Mayo Clinic, Rochester, MN 55905.
Received October 19, 2001;
accepted after revision December 6, 2001.
Supported by National Institutes of Health grant CA75552.
Abstract
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
SUBJECTS AND METHODS. The prototypic elasticity imaging technique consists of a device for generating acoustic shear waves in tissue, an MR imaging—based method for imaging the propagation of these waves, and an algorithm for processing the wave images to generate quantitative images depicting tissue stiffness. After tests with tissue-simulating phantom materials and breast cancer specimens, we used the prototypic breast MR elastography technique to image six healthy women and six patients with known breast cancer.
RESULTS. Acoustic shear waves were clearly visualized in phantoms, breast cancer specimens, healthy volunteers, and patients with breast cancer. The elastograms of the tumor specimens showed focal areas of high shear stiffness. MR elastograms of healthy volunteers revealed moderately heterogeneous mechanical properties, with the shear stiffness of fibroglandular tissue measuring slightly higher than that of adipose tissue. The elastograms of patients with breast cancer showed focal areas of high shear stiffness corresponding to the locations of the known tumors. The mean shear stiffness of breast carcinoma was 418% higher than the mean value of surrounding breast tissues.
CONCLUSION. The results confirm the hypothesis that the prototypic breast MR elastographic technique can quantitatively depict the elastic properties of breast tissues in vivo and reveal high shear elasticity in known breast tumors. Further research is needed to evaluate the potential applications of MR elastography, such as detecting breast carcinoma and characterizing suspicious breast lesions.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
We wondered which currently unexploited tissue properties might be most useful for detecting and characterizing breast cancer if they could be interrogated with an imaging technique. Evidence suggests that tissue elastic properties may represent a promising target [9]. Palpation is a basic clinical examination tool that has been used for centuries to detect tumors in accessible regions of the body. Student physicians learn that the presence of a hard mass in the breast is suspicious for malignancy. The elastic modulus of breast carcinoma specimens has been reported to be five- to 20-fold higher than that of adipose—glandular tissue and fibroadenoma specimens, using laboratory mechanical testing [10] (Sarvazyan A et al., presented at the International Society for Magnetic Resonance Medicine meeting, 1998).
Motivated by these considerations, we have developed a prototypic MR imaging—based method for quantitatively imaging the elastic properties of breast tissues in vivo [11,12,13]. Our main objective was to conduct a preliminary evaluation of the feasibility of MR elastography of the breast by phantom studies with simulated breast tissues, by imaging breast tumor specimens, and by imaging healthy volunteers and a small number of patients with known breast cancer. The main hypothesis of this study was that the prototypic breast MR elastography technique can depict the elastic properties of breast tissues in vivo and show high shear elasticity in known breast tumors.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
An electromechanical driver, controlled by trigger pulses generated by the MR imaging sequence, generates shear waves in the object with displacement amplitudes of typically less than 500 µm at the surface (Fig. 1). A phase-contrast MR imaging sequence incorporates oscillating motion-sensitizing field gradients that are applied synchronously with the mechanical shear waves that traverse the imaged object [15,16,17] (Fig. 2). Cyclic motion of the material that is synchronized with the motion-sensitizing gradients causes a measurable phase shift in the received MR signal.
|
|
The phase shift caused by a propagating mechanical wave with a wave vector
k in a medium at a given frequency (1/T), in the presence of a cyclic
motion-encoding gradient, is given by:
 |
and the displacement
vector 
,the number of
gradient cycles (N), and the period of the gradient waveform
(T). The quantity, 
, is the gyromagnetic ratio,
is the spin position vector, and
is the phase relationship between the motion-encoding gradient and
mechanical excitation. High sensitivity to small-amplitude synchronous motion
can be achieved by accumulating phase shifts over multiple cycles of
mechanical excitation and the motion-sensitizing gradient waveform. Two acquisitions are made for each repetition, in an interleaved fashion, with motion-sensitizing gradients of alternating polarity. The two acquisitions are then subtracted to minimize systematic phase errors while doubling the motion sensitivity [15, 18, 19]. The motion-sensitizing gradients are applied parallel to the dominant shear motion and perpendicular to the direction of initial wave propagation. The resulting images are equivalent to snapshots of the wave field at a particular phase in the wave cycle. Typically, four to eight wave images are obtained, spanning the wave cycle.
For a simple elastic material, the shear stiffness (µ) can be estimated
by measuring the wavelength (
) of the propagating waves according to
the following equation:
 |
is the
density of the material. For our study, the density of soft tissue is assumed
to be equal to that of water, an approximation that is correct to within 7%
[20]. The wave images are processed to estimate the local wavelength at each location in the image using a technique known as local wavelength estimation (Manduca A et al., presented at the Society of Photo-Optical Instrumentation Engineers International Symposium on Medical Imaging meeting, 1996; and Knutsson H et al., presented at the Institute of Electrical and Electronics Engineers International Conference on Image Processing meeting, 1994). The estimates are derived from spatial filters that are a product of radial and directional components and can be considered oriented lognormal quadrature wavelets. This approach is equivalent to solving the lossless Helmholtz equation obtained under the local homogeneity and incompressibility assumptions. The estimates of local wavelength are then processed to compute a quantitative map of shear stiffness using the second equation.
The combined driver and radiofrequency coil developed for our study is illustrated in Figure 1. Volunteers and patients were imaged in a prone position, with the breast positioned between two contact plates, with minimal compression applied only for stabilization. Independent electromechanical drivers were used to generate superoinferior shear motion at the two plates, 180° out of phase with respect to each other to minimize bulk motion. The maximum displacement of the electromechanical driver was 150 µm. The apparatus contains phased array radiofrequency coils located behind each driver plate.
In our study, shear wave imaging was performed with a two-dimensional
gradient-echo MR elastography sequence, in the axial plane, with 3- to 5-mm
section thickness, 64 phase-encoding views, 12- to 16-cm field of view, a TR
range/TE, 100-300/28, and a flip angle of 30-45°. The motion-encoding
gradient was applied in a direction orthogonal to the plane of section. Shear
wave frequencies of 75-300 Hz were used for phantom and in vivo studies. The
electromechanical driver was activated 15-60 msec before the onset of the
first radiofrequency pulse to allow wave penetration into the tissues before
the motion sensitization waveforms were activated. Each wave image required
13-40 sec for acquisition. The sequence was repeated four to eight times with
incremented values of (phase offset between mechanical motion and
motion-encoding gradient pulses). A corresponding T1-weighted image was
acquired for anatomic mapping. The MR elastographic experiments were performed
on a Signa 1.5-T scanner (General Electric Medical Systems, Milwaukee,
WI).
Tissue-simulating breast phantoms were constructed to evaluate the driver and radiofrequency coil device. One phantom (homogeneous) was made to simulate soft, highly attenuating breastlike material using a tissue-simulating gel (7% bovine gelatin). A second homogeneous phantom consisted of stiffer material (2% agarose gel). A third breast cancer phantom consisted of the soft gel with an embedded 2.5-cm-diameter sphere of the stiffer material. The three phantoms were imaged separately.
To assess the feasibility of differentiating breast cancer from surrounding tissues on the basis of elasticity differences, fresh tissue specimens obtained at mastectomy from three patients with invasive ductal breast carcinoma were imaged. The specimens were thick sections through the tumors, averaging 11 mm in thickness and 7 cm in diameter. The specimens were embedded in blocks of tissue-simulating gel for imaging with MR elastography. Each specimen contained a stiff palpable lesion corresponding to a discrete area of decreased signal intensity, 1.5-3 cm in diameter on T1-weighted MR images. Shear waves (100-300 Hz) were applied with polarization parallel to the plane of the tissue sections.
The MR elastography method, driver, and radiofrequency coil design were then tested in 12 women: six healthy volunteers and six with known breast malignancies. All MR elastographic examinations were performed with institutional review board approval and with informed consent. The healthy volunteers ranged from 24 to 38 years, with a mean of 29 years. Breast composition ranged from mixed adipose—fibroglandular to primarily adipose. The patients ranged in age from 43 to 69 years, with a mean of 53 years. The patients participating in this study all had biopsy-proven palpable breast malignancies known to be at least 2 cm in diameter (range, 3-10 cm; mean, 5.7 cm) and located in the anterior two thirds of the breast. Uninvolved breast tissue was primarily adipose. Preliminary MR imaging was performed to ensure that the planes of section of the MR elastographic acquisitions passed through these tumors. The histologic diagnoses were infiltrating ductal carcinoma (n = 5) and infiltrating lobular carcinoma (n = 1). Shear stiffness values in the MR elastograms were evaluated by inspection and by recording the mean values in elliptic regions of interest consisting of at least 50 pixels. Testing for statistical significance among shear stiffness measurements was performed using a one-sided, paired Student's t test.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
|
|
MR elastograms of the three mastectomy specimens revealed focal areas of high shear stiffness, corresponding to the location of the palpable tumor (Fig. 4A,4B). Shear stiffness values were measured in elliptical regions of interest in the tumors and in surrounding tissues. The mean shear stiffness of the three tumors was 36 ± 5 kPa, which was 580% higher than the average measured value for the shear stiffness of surrounding adipose—fibroglandular tissues in the specimens.
|
|
Testing of the technique in the series of six female volunteers showed that the prototypic breast MR elastography driver and coil device provided sufficient mechanical energy to allow visualization of propagating shear waves throughout the breast and adjacent chest wall tissues. The volunteers reported no discomfort from the shear wave driver device. MR elastograms were computed from wave image data in all six volunteer studies. The quantitative elastograms showed modest heterogeneity in the shear stiffness of breast tissues, reporting slightly higher elasticity values in fibroglandular tissues than in adipose tissue (Fig. 5A,5B). The mean value for adipose tissue was 3.3 ± 1.9 kPa, whereas the value for fibroglandular tissue was 7.5 ± 3.6 kPa.
|
|
Shear wave imaging was successfully accomplished in all six patients with breast carcinoma. In each patient, the MR elastograms clearly showed a geographic region of high shear stiffness corresponding to the known position of the breast cancer mass (Figs. 6A,6B,6C,7A,7B,8A,8B). Other visualized breast tissues in this group of patients showed modest heterogeneity similar to that observed in the healthy volunteers. The measured shear stiffness of the tumors ranged from 18 to 94 kPa (mean, 33 kPa) (Fig. 9). The shear stiffness measurements of adipose breast tissue in the breast cancer patients ranged from 4 to 16 kPa (mean, 8 kPa). The difference in the mean values was statistically significant (p < 0.05).
|
|
|
|
|
|
|
|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The availability of a practical and reliable method for assessing the mechanical properties of breast tissues could lead to several possible applications in breast cancer. The technique may have potential as a tool for detection of malignant disease, analogous to the role of conventional palpation but diminishing its well-known limitations. Manual palpation can delineate only superficial lesions that are at least 8-10 mm in diameter. Palpation is a subjective technique with limited sensitivity for distinguishing differences in elasticity. It has been estimated, for instance, that a tumor can be detected by palpation only if its elastic modulus is 150-200% greater than surrounding tissues (Sarvazyan A et al., International Workshop on Interaction of Ultrasound with Biologic Media meeting, 1994).
Another potential role for elasticity imaging is as a tool for lesion characterization. This role is especially relevant given the emerging appreciation of the strengths and limitations of contrast-enhanced MR imaging in the diagnosis of breast cancer. Although this technique has high sensitivity (90-100%) for detecting breast cancer, exceeding that of any other imaging technology [21,22,23,24,25], its specificity may be lower, reportedly ranging between 37% and 80% [21,22,23, 26]. We speculate that an MR imaging—based technique for quantitatively interrogating the mechanical properties of breast tissue might have potential to increase diagnostic specificity in contrast-enhanced MR imaging examinations that reveal focal lesions.
The separation between the distributions of stiffness measurements for tumor and fibroglandular tissue shown in Figure 9 is less than expected on the basis of available ex vivo mechanical testing data [10] (Sarvazyan A et al., presented at the International Workshop on Interaction of Ultrasound with Biologic Media meeting, 1994). Yet, the patients with lower tumor stiffness measurements also tended to have lower shear stiffness in adjacent adipose breast tissue. Therefore, the relative shear stiffness contrast was considerable in all cases in the series, with tumors ranging from 2.3 to 6.8 times the stiffness of adjacent breast tissue. The higher shear stiffness values observed for adipose tissue adjacent to tumors, compared with measurements of adipose tissue in healthy volunteers, may be due to edema or other abnormalities caused by the adjacent tumors. Precise radiologic—pathologic correlation was not possible in this pilot study. The results obtained with the breast cancer phantom (Fig. 3A,3B) provide some insight into the possible underestimation of tumor stiffness values. Although the determinants of spatial resolution in MR elastography have not yet been properly investigated, it appears that the substantial underestimation of the shear stiffness of the focal lesion is due to a type of partial volume averaging with adjacent softer material.
We believe that there is considerable scope for improving the sharpness and fidelity of MR elastography, thereby potentially further increasing the contrast of small lesions. To expedite these pilot studies, the prototypic breast MR elastography implementation used a number of simplifying measures. Specifically, shear wave propagation was imaged only in the plane of section, rather than in three dimensions, and cyclic motion sensitization was applied only in one direction. A simplified inversion algorithm was used to calculate tissue stiffness (local wavelength estimation).
To address these limitations, we can readily modify the MR elastographic acquisition to record shear wave propagation in three dimensions and to record cyclic motion in each voxel along all three axes (Muthupillai R et al., presented at the International Society for Magnetic Resonance in Medicine meeting, 1997). A trade-off between such increased information and total imaging time must be optimized. The availability of three-dimensional, polarization-resolved wave propagation data would permit the use of more sophisticated analytic inversion algorithms, which show considerable potential to provide elasticity images with improved sharpness and fidelity [27].
It was not possible in our study to compare in vivo estimates of tissue stiffness with independent measurements because of a lack of an existing alternative technology for quantitatively measuring tissue elasticity in vivo. It is not known to what extent the mechanical properties of ex vivo tissue specimens compare with those of living tissue in situ. In any case, significant technical challenges are associated with measuring the elastic properties of semisolid biologic tissue specimens using conventional mechanical load—cell testing devices [28].
In summary, this preliminary work shows that it is feasible to use a technique combining MR imaging and acoustic technologies to quantitatively image the mechanical properties of normal breast tissues and breast malignancy in vivo. Considerable scope exists for technical improvement of the method. Further research will be needed to determine the possible performance of an optimized MR elastographic technique in terms of resolution and quantitative accuracy for depicting mechanical properties. Only then will it be possible to investigate the potential value of the technique as a tool for detecting breast cancer or characterizing suspicious breast lesions.
Acknowledgments
We gratefully acknowledge the assistance of Katherine Zahasky and David
Lomas in conducting this research.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  A M BLAMIRE The technology of MRI -- the next 10 years? Br. J. Radiol., August 1, 2008; 81(968): 601 - 617. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. K. Venkatesh, M. Yin, J. F. Glockner, N. Takahashi, P. A. Araoz, J. A. Talwalkar, and R. L. Ehman MR Elastography of Liver Tumors: Preliminary Results Am. J. Roentgenol., June 1, 2008; 190(6): 1534 - 1540. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. A. Hobson, M. Z. Kiss, T. Varghese, A. M. Sommer, M. A. Kliewer, J. A. Zagzebski, T. J. Hall, J. Harter, E. M. Hartenbach, and E. L. Madsen In Vitro Uterine Strain Imaging: Preliminary Results J. Ultrasound Med., July 1, 2007; 26(7): 899 - 908. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Medved, G. M. Newstead, H. Abe, M. A. Zamora, O. I. Olopade, and G. S. Karczmar High Spectral and Spatial Resolution MRI of Breast Lesions: Preliminary Clinical Experience Am. J. Roentgenol., January 1, 2006; 186(1): 30 - 37. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. E. Discher, P. Janmey, and Y.-l. Wang Tissue Cells Feel and Respond to the Stiffness of Their Substrate Science, November 18, 2005; 310(5751): 1139 - 1143. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
