- Email: [email protected]
Magnetic resonance elastography: Feasibility of liver stiffness measurements in healthy volunteers at 3 T
Clinical Radiology 67 (2012) 258e262
Contents lists available at SciVerse ScienceDirect
Clinical Radiology journal homepage: www.clinicalradiologyonline.net
Magnetic resonance elastography: Feasibility of liver stiffness measurements in healthy volunteers at 3 Tq L. Mannelli a, b, *, E. Godfrey a, M.J. Graves a, A.J. Patterson a, P. Beddy a, D. Bowden a, I. Joubert a, A.N. Priest a, D.J. Lomas a a b
Department of Radiology, Addenbrooke’s Hospital and University of Cambridge, Cambridge, UK Department of Radiology, University of Washington, Seattle, WA, USA
article in formation Article history: Received 29 June 2011 Received in revised form 18 August 2011 Accepted 22 August 2011
AIM: To demonstrate the feasibility of obtaining liver stiffness measurements with magnetic resonance elastography (MRE) at 3 T in normal healthy volunteers using the same technique that has been successfully applied at 1.5 T. METHODS AND MATERIALS: The study was approved by the local ethics committee and written informed consent was obtained from all volunteers. Eleven volunteers (mean age 35 9 years) with no history of gastrointestinal, hepatobiliary, or cardiovascular disease were recruited. The magnetic resonance imaging (MRI) protocol included a gradient echo-based MRE sequence using a 60 Hz pneumatic excitation. The MRE images were processed using a local frequency estimation inversion algorithm to provide quantitative stiffness maps. Adequate image quality was assessed subjectively by demonstrating the presence of visible propagating waves within the liver parenchyma underlying the driver location. Liver stiffness values were obtained using manually placed regions of interest (ROI) outlining the liver margins on the gradient echo wave images, which were then mapped onto the corresponding stiffness image. The mean stiffness values from two adjacent sections were recorded. RESULTS: Eleven volunteers underwent MRE. The quality of the MRE images was adequate in all the volunteers. The mean liver stiffness for the group was 2.3 0.38 kPa (ranging from 1.7e2.8 kPa). CONCLUSIONS: This preliminary work using MRE at 3 T in healthy volunteers demonstrates the feasibility of liver stiffness evaluation at 3 T without modification of the approach used at 1.5 T. Adequate image quality and normal MRE values were obtained in all volunteers. The obtained stiffness values were in the range of those reported for healthy volunteers in previous studies at 1.5 T. There was good interobserver reproducibility in the stiffness measurements. Ó 2011 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved.
Introduction Currently all conventional imaging techniques are unable to detect early fibrosis and distinguish the fibrosis grades q Accepted as an abstract at ISMRM 2010, Stockholm, Sweden.
* Guarantor and correspondent: L. Mannelli, Department of Radiology, Harborview Medical Center, 325 9th Avenue, Box 359728, Seattle, WA 98104-2499, USA. Tel.: þ1 2067443561; fax: þ1 2067448560. E-mail address: [email protected] (L. Mannelli).
used by histopathologists.1,2 Several chronic liver diseases progress to liver fibrosis and cirrhosis at a variable rate, which can be influenced by treatment.3,4 Several biomarkers have been proposed as an alternative to liver biopsy for the staging of liver fibrosis, including serum blood markers and ultrasound or magnetic resonance imaging (MRI)-based stiffness measurements.3,4 Magnetic resonance elastography (MRE) is a promising technique for evaluating liver stiffness in chronic liver disease patients.5e12 The current MRE literature is based on measurements performed at
0009-9260/$ e see front matter Ó 2011 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.crad.2011.08.022
L. Mannelli et al. / Clinical Radiology 67 (2012) 258e262
1.5 T.1,2,5e12 The technique relies upon the demonstration of propagating shear waves within the liver using a phasecontrast type gradient echo-based sequence.13 Shear waves propagate in tissue and are not known to be influenced by magnetic field strength. However, the technique is limited by the attenuation of the waves through the tissue and by the overall signal-to-noise ratio (SNR) of the measurements.14e16 MRE at 3 T has previously been published using a spin-echo technique.15 However, the authors had to employ a second harmonic technique in order to obtain a short enough echo time (TE) to account for the reduced T2 of the liver at 3 T. In this study a gradient echo-based MRE sequence was utilized with an inherently short TE. The aims of the present work were (a) to demonstrate the feasibility of obtaining MRE stiffness measurements from normal healthy volunteers using the same gradient echo technique that has been successfully applied at 1.5 T; (b) to evaluate the interobserver variability in stiffness measurements; and (c) to compare the liver stiffness values at 3 T in healthy volunteers with previously published results from MRE at 1.5 T.
Materials and methods Volunteers This study was approved by the local ethics committee and informed consent was obtained from the participants. Eleven healthy volunteers (nine male, two females, mean age 35 9 years ranging from 29e58 years) with no history of gastrointestinal, hepatobiliary, or cardiovascular disease; with no history of splenic trauma; and not receiving any regular medication were recruited. The volunteers underwent MRE between 27 May 2009 and 5 February 2010. The studies were performed in the morning and participants were asked to fast for at least 6 h before the examination.
MRI MRI was performed using a commercial whole-body 3 T MRI system (Signa HDx, GE Healthcare, Waukesha, WI, USA) using an eight-element cardiac receive coil.
Conventional MRI Following initial T2-weighted localizer images, 20 contiguous axial-balanced, gradient-echo images (FIESTA) with 10 mm thick sections were obtained through the upper abdomen encompassing the spleen during a single breath-hold and using the following parameters: 4.4 ms repetition time (TR), 1.7 ms TE, 360 288 mm FOV, number of averages 1, and 192 384 matrix. Based on these images a 19 cm diameter pneumatic driver was placed on the right side over the anterior abdominal wall at the axial level of the midpoint of the cranio-caudal extent of the spleen.
MRE A breath-hold MRE acquisition was performed in the transverse plane.A gradient echo-based MRE sequence was used with the following parameters: 99.9 ms TR, 24.2 ms
259
TE, 360 270 mm FOV, 256 96 matrix, two sections 10 mm thick with a 10 mm gap. For the pneumatic driver a 60 Hz excitation frequency was used. The MRE images were processed using a local frequency estimation (LFE) inversion algorithm that has been previously developed and described,12 in order to obtain stiffness maps. Processing generates several images, including a conventional magnitude image, and a grey-scale stiffness LFE image. It is important to note that the LFE inversion algorithm applies cut-offs to the frequencies utilized to compute the stiffness maps; in the present study the higher spatial frequency limit was 1.25/cm (corresponding to stiffness values<0.5 kPa), the lower spatial frequency limit was 0.08/cm (corresponding to stiffness values >100 kPa).
MRI evaluation Two abdominal MRI radiologists (L.M. and E.G. each with 3 years of experience in MRE measurements) independently reviewed the images using a commercial workstation (Advantage Windows 4.4, GE Healthcare, Buc, France). The radiologists subjectively reviewed the MRE phase images in a cine loop in order to verify the presence of propagating shear waves through the liver parenchyma. The same two radiologists measured the liver stiffness values on the MRE images by manually placing regions of interest (ROI) outlining the liver margins on the two axial sections of the gradient echo magnitude images, which were then mapped onto the matching MRE inversion stiffness images.17 The organ stiffness values in Pascals were calculated as the averaged mean of the two sections.
Statistical analysis The inter-observer agreement between the liver stiffness values was assessed using the intraclass correlation coefficient (ICC).
Results Figs 1 and 2 demonstrate typical MRE images in the same volunteer acquired at 3 T and 1.5 T, respectively. (The 1.5 T images were obtained from a separate study conducted 1 year earlier and are only shown to illustrate wave propagation between 1.5 and 3 T.) Fig 3 displays the MRE values for reader one and reader two. There were no technical issues in imaging any of the volunteers’ livers, and the shear waves originating from the driver and propagating through the liver were well visualized in all the volunteers. For reader 1, the mean liver stiffness for the group was 2341 381 Pa (ranging from 1.7e2.8 kPa). For reader 2, the mean liver stiffness for the group was 2.4 0.4 kPa (ranging from 1.7e2.8 kPa). There was an excellent inter-observer agreement for liver stiffness measurements; the ICC was 0.98 (95% CI 0.939 to 0.996; Fig 3).
260
L. Mannelli et al. / Clinical Radiology 67 (2012) 258e262
Figure 1 A 34-year-old man. Images acquired using a whole-body 3 T MRI system (Signa HDx, GE Healthcare, Milwaukee, USA). Balanced gradient echo (FIESTA) and MRE images with the driver placed anteriorly on the right (over the liver). FIESTA images (a) were used to choose the location for the MRE acquisition (bef). From the gradient echo MRE acquisition magnitude images (b) four sets of post-processed images are displayed: the unwrapped phase shift wave image (c); the wrapped phase shift wave image (d); the MRE inversion stiffness grey-scale image (e); and the stiffness colour map (f). The stiffness measurements were made on the images (e) in which the grey-scale reflects the stiffness in Pascals. ROIs were drawn over the liver perimeter on the gradient echo magnitude images (b) and then copied and pasted on to the matching MRE inversion stiffness images (e). In this volunteer the measured mean liver stiffness was 2.3 0.9 kPa.
Discussion This preliminary work using MRE at 3 T in healthy volunteers demonstrates the feasibility of liver stiffness evaluation at 3 T without modification of the approach used at 1.5 T. No sequence modication was required with respect to previously published MRE techniques at 1.5 T.17 An adequate image quality was obtained in all volunteers, with good visualization of the phase shift waves propagating from the driver through the liver parenchyma. The liver stiffness values were in the range of previously published measurements in healthy volunteers using MRE at 1.5 T8,9,11 and 3 T.15
Staging of liver fibrosis is important in the clinical management of patients with chronic liver disease, and as conventional imaging [ultrasound, computed tomography (CT) and MRI] and blood tests are inadequate at staging liver fibrosis, staging is currently performed by histopathologists after biopsy.1e4 However, liver biopsy is invasive and has several limitations, such as sampling error, patient tolerability and morbidity, and subjective histopathologist evaluation, which make repeat biopsies for monitoring disease relatively impractical.1e4 Several alternative imaging-based approaches have been proposed, using MRI and ultrasound, including diffusion-weighted MRI, ultrasound-based elastography, and MRE.1e12,17e19 Each of these techniques has limitations.
L. Mannelli et al. / Clinical Radiology 67 (2012) 258e262
261
Figure 2 A 30-year-old man. Images acquired using a whole-body 1.5 T MRI system (Signa HDx, GE Healthcare, Milwaukee, USA). Balanced gradient echo (FIESTA) and MRE images with the driver placed anteriorly on the right (over the liver). FIESTA images (a) were used to choose the location for the MRE acquisition (bef). From the gradient echo MRE acquisition magnitude images (b) four sets of post-processed images are displayed: the unwrapped phase shift wave image (c); the wrapped phase shift wave image (d); the MRE inversion stiffness greys-cale image (e); and the stiffness colour map (f). The stiffness measurements were made on the images (e) in which the grey-scale reflects the stiffness in Pascals. ROIs were drawn over the liver perimeter on the gradient echo magnitude images (b) and then copied and pasted on to the matching MRE inversion stiffness images (e). In this volunteer the measured mean liver stiffness was 32.1 1.1 kPa.
Ultrasound-based elastography is also prone to sampling error, it is operator dependent, and may be limited in patients with high body mass index or ascites.19 Diffusion-weighted MRI has produced interesting results but the apparent diffusion coefficient (ADC) cannot accurately discriminate well between patients with different stages of liver fibrosis.1,2,18 MRE does not perform well in patients with iron overload as the low SNR prevents adequate visualization of the shear waves. Despite this, MRE has advantages for longitudinal evaluation of liver fibrosis as it is non-invasive and can differentiate between low and high grades of fibrosis with good sensitivity and specificity. MRE has good sensitivity and specificity at 1.5 T, when validated against liver biopsy.1e12
However, this has been demonstrated at 1.5 T, where it has been validated against liver biopsy,1e12 but there is only limited experience at 3 T. The mean liver stiffness values obtained in this series of healthy volunteers using MRE at 3 T are comparable to those reported at 1.5 T regarding both mean values and standard deviations. The range of stiffness values extends between 1712 and 2836 Pa, which are values within the expected range for liver stiffness in healthy volunteers, according to the published literature at 1.5 T.1e12,17,20,21 Values of liver stiffness in healthy volunteers have some variability in the literature, with normality being considered <2.3 0.4 kPa by most authors.8,22 These discrepancies are likely due to differences in the imaging acquisition
262
L. Mannelli et al. / Clinical Radiology 67 (2012) 258e262
References
Figure 3 BlandeAltman graph showing the excellent inter-observer agreement. Ninety-five percent confidence intervals for the ICC population values: 0.939 < ICC < 0.996.
technique. such as, for example, different transverse waves frequencies, as demonstrated by Asbach et al.22 Multiple investigators have studied the diagnostic ability of liver MRE for detecting hepatic fibrosis, and consistent findings have established that liver MRE has excellent diagnostic accuracy for assessing hepatic fibrosis. MRE has a negative predictive value of 97% for distinguishing normal from fibrotic livers, using a shear stiffness cut-off value of 2.93 kPa.9,23 Steatohepatitis also causes an increase in liver stiffness; the possibility of differentiating fibrosis from steatohepatitis is still object of investigation.24 The use of 3 T for comprehensive liver MRI may prove valuable in the future with the potential to add further functional MRI techniques including spectroscopy, where examination at 3 T has a potential benefit compared to 1.5 T. A limitation of this study is the small sample size and the fact that MRE at 3 T and at 1.5 T were not obtained on the same day for the same group of volunteers. In addition, direct histopathological validation that each volunteer had a normal liver was not obtained, but this was not practical from an ethical perspective. A more formal evaluation of the benefits of 3 T MRE is required and a direct comparison of volunteer examinations at both 1.5 T and 3 T is planned. This study in healthy volunteers has demonstrated that (a) MRE is feasible at 3 T using the same approach used at 1.5 T; (b) there is good interobserver reproducibility in stiffness measurements; and (c) the results obtained lie within the normal range defined at 1.5 T.
Acknowledgments The authors acknowledge funding support from Addenbrooke’s Charitable Trust and the NIHR Cambridge Biomedical Research Centre.
1. Taouli B, Ehman RL, Reeder SB. Advanced MRI methods for assessment of chronic liver disease. AJR Am J Roentgenol 2009;193:14e27. 2. Faria SC, Ganesan K, Mwangi I, et al. MR imaging of liver fibrosis: current state of the art. RadioGraphics 2009;29:1615e35. 3. Manning DS, Afdhal NH. Diagnosis and quantitation of fibrosis. Gastroenterology 2008;134:1670e81. 4. Popov Y, Schuppan D. Targeting liver fibrosis: strategies for development and validation of antifibrotic therapies. Hepatology 2009;50: 1294e306. 5. Oudry J, Chen J, Glaser KJ, et al. Cross-validation of magnetic resonance elastography and ultrasound-based transient elastography: a preliminary phantom study. J Magn Reson Imaging 2009;30:1145e50. 6. Ehman RL. Science to practice: can MR elastography be used to detect early steatohepatitis in fatty liver disease? Radiology 2009;253: 1e3. 7. Venkatesh SK, Yin M, Glockner JF, et al. MR elastography of liver tumors: preliminary results. AJR Am J Roentgenol 2008;190:1534e40. 8. Talwalkar JA, Yin M, Fidler JL, et al. Magnetic resonance imaging of hepatic fibrosis: emerging clinical applications. Hepatology 2008;47: 332e42. 9. Yin M, Talwalkar JA, Glaser KJ. Assessment of hepatic fibrosis with magnetic resonance elastography. Clin Gastroenterol Hepatol 2007;5: 1207e13. 10. Yin M, Woollard J, Wang X, et al. Quantitative assessment of hepatic fibrosis in an animal model with magnetic resonance elastography. Magn Reson Med 2007;58:346e53. re O, Yin M, Dresner MA, et al. MR elastography of the liver: 11. Rouvie preliminary results. Radiology 2006;240:440e8. 12. Manduca A, Oliphant TE, Dresner MA, et al. Magnetic resonance elastography: non-invasive mapping of tissue elasticity. Med Image Anal 2001;5:237e54. 13. Muthupillai R, Lomas DJ, Rossman PJ, et al. Magnetic resonance elastography by direct visualization of propagating acoustic strain waves. Science 1995;269:1854e7. 14. Wen H, Denison TJ, Singerman RW, et al. The intrinsic signal-to-noise ratio in human cardiac imaging at 1.5, 3, and 4T. J Magn Reson 1997;125:65e71. 15. Herzka DA, Kotys MS, Sinkus R, et al. Magnetic resonance elastography in the liver at 3 Tesla using a second harmonic approach. Magn Reson Med 2009;62:284e91. 16. Dunn T, Majumdar S. Comparison of motion encoding waveforms for magnetic resonance elastography at 3T. Conf Proc IEEE Eng Med Biol Soc 2005;7:7405e8. 17. Mannelli L, Godfrey E, Joubert I, et al. MR elastography: spleen stiffness measurements in healthy volunteersdpreliminary experience. AJR Am J Roentgenol 2010;195:387e92. http://www.ncbi.nlm.nih.gov/pubmed/ 20651194. 18. Mannelli L, Kim S, Hajdu CH, et al. Assessment of tumor necrosis of hepatocellular carcinoma after chemoembolization: diffusionweighted and contrast-enhanced MRI with histopathologic correlation of the explanted liver. AJR Am J Roentgenol 2009;193: 1044e52. 19. Vizzutti F, Arena U, Marra F, et al. Elastography for the non-invasive assessment of liver disease: limitations and future developments. Gut 2009;58:157e60. 20. Talwalkar JA, Yin M, Venkatesh S, et al. Feasibility of in vivo MR elastographic splenic stiffness measurements in the assessment of portal hypertension. AJR Am J Roentgenol 2009;193:122e7. 21. Hines CD, Bley TA, Lindstrom MJ, et al. Repeatability of magnetic resonance elastography for quantification of hepatic stiffness. J Magn Reson Imaging 2010;31:725e31. 22. Asbach P, Klatt D, Schlosser B, et al. Viscoelasticity-based staging of hepatic fibrosis with multifrequency MR elastography. Radiology 2010; 257:80e6. 23. Yin M, Chen J, Glaser KJ, et al. Abdominal magnetic resonance elastography. Top Magn Reson Imaging 2009;20:79e87. 24. Chen J, Talwalkar JA, Yin M, et al. Early detection of nonalcoholic steatohepatitis in patients with nonalcoholic fatty liver disease by using MR elastography. Radiology 2011;259:749e56.
Contents lists available at SciVerse ScienceDirect
Clinical Radiology journal homepage: www.clinicalradiologyonline.net
Magnetic resonance elastography: Feasibility of liver stiffness measurements in healthy volunteers at 3 Tq L. Mannelli a, b, *, E. Godfrey a, M.J. Graves a, A.J. Patterson a, P. Beddy a, D. Bowden a, I. Joubert a, A.N. Priest a, D.J. Lomas a a b
Department of Radiology, Addenbrooke’s Hospital and University of Cambridge, Cambridge, UK Department of Radiology, University of Washington, Seattle, WA, USA
article in formation Article history: Received 29 June 2011 Received in revised form 18 August 2011 Accepted 22 August 2011
AIM: To demonstrate the feasibility of obtaining liver stiffness measurements with magnetic resonance elastography (MRE) at 3 T in normal healthy volunteers using the same technique that has been successfully applied at 1.5 T. METHODS AND MATERIALS: The study was approved by the local ethics committee and written informed consent was obtained from all volunteers. Eleven volunteers (mean age 35 9 years) with no history of gastrointestinal, hepatobiliary, or cardiovascular disease were recruited. The magnetic resonance imaging (MRI) protocol included a gradient echo-based MRE sequence using a 60 Hz pneumatic excitation. The MRE images were processed using a local frequency estimation inversion algorithm to provide quantitative stiffness maps. Adequate image quality was assessed subjectively by demonstrating the presence of visible propagating waves within the liver parenchyma underlying the driver location. Liver stiffness values were obtained using manually placed regions of interest (ROI) outlining the liver margins on the gradient echo wave images, which were then mapped onto the corresponding stiffness image. The mean stiffness values from two adjacent sections were recorded. RESULTS: Eleven volunteers underwent MRE. The quality of the MRE images was adequate in all the volunteers. The mean liver stiffness for the group was 2.3 0.38 kPa (ranging from 1.7e2.8 kPa). CONCLUSIONS: This preliminary work using MRE at 3 T in healthy volunteers demonstrates the feasibility of liver stiffness evaluation at 3 T without modification of the approach used at 1.5 T. Adequate image quality and normal MRE values were obtained in all volunteers. The obtained stiffness values were in the range of those reported for healthy volunteers in previous studies at 1.5 T. There was good interobserver reproducibility in the stiffness measurements. Ó 2011 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved.
Introduction Currently all conventional imaging techniques are unable to detect early fibrosis and distinguish the fibrosis grades q Accepted as an abstract at ISMRM 2010, Stockholm, Sweden.
* Guarantor and correspondent: L. Mannelli, Department of Radiology, Harborview Medical Center, 325 9th Avenue, Box 359728, Seattle, WA 98104-2499, USA. Tel.: þ1 2067443561; fax: þ1 2067448560. E-mail address: [email protected] (L. Mannelli).
used by histopathologists.1,2 Several chronic liver diseases progress to liver fibrosis and cirrhosis at a variable rate, which can be influenced by treatment.3,4 Several biomarkers have been proposed as an alternative to liver biopsy for the staging of liver fibrosis, including serum blood markers and ultrasound or magnetic resonance imaging (MRI)-based stiffness measurements.3,4 Magnetic resonance elastography (MRE) is a promising technique for evaluating liver stiffness in chronic liver disease patients.5e12 The current MRE literature is based on measurements performed at
0009-9260/$ e see front matter Ó 2011 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.crad.2011.08.022
L. Mannelli et al. / Clinical Radiology 67 (2012) 258e262
1.5 T.1,2,5e12 The technique relies upon the demonstration of propagating shear waves within the liver using a phasecontrast type gradient echo-based sequence.13 Shear waves propagate in tissue and are not known to be influenced by magnetic field strength. However, the technique is limited by the attenuation of the waves through the tissue and by the overall signal-to-noise ratio (SNR) of the measurements.14e16 MRE at 3 T has previously been published using a spin-echo technique.15 However, the authors had to employ a second harmonic technique in order to obtain a short enough echo time (TE) to account for the reduced T2 of the liver at 3 T. In this study a gradient echo-based MRE sequence was utilized with an inherently short TE. The aims of the present work were (a) to demonstrate the feasibility of obtaining MRE stiffness measurements from normal healthy volunteers using the same gradient echo technique that has been successfully applied at 1.5 T; (b) to evaluate the interobserver variability in stiffness measurements; and (c) to compare the liver stiffness values at 3 T in healthy volunteers with previously published results from MRE at 1.5 T.
Materials and methods Volunteers This study was approved by the local ethics committee and informed consent was obtained from the participants. Eleven healthy volunteers (nine male, two females, mean age 35 9 years ranging from 29e58 years) with no history of gastrointestinal, hepatobiliary, or cardiovascular disease; with no history of splenic trauma; and not receiving any regular medication were recruited. The volunteers underwent MRE between 27 May 2009 and 5 February 2010. The studies were performed in the morning and participants were asked to fast for at least 6 h before the examination.
MRI MRI was performed using a commercial whole-body 3 T MRI system (Signa HDx, GE Healthcare, Waukesha, WI, USA) using an eight-element cardiac receive coil.
Conventional MRI Following initial T2-weighted localizer images, 20 contiguous axial-balanced, gradient-echo images (FIESTA) with 10 mm thick sections were obtained through the upper abdomen encompassing the spleen during a single breath-hold and using the following parameters: 4.4 ms repetition time (TR), 1.7 ms TE, 360 288 mm FOV, number of averages 1, and 192 384 matrix. Based on these images a 19 cm diameter pneumatic driver was placed on the right side over the anterior abdominal wall at the axial level of the midpoint of the cranio-caudal extent of the spleen.
MRE A breath-hold MRE acquisition was performed in the transverse plane.A gradient echo-based MRE sequence was used with the following parameters: 99.9 ms TR, 24.2 ms
259
TE, 360 270 mm FOV, 256 96 matrix, two sections 10 mm thick with a 10 mm gap. For the pneumatic driver a 60 Hz excitation frequency was used. The MRE images were processed using a local frequency estimation (LFE) inversion algorithm that has been previously developed and described,12 in order to obtain stiffness maps. Processing generates several images, including a conventional magnitude image, and a grey-scale stiffness LFE image. It is important to note that the LFE inversion algorithm applies cut-offs to the frequencies utilized to compute the stiffness maps; in the present study the higher spatial frequency limit was 1.25/cm (corresponding to stiffness values<0.5 kPa), the lower spatial frequency limit was 0.08/cm (corresponding to stiffness values >100 kPa).
MRI evaluation Two abdominal MRI radiologists (L.M. and E.G. each with 3 years of experience in MRE measurements) independently reviewed the images using a commercial workstation (Advantage Windows 4.4, GE Healthcare, Buc, France). The radiologists subjectively reviewed the MRE phase images in a cine loop in order to verify the presence of propagating shear waves through the liver parenchyma. The same two radiologists measured the liver stiffness values on the MRE images by manually placing regions of interest (ROI) outlining the liver margins on the two axial sections of the gradient echo magnitude images, which were then mapped onto the matching MRE inversion stiffness images.17 The organ stiffness values in Pascals were calculated as the averaged mean of the two sections.
Statistical analysis The inter-observer agreement between the liver stiffness values was assessed using the intraclass correlation coefficient (ICC).
Results Figs 1 and 2 demonstrate typical MRE images in the same volunteer acquired at 3 T and 1.5 T, respectively. (The 1.5 T images were obtained from a separate study conducted 1 year earlier and are only shown to illustrate wave propagation between 1.5 and 3 T.) Fig 3 displays the MRE values for reader one and reader two. There were no technical issues in imaging any of the volunteers’ livers, and the shear waves originating from the driver and propagating through the liver were well visualized in all the volunteers. For reader 1, the mean liver stiffness for the group was 2341 381 Pa (ranging from 1.7e2.8 kPa). For reader 2, the mean liver stiffness for the group was 2.4 0.4 kPa (ranging from 1.7e2.8 kPa). There was an excellent inter-observer agreement for liver stiffness measurements; the ICC was 0.98 (95% CI 0.939 to 0.996; Fig 3).
260
L. Mannelli et al. / Clinical Radiology 67 (2012) 258e262
Figure 1 A 34-year-old man. Images acquired using a whole-body 3 T MRI system (Signa HDx, GE Healthcare, Milwaukee, USA). Balanced gradient echo (FIESTA) and MRE images with the driver placed anteriorly on the right (over the liver). FIESTA images (a) were used to choose the location for the MRE acquisition (bef). From the gradient echo MRE acquisition magnitude images (b) four sets of post-processed images are displayed: the unwrapped phase shift wave image (c); the wrapped phase shift wave image (d); the MRE inversion stiffness grey-scale image (e); and the stiffness colour map (f). The stiffness measurements were made on the images (e) in which the grey-scale reflects the stiffness in Pascals. ROIs were drawn over the liver perimeter on the gradient echo magnitude images (b) and then copied and pasted on to the matching MRE inversion stiffness images (e). In this volunteer the measured mean liver stiffness was 2.3 0.9 kPa.
Discussion This preliminary work using MRE at 3 T in healthy volunteers demonstrates the feasibility of liver stiffness evaluation at 3 T without modification of the approach used at 1.5 T. No sequence modication was required with respect to previously published MRE techniques at 1.5 T.17 An adequate image quality was obtained in all volunteers, with good visualization of the phase shift waves propagating from the driver through the liver parenchyma. The liver stiffness values were in the range of previously published measurements in healthy volunteers using MRE at 1.5 T8,9,11 and 3 T.15
Staging of liver fibrosis is important in the clinical management of patients with chronic liver disease, and as conventional imaging [ultrasound, computed tomography (CT) and MRI] and blood tests are inadequate at staging liver fibrosis, staging is currently performed by histopathologists after biopsy.1e4 However, liver biopsy is invasive and has several limitations, such as sampling error, patient tolerability and morbidity, and subjective histopathologist evaluation, which make repeat biopsies for monitoring disease relatively impractical.1e4 Several alternative imaging-based approaches have been proposed, using MRI and ultrasound, including diffusion-weighted MRI, ultrasound-based elastography, and MRE.1e12,17e19 Each of these techniques has limitations.
L. Mannelli et al. / Clinical Radiology 67 (2012) 258e262
261
Figure 2 A 30-year-old man. Images acquired using a whole-body 1.5 T MRI system (Signa HDx, GE Healthcare, Milwaukee, USA). Balanced gradient echo (FIESTA) and MRE images with the driver placed anteriorly on the right (over the liver). FIESTA images (a) were used to choose the location for the MRE acquisition (bef). From the gradient echo MRE acquisition magnitude images (b) four sets of post-processed images are displayed: the unwrapped phase shift wave image (c); the wrapped phase shift wave image (d); the MRE inversion stiffness greys-cale image (e); and the stiffness colour map (f). The stiffness measurements were made on the images (e) in which the grey-scale reflects the stiffness in Pascals. ROIs were drawn over the liver perimeter on the gradient echo magnitude images (b) and then copied and pasted on to the matching MRE inversion stiffness images (e). In this volunteer the measured mean liver stiffness was 32.1 1.1 kPa.
Ultrasound-based elastography is also prone to sampling error, it is operator dependent, and may be limited in patients with high body mass index or ascites.19 Diffusion-weighted MRI has produced interesting results but the apparent diffusion coefficient (ADC) cannot accurately discriminate well between patients with different stages of liver fibrosis.1,2,18 MRE does not perform well in patients with iron overload as the low SNR prevents adequate visualization of the shear waves. Despite this, MRE has advantages for longitudinal evaluation of liver fibrosis as it is non-invasive and can differentiate between low and high grades of fibrosis with good sensitivity and specificity. MRE has good sensitivity and specificity at 1.5 T, when validated against liver biopsy.1e12
However, this has been demonstrated at 1.5 T, where it has been validated against liver biopsy,1e12 but there is only limited experience at 3 T. The mean liver stiffness values obtained in this series of healthy volunteers using MRE at 3 T are comparable to those reported at 1.5 T regarding both mean values and standard deviations. The range of stiffness values extends between 1712 and 2836 Pa, which are values within the expected range for liver stiffness in healthy volunteers, according to the published literature at 1.5 T.1e12,17,20,21 Values of liver stiffness in healthy volunteers have some variability in the literature, with normality being considered <2.3 0.4 kPa by most authors.8,22 These discrepancies are likely due to differences in the imaging acquisition
262
L. Mannelli et al. / Clinical Radiology 67 (2012) 258e262
References
Figure 3 BlandeAltman graph showing the excellent inter-observer agreement. Ninety-five percent confidence intervals for the ICC population values: 0.939 < ICC < 0.996.
technique. such as, for example, different transverse waves frequencies, as demonstrated by Asbach et al.22 Multiple investigators have studied the diagnostic ability of liver MRE for detecting hepatic fibrosis, and consistent findings have established that liver MRE has excellent diagnostic accuracy for assessing hepatic fibrosis. MRE has a negative predictive value of 97% for distinguishing normal from fibrotic livers, using a shear stiffness cut-off value of 2.93 kPa.9,23 Steatohepatitis also causes an increase in liver stiffness; the possibility of differentiating fibrosis from steatohepatitis is still object of investigation.24 The use of 3 T for comprehensive liver MRI may prove valuable in the future with the potential to add further functional MRI techniques including spectroscopy, where examination at 3 T has a potential benefit compared to 1.5 T. A limitation of this study is the small sample size and the fact that MRE at 3 T and at 1.5 T were not obtained on the same day for the same group of volunteers. In addition, direct histopathological validation that each volunteer had a normal liver was not obtained, but this was not practical from an ethical perspective. A more formal evaluation of the benefits of 3 T MRE is required and a direct comparison of volunteer examinations at both 1.5 T and 3 T is planned. This study in healthy volunteers has demonstrated that (a) MRE is feasible at 3 T using the same approach used at 1.5 T; (b) there is good interobserver reproducibility in stiffness measurements; and (c) the results obtained lie within the normal range defined at 1.5 T.
Acknowledgments The authors acknowledge funding support from Addenbrooke’s Charitable Trust and the NIHR Cambridge Biomedical Research Centre.
1. Taouli B, Ehman RL, Reeder SB. Advanced MRI methods for assessment of chronic liver disease. AJR Am J Roentgenol 2009;193:14e27. 2. Faria SC, Ganesan K, Mwangi I, et al. MR imaging of liver fibrosis: current state of the art. RadioGraphics 2009;29:1615e35. 3. Manning DS, Afdhal NH. Diagnosis and quantitation of fibrosis. Gastroenterology 2008;134:1670e81. 4. Popov Y, Schuppan D. Targeting liver fibrosis: strategies for development and validation of antifibrotic therapies. Hepatology 2009;50: 1294e306. 5. Oudry J, Chen J, Glaser KJ, et al. Cross-validation of magnetic resonance elastography and ultrasound-based transient elastography: a preliminary phantom study. J Magn Reson Imaging 2009;30:1145e50. 6. Ehman RL. Science to practice: can MR elastography be used to detect early steatohepatitis in fatty liver disease? Radiology 2009;253: 1e3. 7. Venkatesh SK, Yin M, Glockner JF, et al. MR elastography of liver tumors: preliminary results. AJR Am J Roentgenol 2008;190:1534e40. 8. Talwalkar JA, Yin M, Fidler JL, et al. Magnetic resonance imaging of hepatic fibrosis: emerging clinical applications. Hepatology 2008;47: 332e42. 9. Yin M, Talwalkar JA, Glaser KJ. Assessment of hepatic fibrosis with magnetic resonance elastography. Clin Gastroenterol Hepatol 2007;5: 1207e13. 10. Yin M, Woollard J, Wang X, et al. Quantitative assessment of hepatic fibrosis in an animal model with magnetic resonance elastography. Magn Reson Med 2007;58:346e53. re O, Yin M, Dresner MA, et al. MR elastography of the liver: 11. Rouvie preliminary results. Radiology 2006;240:440e8. 12. Manduca A, Oliphant TE, Dresner MA, et al. Magnetic resonance elastography: non-invasive mapping of tissue elasticity. Med Image Anal 2001;5:237e54. 13. Muthupillai R, Lomas DJ, Rossman PJ, et al. Magnetic resonance elastography by direct visualization of propagating acoustic strain waves. Science 1995;269:1854e7. 14. Wen H, Denison TJ, Singerman RW, et al. The intrinsic signal-to-noise ratio in human cardiac imaging at 1.5, 3, and 4T. J Magn Reson 1997;125:65e71. 15. Herzka DA, Kotys MS, Sinkus R, et al. Magnetic resonance elastography in the liver at 3 Tesla using a second harmonic approach. Magn Reson Med 2009;62:284e91. 16. Dunn T, Majumdar S. Comparison of motion encoding waveforms for magnetic resonance elastography at 3T. Conf Proc IEEE Eng Med Biol Soc 2005;7:7405e8. 17. Mannelli L, Godfrey E, Joubert I, et al. MR elastography: spleen stiffness measurements in healthy volunteersdpreliminary experience. AJR Am J Roentgenol 2010;195:387e92. http://www.ncbi.nlm.nih.gov/pubmed/ 20651194. 18. Mannelli L, Kim S, Hajdu CH, et al. Assessment of tumor necrosis of hepatocellular carcinoma after chemoembolization: diffusionweighted and contrast-enhanced MRI with histopathologic correlation of the explanted liver. AJR Am J Roentgenol 2009;193: 1044e52. 19. Vizzutti F, Arena U, Marra F, et al. Elastography for the non-invasive assessment of liver disease: limitations and future developments. Gut 2009;58:157e60. 20. Talwalkar JA, Yin M, Venkatesh S, et al. Feasibility of in vivo MR elastographic splenic stiffness measurements in the assessment of portal hypertension. AJR Am J Roentgenol 2009;193:122e7. 21. Hines CD, Bley TA, Lindstrom MJ, et al. Repeatability of magnetic resonance elastography for quantification of hepatic stiffness. J Magn Reson Imaging 2010;31:725e31. 22. Asbach P, Klatt D, Schlosser B, et al. Viscoelasticity-based staging of hepatic fibrosis with multifrequency MR elastography. Radiology 2010; 257:80e6. 23. Yin M, Chen J, Glaser KJ, et al. Abdominal magnetic resonance elastography. Top Magn Reson Imaging 2009;20:79e87. 24. Chen J, Talwalkar JA, Yin M, et al. Early detection of nonalcoholic steatohepatitis in patients with nonalcoholic fatty liver disease by using MR elastography. Radiology 2011;259:749e56.