Subscribe to RSS
DOI: 10.1055/s-2006-925158
Endoscopic Ultrasound Elastography: the First Step towards Virtual Biopsy? Preliminary Results in 49 Patients
M. Giovannini, M. D.
Endoscopic Unit, Paoli-Calmettes Institute
232 Boulevard Sainte-Marguerite · 13273 Marseilles-Cedex 9 · France
Fax: +33-491223658
Email: hdjchir@marseille.fnclcc.fr
Publication History
Submitted 21 June 2005
Accepted after revision 19 December 2005
Publication Date:
05 May 2006 (online)
Background and Study Aims: It is well known that some diseases, such as cancer, lead to changes in the hardness
of tissue. Sonoelastography, a technique that allows the elasticity of tissue to be
assessed during ultrasound examination, provides the ultrasonographer with important
additional information that can be used for diagnosis. The aim of this study was to
evaluate the ability of endoscopic ultrasound elastography to differentiate between
benign and malignant pancreatic masses and lymph nodes.
Patients and Methods: During a 12-month period, 49 patients underwent endoscopic ultrasound (EUS) examinations
with elastography, conducted by a single endoscopist. Twenty-four patients underwent
evaluation of a pancreatic mass (mean diameter 24.7 ± 11.1mm) and 25 underwent evaluation
of 31 lymph nodes. The mean diameter of the lymph nodes was 19.7 ± 8.6 mm, and they
were found in the cervical area (n = 3), mediastinum (n = 17), celiac arterial trunk
region (n = 5), and aortocaval region (n = 6).
Results: The sonoelastography images of pancreatic masses were interpreted as benign in four
cases and malignant in 20. The sensitivity and specificity of sonoelastography in
the diagnosis of malignant lesions were 100 % and 67 %, respectively. The sonoelastography
images of the lymph nodes were interpreted as showing malignancy in 22 cases, benign
conditions in seven, and indeterminate status in two. The sensitivity and specificity
of sonoelastography for evaluating malignant lymph-node invasion were 100 % and 50
%, respectively.
Conclusions: EUS elastography is potentially capable of further defining the tissue characteristics
of benign and malignant lesions but specifity has to be improved. It can be used to
guide biopsy sampling for diagnosis.
Introduction
Endoscopic ultrasonography (EUS) provides imaging of tumors and enhances the accuracy of TNM staging. It can also provide guidance for fine-needle aspiration (FNA) and biopsies of undiagnosed masses and lymph nodes suspicious for malignant invasion. However, FNA is technically demanding, and multiple puncturing of lymph nodes or masses is sometimes required in order to obtain sufficient tissue for histological assessment. In addition, when several lymph nodes appear suspicious, the choice of which to puncture is not always clear. The current ultrasound criteria for assessing malignant lymph nodes (round, hypoechoic, diameter > 1 cm, and with distinct margins) are helpful for targeting lesions, but there are problems with the specificity of these criteria, as they may overlap with benign lymph nodes [1] [2]. In addition, the differential diagnosis of pancreatic masses is extensive and includes both benign and malignant etiologies, while FNA of the pancreas is associated with a small, but not insignificant, risk of pancreatitis [3]. The technique of sonoelastography may be able to provide more accurate evaluation of masses and lymph nodes before they are aspirated, improving the targeting of lesions for FNA and potentially increasing the diagnostic yield and reducing complications. This paper reports on preliminary experience with sonoelastography at a tertiary-care center with expertise in EUS.
#Patients and Methods
#Theory and Technical Aspects of Sonoelastography
The technique of sonoelastography is based on the fact that some diseases, such as cancer, lead to changes in the hardness of tissue (i. e., what is known as its elasticity modulus). The method is a further development of the well-known fremitus technique in breast ultrasonography [4] [5] [6], during which the patient is asked to hum while color or power Doppler ultrasound imaging is used to examine the breast. Softer portions of the breast vibrate more in response to the humming, while cancers and other firm masses vibrate less. Elastography allows assessment of the elastic properties of tissues by applying slight compression to the tissue and comparing the images obtained before and after compression. The data are then compared using a cross-correlation technique to determine the amount of displacement each small portion of tissue undergoes in response to the compression applied by the ultrasound transducer (Figure [1]) [7] [8] [9]. The elasticity modulus - i. e., the tissue elasticity distribution - can be calculated from the strain and stress on the structures examined. While the strain field can be estimated from the radiofrequency signals returned from tissue structures before and after compression, it is not possible to measure the stress field directly within the tissue. Another problem is that compression of harder tissue structures is often followed by lateral displacement of these structures [10]. To overcome these problems, an extended combined autocorrelation method has been developed, which makes it possible to reconstruct the tissue elasticity of the examined structures on the basis of a three-dimensional finite element model. The new technique allows highly accurate estimation of the tissue elasticity distribution and provides adequate compensation for sideslipping. Elasticity imaging can be carried out with a sonoelastography module incorporated into the Hitachi EUB-8500 system (Hitachi Medical Systems Europe, Zug, Switzerland).

Figure 1 A gelatin test object contained a circular inclusion 1.9 cm in diameter that had the same ultrasound properties as the surrounding medium, but was three times harder. The sonogram (left) does not detect the presence of the inclusion, while the elastogram (center) demonstrates it well. The bright region centered on the inclusion in the elastogram is a stress-concentration artifact predicted from the Algor simulation of the sample at a 45° angle (right). (The test object was created by Dr. T. Hall at the University of Kansas Medical Center.)
Procedure, Technique, and Criteria
Endoscopic ultrasound tissue elasticity imaging is carried out with conventional EUS probes and does not require additional instruments. The vibrations and compressions are provided physiologically by vascular pulsation and respiratory motion. Calculation of the tissue elasticity distribution is carried out in real time, and the examination results are displayed in color superimposed over the conventional B-mode image.
To date, the majority of clinical research involving sonoelastography has focused on the evaluation of breast masses. Three different patterns have been identified in elastograms of breast cancers: a well-defined, very hard (dark) mass or nodule; a moderately hard mass or nodule containing much harder (darker) foci within it; and a very dark or hard central core surrounded by a somewhat softer or less dark peripheral component [6]. With conventional ultrasound or endoscopic ultrasound, fibrosis generally appears as hyperechoic regions with posterior acoustic shadowing (an appearance also seen in cancers); in elastography, however, it generally appears as a uniform, moderately hard region with no distinct foci of increased hardness. Preliminary research in breast tissue elastography has shown that the technique allows correct classification of most benign and malignant masses [6].
Between March 2004 and April 2005, 49 patients underwent EUS examination with sonoelastography by one experienced endoscopist (M.G.). The indications for elastography examination included evaluation of a pancreatic mass (n = 24) and assessment of suspicious lymph nodes (n = 25). The real-time elasticity imaging described in this study was carried out with a sonoelastography module incorporated into the Hitachi EUB-8500 system. Tissue elasticity imaging was performed with a Pentax EG38-UT EUS scope (Pentax Europe Ltd., Hamburg, Germany). The imaging results are displayed in color over the conventional B-mode image, with malignant tissue appearing in blue, fibrosis in green, normal tissue in yellow, and fat in red.
In all cases, EUS-FNA was carried out with a 22-gauge needle (Wilson-Cook Medical, Winston-Salem, North Carolina, USA). In our center, all specimens are routinely examined using the monolayer technique, and although a cytopathologist is not present when the biopsies are taken, the endoscopist assesses the sample to ensure that there is an adequate tissue core, with repeat punctures being carried out if necessary.
The sonoelastography images were read by one endoscopist (M.G.) during the examination, with the conclusions being recorded at the end of the examination before histopathological assessment. Masses or lymph nodes that appeared mostly blue (harder) were considered to be malignant, with other results being regarded as benign. The final diagnosis was based on the histological assessment of the FNA samples and surgical specimens when available.
The results are presented as means plus or minus standard deviation or as medians with ranges, depending on the data distribution.
#Results
#Pancreatic Masses
Twenty-four patients (median age 60, range 39 - 88) underwent EUS examinations with sonoelastography for evaluation of pancreatic masses (mean diameter 24.7 ± 11.1mm). The masses were located in the pancreatic head (n = 12), body (n = 6), and tail (n = 6). The final histological assessment was based on the FNA results in 21 cases and on surgical pathology in three cases. The final diagnoses of malignant masses included adenocarcinoma of the pancreas (n = 14), metastatic renal cancer (n = 2), sarcoma (n = 1), and ovarian cancer (n = 1). Benign masses identified consisted of chronic pancreatitis-related nodules (n = 4), a neuroendocrine tumor (n = 1), and an intrapapillary mucinous tumor (n = 1).
The sonoelastography images of pancreatic masses were interpreted as benign in four cases (Figure [2]) and malignant in 20 (Figures [3], [4]). Two masses were misclassified as malignant on elastography; the first was a neuroendocrine tumor, and the second a benign fibromyoblastic tumor of the pancreas that was surgically resected. The sensitivity and specificity of sonoelastography in the diagnosis of malignant lesions were 100 % and 67 %, respectively.

Figure 2 A fibrous mass in chronic pancreatitis.

Figure 3 An advanced pancreatic adenocarcinoma.

Figure 4 A pancreatic adenocarcinoma arising in a pancreas with chronic pancreatitis.
This experience with elastography and pancreatic masses was subsequently reviewed, and a more refined classification of pancreatic mass sonoelastography images was then developed in which the images are differentiated into five scores (Figure [5]). Score 1 is for a homogeneous hypoechoic area (soft, green), corresponding to the normal pancreatic tissue. In score 2, the elastogram is heterogeneous but still within the soft-tissue range (green, yellow, and red), corresponding to fibrosis. Score 3 is assigned to elastographic images that are largely blue (hard), with minimal heterogeneity, corresponding to a small, early pancreatic adenocarcinoma (less than 25 mm in diameter). In tumors assigned score 4, there is a hypoechoic region in the center, with a green appearance in this small area surrounded by blue or harder tissue; this corresponds to a hypervascular lesion such as a neuroendocrine tumor (Figure [6]) or small pancreatic metastasis. Finally, score 5 is assigned to lesions that are largely blue on the elastogram but with heterogeneity including softer tissue colors (green, red); this represents necrosis, and is seen in advanced pancreatic adenocarcinomas.

Figure 5 Classification of elastography findings, based on the authors’ experience after the initial analysis of accuracy. Score 1: there is distortion over the entire area, representing a normal pancreas. Score 2: the elastogram is heterogeneous but still within the soft-tissue range (green, yellow, and red), representing fibrosis or chronic pancreatitis. Score 3: there is distortion only on the edge of a largely blue (hard) area, with minimal heterogeneity, corresponding to a small, early pancreatic adenocarcinoma. Score 4: there is a hypoechoic region in the center, with a green appearance in this small area surrounded by blue or harder tissue. This represents a hypervascular lesion such as a neuroendocrine tumor or small pancreatic metastasis. Score 5: the lesion is largely blue, but with heterogeneity including softer tissue colors (green, red). This represents necrosis, and is seen in advanced pancreatic adenocarcinomas.

Figure 6 A pancreatic insulinoma.
Lymph Nodes
Twenty-five patients (median age 57, range 16 - 76) underwent EUS examinations with sonoelastography of 31 lymph nodes. The mean diameter of the lymph nodes was 19.7 ± 8.6 mm, and they were found in the cervical area (n = 3), mediastinum (n = 17), celiac arterial trunk region (n = 5), and aortocaval region (n = 6). The final histological assessment was based on FNA and classified the lymph nodes as benign in 14 cases and malignant in 17.
The sonoelastography images of the lymph nodes were interpreted as showing malignancy in 22 cases (Figure [7 ] a), a benign condition in seven cases (Figure [7 ] b), and indeterminate status in two. There were no false-negative findings, but five false-positive ones. The indeterminate cases were due to heterogeneity of the sonoelastography images, and the nodes were both found to be benign on the final histological assessment. The sensitivity and specificity of sonoelastography for evaluating malignant invasion of lymph nodes were 100 % and 50 %, respectively. Six patients underwent sonoelastography of more than one lymph node. In two of these cases, one lymph node was benign and the other malignant, and elastography correctly differentiated between the two.

Figure 7 Elastography of lymph nodes (blue). a Malignant lymph node. b Inflammatory lymph node.
No complications occurred during the study.
#Discussion
Sonoelastography allows the hardness or stiffness of biological tissues to be estimated and imaged using conventional ultrasound instruments with modified software. It is known that some pathological conditions, such as malignant tumors, often produce changes in the mechanical properties of tissue. The elastic characteristics of the tissue appear to be fairly uniform throughout benign lesions. By contrast, cancer lesions grow in a very disorganized way, so that the elastic properties of one area of a malignant tumor may differ significantly from those in another area. The approach used to assess these tissue changes is an extension of the basic principles involved in traditional medical ultrasound imaging. The principle is based on the fact that tissues are distorted slightly when a small displacement is applied externally [8] [9] [11].
The present study is a continuation of our previous research on EUS-guided sonoelastography [12]. The sensitivity of the assessment of both pancreatic masses and lymph nodes was 100 %. Although false-positive findings occurred in both groups of patients and there may be some concern regarding the specificity of sonoelastography in both settings (67 % in the pancreas group, 50 % in the lymph-node group), it should be recalled that the number of benign lesions in this study was relatively small. It is likely that the specificity will improve with more experience and more refined criteria. We have already reviewed the experience with elastography and pancreatic masses and have developed a new, more refined classification of the sonoelastography images (Figure [5]). Further assessment of this classification system is continuing. The results in the six patients with multiple suspicious lymph nodes highlight the potential value of elastography for selecting which lymph node or nodes to puncture, thus potentially reducing puncture-related risks and reducing the procedure time. Although EUS-FNA has the potential to miss microinvasion of malignancy into lymph nodes and is therefore relatively imperfect as a gold standard in the absence of surgical specimens, we consider that it is representative of everyday practice, particularly when it is combined with an adequate follow-up period.
#Conclusions
EUS elastography is a new application in the field of endosonography and appears to be capable of distinguishing between fibrous and benign tissue, on the one hand, and malignant lesions on the other. While these results are very encouraging, further research is needed to further clarify the role of this new technique. Research should focus on defining the criteria required for accurate elastography and on subsequent assessment of the technique, using several endoscopists in a blinded setting. EUS elastography has the potential to enhance the diagnosis and treatment of gastrointestinal tumors.
#Acknowledgment
Dr. L. C. Hookey’s contribution to this study was supported by a Canadian Association of Gastroenterology/Canadian Institutes of Health Research/Solvay Research Fellowship Award.
Dr. H. Schreiber from HITACHI Coorporation for his help and support.
Competing interests: None
In Brief
Elastography uses changes in tissue stiffness, translated into different colors, in an attempt to improve the differential diagnosis between benign and malignant lesions. Twenty-four patients with pancreatic masses and 25 with lymph nodes were tested in a simplified protocol equating blue lesions with malignancy. Good sensitivity but poor specificity was achieved. It remains to be seen whether an improved grading system, as suggested on the basis of this study, might help (see also the Editorial on p. 415).
References
- 1 Bhutani M S, Hawes R H, Hoffman B J. A comparison of the accuracy of echo features during endoscopic ultrasound (EUS) and EUS-guided fine-needle aspiration for diagnosis of malignant lymph node invasion. Gastrointest Endosc. 1997; 45 474-479
- 2 Tamerisa R, Irisawa A, Bhutani M S. Endoscopic ultrasound in the diagnosis, staging, and management of gastrointestinal and adjacent malignancies. Med Clin North Am. 2005; 89 139-158, viii
- 3 Gress F, Michael H, Gelrud D. et al . EUS-guided fine-needle aspiration of the pancreas: evaluation of pancreatitis as a complication. Gastrointest Endosc. 2002; 56 864-867
- 4 Chaudhari M H, Forsberg F, Voodarla A. et al . Breast tumor vascularity identified by contrast enhanced ultrasound and pathology: initial results. Ultrasonics. 2000; 38 105-109
- 5 Fornage B D. Recent advances in breast sonography. JBR-BTR. 2000; 83 75-80
- 6 Garra B S, Cespedes E I, Ophir J. et al . Elastography of breast lesions: initial clinical results. Radiology. 1997; 202 79-86
- 7 Evans D H, McDicken W N. Doppler ultrasound: physics, instrumentation, and signal processing, 2nd ed. New York; Wiley 1999
- 8 Gao L, Parker K J, Lerner R M, Levinson S F. Imaging of the elastic properties of tissue: a review. Ultrasound Med Biol. 1996; 22 959-977
- 9 Ophir J, Cespedes E I, Garra B S. et al . Elastography: ultrasonic imaging of tissue strain and elastic modulus in vivo. Eur J Ultrasound. 1996; 3 49-70
- 10 Doyley M M, Meaney P M, Bamber J C. Evaluation of an iterative reconstruction method for quantitative elastography. Phys Med Biol. 2000; 45 1521-1540
- 11 Rubens D J, Hadley M A, Alam S K. et al . Sonoelasticity imaging of prostate cancer: in vitro results. Radiology. 1995; 195 379-383
- 12 Giovannini M, Bories E, Pesenti C. et al . Sonoelastography guided by endoscopic ultrasound: the first step for virtual biopsy? Results in 14 patients with a pancreatic mass [abstract]. Endoscopy. 2004; 36 (Suppl 1) A43
M. Giovannini, M. D.
Endoscopic Unit, Paoli-Calmettes Institute
232 Boulevard Sainte-Marguerite · 13273 Marseilles-Cedex 9 · France
Fax: +33-491223658
Email: hdjchir@marseille.fnclcc.fr
References
- 1 Bhutani M S, Hawes R H, Hoffman B J. A comparison of the accuracy of echo features during endoscopic ultrasound (EUS) and EUS-guided fine-needle aspiration for diagnosis of malignant lymph node invasion. Gastrointest Endosc. 1997; 45 474-479
- 2 Tamerisa R, Irisawa A, Bhutani M S. Endoscopic ultrasound in the diagnosis, staging, and management of gastrointestinal and adjacent malignancies. Med Clin North Am. 2005; 89 139-158, viii
- 3 Gress F, Michael H, Gelrud D. et al . EUS-guided fine-needle aspiration of the pancreas: evaluation of pancreatitis as a complication. Gastrointest Endosc. 2002; 56 864-867
- 4 Chaudhari M H, Forsberg F, Voodarla A. et al . Breast tumor vascularity identified by contrast enhanced ultrasound and pathology: initial results. Ultrasonics. 2000; 38 105-109
- 5 Fornage B D. Recent advances in breast sonography. JBR-BTR. 2000; 83 75-80
- 6 Garra B S, Cespedes E I, Ophir J. et al . Elastography of breast lesions: initial clinical results. Radiology. 1997; 202 79-86
- 7 Evans D H, McDicken W N. Doppler ultrasound: physics, instrumentation, and signal processing, 2nd ed. New York; Wiley 1999
- 8 Gao L, Parker K J, Lerner R M, Levinson S F. Imaging of the elastic properties of tissue: a review. Ultrasound Med Biol. 1996; 22 959-977
- 9 Ophir J, Cespedes E I, Garra B S. et al . Elastography: ultrasonic imaging of tissue strain and elastic modulus in vivo. Eur J Ultrasound. 1996; 3 49-70
- 10 Doyley M M, Meaney P M, Bamber J C. Evaluation of an iterative reconstruction method for quantitative elastography. Phys Med Biol. 2000; 45 1521-1540
- 11 Rubens D J, Hadley M A, Alam S K. et al . Sonoelasticity imaging of prostate cancer: in vitro results. Radiology. 1995; 195 379-383
- 12 Giovannini M, Bories E, Pesenti C. et al . Sonoelastography guided by endoscopic ultrasound: the first step for virtual biopsy? Results in 14 patients with a pancreatic mass [abstract]. Endoscopy. 2004; 36 (Suppl 1) A43
M. Giovannini, M. D.
Endoscopic Unit, Paoli-Calmettes Institute
232 Boulevard Sainte-Marguerite · 13273 Marseilles-Cedex 9 · France
Fax: +33-491223658
Email: hdjchir@marseille.fnclcc.fr

Figure 1 A gelatin test object contained a circular inclusion 1.9 cm in diameter that had the same ultrasound properties as the surrounding medium, but was three times harder. The sonogram (left) does not detect the presence of the inclusion, while the elastogram (center) demonstrates it well. The bright region centered on the inclusion in the elastogram is a stress-concentration artifact predicted from the Algor simulation of the sample at a 45° angle (right). (The test object was created by Dr. T. Hall at the University of Kansas Medical Center.)

Figure 2 A fibrous mass in chronic pancreatitis.

Figure 3 An advanced pancreatic adenocarcinoma.

Figure 4 A pancreatic adenocarcinoma arising in a pancreas with chronic pancreatitis.

Figure 5 Classification of elastography findings, based on the authors’ experience after the initial analysis of accuracy. Score 1: there is distortion over the entire area, representing a normal pancreas. Score 2: the elastogram is heterogeneous but still within the soft-tissue range (green, yellow, and red), representing fibrosis or chronic pancreatitis. Score 3: there is distortion only on the edge of a largely blue (hard) area, with minimal heterogeneity, corresponding to a small, early pancreatic adenocarcinoma. Score 4: there is a hypoechoic region in the center, with a green appearance in this small area surrounded by blue or harder tissue. This represents a hypervascular lesion such as a neuroendocrine tumor or small pancreatic metastasis. Score 5: the lesion is largely blue, but with heterogeneity including softer tissue colors (green, red). This represents necrosis, and is seen in advanced pancreatic adenocarcinomas.

Figure 6 A pancreatic insulinoma.

Figure 7 Elastography of lymph nodes (blue). a Malignant lymph node. b Inflammatory lymph node.