- Email: [email protected]
EUS elastography of the pancreas: feasibility and pattern description of the normal pancreas, chronic pancreatitis, and focal pancreatic lesions
ORIGINAL ARTICLE: Clinical Endoscopy
EUS elastography of the pancreas: feasibility and pattern description of the normal pancreas, chronic pancreatitis, and focal pancreatic lesions ¨ rer, MD, Lucas Greiner, MD Jan Janssen, MD, Eva Schlo Wuppertal, Germany
Background: Initial clinical applications have shown that US elastography might be able to distinguish tissues because of their specific consistency. Objective: (1) To investigate the feasibility of EUS elastography of the pancreas and (2) to describe elastographic patterns of the normal pancreas and the pancreas affected by inflammatory or focal disease. Design: Prospective single-center study. Setting: Academic center of the University of Witten/Herdecke. Patients: Twenty patients with normal pancreas, 20 patients with chronic pancreatitis, and 33 patients with focal pancreatic lesion, histologically later proven in 32 of these 33 cases. Interventions: Commercially available US equipment was used. The elasticity of tissue was reconstructed in real time within a sample area and was translated into a color scale imaging relative tissue elasticity within this area. Representative loops of at least 20 seconds were recorded regarding each region of interest. Results: Adequate elastographic recordings could be obtained in all 73 patients. Patients with hypoechoic and intermediately echogenic normal pancreas revealed a relatively homogeneous elastographic pattern. Thirty-one focal lesions, including 30 neoplasms and most of the chronically inflamed pancreata had a honeycomb pattern dominated by hard strands. This pattern showed analogies to the histologic structure of 10 resected tumors. Other patients with chronic pancreatitis and those with hyperechoic healthy pancreas had miscellaneous elastographic appearances. Conclusions: EUS elastography of the pancreas is feasible and produces plausible results. The examination of homogeneous tissue is impaired by the relative scale used. Chronic pancreatitis and hard tumors cannot be distinguished by elastography, probably because of their similar fibrous structure. (Gastrointest Endosc 2007;65:971-8.)
EUS has become an indispensable tool in the investigation of pancreatic diseases. It is used to localize small focal lesions, such as pancreatic endocrine tumors,1-3 to define (early) chronic pancreatitis,4-7 to predict operability of tumors,8,9 and to obtain tissue from focal lesions by EUS-guided FNA.10-13 Color flow Doppler implemented in electronic EUS devices is suitable to characterize the vascular pattern of pancreatic tumors. Nevertheless, the
Copyright ª 2007 by the American Society for Gastrointestinal Endoscopy 0016-5107/$32.00 doi:10.1016/j.gie.2006.12.057
exact clinical determination of focal lesions remains challenging, especially in cases when microscopy after EUSguided biopsy is not conclusive. Recently, strain imaging, which is also referred to as elastography, has been presented in clinical studies as a new US-associated technology.14-17 The method is designed to distinguish different tissues because of their specific consistency. Because malignant tumors tend to be harder than benign ones, this new approach could become clinically relevant. Initial clinical experiences have been promising in differentiating, by US elastography, nodules of the breast, the thyroid gland, and the prostate.14-17 Aside from the characterization of parenchymal focal lesions, US elastography (FibroScan;
www.giejournal.org
Volume 65, No. 7 : 2007 GASTROINTESTINAL ENDOSCOPY 971
EUS elastography of the pancreas
Janssen et al
Capsule Summary
TABLE 1. Group demographics Group 1 No. patients
20
Group 2
Group 3
20
33
What is already known on this topic d
Median age (range), y 62.5 (23-80) 49 (24-64) 67 (43-88) Men/women
5/15
7/13
13/20
EUS elastography, designed to distinguish tissues by consistency, has been used to differentiate benign and malignant tissue in the breast, thyroid gland, and prostate, and to assess liver fibrosis.
What this study adds to our knowledge d
In a prospective single-center study of 73 patients undergoing EUS elastography of the pancreas, real-time elastographic recordings were obtained with high reliability, revealing distinct patterns in pancreatic tumors and chronic pancreatitis, although the results were less successful in tissue with a homogeneous consistency.
PATIENTS AND METHODS This prospective study was conducted from March to December 2005 in the gastroenterologic department of the HELIOS Klinikum Wuppertal, an academic center.
Patients
Echosens, Paris, France) is used to assess liver fibrosis in patients at risk.18,19 Data concerning EUS elastography were recently published as preliminary results.20,21 We investigated the feasibility and capability of this new EUS technology to characterize normal pancreatic parenchyma and pancreatic tissue affected by chronic inflammatory or various focal diseases.
EUS elastography within the study was performed during routine EUS, which may have been indicated for any reason. The study was approved by the local ethics committee. All patients gave written informed consent before the examination. The patient demographics are given in Table 1. We defined 3 groups: d Group 1: 20 patients with normal pancreas. Normal pancreas was assumed (1) when patients had no history of pancreatic disease, ie, known episodes of pancreatitis, symptoms of maldigestion, alcohol abuse, or elevated pancreas enzymes; and (2) when the B-mode EUS image did not show pathologic findings, especially none of the criteria used to define chronic pancreatitis (see group 2 below). Diffuse hyperechoic parenchyma (n Z 4) was included within this group. Indications for EUS in this group were exclusion of common bile duct stones (n Z 9), exploration of submucosal (n Z 6) or mucosal (n Z 3) masses of the upper-GI tract, and EUS-guided FNA of mediastinal lymph nodes (n Z 1) or a small liver lesion (n Z 1). d Group 2: 20 patients with chronic pancreatitis. A diagnosis of chronic pancreatitis was accepted when patients revealed a history of recurrent pancreatitis and at least 4 of the 11 EUS criteria as defined by Catalano et al.5 The median number of positive criteria was 6 (range, 4-9) per patient. Images of endoscopic retrograde pancreatography demonstrating ductal changes typical of chronic pancreatitis were available in 11 patients. Patients in this group had to be free from any suspicion of focal pancreatic disease.
972 GASTROINTESTINAL ENDOSCOPY Volume 65, No. 7 : 2007
www.giejournal.org
Figure 1. A, Elastography of the normal pancreas, revealing a relatively homogeneous green appearance according to type 1B. B, Intermittently blue clouds and blue margin effects may occur.
Janssen et al
EUS elastography of the pancreas
TABLE 2. Elastographic types in patients with chronic pancreatitis (group 2) Elastographic pattern
d
Pancreatic head
Pancreatic body
3A/B
16
10
3B/A
2
1
2A/B
–
1
2B/A
1
4
2B/A/C
1
–
1B
–
4
Group 3: 33 patients with focal pancreatic lesions. The final diagnosis had to be proven by histology. The specimens were obtained for clinical reasons, independent of this study, by EUS-guided FNA (n Z 21) or surgery (n Z 11). The technique of EUS-guided FNA is described elsewhere and provides specimens for minihistologic examination.22 In a single patient, the diagnosis of lipoma by CT densitometry was accepted without histologic proof. Three patients were excluded from the study because of inconclusive histologic results.
Examination technique We used linear EUS probes (FG 34ux and FG 38ux; Pentax Europe GmbH, Hamburg, Germany) in connection with the platform EUB 8500 (Hitachi Medical Systems GmbH, Wiesbaden, Germany). The elastography module used is generally available and is integrated into this platform. It enables real-time elastographic recording and presentation. The technology is based on the detection of small structure deformations within the B-mode image caused by compression. The degree of deformation (speckle motion) is used as an indicator for the stiffness of the tissue. Further technical details are described elsewhere.23 The elasticity of tissue within a defined sample area is translated into color, which overlays the real-time B-mode image. The colors that describe hard, intermediate, and soft tissues are blue, green/yellow, and red, respectively. The complete spectrum from blue to red encoding is applied to each elastographic record and indicates the graduation of relative elasticity within this sample area. There is no scale of absolute tissue stiffness. The suitability of the elastographic signal is indicated by a numeric scale within the image. Elastographic and B-mode images are displayed simultaneously side by side. We always used maximal sensitivity for elastographic registration. The sample area was adjusted to the region of interest; the maximal depth of the area was about 3.5 cm. The examinations were performed according to our standard procedure in the given indication,24 by the same endoscopist (J.J.). We never used a balloon to improve the www.giejournal.org
Figure 2. A, In the majority of elastographic registrations in patients with diffuse chronic pancreatitis, the honeycomb pattern type 3A/B was seen. B, Type 2 elastograms with areas of different colors (in this example type 2A/B with predominantly blue areas) were less frequent.
coupling to the wall. The probe was attached to the wall just exerting the amount of pressure appropriate for an optimal B-mode image. The B-mode frequency used in most examinations was 7.5 MHz (range, 5.0-10.0 MHz). The study intervention consisted of recording video loops (at least 20 seconds each) of significant sections of the pancreatic head and body (groups 1 and 2), or the focal lesion (group 3). The examination of the 2 pancreatic regions (groups 1 and 2) was chosen to investigate the possible influence of the surrounding tissue on the elastographic pattern of the pancreas. When examining the pancreatic head, the sample area usually was completely located within the pancreatic head, without covering the surrounding tissue, whereas surrounding tissue was normally included when the pancreatic body was explored. The following parameters were recorded in a protocol: the classification of the patient into 1 of the 3 groups; and Volume 65, No. 7 : 2007 GASTROINTESTINAL ENDOSCOPY 973
EUS elastography of the pancreas
Janssen et al
TABLE 3. Description of the focal pancreatic lesions (group 3) Anatomic position
No.
Maximal diameter (range), mm
All lesions
33
Pancreatic cancer
Elastographic pattern
Head
Body
Tail
1B
3A/B
30 (13-90)
23
8
2
2
31
24
32.5 (18-75)
16
8
–
–
24
Bile-duct cancer
2
20
2
–
–
–
2
Neuroendocrine carcinoma
1
20
1
–
–
–
1
Microcystic adenoma
3
1
–
2
–
3
Insulinoma
1
15
1
–
–
1
–
Lipoma
1
13
1
–
–
1
–
Chronic pancreatitis
1
40
1
–
–
–
1
30 (24-90)
the B-mode description of the pancreas, considering echogenicity, homogeneity, lobulation, septa, outer margin, focal lesions, and ductal changes. In focal lesions, the following parameters were recorded: size, localization, echogenicity, homogeneity, and outer delineability. The description of the elastographic pattern comprised the predominant color, the distribution of colors, and the constancy of the pattern. A comparison between the elastographic image and the histology was performed in 11 patients who underwent surgical resection.
inconstant clouds of different color may cross the region of interest, and the outer margin of the region of interest may be coded as inconstantly divergent. The elastographic colors were marked by letters: A, blue; B, green/yellow; C, red. To clearly describe groups, these letters were added to the type number in declining sequence of the proportion of color they represent within the region of interest.
Group analysis
We observed different elastographic patterns within the region of interest, which were classified as follows: type 1, relatively homogeneous; type 2, areas of 2 or 3 different colors; and type 3, honeycomb pattern. In type 1 images,
In group 1, 16 patients revealed regularly lobulated organs with echopoor or intermediately echogenic parenchyma. The elastographic images of the pancreatic head and body of these 16 patients were characterized by a relatively homogeneous green color interfered with inconstant and mostly blue clouds according to type 1B (Fig. 1). Three patients of this group had homogeneous echorich pancreata without lobulation, 1 patient had an echo-rich pancreatic head and an intermediately echogenic pancreatic body with regular lobulation, which was classified as type 1B in elastography. The 4 echorich pancreatic heads were classified as type 2A/B (n Z 1), type 2B/A (n Z 2), and type 3A/B (n Z 1); the 3 hyperechoic pancreatic bodies were classified as type 2A/B (n Z 2) and type 2B/A (n Z 1). In 2 patients with echorich pancreata, elastographic signals could be obtained up to a 1.5to 2-cm penetration depth only. This kind of limitation was seen in none of the other examinations. The elastographic patterns of the pancreatic head and body did not differ perceptibly within this group. In group 2, the elastographic patterns of patients with chronic pancreatitis were various and are listed in Table 2. The honeycomb type 3 with predominantly hard (blue) strands (type 3A/B) was seen in the head in 16 patients and in the body of the pancreas in 10 patients (Fig. 2A);
974 GASTROINTESTINAL ENDOSCOPY Volume 65, No. 7 : 2007
www.giejournal.org
Statistical analysis A sample size of at least 20 patients per group was chosen as the basis to find the possible elastographic patterns. Recruitment for group 3 was continued up to 33 patients until the last group (group 2) was completed. Elastographic findings within the 3 groups or within subgroups consisting of patients with well-defined pancreatic diseases were described. Sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy of elastographic characteristics for groups and subgroups were calculated, as appropriate.
RESULTS EUS examinations in all 73 patients were complete, and adequate elastographic recordings could be obtained in all patients according to the study design.
Definition of elastographic patterns
Janssen et al
EUS elastography of the pancreas
Figure 3. A, Small carcinoma of the distal common bile duct with an elastographic blue/green honeycomb pattern representing type 3A/B. B, Histology, showing fibrous septa (van Gieson’s stain, orig. mag. 25).
Figure 4. A, Neuroendocrine cancer with type 3A/B elastogram. B, A fibrous framework in a histologic examination (van Gieson’s stain, orig. mag. 25).
type 3B/A was less frequent (pancreatic head, n Z 2; pancreatic body, n Z 1). A minority of patients had type 2 elastograms, with differently colored areas (head, n Z 2; body, n Z 5; Fig. 2B) or the homogeneous type 1B pattern, which was found in the pancreatic body only (n Z 4). The elastograms of the pancreatic head and body were intraindividually not as concordant as in group 1. In group 3, the characteristics of the 33 focal lesions are summarized in Table 3. All pancreatic carcinomas (n Z 24), microcystic adenomas (n Z 3), bile-duct carcinomas (n Z 2), and the neuroendocrine carcinoma (n Z 1) had the same elastographic appearance, ie, the blue/green honeycomb pattern (type 3A/B; Figs. 3 to 5). One patient with a tumorlike hypoechoic enlargement of the pancreatic head caused by chronic pancreatitis also showed the type 3A/B pattern (Fig. 6). Only the insulinoma (Fig. 7) and the lipoma had a different pattern and were coded homogeneously green (type 1B).
Based on all 113 registrations, sensitivity, specificity, PPV, NPV, and accuracy are (1) for type 1B elastogram indicating normal healthy pancreas: 82.5%, 91.8%, 84.6%, 90.5%, 88.5%; (2) for type 3A/B, indicating chronic pancreatitis: 65.9%, 56.9%, 46.6%, 74.5%, 60.2%; and (3) for type 3A/B indicating neoplasia: 93.8%, 65.4%, 51.7%, 96.5%, 73.5%, respectively. The study design did not primarily aim at the comparison of the elastographic with histologic images. Nevertheless, it was striking that the histologic report of the 10 surgically resected tumors with type 3A/B elastographic appearance stressed a significant desmoplastic or sclerosing reaction in the form of strands (pancreatic carcinoma [n Z 3], bile-duct carcinoma [n Z 2], microcystic adenoma [n Z 3], neuroendocrine carcinoma [n Z 1], chronic pancreatitis [n Z 1]) (Figs. 3 to 6). Only the resected insulinoma with the type 1B elastogram did not reveal any fibrous tissue (Fig. 7). The specimens obtained by
www.giejournal.org
Volume 65, No. 7 : 2007 GASTROINTESTINAL ENDOSCOPY 975
EUS elastography of the pancreas
Figure 5. Microcystic adenoma. A, The elastographic appearance corresponding to type 3A/B. B, Histology, showing fibrous strands surrounding the cystic compartments (van Gieson’s stain, orig. mag. 25).
EUS-guided FNA were too small to estimate the amount of desmoplastic reaction in these 21 tumors, which were not surgically resected.
Janssen et al
Figure 6. Enlarged pancreatic head caused by chronic pancreatitis. The resection was performed, because malignancy was clinically suspected. A, The elastographic pattern (type 3A/B) is identical with all malignomas in this study. B, Histology, revealing the fibrous framework (van Gieson’s stain, orig. mag. 25).
The aim of this study was to test a commercially available US platform with an integrated elastography module as to its applicability in terms of pancreatic EUS. Our data show that real-time elastographic recordings can be obtained with high reliability. Impulses generated by the body itself, ie, arterial pulsations and respiratory movements, are sufficient to deform the pancreas and thus to allow strain calculations. Neither external compression nor major manipulation with the probe, which was used as a fixed counterpart, was needed. We did not change our examination habits during the study and, therefore, missed
testing the potential influence of a mounted balloon, which might absorb the mechanical impulses and lessen the organ deformations. The depth of the elastographic registration is limited the same way as in B-mode EUS, as shown by 2 patients with hyperechoic pancreata. If the resolution quality of the B-mode image is too low, subsequent calculation of the elastographic image will be impossible. The elastographic-type assignment to the head and body was intraindividually concordant in group 1, indicating that the influence of the surrounding tissue on the elastogram was negligible. The main requirement for the clinical use of elastography was to enable the characterization of tumors with high sensitivity and specificity, and thus to improve the identification of relevant, mostly malignant lesions for biopsy and/or surgery. To assess the elastographic property to discriminate
976 GASTROINTESTINAL ENDOSCOPY Volume 65, No. 7 : 2007
www.giejournal.org
DISCUSSION
Janssen et al
EUS elastography of the pancreas
malignant from benign disease, it is mandatory to study the elastographic patterns of the normal pancreas and the pancreas affected by chronic pancreatitis as well. EUS elastography of the normal pancreas, with the exception of hyperechoic pancreata, showed a relatively homogeneous green image of the organ (type 1B), providing satisfactory diagnostic accuracy (88.5%). The interpretation of these images was impaired by the fact that the elastographic image reconstruction always uses the complete color spectrum from blue to red within the sample area, even if the tissue under observation is almost homogeneous. This phenomenon causes the inconstant cloudlike and margin color effects. The presentation of homogeneous tissue by 1 homogeneous color requires an absolute scale of the elastographic tissue properties. Because the power and the extent of the arterial excursions that cause structure deformation and the distance between impulse donator and region of interest may vary from image to image, this aim could only be achieved by calibrating each image by means of a structure with known stiffness included within the sample area. This precondition can hardly be fulfilled. Nevertheless, the pattern of the normal pancreas as relatively homogeneous green is striking in real-time sequences in contrast to the loops obtained in groups 2 and 3. Single image interpretation might be misleading and must generally be avoided. It is questionable whether diffuse hyperechoic pancreata as seen in 4 patients were correctly classified in group 1 as controls. Although this finding remains without consequence in daily practice, the possibly underlying conditions, lipomatosis, and fibrosis,25 will have effects on elasticity, and an overlap toward chronic pancreatitis may be present. The elastographic pattern of most pancreatic tumors was monotonous and characterized by a blue/green honeycomb pattern (type 3A/B). Regarding the histologic results of the patients whose tumors were surgically resected, the corresponding amount of fibrous strands characterizing desmoplastic pancreatic carcinomas and microcystic adenomas26 caused the stiffness of the tumor and probably the concordant elastographic appearance. Only tumors without sclerosis (insulinoma and lipoma) were coded homogeneously green, like normal pancreatic tissue (type 1B). Chronic pancreatitis is a sclerosing disease and comprises stone formation and calcifications.27 The sclerosis arises within the interlobular septa, forming a strong fibrous framework, which corresponds to the above-discussed hard tumors and probably caused the similar elastographic pattern in patients with chronic pancreatitis (mostly type 3A/B) and malignant tumors. Thus, the specificity and PPV for the elastographic diagnosis of chronic pancreatitis (56.9%, 46.6%, respectively) and pancreatic neoplasms (65.4%, 51.7%, respectively) were low. Because the sufficient distinction between malignant tumor and chronic pancreatitis cannot be achieved by EUS elastography, this
method will not help to target malignant lesions within regions of chronic pancreatitis and improve results of EUS-guided FNA in problematic cases. In contrast to the hard tumors under investigation, the elastographic appearance of chronic pancreatitis varies to some degree. This is probably caused by the inhomogeneous affection of the organs and the different stages of the disease. However, our study was not designed to correlate the severity of chronic pancreatitis with elastographic appearances. We concluded that the module that we tested for EUS elastography of the pancreas worked reliably and brought about plausible results. The relative scale used was suitable to reveal distinct patterns in pancreatic tumors and in chronic pancreatitis. It was less suitable to examine tissue with homogeneous consistency. Only an absolute scale with an adjustable range could solve this problem. Contrary to the optimistic expectation by Giovannini
www.giejournal.org
Volume 65, No. 7 : 2007 GASTROINTESTINAL ENDOSCOPY 977
Figure 7. Insulinoma of the pancreatic head. A, The elastogram is homogeneously green. B, The tumor does not show any fibrosis at histologic examination. (van Gieson’s stain, orig. mag. 25).
EUS elastography of the pancreas
et al,20 we are convinced that EUS elastography is not able to distinguish chronic pancreatitis from various types of neoplastic and, especially, malignant tumors, probably because of the similarities in their mechanical structure. Maybe EUS elastography is suitable for the early diagnosis of chronic pancreatitis and to assess the amount of fibrous tissue in chronic pancreatitis by quantifying the proportion of hard tissue within the sample area. Further research in this field is needed. DISCLOSURES None of the authors has any disclosures to make.
REFERENCES 1. Pitre J, Soubrane O, Dousset B, et al. Pancreatic echo-endoscopy and preoperative localization of insulinomas. Ann Chir 1998;52:369-73. 2. Ardengh JC, Rosenbaum P, Ganc AJ, et al. Role of EUS in the preoperative localization of insulinomas compared with spiral CT. Gastrointest Endosc 2000;51:552-5. 3. Langer P, Kann PH, Fendrich V, et al. Prospective evaluation of imaging procedures for the detection of pancreaticoduodenal endocrine tumors in patients with multiple endocrine neoplasia type I. World J Surg 2004;28:1317-22. 4. Wiersema MJ, Hawes RH, Lehman GA, et al. Prospective evaluation of endoscopic ultrasonography and endoscopic retrograde cholangiopancreatography in patients with chronic abdominal pain of suspected pancreatic origin. Endoscopy 1993;25:555-64. 5. Catalano MF, Lahoti S, Geenen JE, et al. Prospective evaluation of endoscopic ultrasonography, endoscopic retrograde pancreatography, and secretin test in the diagnosis of chronic pancreatitis. Gastrointest Endosc 1998;48:11-7. 6. Sahai AV, Zimmerman M, Aabakken L, et al. Prospective assessment of the ability of endoscopic ultrasound to diagnose, exclude, or establish the severity of chronic pancreatitis found by endoscopic retrograde cholangiopancreaticography. Gastrointest Endosc 1998;48:18-25. 7. Kahl S, Glasbrenner B, Leodolter A, et al. EUS in the diagnosis of early chronic pancreatitis: a prospective follow-up study. Gastrointest Endosc 2002;55:507-11. 8. Gress FG, Hawes RH, Savides TJ, et al. Role of EUS in the preoperative staging of pancreatic cancer: a large single-center experience. Gastrointest Endosc 1999;50:786-91. 9. Soriano A, Castells A, Ayuso C, et al. Preoperative staging and tumor resectability assessment of pancreatic cancer: prospective study comparing endoscopic ultrasonography, helical computed tomography, magnetic resonance imaging, and angiography. Am J Gastroenterol 2004;99:492-501. 10. Fritscher-Ravens A, Izbicki JR, Sriram PV, et al. Endosonographyguided, fine-needle aspiration cytology extending the indication for organ-preserving pancreatic surgery. Am J Gastroenterol 2000;95: 2255-60. 11. Harewood GC, Wiersema MJ. Endosonography-guided fine needle aspiration biopsy in the evaluation of pancreatic masses. Am J Gastroenterol 2002;97:1386-91.
978 GASTROINTESTINAL ENDOSCOPY Volume 65, No. 7 : 2007
Janssen et al 12. Eloubeidi MA, Chen VK, Eltoum IA, et al. Endoscopic ultrasound-guided fine needle aspiration biopsy of patients with suspected pancreatic cancer: diagnostic accuracy and acute and 30-day complications. Am J Gastroenterol 2003;98:2663-8. 13. DeWitt J, Jowell P, Leblanc J, et al. EUS-guided FNA of pancreatic metastases: a multicenter experience. Gastrointest Endosc 2005;61:697-9. 14. Garra BS, Cespedes EI, Ophir J, et al. Elastography of breast lesions: initial clinical results. Radiology 1997;202:79-86. 15. Moon KM, Chang RF, Chen CJ, et al. Solid breast masses: classification with computer-aided analysis of continuous US images obtained with probe compression. Radiology 2005;236:458-64. 16. Ko¨nig K, Scheipers U, Pesavento A, et al. Initial experiences with realtime elastography guided biopsies of the prostate. J Urol 2005;174: 115-7. 17. Lyshchik A, Higashi T, Ryo A, et al. Thyroid gland tumor diagnosis at US elastography. Radiology 2005;237:202-11. 18. Caste´ra L, Vergniol J, Foucher J, et al. Prospective comparison of transient elastography, fibrotest, APRI, and liver biopsy for the assessment of fibrosis in chronic hepatitis C. Gastroenterology 2005;128: 343-50. 19. Foucher J, Chanteloup E, Vergniol J, et al. Diagnosis of cirrhosis by transient elastography (FibroScan): a prospective study. Gut 2006;55: 403-8. 20. Giovannini M, Hookey LC, Bories E, et al. Endoscopic ultrasound elastography: the first step towards virtual biopsy? Preliminary results in 49 patients. Endoscopy 2006;38:344-8. 21. Saftoiu A, Vilman P. Endoscopic ultrasound elastography: a new imaging technique for the visualization of tissue elasticity distribution. J Gastrointest Liver Dis 2006;15:161-5. 22. Janssen J, Johanns W, Luis W, et al. Clinical value of endoscopic ultrasound-guided transesophageal fine needle puncture of mediastinal lesions [German]. Dtsch Med Wochenschr 1998;123:1402-9. 23. Frey H. Real-time elastography. A new ultrasound procedure for the reconstruction of tissue elasticity [German]. Radiologe 2003;43: 850-5. 24. Janssen J, Greiner L. Examination technique and general principles. In: Ro¨sch T, Will U, Chang KJ, editors. Longitudinal endosonography: atlas and manual for use in the upper gastrointestinal tract. New York: Springer-Verlag; 2001. p. 11-9. 25. Marks WM, Filly RA, Callen PW. Ultrasonic evaluation of normal pancreatic echogenicity and its relationship to fat deposition. Radiology 1980;137:475-9. 26. Solcia E, Capella C, Klo¨ppel G. Tumors of the pancreas. In: Rosai J, Sobin LH, editors. Atlas of tumor pathology. 3rd series, fascicle 20. Washington DC: Armed Forces Institute of Pathology; 1997. p. 31-7, 64-102. 27. Klo¨ppel G, Detlefsen S, Feyerabend B. Fibrosis of the pancreas: the initial tissue damage and the resulting pattern. Virchows Arch 2004; 445:1-8.
Received February 24, 2006. Accepted December 26, 2006. Current affiliations: Medical Clinic 2 (J.J., L.G.), Institute for Pathology (E.S.), HELIOS Klinikum Wuppertal, University of Witten/Herdecke, Wuppertal, Germany. Reprint requests: Jan Janssen, MD, Medical Clinic 2, HELIOS Klinikum Wuppertal, Heusnerstrasse 40, D-42283 Wuppertal, Germany.
www.giejournal.org
EUS elastography of the pancreas: feasibility and pattern description of the normal pancreas, chronic pancreatitis, and focal pancreatic lesions ¨ rer, MD, Lucas Greiner, MD Jan Janssen, MD, Eva Schlo Wuppertal, Germany
Background: Initial clinical applications have shown that US elastography might be able to distinguish tissues because of their specific consistency. Objective: (1) To investigate the feasibility of EUS elastography of the pancreas and (2) to describe elastographic patterns of the normal pancreas and the pancreas affected by inflammatory or focal disease. Design: Prospective single-center study. Setting: Academic center of the University of Witten/Herdecke. Patients: Twenty patients with normal pancreas, 20 patients with chronic pancreatitis, and 33 patients with focal pancreatic lesion, histologically later proven in 32 of these 33 cases. Interventions: Commercially available US equipment was used. The elasticity of tissue was reconstructed in real time within a sample area and was translated into a color scale imaging relative tissue elasticity within this area. Representative loops of at least 20 seconds were recorded regarding each region of interest. Results: Adequate elastographic recordings could be obtained in all 73 patients. Patients with hypoechoic and intermediately echogenic normal pancreas revealed a relatively homogeneous elastographic pattern. Thirty-one focal lesions, including 30 neoplasms and most of the chronically inflamed pancreata had a honeycomb pattern dominated by hard strands. This pattern showed analogies to the histologic structure of 10 resected tumors. Other patients with chronic pancreatitis and those with hyperechoic healthy pancreas had miscellaneous elastographic appearances. Conclusions: EUS elastography of the pancreas is feasible and produces plausible results. The examination of homogeneous tissue is impaired by the relative scale used. Chronic pancreatitis and hard tumors cannot be distinguished by elastography, probably because of their similar fibrous structure. (Gastrointest Endosc 2007;65:971-8.)
EUS has become an indispensable tool in the investigation of pancreatic diseases. It is used to localize small focal lesions, such as pancreatic endocrine tumors,1-3 to define (early) chronic pancreatitis,4-7 to predict operability of tumors,8,9 and to obtain tissue from focal lesions by EUS-guided FNA.10-13 Color flow Doppler implemented in electronic EUS devices is suitable to characterize the vascular pattern of pancreatic tumors. Nevertheless, the
Copyright ª 2007 by the American Society for Gastrointestinal Endoscopy 0016-5107/$32.00 doi:10.1016/j.gie.2006.12.057
exact clinical determination of focal lesions remains challenging, especially in cases when microscopy after EUSguided biopsy is not conclusive. Recently, strain imaging, which is also referred to as elastography, has been presented in clinical studies as a new US-associated technology.14-17 The method is designed to distinguish different tissues because of their specific consistency. Because malignant tumors tend to be harder than benign ones, this new approach could become clinically relevant. Initial clinical experiences have been promising in differentiating, by US elastography, nodules of the breast, the thyroid gland, and the prostate.14-17 Aside from the characterization of parenchymal focal lesions, US elastography (FibroScan;
www.giejournal.org
Volume 65, No. 7 : 2007 GASTROINTESTINAL ENDOSCOPY 971
EUS elastography of the pancreas
Janssen et al
Capsule Summary
TABLE 1. Group demographics Group 1 No. patients
20
Group 2
Group 3
20
33
What is already known on this topic d
Median age (range), y 62.5 (23-80) 49 (24-64) 67 (43-88) Men/women
5/15
7/13
13/20
EUS elastography, designed to distinguish tissues by consistency, has been used to differentiate benign and malignant tissue in the breast, thyroid gland, and prostate, and to assess liver fibrosis.
What this study adds to our knowledge d
In a prospective single-center study of 73 patients undergoing EUS elastography of the pancreas, real-time elastographic recordings were obtained with high reliability, revealing distinct patterns in pancreatic tumors and chronic pancreatitis, although the results were less successful in tissue with a homogeneous consistency.
PATIENTS AND METHODS This prospective study was conducted from March to December 2005 in the gastroenterologic department of the HELIOS Klinikum Wuppertal, an academic center.
Patients
Echosens, Paris, France) is used to assess liver fibrosis in patients at risk.18,19 Data concerning EUS elastography were recently published as preliminary results.20,21 We investigated the feasibility and capability of this new EUS technology to characterize normal pancreatic parenchyma and pancreatic tissue affected by chronic inflammatory or various focal diseases.
EUS elastography within the study was performed during routine EUS, which may have been indicated for any reason. The study was approved by the local ethics committee. All patients gave written informed consent before the examination. The patient demographics are given in Table 1. We defined 3 groups: d Group 1: 20 patients with normal pancreas. Normal pancreas was assumed (1) when patients had no history of pancreatic disease, ie, known episodes of pancreatitis, symptoms of maldigestion, alcohol abuse, or elevated pancreas enzymes; and (2) when the B-mode EUS image did not show pathologic findings, especially none of the criteria used to define chronic pancreatitis (see group 2 below). Diffuse hyperechoic parenchyma (n Z 4) was included within this group. Indications for EUS in this group were exclusion of common bile duct stones (n Z 9), exploration of submucosal (n Z 6) or mucosal (n Z 3) masses of the upper-GI tract, and EUS-guided FNA of mediastinal lymph nodes (n Z 1) or a small liver lesion (n Z 1). d Group 2: 20 patients with chronic pancreatitis. A diagnosis of chronic pancreatitis was accepted when patients revealed a history of recurrent pancreatitis and at least 4 of the 11 EUS criteria as defined by Catalano et al.5 The median number of positive criteria was 6 (range, 4-9) per patient. Images of endoscopic retrograde pancreatography demonstrating ductal changes typical of chronic pancreatitis were available in 11 patients. Patients in this group had to be free from any suspicion of focal pancreatic disease.
972 GASTROINTESTINAL ENDOSCOPY Volume 65, No. 7 : 2007
www.giejournal.org
Figure 1. A, Elastography of the normal pancreas, revealing a relatively homogeneous green appearance according to type 1B. B, Intermittently blue clouds and blue margin effects may occur.
Janssen et al
EUS elastography of the pancreas
TABLE 2. Elastographic types in patients with chronic pancreatitis (group 2) Elastographic pattern
d
Pancreatic head
Pancreatic body
3A/B
16
10
3B/A
2
1
2A/B
–
1
2B/A
1
4
2B/A/C
1
–
1B
–
4
Group 3: 33 patients with focal pancreatic lesions. The final diagnosis had to be proven by histology. The specimens were obtained for clinical reasons, independent of this study, by EUS-guided FNA (n Z 21) or surgery (n Z 11). The technique of EUS-guided FNA is described elsewhere and provides specimens for minihistologic examination.22 In a single patient, the diagnosis of lipoma by CT densitometry was accepted without histologic proof. Three patients were excluded from the study because of inconclusive histologic results.
Examination technique We used linear EUS probes (FG 34ux and FG 38ux; Pentax Europe GmbH, Hamburg, Germany) in connection with the platform EUB 8500 (Hitachi Medical Systems GmbH, Wiesbaden, Germany). The elastography module used is generally available and is integrated into this platform. It enables real-time elastographic recording and presentation. The technology is based on the detection of small structure deformations within the B-mode image caused by compression. The degree of deformation (speckle motion) is used as an indicator for the stiffness of the tissue. Further technical details are described elsewhere.23 The elasticity of tissue within a defined sample area is translated into color, which overlays the real-time B-mode image. The colors that describe hard, intermediate, and soft tissues are blue, green/yellow, and red, respectively. The complete spectrum from blue to red encoding is applied to each elastographic record and indicates the graduation of relative elasticity within this sample area. There is no scale of absolute tissue stiffness. The suitability of the elastographic signal is indicated by a numeric scale within the image. Elastographic and B-mode images are displayed simultaneously side by side. We always used maximal sensitivity for elastographic registration. The sample area was adjusted to the region of interest; the maximal depth of the area was about 3.5 cm. The examinations were performed according to our standard procedure in the given indication,24 by the same endoscopist (J.J.). We never used a balloon to improve the www.giejournal.org
Figure 2. A, In the majority of elastographic registrations in patients with diffuse chronic pancreatitis, the honeycomb pattern type 3A/B was seen. B, Type 2 elastograms with areas of different colors (in this example type 2A/B with predominantly blue areas) were less frequent.
coupling to the wall. The probe was attached to the wall just exerting the amount of pressure appropriate for an optimal B-mode image. The B-mode frequency used in most examinations was 7.5 MHz (range, 5.0-10.0 MHz). The study intervention consisted of recording video loops (at least 20 seconds each) of significant sections of the pancreatic head and body (groups 1 and 2), or the focal lesion (group 3). The examination of the 2 pancreatic regions (groups 1 and 2) was chosen to investigate the possible influence of the surrounding tissue on the elastographic pattern of the pancreas. When examining the pancreatic head, the sample area usually was completely located within the pancreatic head, without covering the surrounding tissue, whereas surrounding tissue was normally included when the pancreatic body was explored. The following parameters were recorded in a protocol: the classification of the patient into 1 of the 3 groups; and Volume 65, No. 7 : 2007 GASTROINTESTINAL ENDOSCOPY 973
EUS elastography of the pancreas
Janssen et al
TABLE 3. Description of the focal pancreatic lesions (group 3) Anatomic position
No.
Maximal diameter (range), mm
All lesions
33
Pancreatic cancer
Elastographic pattern
Head
Body
Tail
1B
3A/B
30 (13-90)
23
8
2
2
31
24
32.5 (18-75)
16
8
–
–
24
Bile-duct cancer
2
20
2
–
–
–
2
Neuroendocrine carcinoma
1
20
1
–
–
–
1
Microcystic adenoma
3
1
–
2
–
3
Insulinoma
1
15
1
–
–
1
–
Lipoma
1
13
1
–
–
1
–
Chronic pancreatitis
1
40
1
–
–
–
1
30 (24-90)
the B-mode description of the pancreas, considering echogenicity, homogeneity, lobulation, septa, outer margin, focal lesions, and ductal changes. In focal lesions, the following parameters were recorded: size, localization, echogenicity, homogeneity, and outer delineability. The description of the elastographic pattern comprised the predominant color, the distribution of colors, and the constancy of the pattern. A comparison between the elastographic image and the histology was performed in 11 patients who underwent surgical resection.
inconstant clouds of different color may cross the region of interest, and the outer margin of the region of interest may be coded as inconstantly divergent. The elastographic colors were marked by letters: A, blue; B, green/yellow; C, red. To clearly describe groups, these letters were added to the type number in declining sequence of the proportion of color they represent within the region of interest.
Group analysis
We observed different elastographic patterns within the region of interest, which were classified as follows: type 1, relatively homogeneous; type 2, areas of 2 or 3 different colors; and type 3, honeycomb pattern. In type 1 images,
In group 1, 16 patients revealed regularly lobulated organs with echopoor or intermediately echogenic parenchyma. The elastographic images of the pancreatic head and body of these 16 patients were characterized by a relatively homogeneous green color interfered with inconstant and mostly blue clouds according to type 1B (Fig. 1). Three patients of this group had homogeneous echorich pancreata without lobulation, 1 patient had an echo-rich pancreatic head and an intermediately echogenic pancreatic body with regular lobulation, which was classified as type 1B in elastography. The 4 echorich pancreatic heads were classified as type 2A/B (n Z 1), type 2B/A (n Z 2), and type 3A/B (n Z 1); the 3 hyperechoic pancreatic bodies were classified as type 2A/B (n Z 2) and type 2B/A (n Z 1). In 2 patients with echorich pancreata, elastographic signals could be obtained up to a 1.5to 2-cm penetration depth only. This kind of limitation was seen in none of the other examinations. The elastographic patterns of the pancreatic head and body did not differ perceptibly within this group. In group 2, the elastographic patterns of patients with chronic pancreatitis were various and are listed in Table 2. The honeycomb type 3 with predominantly hard (blue) strands (type 3A/B) was seen in the head in 16 patients and in the body of the pancreas in 10 patients (Fig. 2A);
974 GASTROINTESTINAL ENDOSCOPY Volume 65, No. 7 : 2007
www.giejournal.org
Statistical analysis A sample size of at least 20 patients per group was chosen as the basis to find the possible elastographic patterns. Recruitment for group 3 was continued up to 33 patients until the last group (group 2) was completed. Elastographic findings within the 3 groups or within subgroups consisting of patients with well-defined pancreatic diseases were described. Sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy of elastographic characteristics for groups and subgroups were calculated, as appropriate.
RESULTS EUS examinations in all 73 patients were complete, and adequate elastographic recordings could be obtained in all patients according to the study design.
Definition of elastographic patterns
Janssen et al
EUS elastography of the pancreas
Figure 3. A, Small carcinoma of the distal common bile duct with an elastographic blue/green honeycomb pattern representing type 3A/B. B, Histology, showing fibrous septa (van Gieson’s stain, orig. mag. 25).
Figure 4. A, Neuroendocrine cancer with type 3A/B elastogram. B, A fibrous framework in a histologic examination (van Gieson’s stain, orig. mag. 25).
type 3B/A was less frequent (pancreatic head, n Z 2; pancreatic body, n Z 1). A minority of patients had type 2 elastograms, with differently colored areas (head, n Z 2; body, n Z 5; Fig. 2B) or the homogeneous type 1B pattern, which was found in the pancreatic body only (n Z 4). The elastograms of the pancreatic head and body were intraindividually not as concordant as in group 1. In group 3, the characteristics of the 33 focal lesions are summarized in Table 3. All pancreatic carcinomas (n Z 24), microcystic adenomas (n Z 3), bile-duct carcinomas (n Z 2), and the neuroendocrine carcinoma (n Z 1) had the same elastographic appearance, ie, the blue/green honeycomb pattern (type 3A/B; Figs. 3 to 5). One patient with a tumorlike hypoechoic enlargement of the pancreatic head caused by chronic pancreatitis also showed the type 3A/B pattern (Fig. 6). Only the insulinoma (Fig. 7) and the lipoma had a different pattern and were coded homogeneously green (type 1B).
Based on all 113 registrations, sensitivity, specificity, PPV, NPV, and accuracy are (1) for type 1B elastogram indicating normal healthy pancreas: 82.5%, 91.8%, 84.6%, 90.5%, 88.5%; (2) for type 3A/B, indicating chronic pancreatitis: 65.9%, 56.9%, 46.6%, 74.5%, 60.2%; and (3) for type 3A/B indicating neoplasia: 93.8%, 65.4%, 51.7%, 96.5%, 73.5%, respectively. The study design did not primarily aim at the comparison of the elastographic with histologic images. Nevertheless, it was striking that the histologic report of the 10 surgically resected tumors with type 3A/B elastographic appearance stressed a significant desmoplastic or sclerosing reaction in the form of strands (pancreatic carcinoma [n Z 3], bile-duct carcinoma [n Z 2], microcystic adenoma [n Z 3], neuroendocrine carcinoma [n Z 1], chronic pancreatitis [n Z 1]) (Figs. 3 to 6). Only the resected insulinoma with the type 1B elastogram did not reveal any fibrous tissue (Fig. 7). The specimens obtained by
www.giejournal.org
Volume 65, No. 7 : 2007 GASTROINTESTINAL ENDOSCOPY 975
EUS elastography of the pancreas
Figure 5. Microcystic adenoma. A, The elastographic appearance corresponding to type 3A/B. B, Histology, showing fibrous strands surrounding the cystic compartments (van Gieson’s stain, orig. mag. 25).
EUS-guided FNA were too small to estimate the amount of desmoplastic reaction in these 21 tumors, which were not surgically resected.
Janssen et al
Figure 6. Enlarged pancreatic head caused by chronic pancreatitis. The resection was performed, because malignancy was clinically suspected. A, The elastographic pattern (type 3A/B) is identical with all malignomas in this study. B, Histology, revealing the fibrous framework (van Gieson’s stain, orig. mag. 25).
The aim of this study was to test a commercially available US platform with an integrated elastography module as to its applicability in terms of pancreatic EUS. Our data show that real-time elastographic recordings can be obtained with high reliability. Impulses generated by the body itself, ie, arterial pulsations and respiratory movements, are sufficient to deform the pancreas and thus to allow strain calculations. Neither external compression nor major manipulation with the probe, which was used as a fixed counterpart, was needed. We did not change our examination habits during the study and, therefore, missed
testing the potential influence of a mounted balloon, which might absorb the mechanical impulses and lessen the organ deformations. The depth of the elastographic registration is limited the same way as in B-mode EUS, as shown by 2 patients with hyperechoic pancreata. If the resolution quality of the B-mode image is too low, subsequent calculation of the elastographic image will be impossible. The elastographic-type assignment to the head and body was intraindividually concordant in group 1, indicating that the influence of the surrounding tissue on the elastogram was negligible. The main requirement for the clinical use of elastography was to enable the characterization of tumors with high sensitivity and specificity, and thus to improve the identification of relevant, mostly malignant lesions for biopsy and/or surgery. To assess the elastographic property to discriminate
976 GASTROINTESTINAL ENDOSCOPY Volume 65, No. 7 : 2007
www.giejournal.org
DISCUSSION
Janssen et al
EUS elastography of the pancreas
malignant from benign disease, it is mandatory to study the elastographic patterns of the normal pancreas and the pancreas affected by chronic pancreatitis as well. EUS elastography of the normal pancreas, with the exception of hyperechoic pancreata, showed a relatively homogeneous green image of the organ (type 1B), providing satisfactory diagnostic accuracy (88.5%). The interpretation of these images was impaired by the fact that the elastographic image reconstruction always uses the complete color spectrum from blue to red within the sample area, even if the tissue under observation is almost homogeneous. This phenomenon causes the inconstant cloudlike and margin color effects. The presentation of homogeneous tissue by 1 homogeneous color requires an absolute scale of the elastographic tissue properties. Because the power and the extent of the arterial excursions that cause structure deformation and the distance between impulse donator and region of interest may vary from image to image, this aim could only be achieved by calibrating each image by means of a structure with known stiffness included within the sample area. This precondition can hardly be fulfilled. Nevertheless, the pattern of the normal pancreas as relatively homogeneous green is striking in real-time sequences in contrast to the loops obtained in groups 2 and 3. Single image interpretation might be misleading and must generally be avoided. It is questionable whether diffuse hyperechoic pancreata as seen in 4 patients were correctly classified in group 1 as controls. Although this finding remains without consequence in daily practice, the possibly underlying conditions, lipomatosis, and fibrosis,25 will have effects on elasticity, and an overlap toward chronic pancreatitis may be present. The elastographic pattern of most pancreatic tumors was monotonous and characterized by a blue/green honeycomb pattern (type 3A/B). Regarding the histologic results of the patients whose tumors were surgically resected, the corresponding amount of fibrous strands characterizing desmoplastic pancreatic carcinomas and microcystic adenomas26 caused the stiffness of the tumor and probably the concordant elastographic appearance. Only tumors without sclerosis (insulinoma and lipoma) were coded homogeneously green, like normal pancreatic tissue (type 1B). Chronic pancreatitis is a sclerosing disease and comprises stone formation and calcifications.27 The sclerosis arises within the interlobular septa, forming a strong fibrous framework, which corresponds to the above-discussed hard tumors and probably caused the similar elastographic pattern in patients with chronic pancreatitis (mostly type 3A/B) and malignant tumors. Thus, the specificity and PPV for the elastographic diagnosis of chronic pancreatitis (56.9%, 46.6%, respectively) and pancreatic neoplasms (65.4%, 51.7%, respectively) were low. Because the sufficient distinction between malignant tumor and chronic pancreatitis cannot be achieved by EUS elastography, this
method will not help to target malignant lesions within regions of chronic pancreatitis and improve results of EUS-guided FNA in problematic cases. In contrast to the hard tumors under investigation, the elastographic appearance of chronic pancreatitis varies to some degree. This is probably caused by the inhomogeneous affection of the organs and the different stages of the disease. However, our study was not designed to correlate the severity of chronic pancreatitis with elastographic appearances. We concluded that the module that we tested for EUS elastography of the pancreas worked reliably and brought about plausible results. The relative scale used was suitable to reveal distinct patterns in pancreatic tumors and in chronic pancreatitis. It was less suitable to examine tissue with homogeneous consistency. Only an absolute scale with an adjustable range could solve this problem. Contrary to the optimistic expectation by Giovannini
www.giejournal.org
Volume 65, No. 7 : 2007 GASTROINTESTINAL ENDOSCOPY 977
Figure 7. Insulinoma of the pancreatic head. A, The elastogram is homogeneously green. B, The tumor does not show any fibrosis at histologic examination. (van Gieson’s stain, orig. mag. 25).
EUS elastography of the pancreas
et al,20 we are convinced that EUS elastography is not able to distinguish chronic pancreatitis from various types of neoplastic and, especially, malignant tumors, probably because of the similarities in their mechanical structure. Maybe EUS elastography is suitable for the early diagnosis of chronic pancreatitis and to assess the amount of fibrous tissue in chronic pancreatitis by quantifying the proportion of hard tissue within the sample area. Further research in this field is needed. DISCLOSURES None of the authors has any disclosures to make.
REFERENCES 1. Pitre J, Soubrane O, Dousset B, et al. Pancreatic echo-endoscopy and preoperative localization of insulinomas. Ann Chir 1998;52:369-73. 2. Ardengh JC, Rosenbaum P, Ganc AJ, et al. Role of EUS in the preoperative localization of insulinomas compared with spiral CT. Gastrointest Endosc 2000;51:552-5. 3. Langer P, Kann PH, Fendrich V, et al. Prospective evaluation of imaging procedures for the detection of pancreaticoduodenal endocrine tumors in patients with multiple endocrine neoplasia type I. World J Surg 2004;28:1317-22. 4. Wiersema MJ, Hawes RH, Lehman GA, et al. Prospective evaluation of endoscopic ultrasonography and endoscopic retrograde cholangiopancreatography in patients with chronic abdominal pain of suspected pancreatic origin. Endoscopy 1993;25:555-64. 5. Catalano MF, Lahoti S, Geenen JE, et al. Prospective evaluation of endoscopic ultrasonography, endoscopic retrograde pancreatography, and secretin test in the diagnosis of chronic pancreatitis. Gastrointest Endosc 1998;48:11-7. 6. Sahai AV, Zimmerman M, Aabakken L, et al. Prospective assessment of the ability of endoscopic ultrasound to diagnose, exclude, or establish the severity of chronic pancreatitis found by endoscopic retrograde cholangiopancreaticography. Gastrointest Endosc 1998;48:18-25. 7. Kahl S, Glasbrenner B, Leodolter A, et al. EUS in the diagnosis of early chronic pancreatitis: a prospective follow-up study. Gastrointest Endosc 2002;55:507-11. 8. Gress FG, Hawes RH, Savides TJ, et al. Role of EUS in the preoperative staging of pancreatic cancer: a large single-center experience. Gastrointest Endosc 1999;50:786-91. 9. Soriano A, Castells A, Ayuso C, et al. Preoperative staging and tumor resectability assessment of pancreatic cancer: prospective study comparing endoscopic ultrasonography, helical computed tomography, magnetic resonance imaging, and angiography. Am J Gastroenterol 2004;99:492-501. 10. Fritscher-Ravens A, Izbicki JR, Sriram PV, et al. Endosonographyguided, fine-needle aspiration cytology extending the indication for organ-preserving pancreatic surgery. Am J Gastroenterol 2000;95: 2255-60. 11. Harewood GC, Wiersema MJ. Endosonography-guided fine needle aspiration biopsy in the evaluation of pancreatic masses. Am J Gastroenterol 2002;97:1386-91.
978 GASTROINTESTINAL ENDOSCOPY Volume 65, No. 7 : 2007
Janssen et al 12. Eloubeidi MA, Chen VK, Eltoum IA, et al. Endoscopic ultrasound-guided fine needle aspiration biopsy of patients with suspected pancreatic cancer: diagnostic accuracy and acute and 30-day complications. Am J Gastroenterol 2003;98:2663-8. 13. DeWitt J, Jowell P, Leblanc J, et al. EUS-guided FNA of pancreatic metastases: a multicenter experience. Gastrointest Endosc 2005;61:697-9. 14. Garra BS, Cespedes EI, Ophir J, et al. Elastography of breast lesions: initial clinical results. Radiology 1997;202:79-86. 15. Moon KM, Chang RF, Chen CJ, et al. Solid breast masses: classification with computer-aided analysis of continuous US images obtained with probe compression. Radiology 2005;236:458-64. 16. Ko¨nig K, Scheipers U, Pesavento A, et al. Initial experiences with realtime elastography guided biopsies of the prostate. J Urol 2005;174: 115-7. 17. Lyshchik A, Higashi T, Ryo A, et al. Thyroid gland tumor diagnosis at US elastography. Radiology 2005;237:202-11. 18. Caste´ra L, Vergniol J, Foucher J, et al. Prospective comparison of transient elastography, fibrotest, APRI, and liver biopsy for the assessment of fibrosis in chronic hepatitis C. Gastroenterology 2005;128: 343-50. 19. Foucher J, Chanteloup E, Vergniol J, et al. Diagnosis of cirrhosis by transient elastography (FibroScan): a prospective study. Gut 2006;55: 403-8. 20. Giovannini M, Hookey LC, Bories E, et al. Endoscopic ultrasound elastography: the first step towards virtual biopsy? Preliminary results in 49 patients. Endoscopy 2006;38:344-8. 21. Saftoiu A, Vilman P. Endoscopic ultrasound elastography: a new imaging technique for the visualization of tissue elasticity distribution. J Gastrointest Liver Dis 2006;15:161-5. 22. Janssen J, Johanns W, Luis W, et al. Clinical value of endoscopic ultrasound-guided transesophageal fine needle puncture of mediastinal lesions [German]. Dtsch Med Wochenschr 1998;123:1402-9. 23. Frey H. Real-time elastography. A new ultrasound procedure for the reconstruction of tissue elasticity [German]. Radiologe 2003;43: 850-5. 24. Janssen J, Greiner L. Examination technique and general principles. In: Ro¨sch T, Will U, Chang KJ, editors. Longitudinal endosonography: atlas and manual for use in the upper gastrointestinal tract. New York: Springer-Verlag; 2001. p. 11-9. 25. Marks WM, Filly RA, Callen PW. Ultrasonic evaluation of normal pancreatic echogenicity and its relationship to fat deposition. Radiology 1980;137:475-9. 26. Solcia E, Capella C, Klo¨ppel G. Tumors of the pancreas. In: Rosai J, Sobin LH, editors. Atlas of tumor pathology. 3rd series, fascicle 20. Washington DC: Armed Forces Institute of Pathology; 1997. p. 31-7, 64-102. 27. Klo¨ppel G, Detlefsen S, Feyerabend B. Fibrosis of the pancreas: the initial tissue damage and the resulting pattern. Virchows Arch 2004; 445:1-8.
Received February 24, 2006. Accepted December 26, 2006. Current affiliations: Medical Clinic 2 (J.J., L.G.), Institute for Pathology (E.S.), HELIOS Klinikum Wuppertal, University of Witten/Herdecke, Wuppertal, Germany. Reprint requests: Jan Janssen, MD, Medical Clinic 2, HELIOS Klinikum Wuppertal, Heusnerstrasse 40, D-42283 Wuppertal, Germany.
www.giejournal.org