- Email: [email protected]
Acoustic radiation force impulse shear wave elastography (ARFI) of acute and chronic pancreatitis and pancreatic tumor
European Journal of Radiology 85 (2016) 2211–2216
Contents lists available at ScienceDirect
European Journal of Radiology journal homepage: www.elsevier.com/locate/ejrad
Acoustic radiation force impulse shear wave elastography (ARFI) of acute and chronic pancreatitis and pancreatic tumor Ruediger S. Goertz (MD) ∗ , Johanna Schuderer, Deike Strobel (MD), Lukas Pfeifer (MD), Markus F. Neurath (MD) (Prof.), Dane Wildner (MD) Department of Internal Medicine 1, University Erlangen-Nuernberg, Ulmenweg 18, 91054 Erlangen, Germany
a r t i c l e
i n f o
Article history: Received 9 May 2016 Received in revised form 19 September 2016 Accepted 18 October 2016 Keywords: Pancreas Pancreatitis Carcinoma ARFI Elastography
a b s t r a c t Introduction: Acoustic Radiation Force Impulse (ARFI) elastography evaluates tissue stiffness noninvasively and has rarely been applied to pancreas examinations so far. In a prospective and retrospective analysis, ARFI shear wave velocities of healthy parenchyma, pancreatic lipomatosis, acute and chronic pancreatitis, adenocarcinoma and neuroendocrine tumor (NET) of the pancreas were evaluated and compared. Material and methods: In 95 patients ARFI elastography of the pancreatic head, and also of the tail for a specific group, was analysed retrospectively. Additionally, prospectively in 100 patients ARFI was performed in the head and tail of the pancreas. Results: A total of 195 patients were included in the study. Healthy parenchyma (n = 21) and lipomatosis (n = 30) showed similar shear wave velocities of about 1.3 m/s. Acute pancreatitis (n = 35), chronic pancreatitis (n = 53) and adenocarcinoma (n = 52) showed consecutively increasing ARFI values, respectively. NET (n = 4) revealed the highest shear wave velocities amounting to 3.62 m/s. ARFI elastography showed relevant differences between acute pancreatitis and chronic pancreatitis or adenocarcinoma. With a cut-off value of 1.74 m/s for the diagnosis of a malignant disease the sensitivity was 91.1% whereas the specificity amounted to 60.4%. Conclusion: ARFI shear wave velocities present differences in various pathologies of the pancreas. Acute and chronic pancreatitis as well as neoplastic lesions show high ARFI values. Very high elasticity values may indicate malignant disease of the pancreas. However, there is a considerable overlap between the entities. © 2016 Elsevier Ireland Ltd. All rights reserved.
1. Introduction Ultrasound is the most commonly used initial method to evaluate abdominal pain or pathologic laboratory data. Furthermore, it is utilized in screening situations. During transabdominal ultrasound, the pancreas is regularly visualized and abnormalities, such
Abbreviations: ARFI, acoustic radiation force impulse; AUROC, area under the receiver operating characteristic; BMI, body mass index; CT, computed tomography; ERCP, endoscopic retrograde cholangiopancreaticography; EUS, endoscopic ultrasound; IQR, interquartile range; MRE, magnetic resonance elastography; MRI, magnetic resonance imaging; NET, neuroendocrine tumor; SD, standard deviation. ∗ Corresponding author. E-mail addresses: [email protected] (R.S. Goertz), Johanna@schuderer-floss.de (J. Schuderer), [email protected] (D. Strobel), [email protected] (L. Pfeifer), [email protected] (M.F. Neurath), [email protected] (D. Wildner). http://dx.doi.org/10.1016/j.ejrad.2016.10.019 0720-048X/© 2016 Elsevier Ireland Ltd. All rights reserved.
as pancreatic lipomatosis, acute or chronic pancreatitis and pancreatic lesions, can be delineated. In the daily routine, differentiation between early diseases and normal findings is often challenging. Elastographic ultrasound techniques are suitable to assess elastic properties of a tissue non-invasively. Hard focal pancreatic lesions occurring on elastography during endoscopic ultrasound (EUS) are suspicious for pancreatic carcinoma [1], however differentiation from chronic pancreatitis is not regularly possible [2]. Nevertheless, elastography during EUS can be useful as a complementary tool for early detection of chronic pancreatitis or pancreatic cancer, for the characterization of focal pancreatic lesions and for guiding biopsy in hard areas [3–5]. The peak shear wave elastography Acoustic Radiation Force Impulse (ARFI) cannot be performed by EUS up to now. The ARFI technique uses transabdominal short-duration acoustic radiation forces to generate localized displacements in tissue. These displacements result in shear-wave propagation away from the region of excitation and are tracked using ultrasonic
2212
R.S. Goertz et al. / European Journal of Radiology 85 (2016) 2211–2216
correlation-based methods. The tissue response to these acoustic forces can be monitored both spatially and temporally. Displacement magnitude is inversely proportional and shear-wave velocity is directly proportional to local tissue stiffness. A single transducer is used both for applying localized radiation forces within tissue for short time periods and for tracking the resulting tissue displacements. Results are expressed in meters per second (m/s), this shear-wave propagation velocity is proportional to the square root of tissue elasticity. Recently, studies using the transabdominal approach by ARFI elastography evaluated different abdominal organs [6]. Especially in pancreatic diseases, such as acute, chronic pancreatitis or pancreatic carcinoma, relevant higher shear wave velocities than in healthy pancreas parenchyma are described [7–9]. This study aimed at analyzing the feasibility and performance of ARFI shear wave elastography in various pancreatic diseases, such as pancreatic lipomatosis, acute pancreatitis, chronic pancreatitis and pancreatic masses, respectively. For that purpose, a retrospective and prospective evaluation of pancreatic ARFI data was performed including additively healthy pancreas and neuroendocrine tumors (NET). 2. Material and methods Between September 2009 and August 2012 pancreas elastography results of healthy organs, acute and chronic pancreatitis, adenocarcinoma and NET of 95 patients were retrospectively analyzed. Then between February 2014 and June 2015 100 patients with hyperechoic pancreas, acute pancreatitis, chronic pancreatitis or pancreatic carcinoma were prospectively enrolled including measurement depth and levels of CRP and lipase. All patients were evaluated for demographic data (age, gender, body mass index (BMI)). The study complies with the ethics guidelines of the Helsinki Declaration and informed consent has been obtained from each patient. 2.1. ARFI measurements The ultrasound system Acuson S2000 (Siemens Medical Solutions, Erlangen, Germany) was used in the Virtual Touch Tissue Quantification mode to perform an abdominal sonography with ARFI measurements. In the prospective evaluation a region of interest (10 × 5 mm) was set to the head/corpus area as well as to the corpus/tail of the pancreas during real-time B-mode imaging (Fig. 1a and b). In patients with adenocarcinoma, the lesion and the surrounding parenchyma were assessed by ARFI. In the retrospective analysis, ARFI measurements were recorded only in the head, except for the healthy group in which we evaluated both locations (head and tail). All measurements were conducted in an axial plane from an epigastric position during a relaxed pause in breathing. For the retrospective evaluation an inclusion criterion of at least 4 measurements was applied (by use of transducer 4C1). Within the prospective evaluation process 10 measurements were performed by means of the curved array transducer 6C1HD. Measurements were performed by two experienced operators (DEGUM qualification level 2). In the prospective group the number of failed measurements (“X.XX” m/s) was recorded. 2.2. Confirmation of diagnosis Healthy pancreas was assumed due to lack of epigastric symptoms or diarrhea, normal levels of lipase and normal transabdominal ultrasound. Diagnosis of pancreatic lipomatosis was based upon hyperechoic parenchyma on B-mode ultrasound plus evidence of hepatic steatosis (hyperechoic liver compared to the
Fig. 1. Axial plane of the epigastric region: measurement of ARFI elastography of a acute pancreatitis in the corpus/tail and b adenocarcinoma in the head of the pancreas.
kidney parenchyma, dorsal attenuation and/or blurring vessel borders). Patients with acute pancreatitis presented with typical epigastric pain and elevation of lipase (at least 3x the upper limit of normal). Chronic pancreatitis was diagnosed by the medical history and typical abdominal ultrasound findings (irregular pancreatic duct or surface or calcifications) or results of further imaging techniques EUS, magnetic resonance imaging (MRI) or endoscopic retrograde cholangiopancreatography (ERCP), according to general recommendations [10]: pathological side branches, cystic changes, heterogeneous parenchymal structure, calcifications and duct irregularities or strictures. Malignant disease (adenocarcinoma or neuroendocrine carcinoma) has been proved by histopathology in every case. 2.3. Statistical analysis Clinical and laboratory characteristics of patients as well as the ARFI values were expressed as the mean ± standard deviation (SD) and the range indicated in brackets. The interquartile range (IQR) is a measure of statistical dispersion, expressing the difference between the upper and lower quartiles. The Spearman’s correlation coefficient (r) was used for analysis of relationships between variables and the t-test to compare means. The results were illustrated in tables or in a boxplot, with median as a thick line passing through each box, which represents interquartile range within which 50% of the values are located. Error bars mark minimum and maximum values (range). Small asterisks mark outliners. For calculation of the cut-off ARFI value the Youden-Index was applied and the diagnostic performance of the area under the receiver operating characteristic (AUROC) was calculated. A p-value <0.05 indicated a significant correlation or difference. All reported p-values are two-sided. Sta-
R.S. Goertz et al. / European Journal of Radiology 85 (2016) 2211–2216
tistical analysis was performed using the Statistical Package for the Social Sciences (version 19.0.0.1, IBM SPSS statistics, New York, USA). 3. Results A total of 100 patients (41 female, 59 male, mean age 59 years, BMI 26.6 kg/m2 ) were evaluated in the prospective group and 95 patients (45 female, 50 male, mean age 58 years, BMI 25.5 kg/m2 ) in the course of the retrospective analysis. 3.1. Retrospective evaluation Results of the retrospective analysis are demonstrated in Table 1. On average 7.4 ± 3.1 ARFI measurements were performed, data (n = 5) for measurement depth was missing. Concerning the head and tail in healthy parenchyma, there was no ARFI difference. Mean ARFI shear wave velocities and SD were higher in acute, chronic pancreatitis and adenocarcinoma than in healthy parenchyma (p = 0.002, p < 0.001 and p < 0.001, respectively). Acute pancreatitis showed lower ARFI values than those of chronic pancreatitis (p = 0.13) and adenocarcinoma (p = 0.005). ARFI measurements have even shown a significant difference between chronic pancreatitis and adenocarcinoma (p = 0.045). In addition to the already cited pancreatic diseases, it was possible to assess high ARFI values of four NET. These were significantly higher than those of adenocarcinomas (p = 0.035). Tumor size was 4.3 ± 1.6 cm for adenocarcinoma and 5.3 ± 2.8 cm for NET. In patients with healthy parenchyma the correlations of shear wave velocities with age, measurement depth and BMI or SD with age showed no significance. ARFI elastography of normal pancreas appeared to be slightly lower in men (1.2 ± 0.16 m/s) than in women (1.30 ± 0.25 m/s), without statistical significance (p = 0.222). 3.2. Prospective evaluation In the prospective group the patients with pancreatic carcinoma showed a higher mean age than those without pancreatic adenocarcinoma (p = 0.0001). Patients with pancreatic lipomatosis and fatty liver had a higher BMI than those with acute pancreatitis – p = 0.0004 (see Table 2). Etiologically, acute pancreatitis in 21 patients was biliary (n = 10), alcoholic intoxication (n = 5), autoimmunity (n = 2), post-ERCP (n = 1) and unknown (n = 3). The etiology of chronic pancreatitis was alcoholic (n = 15), idiopathic (n = 8), pancreas divisum (n = 3) or others (hereditary, autoimmune). Imaging in chronic pancreatitis was based on abdominal ultrasound and/or CT (n = 10), ERCP (n = 9), EUS (n = 6) or MRI (n = 2). The pancreatic adenocarcinoma had a mean size of 3.7 ± 1.4 cm (range 2.0–8.3). In 6/21 cases ARFI could not visualize or measure pancreatic parenchyma in the tail nearby carcinoma. Table 3 gives an overview of the detailed shear wave velocities of the different pancreas entities. Subsequently pancreatic lipomatosis, acute pancreatitis, chronic pancreatitis and adenocarcinoma show rising ARFI values with higher SD. The differences of mean ARFI shear wave velocities between the lower values of pancreatic lipomatosis and the higher values of acute pancreatitis or of chronic pancreatitis or of adenocarcinoma were statistically significant (p = 0.018, p < 0.001 and p < 0.001, respectively). Moreover, ARFI values of acute pancreatitis were significantly lower, in comparison to those of chronic pancreatitis (p = 0.02) and adenocarcinoma (p < 0.001). On the other hand, ARFI measurements between chronic pancreatitis and adenocarcinoma showed no significant differences (p = 0.43). 19/28 patients with chronic pancreatitis had calcifications. No difference of the respective ARFI values between calcified and non-calcified chronic pancreatitis
2213
was found (p = 0.306). There were no differences in measurements between head or tail in lipomatosis and acute or chronic inflammation of the pancreas (p = 0.46, p = 0.22 and p = 0.46, respectively). During the evaluation of pancreatic adenocarcinoma, in 15 patients, it was possible to evaluate both sites: ARFI elastography of adenocarcinomas (2.35 ± 0.52 m/s) showed always higher values than the surrounding parenchyma (1.53 ± 0.45 m/s) – p = 0.0001 (Table 4). The mean percentage of failed measurements in the head was 7.4%, 8.0%, 14.0% and 25.5% for pancreatic lipomatosis, acute pancreatitis, chronic pancreatitis and pancreatic carcinoma, respectively. The ARFI values of acute pancreatitis showed no significant correlation with lipase levels (r = −0.252 and p = 0.271). In patients with lipomatosis of the pancreas, there was no correlation of ARFI elastography in head and tail with measurement depth, BMI or age. ARFI shear wave velocities between normal and pancreatic lipomatosis were not significant (p = 0.763). A cut-off analysis of all 195 patients (see Fig. 2) showed a cutoff value of 1.74 m/s for the diagnosis of a malignant disease, with a sensitivity of 91.1% and a specificity of 60.4% (AUROC 78.6% with a 95%-CI 71.6–85.6%). Only malignancy reached ARFI values being >3.95 m/s (n = 6/56 malignant lesions).
4. Discussion ARFI shear wave velocity measurement is a quantitative elastographic method that quantifies elastic properties of tissue non-invasively. Shear wave velocities of the liver correlate well with hepatic fibrosis and cirrhosis [11], reflecting histopathologic architectural changes of liver tissue. Normal healthy pancreas contains lobularly arranged glandular cells with endocrine and exocrine function. In pancreatic lipomatosis, there is an increase of interstitial fatty tissue. Acute pancreatitis leads to infiltration by granulo-lymphocytic cells and disseminated necrosis of acinus cells. Chronic inflammation leads to fibrotic changes of the pancreatic parenchyma and to calcifications. In pancreatic adenocarcinoma, the normal architecture is destroyed by proliferation of epithelial gland cells, concomitantly with a surrounding desmoplastic reaction which result in stiff connective tissue. Pancreatic NET mainly consist of endocrine main cells. Initially, semi-quantitative elastography by EUS was studied to diagnose malignant disease of the pancreas [1]. Due to high sensitivity of endoscopic elastography the latter was thought to be a promising supplement to fine needle aspiration [3]. Recently, magnetic resonance elastography (MRE) showed promising results with regard to its ability to reproduce stiffness measurements throughout the whole pancreas [12]. Using a 3T MRE technique Itoh et al. [13] presented increasing higher stiffness values of the head, body and tail of the pancreas in healthy volunteers. 7 patients with pancreatic carcinoma showed a stiffness amounting to 6.06 kPa and thus being twice as high. In our study of transcutaneous pancreatic ARFI elastography, we have managed to highlight differences in various pancreatic conditions, which reflect these above-mentioned histopathologic changes of the pancreas. In contrast to healthy parenchyma, pancreatic lipomatosis, acute pancreatitis, chronic pancreatitis and adenocarcinoma showed higher ARFI shear wave velocities with increasing SD. ARFI measurements of NET have revealed the highest values. The number of measurement failure increased gradually for pancreatic lipomatosis, acute pancreatitis, chronic pancreatitis and pancreatic carcinoma, with the latter having the highest frequency of measurement failure. Normal ARFI values of the pancreas are reported to be between 1.22 and 1.30 m/s.(6, 8, 14) According to our findings, the measurements of the pancreatic body were slightly higher, however without statistical significance. An analysis of influencing factors on
2214
R.S. Goertz et al. / European Journal of Radiology 85 (2016) 2211–2216
Table 1 Retrospective analysis of ARFI shear wave velocities in patients with healthy parenchyma, acute and chronic pancreatitis, pancreatic carcinoma or neuroendocrine tumor (NET).
Sex [m/w] Age [years] BMI [kg/m2 ] Location ARFI [m/s] SD Depth[cm]
Healthy
Acute pancreatitis
Chronic pancreatitis
Adenocarcinoma
Neuroendocrine tumor
9/12 47.3 (18–82) 25.0 Head 1.3 ± 0.23 (0.97–1.78) 0.33 5.6 ± 1.4
7/7 47.2 (18–69) 27.3 Head 1.79 ± 0.62 (1.06–2.87) 0.51 5.9 ± 1.6
16/9 57.5 (20–82) 24.2 Head 2.13 ± 0.68 (1.02–3.24) 0.62 4.5 ± 1.5
14/17 70.9 (55–86) 26.5 Tumor 2.58 ± 0.91 (0.79–4.19) 0.68 5.1 ± 1.2
4/0 67.5 (58–75) 23.8 Tumor 3.62 ± 0.63 (2.73–4.11) 0.57 4.2 ± 1.2
Tail 1.34 ± 0.23 (0.96–1.92) 0.29 5.3 ± 1.0
Table 2 Patient data of all prospectively evaluated patients with pancreatic disease. Pancreatic lipomatosis
Acute pancreatitis
Chronic pancreatitis
Adenocarcinoma
CRP [<5 mg/l]
19/11 56.8 ± 12.8 (29–79) 31.2 ± 5.1 (22.7–45.6) –
Lipase [<60 U/l]
–
10/11 55.9 ± 21.6 (23–89) 26.0 ± 4.3 (18.3–33.9) 122.7 ± 126 (1–432) 2944.7 ± 4636 (194–20803)
19/9 52.9 ± 11.4 (31–78) 24.0 ± 3.3 (18.4–30.9) 50.7 ± 79 (0.5–309) 936 ± 2484 (1.4–11787)
11/10 71.6 ± 8.1 (50–84) 24.6 ± 3.8 (17.3–31.6) 60.2 ± 77 (1.3–329) 107.9 ± 286 (0–1308)
Sex [m/w] Age [years] BMI [kg/m2 ]
Table 3 Prospective ARFI shear wave velocities in the pancreatic head and tail of pancreatic lipomatosis, acute and chronic pancreatitis and pancreatic adenocarcinoma. Pancreatic lipomatosis
Acute pancreatitis
Chronic pancreatitis
Adenocarcinoma
m/w
30
21
28
21
Location
Head
Tail
Head
Tail
Head
Tail
Tumor
ARFI [m/s]
1.33 ± 0.35 (0.95–2.59) 0.36 1.22 0.53 ± 0.25 (0.14–0.99) 7.1 ± 1.1 (4.1–8.0)
1.40 ± 0.25 (1.01–2.06) 0.32 1.36 0.44 ± 0.19 (0.12–0.83) 6.8 ± 1.3 (3.4–8.0)
1.74 ± 0.71 (0.87–3.62) 0.50 1.63 0.76 ± 0.47 (0.19–1.7) 5.9 ± 1.9 (4.5–8.0)
1.53 ± 0.61 (1.03–3.2) 0.44 1.49 0.57 ± 0.33 (0.19–1.35) 5.9 ± 1.6 (3.3–8.0)
2.21 ± 0.65 (1.14–3.92) 0.62 2.19 0.95 ± 0.49 (0.22–1.91) 4.9 ± 1.0 (3.3–6.8)
2.06 ± 0.61 (0.95–3.39) 0.57 1.98 0.87 ± 0.46 (0.24–2.58) 4.7 ± 7.9 (3.1–6.1)
2.35 ± 0.58 (1.4–3.89) 0.74 2.3 1.04 ± 0.4 (0.46–1.89) 5.5 ± 1.2 (2.8–8.0)
SD Median IQR depth [cm]
Table 4 Difference between prospective ARFI shear wave velocities of pancreatic adenocarcinoma (n = 15) and peritumoral parenchyma (p = 0.0001).
ARFI [m/s] SD IQR
Pancreatic adenocarcinoma
Peritumoral parenchyma
2.35 ± 0.52 (1.4–3.45) 0.76 ± 0.18 (0.43–0.97) 1.08 ± 0.42 (0.62–1.89)
1.53 ± 0.45 (0.91–2.28) 0.47 ± 0.17 (0.21–0.79) 0.62 ± 0.36 (0.21–1.37)
pancreatic ARFI elastography of 234 healthy volunteers showed a significant difference between genders (examination of men resulting in lower values than that of women) and increasing stiffness with age – in contrast to the non-significant results we obtained with regard to these issues. Alcohol or smoking demonstrated no significant effect [14]. Another study by Kawada [15] evaluating factors associated with shear wave velocity mentioned that older age or increased alcohol intake led to higher ARFI shear wave velocities, whereas a brighter (i.e. lipomatotic) pancreas or a deeper depth of ROI showed lower shear wave velocities. ARFI measurements performed in the tail failed more often than it did in the corpus or head region. Mateen et al. [16] performed ARFI elastographic quantification in 52 healthy controls, 68 acute pancreatitis and 46 chronic pancreatitis. A single high ARFI value of 2.21 m/s was obtained even in the healthy group, however the healthy mean was nearly equal to
that of the patients with a chronic pancreatitis (1.25 m/s). Thus, the chronic pancreatitis values were obviously lower than they were in our study (2.21 m/s). Only the acute group had significantly higher ARFI shear wave velocities the mean of which amounts to 3.28 m/s, and showed individual mean values even being above 4.95 m/s, resulting in a high sensitivity and specificity (>90%, respectively) for the diagnosis of an acute pancreatitis in this study. The number of measurements was not recorded. 46 patients with chronic pancreatitis in comparison to 52 healthy volunteers were examined by Yashima et al. [8], who applied ARFI elastography in the head, body and tail of the pancreas. Chronic pancreatitis showed significantly higher ARFI values in the three pancreatic regions (1.65 to 2.09 m/s with high SD) than in the normal pancreas. Failed measurements were recorded in the tail of chronic pancreatitis in up to 65% of the patients. The inhomogeneity of the tissue in chronic pancreatitis and adenocarcinoma might account to that fact. When measuring the head and body, the measurement was not successful in 10% of the patients regularly. In our study, 68% of the patients with chronic pancreatitis suffered from calcifications. Park et al. [7] examined focal solid pancreatic lesions in 26 patients by applying ARFI tissue quantification. The difference between benignity (2.4 ± 1.1 m/s) and malignancy (3.3 ± 1.0 m/s) was not statistically significant, but the relative stiffness of malignant lesions in contrast to the adjacent parenchyma was sig-
R.S. Goertz et al. / European Journal of Radiology 85 (2016) 2211–2216
2215
Fig. 2. Boxplot of pancreatic ARFI measurements in all 195 patients with healthy parenchyma (n = 21), pancreatic lipomatosis (n = 30), acute (n = 35) or chronic (n = 53) pancreatitis, adenocarcinoma (n = 52) or NET (n = 4) of the pancreas.
nificantly higher than that of benign lesions (p = 0.011). This mean “malignant relative stiffness” being 1.5 m/s, determined by Park in pancreatic adenocarcinoma, is higher than that found out by our team, as we determined it being 0.82 m/s only. Furthermore in a different study, a significant ARFI difference between normal pancreas (n = 40) and adenocarcinoma (n = 62) was found being 1.17 m/s versus 2.74 m/s [17]. Interestingly, it was also emphasized that high ARFI values above 4.0 m/s obtained by more than one measurement were already mainly associated with adenocarcinoma. The three NET examined in this study were subsumed to non-adenocarcinoma and had, on average, equal ARFI values in comparison with adenocarcinomas, however showing a lower maximum value. In synopsis with our data, NET and adenocarcinoma show higher ARFI elastography than normal tissue, but cannot be distinguished from each other. Chronic pancreatitis and malignancy have higher ARFI values than normal tissue has. Their risk factors, i.e. alcohol consumption and patients’ age as an expression of duration of pancreatic injury leading to fibrosis, are also associated with higher values. The main limitations of our study are the non-consecutive inclusion of patients in the prospective group and the somewhat lower total number of measurements (mean 7.4) performed in the retrospective group. Besides, due to its rarity, only 4 NET could be investigated. Not all results were congruent between the retrospective and prospective analysis. In the retrospective group a significant ARFI difference between chronic pancreatitis and adenocarcinoma was demonstrated, whereas in the prospective group was not. Thus, ARFI needs a careful examination in order to achieve reproducible results. There might be some elements that affect ARFI measurements (i.e. probe frequency, the orientation of the ROI, the number of measurements). In analogy to that, a different point shear wave technique might not reveal identical results. In conclusion, it is possible to perform ARFI elastography in the head and tail of the pancreas. ARFI shear wave velocities increase
in acute and chronic pancreatitis as well as in neoplastic lesions of the pancreas. Very high ARFI values (>3.95 m/s) indicate malignant disease of the pancreas. However, there is a considerable overlap between the entities.
Conflict of interest None.
References [1] Q. Pei, X. Zou, X. Zhang, et al., Diagnostic value of EUS elastography in differentiation of benign and malignant solid pancreatic masses: a meta-analysis, Pancreatology 12 (2012) 402–408. [2] J. Janssen, I. Papavassiliou, Effect of aging and diffuse chronic pancreatitis on pancreas elasticity evaluated using semiquantitative EUS elastography, Ultraschall Med. 35 (2014) 253–258. [3] D. Cosgrove, F. Piscaglia, J. Bamber, et al., EFSUMB guidelines and recommendations on the clinical use of ultrasound elastography. Part 2: clinical applications, Ultraschall Med. 34 (2013) 238–253. [4] N. Kawada, S. Tanaka, Elastography for the pancreas: current status and future perspective, World J. Gastroenterol. 22 (2016) 3712–3724. [5] M. D’Onofrio, S. Crosara, R. De Robertis, et al., Elastography of the pancreas, Eur. J. Radiol. 83 (2014) 415–419. [6] R.S. Goertz, K. Amann, R. Heide, et al., An abdominal and thyroid status with acoustic radiation force impulse elastometry–a feasibility study: acoustic radiation force impulse elastometry of human organs, Eur. J. Radiol. 80 (2011) e226–e230. [7] M.K. Park, J. Jo, H. Kwon, et al., Usefulness of acoustic radiation force impulse elastography in the differential diagnosis of benign and malignant solid pancreatic lesions, Ultrasonography (Seoul, Korea) 33 (2014) 26–33. [8] Y. Yashima, N. Sasahira, H. Isayama, et al., Acoustic radiation force impulse elastography for noninvasive assessment of chronic pancreatitis, J. Gastroenterol. 47 (2012) 427–432. [9] C. Goya, C. Hamidi, S. Hattapoglu, et al., Use of acoustic radiation force impulse elastography to diagnose acute pancreatitis at hospital admission: comparison with sonography and computed tomography, J. Ultrasound Med. 33 (2014) 1453–1460. [10] A.G. Schreyer, M. Jung, J.F. Riemann, et al., S3 guideline for chronic pancreatitis – diagnosis, classification and therapy for the radiologist, RoFo:
2216
[11]
[12]
[13]
[14]
R.S. Goertz et al. / European Journal of Radiology 85 (2016) 2211–2216
Fortschritte auf dem Gebiete der Rontgenstrahlen und der Nuklearmedizin 186 (2014) 1002–1008. S. Bota, H. Herkner, I. Sporea, et al., Meta-analysis: ARFI elastography versus transient elastography for the evaluation of liver fibrosis, Liver Int. 33 (2013) 1138–1147. Y. Shi, K.J. Glaser, S.K. Venkatesh, E.I. Ben-Abraham, R.L. Ehman, Feasibility of using 3D MR elastography to determine pancreatic stiffness in healthy volunteers, J. Magn. Reson. Imaging: JMRI 41 (2015) 369–375. Y. Itoh, Y. Takehara, T. Kawase, et al., Feasibility of magnetic resonance elastography for the pancreas at 3T, J. Magn. Reson. Imaging: JMRI 43 (2016) 384–390. S. Stumpf, H. Jaeger, T. Graeter, et al., Influence of age, sex, body mass index, alcohol, and smoking on shear wave velocity (p-SWE) of the pancreas, Abdom. Radiol. (New York) (2016).
[15] N. Kawada, S. Tanaka, H. Uehara, et al., Potential use of point shear wave elastography for the pancreas: a single center prospective study, Eur. J. Radiol. 83 (2014) 620–624. [16] M.A. Mateen, K.A. Muheet, R.J. Mohan, et al., Evaluation of ultrasound based acoustic radiation force impulse (ARFI) and eSie touch sonoelastography for diagnosis of inflammatory pancreatic diseases, JOP: J. Pancreas 13 (2012) 36–44. [17] M. D’Onofrio, R. De Robertis, S. Crosara, et al., Acoustic radiation force impulse with shear wave speed quantification of pancreatic masses: a prospective study, Pancreatology 16 (2016) 106–109.
Contents lists available at ScienceDirect
European Journal of Radiology journal homepage: www.elsevier.com/locate/ejrad
Acoustic radiation force impulse shear wave elastography (ARFI) of acute and chronic pancreatitis and pancreatic tumor Ruediger S. Goertz (MD) ∗ , Johanna Schuderer, Deike Strobel (MD), Lukas Pfeifer (MD), Markus F. Neurath (MD) (Prof.), Dane Wildner (MD) Department of Internal Medicine 1, University Erlangen-Nuernberg, Ulmenweg 18, 91054 Erlangen, Germany
a r t i c l e
i n f o
Article history: Received 9 May 2016 Received in revised form 19 September 2016 Accepted 18 October 2016 Keywords: Pancreas Pancreatitis Carcinoma ARFI Elastography
a b s t r a c t Introduction: Acoustic Radiation Force Impulse (ARFI) elastography evaluates tissue stiffness noninvasively and has rarely been applied to pancreas examinations so far. In a prospective and retrospective analysis, ARFI shear wave velocities of healthy parenchyma, pancreatic lipomatosis, acute and chronic pancreatitis, adenocarcinoma and neuroendocrine tumor (NET) of the pancreas were evaluated and compared. Material and methods: In 95 patients ARFI elastography of the pancreatic head, and also of the tail for a specific group, was analysed retrospectively. Additionally, prospectively in 100 patients ARFI was performed in the head and tail of the pancreas. Results: A total of 195 patients were included in the study. Healthy parenchyma (n = 21) and lipomatosis (n = 30) showed similar shear wave velocities of about 1.3 m/s. Acute pancreatitis (n = 35), chronic pancreatitis (n = 53) and adenocarcinoma (n = 52) showed consecutively increasing ARFI values, respectively. NET (n = 4) revealed the highest shear wave velocities amounting to 3.62 m/s. ARFI elastography showed relevant differences between acute pancreatitis and chronic pancreatitis or adenocarcinoma. With a cut-off value of 1.74 m/s for the diagnosis of a malignant disease the sensitivity was 91.1% whereas the specificity amounted to 60.4%. Conclusion: ARFI shear wave velocities present differences in various pathologies of the pancreas. Acute and chronic pancreatitis as well as neoplastic lesions show high ARFI values. Very high elasticity values may indicate malignant disease of the pancreas. However, there is a considerable overlap between the entities. © 2016 Elsevier Ireland Ltd. All rights reserved.
1. Introduction Ultrasound is the most commonly used initial method to evaluate abdominal pain or pathologic laboratory data. Furthermore, it is utilized in screening situations. During transabdominal ultrasound, the pancreas is regularly visualized and abnormalities, such
Abbreviations: ARFI, acoustic radiation force impulse; AUROC, area under the receiver operating characteristic; BMI, body mass index; CT, computed tomography; ERCP, endoscopic retrograde cholangiopancreaticography; EUS, endoscopic ultrasound; IQR, interquartile range; MRE, magnetic resonance elastography; MRI, magnetic resonance imaging; NET, neuroendocrine tumor; SD, standard deviation. ∗ Corresponding author. E-mail addresses: [email protected] (R.S. Goertz), Johanna@schuderer-floss.de (J. Schuderer), [email protected] (D. Strobel), [email protected] (L. Pfeifer), [email protected] (M.F. Neurath), [email protected] (D. Wildner). http://dx.doi.org/10.1016/j.ejrad.2016.10.019 0720-048X/© 2016 Elsevier Ireland Ltd. All rights reserved.
as pancreatic lipomatosis, acute or chronic pancreatitis and pancreatic lesions, can be delineated. In the daily routine, differentiation between early diseases and normal findings is often challenging. Elastographic ultrasound techniques are suitable to assess elastic properties of a tissue non-invasively. Hard focal pancreatic lesions occurring on elastography during endoscopic ultrasound (EUS) are suspicious for pancreatic carcinoma [1], however differentiation from chronic pancreatitis is not regularly possible [2]. Nevertheless, elastography during EUS can be useful as a complementary tool for early detection of chronic pancreatitis or pancreatic cancer, for the characterization of focal pancreatic lesions and for guiding biopsy in hard areas [3–5]. The peak shear wave elastography Acoustic Radiation Force Impulse (ARFI) cannot be performed by EUS up to now. The ARFI technique uses transabdominal short-duration acoustic radiation forces to generate localized displacements in tissue. These displacements result in shear-wave propagation away from the region of excitation and are tracked using ultrasonic
2212
R.S. Goertz et al. / European Journal of Radiology 85 (2016) 2211–2216
correlation-based methods. The tissue response to these acoustic forces can be monitored both spatially and temporally. Displacement magnitude is inversely proportional and shear-wave velocity is directly proportional to local tissue stiffness. A single transducer is used both for applying localized radiation forces within tissue for short time periods and for tracking the resulting tissue displacements. Results are expressed in meters per second (m/s), this shear-wave propagation velocity is proportional to the square root of tissue elasticity. Recently, studies using the transabdominal approach by ARFI elastography evaluated different abdominal organs [6]. Especially in pancreatic diseases, such as acute, chronic pancreatitis or pancreatic carcinoma, relevant higher shear wave velocities than in healthy pancreas parenchyma are described [7–9]. This study aimed at analyzing the feasibility and performance of ARFI shear wave elastography in various pancreatic diseases, such as pancreatic lipomatosis, acute pancreatitis, chronic pancreatitis and pancreatic masses, respectively. For that purpose, a retrospective and prospective evaluation of pancreatic ARFI data was performed including additively healthy pancreas and neuroendocrine tumors (NET). 2. Material and methods Between September 2009 and August 2012 pancreas elastography results of healthy organs, acute and chronic pancreatitis, adenocarcinoma and NET of 95 patients were retrospectively analyzed. Then between February 2014 and June 2015 100 patients with hyperechoic pancreas, acute pancreatitis, chronic pancreatitis or pancreatic carcinoma were prospectively enrolled including measurement depth and levels of CRP and lipase. All patients were evaluated for demographic data (age, gender, body mass index (BMI)). The study complies with the ethics guidelines of the Helsinki Declaration and informed consent has been obtained from each patient. 2.1. ARFI measurements The ultrasound system Acuson S2000 (Siemens Medical Solutions, Erlangen, Germany) was used in the Virtual Touch Tissue Quantification mode to perform an abdominal sonography with ARFI measurements. In the prospective evaluation a region of interest (10 × 5 mm) was set to the head/corpus area as well as to the corpus/tail of the pancreas during real-time B-mode imaging (Fig. 1a and b). In patients with adenocarcinoma, the lesion and the surrounding parenchyma were assessed by ARFI. In the retrospective analysis, ARFI measurements were recorded only in the head, except for the healthy group in which we evaluated both locations (head and tail). All measurements were conducted in an axial plane from an epigastric position during a relaxed pause in breathing. For the retrospective evaluation an inclusion criterion of at least 4 measurements was applied (by use of transducer 4C1). Within the prospective evaluation process 10 measurements were performed by means of the curved array transducer 6C1HD. Measurements were performed by two experienced operators (DEGUM qualification level 2). In the prospective group the number of failed measurements (“X.XX” m/s) was recorded. 2.2. Confirmation of diagnosis Healthy pancreas was assumed due to lack of epigastric symptoms or diarrhea, normal levels of lipase and normal transabdominal ultrasound. Diagnosis of pancreatic lipomatosis was based upon hyperechoic parenchyma on B-mode ultrasound plus evidence of hepatic steatosis (hyperechoic liver compared to the
Fig. 1. Axial plane of the epigastric region: measurement of ARFI elastography of a acute pancreatitis in the corpus/tail and b adenocarcinoma in the head of the pancreas.
kidney parenchyma, dorsal attenuation and/or blurring vessel borders). Patients with acute pancreatitis presented with typical epigastric pain and elevation of lipase (at least 3x the upper limit of normal). Chronic pancreatitis was diagnosed by the medical history and typical abdominal ultrasound findings (irregular pancreatic duct or surface or calcifications) or results of further imaging techniques EUS, magnetic resonance imaging (MRI) or endoscopic retrograde cholangiopancreatography (ERCP), according to general recommendations [10]: pathological side branches, cystic changes, heterogeneous parenchymal structure, calcifications and duct irregularities or strictures. Malignant disease (adenocarcinoma or neuroendocrine carcinoma) has been proved by histopathology in every case. 2.3. Statistical analysis Clinical and laboratory characteristics of patients as well as the ARFI values were expressed as the mean ± standard deviation (SD) and the range indicated in brackets. The interquartile range (IQR) is a measure of statistical dispersion, expressing the difference between the upper and lower quartiles. The Spearman’s correlation coefficient (r) was used for analysis of relationships between variables and the t-test to compare means. The results were illustrated in tables or in a boxplot, with median as a thick line passing through each box, which represents interquartile range within which 50% of the values are located. Error bars mark minimum and maximum values (range). Small asterisks mark outliners. For calculation of the cut-off ARFI value the Youden-Index was applied and the diagnostic performance of the area under the receiver operating characteristic (AUROC) was calculated. A p-value <0.05 indicated a significant correlation or difference. All reported p-values are two-sided. Sta-
R.S. Goertz et al. / European Journal of Radiology 85 (2016) 2211–2216
tistical analysis was performed using the Statistical Package for the Social Sciences (version 19.0.0.1, IBM SPSS statistics, New York, USA). 3. Results A total of 100 patients (41 female, 59 male, mean age 59 years, BMI 26.6 kg/m2 ) were evaluated in the prospective group and 95 patients (45 female, 50 male, mean age 58 years, BMI 25.5 kg/m2 ) in the course of the retrospective analysis. 3.1. Retrospective evaluation Results of the retrospective analysis are demonstrated in Table 1. On average 7.4 ± 3.1 ARFI measurements were performed, data (n = 5) for measurement depth was missing. Concerning the head and tail in healthy parenchyma, there was no ARFI difference. Mean ARFI shear wave velocities and SD were higher in acute, chronic pancreatitis and adenocarcinoma than in healthy parenchyma (p = 0.002, p < 0.001 and p < 0.001, respectively). Acute pancreatitis showed lower ARFI values than those of chronic pancreatitis (p = 0.13) and adenocarcinoma (p = 0.005). ARFI measurements have even shown a significant difference between chronic pancreatitis and adenocarcinoma (p = 0.045). In addition to the already cited pancreatic diseases, it was possible to assess high ARFI values of four NET. These were significantly higher than those of adenocarcinomas (p = 0.035). Tumor size was 4.3 ± 1.6 cm for adenocarcinoma and 5.3 ± 2.8 cm for NET. In patients with healthy parenchyma the correlations of shear wave velocities with age, measurement depth and BMI or SD with age showed no significance. ARFI elastography of normal pancreas appeared to be slightly lower in men (1.2 ± 0.16 m/s) than in women (1.30 ± 0.25 m/s), without statistical significance (p = 0.222). 3.2. Prospective evaluation In the prospective group the patients with pancreatic carcinoma showed a higher mean age than those without pancreatic adenocarcinoma (p = 0.0001). Patients with pancreatic lipomatosis and fatty liver had a higher BMI than those with acute pancreatitis – p = 0.0004 (see Table 2). Etiologically, acute pancreatitis in 21 patients was biliary (n = 10), alcoholic intoxication (n = 5), autoimmunity (n = 2), post-ERCP (n = 1) and unknown (n = 3). The etiology of chronic pancreatitis was alcoholic (n = 15), idiopathic (n = 8), pancreas divisum (n = 3) or others (hereditary, autoimmune). Imaging in chronic pancreatitis was based on abdominal ultrasound and/or CT (n = 10), ERCP (n = 9), EUS (n = 6) or MRI (n = 2). The pancreatic adenocarcinoma had a mean size of 3.7 ± 1.4 cm (range 2.0–8.3). In 6/21 cases ARFI could not visualize or measure pancreatic parenchyma in the tail nearby carcinoma. Table 3 gives an overview of the detailed shear wave velocities of the different pancreas entities. Subsequently pancreatic lipomatosis, acute pancreatitis, chronic pancreatitis and adenocarcinoma show rising ARFI values with higher SD. The differences of mean ARFI shear wave velocities between the lower values of pancreatic lipomatosis and the higher values of acute pancreatitis or of chronic pancreatitis or of adenocarcinoma were statistically significant (p = 0.018, p < 0.001 and p < 0.001, respectively). Moreover, ARFI values of acute pancreatitis were significantly lower, in comparison to those of chronic pancreatitis (p = 0.02) and adenocarcinoma (p < 0.001). On the other hand, ARFI measurements between chronic pancreatitis and adenocarcinoma showed no significant differences (p = 0.43). 19/28 patients with chronic pancreatitis had calcifications. No difference of the respective ARFI values between calcified and non-calcified chronic pancreatitis
2213
was found (p = 0.306). There were no differences in measurements between head or tail in lipomatosis and acute or chronic inflammation of the pancreas (p = 0.46, p = 0.22 and p = 0.46, respectively). During the evaluation of pancreatic adenocarcinoma, in 15 patients, it was possible to evaluate both sites: ARFI elastography of adenocarcinomas (2.35 ± 0.52 m/s) showed always higher values than the surrounding parenchyma (1.53 ± 0.45 m/s) – p = 0.0001 (Table 4). The mean percentage of failed measurements in the head was 7.4%, 8.0%, 14.0% and 25.5% for pancreatic lipomatosis, acute pancreatitis, chronic pancreatitis and pancreatic carcinoma, respectively. The ARFI values of acute pancreatitis showed no significant correlation with lipase levels (r = −0.252 and p = 0.271). In patients with lipomatosis of the pancreas, there was no correlation of ARFI elastography in head and tail with measurement depth, BMI or age. ARFI shear wave velocities between normal and pancreatic lipomatosis were not significant (p = 0.763). A cut-off analysis of all 195 patients (see Fig. 2) showed a cutoff value of 1.74 m/s for the diagnosis of a malignant disease, with a sensitivity of 91.1% and a specificity of 60.4% (AUROC 78.6% with a 95%-CI 71.6–85.6%). Only malignancy reached ARFI values being >3.95 m/s (n = 6/56 malignant lesions).
4. Discussion ARFI shear wave velocity measurement is a quantitative elastographic method that quantifies elastic properties of tissue non-invasively. Shear wave velocities of the liver correlate well with hepatic fibrosis and cirrhosis [11], reflecting histopathologic architectural changes of liver tissue. Normal healthy pancreas contains lobularly arranged glandular cells with endocrine and exocrine function. In pancreatic lipomatosis, there is an increase of interstitial fatty tissue. Acute pancreatitis leads to infiltration by granulo-lymphocytic cells and disseminated necrosis of acinus cells. Chronic inflammation leads to fibrotic changes of the pancreatic parenchyma and to calcifications. In pancreatic adenocarcinoma, the normal architecture is destroyed by proliferation of epithelial gland cells, concomitantly with a surrounding desmoplastic reaction which result in stiff connective tissue. Pancreatic NET mainly consist of endocrine main cells. Initially, semi-quantitative elastography by EUS was studied to diagnose malignant disease of the pancreas [1]. Due to high sensitivity of endoscopic elastography the latter was thought to be a promising supplement to fine needle aspiration [3]. Recently, magnetic resonance elastography (MRE) showed promising results with regard to its ability to reproduce stiffness measurements throughout the whole pancreas [12]. Using a 3T MRE technique Itoh et al. [13] presented increasing higher stiffness values of the head, body and tail of the pancreas in healthy volunteers. 7 patients with pancreatic carcinoma showed a stiffness amounting to 6.06 kPa and thus being twice as high. In our study of transcutaneous pancreatic ARFI elastography, we have managed to highlight differences in various pancreatic conditions, which reflect these above-mentioned histopathologic changes of the pancreas. In contrast to healthy parenchyma, pancreatic lipomatosis, acute pancreatitis, chronic pancreatitis and adenocarcinoma showed higher ARFI shear wave velocities with increasing SD. ARFI measurements of NET have revealed the highest values. The number of measurement failure increased gradually for pancreatic lipomatosis, acute pancreatitis, chronic pancreatitis and pancreatic carcinoma, with the latter having the highest frequency of measurement failure. Normal ARFI values of the pancreas are reported to be between 1.22 and 1.30 m/s.(6, 8, 14) According to our findings, the measurements of the pancreatic body were slightly higher, however without statistical significance. An analysis of influencing factors on
2214
R.S. Goertz et al. / European Journal of Radiology 85 (2016) 2211–2216
Table 1 Retrospective analysis of ARFI shear wave velocities in patients with healthy parenchyma, acute and chronic pancreatitis, pancreatic carcinoma or neuroendocrine tumor (NET).
Sex [m/w] Age [years] BMI [kg/m2 ] Location ARFI [m/s] SD Depth[cm]
Healthy
Acute pancreatitis
Chronic pancreatitis
Adenocarcinoma
Neuroendocrine tumor
9/12 47.3 (18–82) 25.0 Head 1.3 ± 0.23 (0.97–1.78) 0.33 5.6 ± 1.4
7/7 47.2 (18–69) 27.3 Head 1.79 ± 0.62 (1.06–2.87) 0.51 5.9 ± 1.6
16/9 57.5 (20–82) 24.2 Head 2.13 ± 0.68 (1.02–3.24) 0.62 4.5 ± 1.5
14/17 70.9 (55–86) 26.5 Tumor 2.58 ± 0.91 (0.79–4.19) 0.68 5.1 ± 1.2
4/0 67.5 (58–75) 23.8 Tumor 3.62 ± 0.63 (2.73–4.11) 0.57 4.2 ± 1.2
Tail 1.34 ± 0.23 (0.96–1.92) 0.29 5.3 ± 1.0
Table 2 Patient data of all prospectively evaluated patients with pancreatic disease. Pancreatic lipomatosis
Acute pancreatitis
Chronic pancreatitis
Adenocarcinoma
CRP [<5 mg/l]
19/11 56.8 ± 12.8 (29–79) 31.2 ± 5.1 (22.7–45.6) –
Lipase [<60 U/l]
–
10/11 55.9 ± 21.6 (23–89) 26.0 ± 4.3 (18.3–33.9) 122.7 ± 126 (1–432) 2944.7 ± 4636 (194–20803)
19/9 52.9 ± 11.4 (31–78) 24.0 ± 3.3 (18.4–30.9) 50.7 ± 79 (0.5–309) 936 ± 2484 (1.4–11787)
11/10 71.6 ± 8.1 (50–84) 24.6 ± 3.8 (17.3–31.6) 60.2 ± 77 (1.3–329) 107.9 ± 286 (0–1308)
Sex [m/w] Age [years] BMI [kg/m2 ]
Table 3 Prospective ARFI shear wave velocities in the pancreatic head and tail of pancreatic lipomatosis, acute and chronic pancreatitis and pancreatic adenocarcinoma. Pancreatic lipomatosis
Acute pancreatitis
Chronic pancreatitis
Adenocarcinoma
m/w
30
21
28
21
Location
Head
Tail
Head
Tail
Head
Tail
Tumor
ARFI [m/s]
1.33 ± 0.35 (0.95–2.59) 0.36 1.22 0.53 ± 0.25 (0.14–0.99) 7.1 ± 1.1 (4.1–8.0)
1.40 ± 0.25 (1.01–2.06) 0.32 1.36 0.44 ± 0.19 (0.12–0.83) 6.8 ± 1.3 (3.4–8.0)
1.74 ± 0.71 (0.87–3.62) 0.50 1.63 0.76 ± 0.47 (0.19–1.7) 5.9 ± 1.9 (4.5–8.0)
1.53 ± 0.61 (1.03–3.2) 0.44 1.49 0.57 ± 0.33 (0.19–1.35) 5.9 ± 1.6 (3.3–8.0)
2.21 ± 0.65 (1.14–3.92) 0.62 2.19 0.95 ± 0.49 (0.22–1.91) 4.9 ± 1.0 (3.3–6.8)
2.06 ± 0.61 (0.95–3.39) 0.57 1.98 0.87 ± 0.46 (0.24–2.58) 4.7 ± 7.9 (3.1–6.1)
2.35 ± 0.58 (1.4–3.89) 0.74 2.3 1.04 ± 0.4 (0.46–1.89) 5.5 ± 1.2 (2.8–8.0)
SD Median IQR depth [cm]
Table 4 Difference between prospective ARFI shear wave velocities of pancreatic adenocarcinoma (n = 15) and peritumoral parenchyma (p = 0.0001).
ARFI [m/s] SD IQR
Pancreatic adenocarcinoma
Peritumoral parenchyma
2.35 ± 0.52 (1.4–3.45) 0.76 ± 0.18 (0.43–0.97) 1.08 ± 0.42 (0.62–1.89)
1.53 ± 0.45 (0.91–2.28) 0.47 ± 0.17 (0.21–0.79) 0.62 ± 0.36 (0.21–1.37)
pancreatic ARFI elastography of 234 healthy volunteers showed a significant difference between genders (examination of men resulting in lower values than that of women) and increasing stiffness with age – in contrast to the non-significant results we obtained with regard to these issues. Alcohol or smoking demonstrated no significant effect [14]. Another study by Kawada [15] evaluating factors associated with shear wave velocity mentioned that older age or increased alcohol intake led to higher ARFI shear wave velocities, whereas a brighter (i.e. lipomatotic) pancreas or a deeper depth of ROI showed lower shear wave velocities. ARFI measurements performed in the tail failed more often than it did in the corpus or head region. Mateen et al. [16] performed ARFI elastographic quantification in 52 healthy controls, 68 acute pancreatitis and 46 chronic pancreatitis. A single high ARFI value of 2.21 m/s was obtained even in the healthy group, however the healthy mean was nearly equal to
that of the patients with a chronic pancreatitis (1.25 m/s). Thus, the chronic pancreatitis values were obviously lower than they were in our study (2.21 m/s). Only the acute group had significantly higher ARFI shear wave velocities the mean of which amounts to 3.28 m/s, and showed individual mean values even being above 4.95 m/s, resulting in a high sensitivity and specificity (>90%, respectively) for the diagnosis of an acute pancreatitis in this study. The number of measurements was not recorded. 46 patients with chronic pancreatitis in comparison to 52 healthy volunteers were examined by Yashima et al. [8], who applied ARFI elastography in the head, body and tail of the pancreas. Chronic pancreatitis showed significantly higher ARFI values in the three pancreatic regions (1.65 to 2.09 m/s with high SD) than in the normal pancreas. Failed measurements were recorded in the tail of chronic pancreatitis in up to 65% of the patients. The inhomogeneity of the tissue in chronic pancreatitis and adenocarcinoma might account to that fact. When measuring the head and body, the measurement was not successful in 10% of the patients regularly. In our study, 68% of the patients with chronic pancreatitis suffered from calcifications. Park et al. [7] examined focal solid pancreatic lesions in 26 patients by applying ARFI tissue quantification. The difference between benignity (2.4 ± 1.1 m/s) and malignancy (3.3 ± 1.0 m/s) was not statistically significant, but the relative stiffness of malignant lesions in contrast to the adjacent parenchyma was sig-
R.S. Goertz et al. / European Journal of Radiology 85 (2016) 2211–2216
2215
Fig. 2. Boxplot of pancreatic ARFI measurements in all 195 patients with healthy parenchyma (n = 21), pancreatic lipomatosis (n = 30), acute (n = 35) or chronic (n = 53) pancreatitis, adenocarcinoma (n = 52) or NET (n = 4) of the pancreas.
nificantly higher than that of benign lesions (p = 0.011). This mean “malignant relative stiffness” being 1.5 m/s, determined by Park in pancreatic adenocarcinoma, is higher than that found out by our team, as we determined it being 0.82 m/s only. Furthermore in a different study, a significant ARFI difference between normal pancreas (n = 40) and adenocarcinoma (n = 62) was found being 1.17 m/s versus 2.74 m/s [17]. Interestingly, it was also emphasized that high ARFI values above 4.0 m/s obtained by more than one measurement were already mainly associated with adenocarcinoma. The three NET examined in this study were subsumed to non-adenocarcinoma and had, on average, equal ARFI values in comparison with adenocarcinomas, however showing a lower maximum value. In synopsis with our data, NET and adenocarcinoma show higher ARFI elastography than normal tissue, but cannot be distinguished from each other. Chronic pancreatitis and malignancy have higher ARFI values than normal tissue has. Their risk factors, i.e. alcohol consumption and patients’ age as an expression of duration of pancreatic injury leading to fibrosis, are also associated with higher values. The main limitations of our study are the non-consecutive inclusion of patients in the prospective group and the somewhat lower total number of measurements (mean 7.4) performed in the retrospective group. Besides, due to its rarity, only 4 NET could be investigated. Not all results were congruent between the retrospective and prospective analysis. In the retrospective group a significant ARFI difference between chronic pancreatitis and adenocarcinoma was demonstrated, whereas in the prospective group was not. Thus, ARFI needs a careful examination in order to achieve reproducible results. There might be some elements that affect ARFI measurements (i.e. probe frequency, the orientation of the ROI, the number of measurements). In analogy to that, a different point shear wave technique might not reveal identical results. In conclusion, it is possible to perform ARFI elastography in the head and tail of the pancreas. ARFI shear wave velocities increase
in acute and chronic pancreatitis as well as in neoplastic lesions of the pancreas. Very high ARFI values (>3.95 m/s) indicate malignant disease of the pancreas. However, there is a considerable overlap between the entities.
Conflict of interest None.
References [1] Q. Pei, X. Zou, X. Zhang, et al., Diagnostic value of EUS elastography in differentiation of benign and malignant solid pancreatic masses: a meta-analysis, Pancreatology 12 (2012) 402–408. [2] J. Janssen, I. Papavassiliou, Effect of aging and diffuse chronic pancreatitis on pancreas elasticity evaluated using semiquantitative EUS elastography, Ultraschall Med. 35 (2014) 253–258. [3] D. Cosgrove, F. Piscaglia, J. Bamber, et al., EFSUMB guidelines and recommendations on the clinical use of ultrasound elastography. Part 2: clinical applications, Ultraschall Med. 34 (2013) 238–253. [4] N. Kawada, S. Tanaka, Elastography for the pancreas: current status and future perspective, World J. Gastroenterol. 22 (2016) 3712–3724. [5] M. D’Onofrio, S. Crosara, R. De Robertis, et al., Elastography of the pancreas, Eur. J. Radiol. 83 (2014) 415–419. [6] R.S. Goertz, K. Amann, R. Heide, et al., An abdominal and thyroid status with acoustic radiation force impulse elastometry–a feasibility study: acoustic radiation force impulse elastometry of human organs, Eur. J. Radiol. 80 (2011) e226–e230. [7] M.K. Park, J. Jo, H. Kwon, et al., Usefulness of acoustic radiation force impulse elastography in the differential diagnosis of benign and malignant solid pancreatic lesions, Ultrasonography (Seoul, Korea) 33 (2014) 26–33. [8] Y. Yashima, N. Sasahira, H. Isayama, et al., Acoustic radiation force impulse elastography for noninvasive assessment of chronic pancreatitis, J. Gastroenterol. 47 (2012) 427–432. [9] C. Goya, C. Hamidi, S. Hattapoglu, et al., Use of acoustic radiation force impulse elastography to diagnose acute pancreatitis at hospital admission: comparison with sonography and computed tomography, J. Ultrasound Med. 33 (2014) 1453–1460. [10] A.G. Schreyer, M. Jung, J.F. Riemann, et al., S3 guideline for chronic pancreatitis – diagnosis, classification and therapy for the radiologist, RoFo:
2216
[11]
[12]
[13]
[14]
R.S. Goertz et al. / European Journal of Radiology 85 (2016) 2211–2216
Fortschritte auf dem Gebiete der Rontgenstrahlen und der Nuklearmedizin 186 (2014) 1002–1008. S. Bota, H. Herkner, I. Sporea, et al., Meta-analysis: ARFI elastography versus transient elastography for the evaluation of liver fibrosis, Liver Int. 33 (2013) 1138–1147. Y. Shi, K.J. Glaser, S.K. Venkatesh, E.I. Ben-Abraham, R.L. Ehman, Feasibility of using 3D MR elastography to determine pancreatic stiffness in healthy volunteers, J. Magn. Reson. Imaging: JMRI 41 (2015) 369–375. Y. Itoh, Y. Takehara, T. Kawase, et al., Feasibility of magnetic resonance elastography for the pancreas at 3T, J. Magn. Reson. Imaging: JMRI 43 (2016) 384–390. S. Stumpf, H. Jaeger, T. Graeter, et al., Influence of age, sex, body mass index, alcohol, and smoking on shear wave velocity (p-SWE) of the pancreas, Abdom. Radiol. (New York) (2016).
[15] N. Kawada, S. Tanaka, H. Uehara, et al., Potential use of point shear wave elastography for the pancreas: a single center prospective study, Eur. J. Radiol. 83 (2014) 620–624. [16] M.A. Mateen, K.A. Muheet, R.J. Mohan, et al., Evaluation of ultrasound based acoustic radiation force impulse (ARFI) and eSie touch sonoelastography for diagnosis of inflammatory pancreatic diseases, JOP: J. Pancreas 13 (2012) 36–44. [17] M. D’Onofrio, R. De Robertis, S. Crosara, et al., Acoustic radiation force impulse with shear wave speed quantification of pancreatic masses: a prospective study, Pancreatology 16 (2016) 106–109.