Clinical Impact of EUS Elastography Mean Strain Histograms (SH) and Contrast Peak-enhancement in Focal Pancreatic Masses and Lymph Nodes

1. Background Endoscopic ultrasound (EUS) is a technique with a major clinical impact in digestive diseases, determining a change in the diagnosis and management of more than half of examined patients [1]. Recent advances in EUS-FNA techniques, but also the development of real-time EUS elastography and contrast-enhancement, allowed a better characterization of focal pancreatic masses, with possible implications in the management of patients with negative EUS-FNA and a strong suspicion of malignancy.

1.a Elastography Elastography is a recent ultrasound method used for the reconstruction of tissue elasticity distribution in real-time [2]. The method allows the calculation of the elasticity modulus, consequently showing differences in tissue hardness patterns that are determined by diseases. The main intended use is to differentiate between benign and malignant focal lesions based on the significantly smaller strain of the latter [3]. Second generation elastography introduces strain ratio (SR) and strain histogram (SH) as reproducible parametric measurements that retrieve numerical values in real time, adding quantification possibilities to the technique [4]. Elastography typically estimates the axial strain (along the direction of insonification / compression) by analyzing ultrasonic signals obtained with standard ultrasonographic systems - the RF signals returned from tissue structures before and after slight compression (about 1%) are compared [5]. Tissue elastography can be easily performed with conventional probes, including the linear EUS probes used for the examination of the pancreas and/or lymph nodes. The calculation of tissue elasticity distribution is performed in real-time under freehand compression and the examination results are represented as transparent overlay colour images overimposed on the conventional gray-scale B-mode images [6]. Thus, this method allows the characterization of many tumors, because they are stiffer than normal tissues. Ultrasound elastography was previously used for the diagnosis of non-digestive as well as digestive tumors: breast lesions [7], prostate cancer [8], thyroid nodules [9], rectal tumors [10]. Regarding the diagnosis of pancreatic focal masses, some authors could not differentiate between malignancy and benign tumors or chronic pancreatitis [11], while others have obtained good results, especially when using computer assisted means of evaluation like hue histogram analysis [12] and artificial neural networks [13]. More recently, lymph node involvement of several tumors has been succesfully determined using this method: esophagus [14], oral squamous cell carcinoma [15], breast cancer [16].

Since elastography images and movies represent a qualitative type output that entails a subjective interpretation by the examiner, human bias is always susceptible to interfere with the results and diagnoses, due to color perception errors, moving artifacts, or possible selection bias induced by the analysis of still images. More objective, computer-assisted semi-quantitative means of interpreting the results were developed, but these have the disadvantage of being labor-intensive and using third-party software that cannot be used in real time [17]. Second generation elastography introduces strain ratio (SR) and strain histogram (SH) as two reproducible measurements that retrieve numerical values in real time, thus greatly reducing the human bias without the need for third-party software [4]. SR calculates the relative strain between two regions of interest (ROI) (normal and pathological). SH measures the strain values of elemental areas inside a ROI and divides the measurement range into intervals; if the strain value of an element falls into an interval, its initial area normalized by the initial total surface area is added to the running total of that interval; the total values of each interval are used to produce a graph and an average value. Both SR and SH have already been used in vivo for pancreatic masses or lymph nodes, with promising results [18].

1. b Contrast-enhancement Ultrasound contrast agents in conjunction with contrast specific imaging techniques are increasingly accepted in clinical use for diagnostic imaging [19]. The study of the pancreas is a new and promising application of contrast-enhanced ultrasound (CE-US), including contrast-enhanced endoscopic ultrasound (CE-EUS). The technique is not indicated to improve the detection of pancreatic lesions, but to improve the delineation and differential diagnosis of pancreatic lesions [20-23]. One of the fluoro-gas-containing contrast agents used in CE-US and CE-EUS is Sonovue®, which consists of phospholipids-stabilized bubbles of sulfurhexafluoride (SF6) [24]. Sonovue® is isotonic, stable and resistant to pressure, with a viscosity similar to blood. It does not diffuse into the extravascular compartment remaining within the blood vessels until the gas dissolves and is eliminated in the expired air (blood pool contrast agent) [25]. The safety profile of SonoVue showed a very low incidence of side effects; it is not nephrotoxic and the incidence of severe hypersensitivity is similar to other magnetic resonance imaging contrast agents. Moreover, Sono-Vue is approved for clinical use in EU countries. The blood supply of the pancreas is entirely arterial, making contrast-enhanced examinations feasible and readily available. Based on the European Federation Societies of Ultrasound in Medicine and Biology guidelines and recommendations, updated in 2008, two phases were defined for CE-US and CE-EUS of the pancreas: an early/arterial phase (starting from 10 to 30 seconds) and a venous/late phase (from 30 to 120 seconds) [19]. Distinguishing pancreatic adenocarcinoma from other pancreatic masses remains challenging with current imaging techniques [22-27]. The specificity of the discrimination between benign and malignant focal pancreatic lesions was found to be 93.3% using power Doppler contrast-enhanced EUS (PD-CE-EUS) compared with 83.3% for conventional EUS [26]. The hypovascular aspect of lesions under PD-CE-EUS seemed highly sensitive and specific (higher than 90%) for adenocarcinoma in several published studies [22-27]. During PD-CE-EUS examinations the ultrasound frequency returned to the transducer is the same with that transmitted, but the method is associated with artifacts resulting from turbulent flow (flash and overpainting) [28]. At CE-EUS, ductal adenocarcinoma is typically hypoenhanced compared to the adjacent pancreatic tissue in all phases [19]. Furthermore, the lesion size and margins are better visualized, as well as the relationship with peripancreatic arteries and veins. Focal lesions in chronic pancreatitis are reported to have similar or hyper enhancement features as compared to the normal pancreatic parenchyma [19]. Dedicated contrast-enhanced harmonic EUS techniques (based on a low mechanical index) are recently available in new EUS systems. The harmonic frequencies returned during CEH-EUS are different from those transmitted by the transducer and are the result of non-linear oscillations of the microbubbles [24]. The image obtained is an addition of the signal created by the distortion of the microbubbles and the tissue-derived signal. This can be optimized by using low MI, which allows minimum bubble destruction and complete "subtraction" of the tissue derived signal, obtaining a high resolution continuous real-time assessment of the microvascularization during the contrast uptake period (real-time perfusion imaging) [29-31]. CEH-EUS allows a more precise location of vascular structures within the parenchyma and focal abnormalities, with better delineation of pancreatic lesions than EUS, especially in the cases where air or fat causes artifacts and insufficient visualization of the pancreatic parenchyma. An initial pilot study described an experimental technique of CEH-EUS based on a linear prototype EUS scope, a low mechanical index (0.08 - 0.25) and a 2nd generation contrast agent (Sono-Vue), which allowed the visualization of early arterial phase and late parenchymal phase enhancement of the pancreas [32]. Another pilot study demonstrated both real-time continuous images of finely branching vessels of the pancreas and intermittent homogenous parenchymal perfusion images, by using a radial prototype EUS scope, a low mechanical index (0.4) and a 2nd generation contrast agent (Sono-Vue) [33]. Several other research groups already reported the feasibility of CEH-EUS with low mechanical index [34-36]. The sensitivity, specificity and accuracy for diagnosing pancreatic adenocarcinoma were 88%, 89%, and 88.5% in one study [34] and 80%, 91.7%, and 86% in the other study [33]. However, the data is still limited and a prospective, multicentric blinded study would certainly be necessary. The study protocol is based on a multi-center semi-quantitative approach of EUS elastography data in combination with contrast-enhanced EUS, consisting of measuring SR and SH for focal pancreatic masses and lymph nodes, as well as several parameters of CE-EUS based on time-intensity-curve (TIC) analysis. A number of parameters must be taken into consideration, as the ROIs are still manually selected by the user. The aim of the study is to establish an EUS based diagnostic algorithm in patients with pancreatic masses and lymph nodes, with negative or inconclusive cytopathology after EUS-FNA, based on previously published results and cut-offs of elastography and contrast-enhancement. The proposed algorithm of sequential use of real-time elastography, followed by contrast-enhanced EUS could be a good clinical tool to help select the patients with possible pancreatic adenocarcinoma or malignant lymph nodes, in the setting of patients with negative EUS-FNA results.
2. Aims of the study The aim of the study is to assess quantitative elastography and contrast-enhancement parameters during EUS examinations of focal pancreatic masses and lymph nodes, to standardize an algorithm for their use and to consequently differentiate benign vs malignant pancreatic masses and evaluate lymph node involvement in a prospective multicenter design.
3. Patients and methods The study design is prospective, blinded and multi-center, comparing endoscopic ultrasound elastography (EG-EUS) and contrast-enhnecement (CE-EUS) results for the characterization of focal pancreatic masses and lymph nodes by using parametric measurements, in comparison with the gold standard represented by pathology. The study will be performed with the approval of the institutional board (ethical committee) review of each center. The complete study protocol and particpating centers will be uploaded on ClinicalTrials.gov, the registry of federally and privately supported clinical trials conducted in the United States and around the world. Inclusion criteria

• Patients diagnosed with solid pancreatic tumor masses, with cytological / histo-logical confirmation
• Patients with or without suspected lymph node involvement are eligible
• Age 18 to 90 years old, men or women
• Signed informed consent for EG-EUS, CE-EUS and FNA biopsy Exclusion criteria
• Prior surgical treatment with curative intent or chemo-radiotherapy
• Patients diagnosed with mucin producing tumors, pancreatic cystic tumors, etc.

1. Data collection

• Personal data (name, surname, age, admission date, SSN, diagnosis at admission)

1. Imaging tests

• All patients with a suspicion of pancreatic masses or lymph nodes should undergo EUS, with sequential EG-EUS and CE-EUS
• EUS with EUS-guided FNA and elastography
• Protocol of EUS with EUS-FNA should include linear EUS instruments with complete examinations of the pancreas.
• Tumor characteristics (echogenicity, echostructure, size) will be described as well as presence / absence of power Doppler signals.
• EUS-FNA will be performed in all pancreatic masses with at least three passes
• All examiners should be blinded for the results of pathology
• EG-EUS procedure:
• EUS-EG will be performed during usual EUS examinations, with two movies of 10 seconds recorded on the embedded HDD in order to minimize variability and to increase repeatability of acquisition.
• A two panel image with the usual conventional gray-scale B-mode EUS image on the right side and with the elastography image on the left side will be used. The same parameters will be set-up in all systems used: e-dynamic range 2, persistence 3, etc.
• The region of interest for EUS-EG will be preferably larger than the focal mass (approximately 50%-50%), in order to include the surrounding structures. If the focal mass is larger than 3 cm, part of the mass will be included in the ROI, as well as the surrounding structures (preferably avoiding large vessels). Very large ROI for the elastography calculations will be avoided due to the appearance of side artifacts.
• The following pre-settings will be used in all centers: elastography colour map 1, frame rejection 2, noise rejection 2, persistence 3, dynamic rage 4, smoothing 2, blend 50%.
• SR and SH will be measured; with three measurements made and recorded on the embedded HDD. For SR, the reference area should be placed at the same level with the lesion, if possible.
• CE-EUS procedure:
• A two panel image with the usual conventional gray-scale B-mode EUS image on the right side and with the contrast harmonic image on the left side will be used, according to pre-established presets.
• The starting point of the timer will be considered the moment of intravenous contrast injection (Sonovue 4.8 mL).
• CE-EUS will be performed during usual EUS examinations, with the whole movie (T0-T120s) recorded on the embedded HDD of the ultrasound system, for later analysis.
• A low mechanical index procedure (dynamic wide-band contrast harmo¬nic imaging mode) will be used, with a mechanical index of 0.2 and corresponding powers.
• The following pre-settings will be used in all centers: contrast mode dCHI-W, WPI-R/P (resolution / penetration for superficial vs deep structures), mechanical index (variable starting with 0.1, preferred 0.2), MI gray-scale (0.03), grey map 4, AGC 0, R-filter C, persistence 2, dynamic range 50, B-colour 21, smoothing 3, gamma curve linear.
• In order to minimize human bias, all post-processing and computer analysis of digital movies will be performed within the coordinating IT Center, with all programmers and statisticians being blinded to the clinical, imaging and pathological data.
• Off-line analysis of time-intensity curves will be performed using Vue-Box, which yields the following quantitative parameters: Peak Enhancement (PE), Wash-in Area Under the Curve (Wi-AUC), Rise Time (RT), mean Transit Time (mTT), Time To Peak (TTP), Wash-in Rate (WiR) and Wash-in Perfusion Index (WiPI). The software also provides referenced values (expressed in percentages), aligning the set of values for the tumor ROI to the parenchymal ones.

1. Final diagnosis

• A positive cytological diagnosis will be taken as a final proof of malignancy of the pancreas mass or lymph node. The diagnoses obtained by EUS-FNA will be further verified either by surgery or during a clinical follow-up of at least 6 months.
• The diagnosis of chronic pancreatitis will be based on the clinical information (history of alcohol abuse, previous diagnosis of chronic pancreatitis or diabetes mellitus), as well as a combination of imaging methods (ultrasound, CT and EUS). At least four criteria of chronic pancreatitis during EUS will be considered for the positive diagnosis. The diagnosis of chronic pseudotumoral pancreatitis or benign lymph nodes will always be confirmed by surgery or by a follow-up of at least six months used to exclude malignancy in the patients that will not be operated on.
• Pathology samples obtained from duodeno-pancreatectomies or caudal pancreatecto-mies done with curative intent, as well as microhistological fragments obtained through EUS-FNA biopsy will be processed by paraffin embedding with usual stainings (haematoxylin-eosin), with subsequent immune-histochemistry at the discretion of the participating centers pathologists in order to exclude neuroendocrine tumors / pancreatic metastases.
• The patients will be followed-up for at least six months through clinical examination, biological exams and transabdominal ultrasound, eventually with a repeat spiral CT / EUS after six months.

1. Statistical analysis

• Descriptive statistics
• All results will be expressed as mean ± standard deviation (SD). Differences between the patients with pancreatic cancer and chronic pancreatitis will performed by the two-sample t-test (two independent samples). Since this parametric method makes assumptions about normality and similar variances, we will initially perform both the Kolmogorov-Smirnov and Shapiro-Wilk W normality tests and verify the equality of variances assumption with the F test. In the case of the two-sample t-test, we will also perform the non-parametric alternative given by the Mann-Whitney U test, since in some instances it may even offer greater power to reject the null hypothesis than the t-test.
• Since with more than two groups of observations it is far better to use a single analysis that enables us to look at all the data in the same time, we will also perform the one-way analysis of variance (ANOVA) method with the same baseline assumptions. A p-value less than 0.05 will be considered as statistically significant.
• Sensitivity, specificity, positive predictive value, negative predictive value and accuracy of EG-EUS and CE-EUS will be determined in comparison with the final diagnosis. Also, the sensitivity, specificity, positive predictive value, negative predictive value and accuracy for the subgroup of patients with negative EUS-FNA and a positive diagnosisi of malignancy during ensuing follow-up will be calculated separately.

1. Power analysis

• The estimated number of patients included is at least 210, based on at least 10 centers with at least 20 patients each, which will be enrolled in a 12 months period. The power analysis was based on the following assumption: in order to use the powerful t-test for independent samples, a sample size equaling 105 patients in each group is sufficient to provide 95% statistical power to detect a difference of 5% in mean, for a type I error alpha = 0.05, and the population standard deviation = 10%.
• The difference in mean was based on previously published data which report an accuracy of approximately 80-85% for EUS-FNA, and 90% for EG-EUS and/or CE-EUS.