EUS in inflammatory diseases of the pancreas

Key points

EUS imaging is subjective and often nonspecific in inflammatory diseases of the pancreas; therefore, the clinical history and presentation are important when making a final diagnosis.
EUS fine-needle aspiration (FNA) in inflammatory diseases of the pancreas is predominately used to exclude/diagnose superimposed malignant processes; the role of fine-needle biopsy (FNB) remains to be defined.
Complementary imaging modalities such as elastography and contrast-enhanced imaging are in the experimental phase and are not recommended for routine use in the evaluation of patients with inflammatory diseases of the pancreas.

Introduction

Some of the most common pancreatic disorders that a gastroenterologist encounters are inflammatory processes, particularly acute and chronic pancreatitis. Imaging plays a key role in the diagnosis and management of these disorders. Due to the proximity to the pancreas and high-resolution imaging, endoscopic ultrasound (EUS) provides an ideal technique for evaluating the entire pancreas. The capability to obtain aspirates for cytological evaluation and core biopsies for histologic examination allows for definitive diagnosis of benign and malignant pancreatic processes. Its role in the management of these common inflammatory disorders is evolving and becoming further defined.

Although much less common, autoimmune pancreatitis (AIP) and benign masses of the pancreas often provide diagnostic challenges for the gastroenterologist. Therefore, novel and complementary imaging modalities that can be performed at the time of EUS, namely elastography and contrast-enhanced EUS, are increasingly being used, although with still uncertain utility. This chapter focuses on the role of EUS imaging and tissue sampling for evaluating patients with suspected inflammatory diseases of the pancreas.

EUS in the nondiseased pancreas

Endosonographically, the nondiseased pancreatic parenchyma typically has a homogeneous salt and pepper appearance with a thin anechoic tubular structure that runs in the middle of the gland corresponding to the main pancreatic duct (MPD) ( Fig. 12.1 , ). The outer contour of the pancreas is normally smooth without significant lobularity. The width of the pancreas in the head, body, and tail measures approximately 19 mm, 13 mm, and 12 mm, respectively. The MPD diameter measures on average 2.2 mm in the head, 1.5 mm in the body, and 1 mm in the tail of the pancreas.

• Fig. 12.1, EUS demonstrates the normal salt and pepper appearance in asymptomatic patients without pancreatic pathology who underwent EUS for alternate indications. (See Video 12.1 .)

The ventral anlage appears as a focal hypoechoic area in the head of the pancreas and can be seen in up to 75% of patients on EUS. Since the initial reports, the introduction of improved EUS processors and transducers now allows detection in nearly all patients. It is important to not mistake the ventral anlage for a pancreatic mass. Conversely, it is also important to remember that the normally appearing hypoechoic ventral anlage can mask the presence of a small neoplasia or other pathology. When occult pathology is suspected in this region, one may consider careful examination of the ventral anlage from both the bulb and second portion of the duodenum. In addition, use of the cine function and magnification can further aid careful examination.

Unfortunately, the pancreas often demonstrates an altered appearance, possessing some chronic pancreatitis-like features even in the absence of any apparent disease. In patients undergoing EUS for indications unrelated to pancreaticobiliary disease, at least one parenchymal and/or ductal abnormality can be seen in up to 28% of patients. The likelihood of EUS imaging abnormalities increases with patient age, particularly in patients >60 years old, and male gender. The most common abnormality is hyperechoic stranding. Other studies have shown that alcohol consumption at low levels increases the likelihood of detecting hyperechoic foci, MPD dilation, and ductal wall hyperechogenicity, while smoking increases the finding of hyperechoic foci. Small visible side branches <1 mm can also be seen in patients without pancreaticobiliary symptoms.

Chronic pancreatitis

Brief overview

Chronic pancreatitis (CP) results from progressive inflammation and scarring of the pancreas. Patients may initially present with recurrent acute pancreatitis and occasionally patients presenting with their first known episode of acute pancreatitis may already have structural and functional changes secondary to chronic pancreatitis. Pain, classically located in the epigastrium with radiation to the back and worsened by eating, is the most common symptom. When more than 85% to 90% of the pancreas is affected, patients may also manifest exocrine insufficiency (with steatorrhea, weight loss, and fat-soluble vitamin deficiency) and/or endocrine insufficiency (with glucose intolerance or diabetes mellitus).

Although the diagnosis of CP is straightforward in patients with severe disease manifested by the presence of calcifications and a dilated MPD, the diagnosis is far more challenging in early disease stages. Furthermore, while pancreatic calcifications most often result from CP, calcifications may be seen in other forms of pancreatic pathology including neuroendocrine tumors, intraductal papillary mucinous neoplasms, mucinous cystic neoplasia, serous cystadenoma, and some pancreatic adenocarcinomas. The pattern and extent of calcification often provides a diagnostic clue as patients with CP often manifest MPD and parenchymal calcification, the latter of which is often actually contained within peripheral side branches. In addition, a dilated MPD may result from other important pathologies including an obstructing neoplasia or main duct IPMN rather than CP.

The diagnosis of CP relies on a combination of symptoms, noninvasive radiographic imaging (CT, MRI/MRCP), and/or EUS. In patients with documented CP, the most common CT findings (in decreasing order of frequency) includes a dilated MPD with secondary radicals (68%), pancreatic parenchymal atrophy (54%), pancreatic calcifications (50%), fluid collections (30%), focal pancreatic enlargement (30%), biliary ductal dilation (29%), alterations in peripancreatic fat (16%), and a normal pancreas (7%). Unfortunately, these CT findings are not specific for CP. CT is also used to help detect complications of CP and exclude other diseases associated with abdominal pain. MRI/MRCP, particularly with secretin stimulation, has improved sensitivity over CT for the detection of early CP. Findings on MRI/MRCP include MPD dilation, side-branch abnormalities, strictures, intraductal stones, intraparenchymal cyst formation, parenchymal atrophy, and abnormal decreased signal intensity on T1-weighted images of the pancreas with delayed and limited enhancement after contrast administration. MPD compliance with distention (normal is approximately 1 mm) and side-branch abnormalities after secretin stimulation has enhanced the ability of MRI to detect early CP.

Role of EUS

The American Pancreatic Association (APA) recommends a step-wise approach to diagnose CP ( Fig. 12.2 ). In patients with compatible clinical signs and symptoms, the APA initially recommends a CT that, if nondiagnostic, should be followed by MRI/MRCP with secretin stimulation. The society recommends EUS only when the results of noninvasive imaging are nondiagnostic or inconclusive. If the EUS is inconclusive, the APA recommends pancreatic function testing with secretin administration. However, it should be noted that pancreatic function testing has a narrow role due to the limited availability, poor patient tolerance, and suboptimal diagnostic accuracy. Finally, the use of diagnostic ERCP may be considered only after the aforementioned evaluation when the diagnosis remains indeterminate. As the various clinical and imaging features of CP are often nonspecific and commonly seen in patients without pancreatic disease, it is important to interpret the findings with caution and in the appropriate clinical setting. This need is even greater when contemplating therapeutic interventions, for which EUS may have a role ( Figs. 12.3 to 12.5 , and ). Similarly, the American College of Gastroenterology (ACG) Clinical Guidelines on chronic pancreatitis recommend the use of cross-sectional imaging (CT or MRI) as first-line tests, followed by EUS if the noninvasive testing is nondiagnostic.

• Fig. 12.2, Proposed algorithm for the diagnosis of chronic pancreatitis.

• Fig. 12.3, A 48-year-old male presented with clinical and CT evidence of chronic pancreatitis with a large pancreatic head duct stone and upstream ductal dilatation. Three prior efforts at ERCP failed and the patient was referred for further evaluation and stone clearance. CT revealed a large stone (orange arrow) within the main pancreatic duct within the pancreatic head with upstream ductal dilatation (green arrow). During three prior ERCPs at the referring center there was failure to cannulate the pancreatic duct.

• Fig. 12.4, A 48-year-old male presented with clinical and CT evidence of chronic pancreatitis with a large pancreatic head duct stone and upstream ductal dilatation. Three prior efforts at ERCP failed and the patient was referred for further evaluation and stone clearance. EUS demonstrated a thickened bile duct wall suggestive of downstream obstruction (left image). Continued imaging revealed a remote location of the minor papilla, relative to the major papilla, initially establishing the diagnosis of pancreas divisum that had been missed on prior cross-sectional imaging (right image). The large stone was seen in the accessory duct. (See Video 12.2 .)

• Fig. 12.5, A 48-year-old male presented with clinical and CT evidence of chronic pancreatitis with a large pancreatic head duct stone and upstream ductal dilatation. Three prior efforts at ERCP failed and the patient was referred for further evaluation and stone clearance. EUS demonstrated typical features of chronic pancreatitis including a dilated MPD (upper left), dilated side branches (upper right), and parenchymal atrophy (bottom). It is likely that some of the features were secondarily induced by the large stone and resulting obstructive changes. (See Video 12.3 .)

EUS imaging characteristics

EUS criteria for CP

Conventional EUS criteria for the diagnosis of CP rely on the evaluation of nine features. The four parenchymal features include hyperechoic foci (distinct 1 to 2 mm hyperechoic points), hyperechoic strands (hyperechoic irregular lines >3 mm), lobularity (2 to 5 mm lobules), and cysts (thin-walled hypoechoic structures >2 mm within the confines of the parenchyma) while the five ductal features include MPD dilation (>3 mm in the head, >2 mm in the body, and >1 mm in the tail of the pancreas), ductal irregularity, hyperechoic duct margins, visible side branches, and intraductal stones ^, ( Figs. 12.6 to 12.9 , and ). The ideal cutoff for the number of EUS criteria needed to diagnose CP varies among endosonographers, institutions, and study protocols. Many consider the finding of 1 to 2 criteria as indicative of a normal pancreas, 3 to 4 criteria indicate early CP, and ≥5 criteria consistent with CP. Understandably, the higher the threshold required for diagnosis, the lower the sensitivity and higher the specificity of the criteria. The International consensus guidelines report that the presence of at least 5 criteria and 2 or fewer criteria strongly suggests or refutes the diagnosis CP.

• Fig. 12.6, EUS demonstrates a long benign stricture in a patient with chronic pancreatitis.

• Fig. 12.7, EUS in a patient with an acute exacerbation of chronic pancreatitis. In addition to the calcification (upper left), there are pancreatic (upper right) and peripancreatic (lower) inflammatory changes. (See Video 12.4 .)

• Fig. 12.8, EUS highlights the features of chronic pancreatitis, including hyperechoic foci (orange arrows), hyperechoic strands (green arrows), and a hyperechoic main pancreatic duct wall (yellow arrow). The EUS suspicion of mild or early features of chronic pancreatitis was confirmed with core biopsies.

• Fig. 12.9, EUS images are shown of a patient who was confirmed on core biopsy to have mild or early features of chronic pancreatitis. (See Video 12.5 .)

Some have adopted the use of an alternate EUS-based classification system to diagnose CP (the Rosemont classification), which was developed from a consensus of internationally recognized endosonographers. The criteria are divided into major and minor criteria based on their perceived accuracy for diagnosing CP. In addition, the three major criteria are subdivided into major A and major B features depending on their predictive diagnostic accuracy. The major criteria include: hyperechoic foci with shadowing (major A), MPD calculi (major A), and lobularity with honeycombing (major B). The eight minor criteria include: lobularity without honeycombing, hyperechoic foci without shadowing, cysts, stranding, irregular MPD contour, dilated side branches, MPD dilation, and a hyperechoic MPD margin. Table 12.1 defines each criterion. Using these major and minor criteria, the Rosemont classification establishes a diagnosis that is “consistent with CP” if any of the following is present: 1 major A feature + ≥3 minor features, 1 major A feature + major B feature, or 2 major A features. EUS examinations “suggestive of CP” include the following: 1 major A + <3 minor features, major B and ≥3 minor features, or ≥5 minor features. Those classified as “indeterminate for CP” include: >2 minor features, <5 minor features without major features, or major B feature + <3 minor features. A “normal” result is one that has ≤2 minor features, excluding cysts, dilated MPD and side branches, hyperechoic foci without shadowing, and major features. Classification of those “suggestive” or “consistent” with CP has been shown to be a predictor of progression from minimal change early pancreatitis to confirmatory CP by imaging or histology.

TABLE 12.1

Definition of Each Rosemont Feature

Parenchymal Features
Hyperechoic foci with shadowing	≥3 echogenic structures ≥2 mm in length and width that shadow
Lobularity	Well-circumscribed, ≥5 mm structures with enhancing rim and relatively echo-poor center; ≥3 lobules in the body or tail need to be present With honeycombing: contiguous ≥3 lobules Without honeycombing: noncontiguous lobules
Hyperechoic foci without shadowing	≥3 echogenic structures ≥2 mm in length and width with no shadowing
Cysts	Anechoic, round/elliptical structures with/without septations
Stranding	≥3 hyperechoic lines of ≥3 mm in length in at least 2 different directions with respect to the imaged plane
Ductal Features
MPD calculi	Echogenic structure(s) within the MPD with acoustic shadowing
Irregular MPD contour	Uneven or irregular outline and ectatic course of the MPD; only assessed in the pancreatic body and tail
Dilated side branches	≥3 tubular anechoic structures each measuring ≥1 mm in width, budding from the MPD; only assess from the pancreatic body and tail
MPD dilation	≥3.5 mm in the body or ≥1.5 mm in the tail
Hyperechoic MPD margin	Echogenic, distinct structure greater than 50% of the entire MPD (on both the proximal and distal border) in the body and tail

In 2019, the Japan Pancreas Society updated their criteria for diagnosing early chronic pancreatitis. The EUS criteria were simplified in this update with the need for two out of four features at EUS including at least A or B: (A) hyperechoic (nonshadowing) foci or strands, (B) lobularity, (C) hyperechoic MPD margin, or (D) dilated side branches. The other imaging criterion was irregular dilation of more than three side branches by MRCP or ERCP. Also to complement criteria, they suggest consideration of clinical symptoms and/or heavy alcohol consumption as additional diagnostic tools for diagnosing chronic pancreatitis.

Comparison of conventional and rosemont classification

As expected based on the definitions for each feature and multiple means in which the grouped criteria can be analyzed, the Rosemont classification is more stringent than the conventional classification and is more difficult to remember and employ in clinical practice. Comparative studies demonstrate that when using a cutoff of 3 criteria for the conventional classification more patients are diagnosed with CP; however, when using a cutoff of five features there is no significant difference in the number of patients diagnosed with CP using conventional classification as compared to the Rosemont classification when combining both “consistent with” and “suggestive of” CP. The Rosemont classification of “consistent with CP” is the most stringent threshold for diagnosing CP, even more than a cutoff of five features in the conventional means of classification.

Correlation of EUS findings and surgical histopathology

In a study assessing conventional EUS criteria in patients who later underwent pancreatic surgery, the presence of ≥4 criteria was the best predictor of histologic CP with a sensitivity, specificity, and accuracy of 90.5%, 85.7%, and 88.1%, respectively. There was also excellent correlation between the number of EUS features and histological fibrosis score ( r = 0.85, P < .0001). In particular, hyperechoic foci ( P < .0001), hyperechoic strands ( P > .001), lobularity ( P = .04), stones ( P < .001), dilated MPD ( P < .0001), irregular MPD ( P < .0001), irregular side branches ( P < .001), and hyperechoic MPD margins ( P = .03) were all significantly associated with fibrosis on histology. A dilated or irregular MPD had the highest sensitivity, specificity, and accuracy in determining the presence of fibrosis.

Another study also demonstrated that the presence of ≥4 EUS criteria predicted the presence of CP in patients who underwent total pancreatectomy with islet autotransplantation for noncalcific CP. However, the sensitivity, specificity, and accuracy in determining a fibrosis score ≥6 in the resected specimen was lower than the previous study at 61%, 75%, and 63%. They found a poor, but significant, correlation between EUS features and degree of fibrosis ( r = 0.24, P < .05). None of the individual conventional features was significantly predictive of the presence of fibrosis on univariate analysis. When linear regression was performed after adjusting for age, sex, BMI, smoking, and alcohol exposure, only MPD irregularity ( P = .02) was found to be predictive of CP. As expected, in another study that utilized a lower cutoff for the fibrosis score on surgically resected specimens (fibrosis score ≥2), the presence of ≥4 EUS criteria was shown to have a higher sensitivity and specificity of 84% and 100%, respectively.

The aforementioned study methodologies and results highlight a notable difficulty using the fibrosis score, namely the lack of consensus regarding the appropriate threshold of this histological criterion for diagnosing CP. Also, the use of a fibrosis score alone as a diagnostic gold standard for CP is problematic in that it may simply reflect the presence of “bland” fibrosis, which is fibrosis seen in the absence of parenchymal destruction or inflammation that is often present in asymptomatic patients without endocrine or exocrine dysfunction. Such “bland” fibrosis has been reported in the setting of alcoholism, advanced age, male gender, obesity, and cigarette smoking. This finding may be detected in patients without any evidence of CP with normal pancreatic imaging, function, and histology. However, some debate whether bland fibrosis indicates the presence of very early disease that is prone to progress in severity over time. This theory is unlikely to account entirely for this situation, given the unmatched finding of bland fibrosis in up to 60% versus the lifetime risk of 2% to 5% of CP in alcoholics. Whether bland fibrosis represents an early stage of CP and/or separate entity is unclear. Caution is needed when performing EUS because this clinically occult bland fibrosis may be indistinguishable from CP and can result in overdiagnosis ( Figs. 12.10 and 12.11 , ).

• Fig. 12.10, EUS images are shown of a patient who was thought based on EUS imaging to have moderately severe chronic pancreatitis. However, core biopsy demonstrated a large quantity of completely normal pancreas. This patient highlights the limitations that can occur with EUS imaging. (See Video 12.6 .)

• Fig. 12.11, Pathology reveals the findings of mild (upper left), moderate (upper right), and severe (bottom) chronic pancreatitis.

Certain individual and grouped criteria have been found to correlate with surgical pancreatic histology. In a study evaluating 100 patients with suspected CP who underwent EUS followed within 1 year by pancreatic resection, lobularity with honeycombing, hyperechoic foci with shadowing, dilated MPD, irregular MPD, and dilated side branches were each associated with histologically diagnosed severe CP. The highest odds ratio between EUS features and the histopathologic findings included hyperechoic foci with shadowing in the head of the pancreas and large duct diameter (OR 10.9 [95% CI 2.9 – 40.5]), hyperechoic foci with shadowing in the head of the pancreas and calcifications (OR 8.8 [95% CI 2.6 – 28]), pancreatic head cysts and pseudocysts (OR 12.9 [95% CI 3.2 – 52.3]), MPD dilation in the head of the pancreas and large duct distortion (OR 12.8 [95% CI 2.6 – 62.9]), dilated side branches in the head of the pancreas and large duct distortion (OR 6.4 [95% CI 1.9 – 22]), lobularity with honeycombing in the body or tail of the pancreas and large duct distortion (OR 6.2 [95% CI 1.3 – 30.2]), and cysts in the body or tail of the pancreas and pseudocysts (OR 32 [95% CI 4.6 – 222.6]). Therefore, although the standard and Rosemont classification systems rely on evaluation of the body and tail of the pancreas, this study shows that findings in the head of the pancreas may also be important in the diagnosis of CP. However, until confirmatory data are generated, we encourage caution if adopting this practice given the common presence of altered EUS morphology within the pancreatic head even among patients without any clinical evidence of pancreatic pathology.

Interobserver variability

One of the major limitations when using EUS to diagnose CP is the suboptimal interobserver variability (IOA). Using conventional criteria, IOA was determined between 11 expert endosonographers who reviewed EUS videotapes from 33 patients with suspected CP and 12 controls. The kappa was moderately good for the overall agreement on the presence of CP (K = 0.45). In regard to individual features, only MPD dilation (K = 0.6) and lobularity (K = 0.51) had good agreement; the remaining 7 features had poor IOA. Another study assessed same-day back-to-back EUS examinations by 2 different endosonographers on 24 patients without any evidence of pancreaticobiliary disease. Despite the lack of any clinical or imaging evidence of CP, 32% of patients had hyperechoic strands, 30% had hyperechoic ductal walls, 16% had hyperechoic foci, 14% had a dilated MPD, 9% had lobularity, and 5% had parenchymal cysts. The IOA between these two endosonographers was good for hyperechoic strands and parenchymal cysts, moderate for lobularity, dilated MPD, and hyperechoic foci, and fair for hyperechoic foci.

A multicenter study that compared IOA between expert endosonographers using both the conventional criteria and Rosemont classification found that there was moderate agreement (K = 0.54 [95% CI 0.44 – 0.66]) and substantial agreement (K = 0.65 [95% CI 0.52 – 0.77]), respectively. However, there was no statistically significant difference between classification systems. Other studies have similarly shown that the Rosemont classification does not improve IOA compared to the conventional classification. ^,

Comparing two Japanese diagnostic criteria in 2009 and 2019 (which simplified the EUS findings as mentioned above), the IOA was better for the simplified version. This suggests that use of a more abbreviated and clearly defined list of criteria may be of more benefit to endosonographers when evaluating chronic pancreatitis.

EUS sampling

The main role of EUS FNA in CP is distinguishing focal (pseudotumoral) CP from pancreatic adenocarcinoma (PaC) and other neoplasias. PaC is found more commonly in patients with underlying CP than the general population with a cumulative risk of 1.8% (95% CI 1% – 2.6%) at 10 years and 4% (95% CI 2% – 5.9%) at 20 years. Some have found that EUS imaging alone may be insufficient in differentiating between the two. EUS FNA of pancreatic masses has a lower sensitivity in the presence of CP and more passes may be required to establish a diagnosis. ^, A meta-analysis found a pooled sensitivity of EUS FNA in diagnosing malignancy in the absence of CP compared to CP patients (91.5% vs. 63.5%, respectively, OR 5.5 [95% CI 2.9 – 10.2]) and suggested that fine-needle biopsy (FNB) be considered in this patient population. The lower diagnostic sensitivity may result from incorrectly targeted biopsies due to the inability to discern the tumor from peritumoral CP changes with EUS imaging. In addition, cytological interpretation is more challenging when specimens contain limited malignant material along with nontumoral CP material. For more details, see the section on Benign Pancreatic Masses below.

Image-enhancing techniques in EUS

EUS elastography

EUS elastography has been shown in several studies to enhance the EUS diagnostic accuracy of CP, with its objectivity helping overcome the limitations of EUS interobserver variability. Elastography assesses tissue stiffness by applying slight compression and comparing images before and after compression to determine the degree of tissue displacement. In patients with known or suspected CP, there was a negative correlation with mean value and number of Rosemont features on EUS ( r = – 0.59, P < .001). A mean elastography value of 90.1 ± 19.3, 73.2 ± 10.6, 63.7 ± 14.2, and 56.1 ± 13.6 was found in patients meeting the Rosemont criteria for normal, indeterminate for CP, suggestive of CP, and consistent with CP categories ( P < .001 for differences between each stage), respectively. A mean strain ratio (SR; quotient B/A ratio with area A corresponding to the largest possible area of the pancreatic parenchyma and area B as the reference area corresponding to the normal surrounding gut wall) taken from an average of the head, body, and tail was found to have a direct linear correlation to the number of EUS Rosemont criteria ( r = 0.813, P < .0001) with an area under the ROC curve of 0.949 (95% CI 0.916 – 0.982).

EUS elastography has also been used to predict exocrine insufficiency in patients with CP. The 13(C)-mixed triglyceride breath test was used to diagnose exocrine insufficiency in 35 patients (30.4%) diagnosed with CP by MRI/MRCP and EUS. A higher SR (quotient B/A ratio corresponding to the pancreatic parenchyma for area A and a soft peripancreatic reference as area B) of 4.89 (95% CI 4.36 – 5.41) was found in patients with pancreatic insufficiency than those with normal breath tests (2.99 [95% CI 2.82 – 3.16, P < .001]). In addition, patients with a SR >5.5 were 92.8% likely to have pancreatic insufficiency.

EUS shear-wave measurement

EUS shear-wave measurement (EUS-SWM) is thought to provide a more objective absolute value of the pancreatic stiffness than elastography. EUS-SWM is generated by applying a radiation force (push pulse) to a selected region of interest, thereby generating shear waves that propagate perpendicular to initial transmitted ultrasound beam. The tissue elasticity is then evaluated by detection (or tracking) pulses which allow calculation of the propagation velocity of the generated shear waves. The elastic modulus, expressed as kPa, is generated and serves as a measure of the tissue’s elasticity. Harder and more stiff tissues will allow shear waves to move more quickly with faster measured velocities. In one study, the EUS-SWM positively correlated with Rosemont criteria stages ( r = 0.81) and the number of EUS features ( r = 0.72). More research is required for each of these image-enhancing techniques to verify their accuracy and potential role in clinical practice.

You're Reading a Preview

Become a Clinical Tree membership for Full access and enjoy Unlimited articles

Become membership

If you are a member. Log in here