Prevalence of incidental pancreatic cyst on upper endoscopic ultrasound

Introduction

A notable increase in the diagnosis of pancreatic cystic lesions (PCLs) has been reported. With the increasing use of computed tomography (CT) and magnetic resonance imaging (MRI), and improvements in imaging resolution [1], the detection rate of incidental pancreatic cysts, typically small [2,3], has improved.

PCL are a heterogeneous group of tumors, ranging from benign to malignant [4], which includes the common precursor lesions for pancreatic cancer: i.e., intraductal papillary mucinous neoplasms (IPMNs) and mucinous cystic neoplasms (MCNs) [5]. Therefore, identifying these types of PCL could aid in the prevention or treatment of pancreatic cancer at an early stage. PCLs, particularly IPMNs, are the typical lesions found in patients who are at high risk of pancreatic cancer and undergo screening programs [6,7]. Moreover, PCL prevalence varies depending on the imaging method used (CT 3% and MRI 20% in unselected patients) [8,9]. In autopsy studies, the prevalence may be as high as 24.3%, with an incidence that increases with age [10]. Recently, a study reported a prevalence of 9.4% in patients who underwent endoscopic ultrasound (EUS) for a cause unrelated to the pancreas [11]. However, the study failed to exclude patients with a history of acute pancreatitis, in which PCL could be either the cause or the consequence. Thus, the findings may not reflect the true prevalence of incidental PCL.

This study employed EUS, which detects more lesions than MRI; hence, a higher prevalence of cysts than that observed in previous studies was expected [6]. We hypothesized that the prevalence of incidental pancreatic cysts would be greater with EUS than with CT or MRI. We aimed to determine the prevalence of incidental pancreatic cysts in patients who underwent EUS without a known pancreatic abnormality, and to identify the factors associated with the presence of incidental cysts.

Patients and methods

Results

During the study period, 351 upper EUS procedures were performed to evaluate a condition unrelated to the pancreas. A total of 53 patients were excluded because of previous acute pancreatitis (n=30), evidence of acute pancreatitis on EUS (n=9), previous upper gastrointestinal tract surgery (n=3), incomplete pancreatic examination because of luminal obstruction or stenosis (n=8), and previous abnormal pancreatic imaging (n=3). Consequently, 298 patients were included in the study: 130 (43.6%) men and 168 (56.4%) women, age 61±15 years (range 23-92 years). Two hundred forty-seven (82.9%) and 51 (17.1%) EUS procedures were performed in Hospital A and Hospital B, respectively. Table 1 shows the patients’ characteristics and indications for EUS.

Table 1 Demographic, clinical and imaging characteristics of patients

Eighty-four percent of patients did not consume alcohol on a regular basis and 66% were non-smokers. Previous abdominal imaging had been performed in 216 (72.5%) patients (51% had CT, 39.3% MRI, and 17.8% both). None of these imaging tests detected the presence of cysts in the pancreas. The most common indication for EUS was biliary tract disease evaluation (59.7%). In addition, pancreatic cysts were found in 64 patients (21.5%; 16.9-26.6%), of whom 51 (79.7%) had unique cysts and 13 (20.3%) had multiple pancreatic cysts (median 2, range 2-8). None had features of chronic pancreatitis on EUS. The size of the cysts was 6.3±3.7 mm (range 3-25 mm). Only 6 (2%) patients had cysts >10 mm and only 3 (1%) had cysts >15 mm. Most of these cysts (68.8%) were in the body of the pancreas. Communication with the pancreatic duct was seen on EUS in 15 (23.4%) patients. Only 2 (3.1%) patients had main pancreatic duct dilatation (3.7 and 3.9 mm). Sixteen (5.4%) had cysts compatible with BD-IPMN; the rest were classified as non-specific. The characteristics of these cysts are presented in Table 2.

Table 2 Characteristics of incidental pancreatic cysts (n=64)

Patients with cysts were significantly older than those without (67.2 vs. 59.3 years, P<0.001). The prevalence of cysts was significantly higher in patients >50 years (27%) and lower in patients <50 years (4.2%) (P<0.001). In the univariate analysis, no differences were found in the other variables analyzed (Table 3).

Table 3 Univariate analysis of risk factors for pancreatic cyst

The prevalence of cysts was similar in both hospitals (21.5% vs. 21.6%, P=0.99) and no differences were found in the other characteristics. Although the difference was not statistically significant (P=0.076), a linear echoendoscope detected cysts in a greater proportion of patients than did a radial echoendoscope (26.3% vs. 17.3%). In patients where both echoendoscopes were used, cysts were found in 33.3%.

EUS-FNA was performed in two patients. The first case was a 57-year-old woman who underwent EUS for biliary tract disease. Multiple cysts were detected; the largest (25 mm) was located in the body of the pancreas and was connected to the main pancreatic duct (Fig. 1). A 7-mm mural nodule was also detected without enhancement after contrast administration (SonoVue, Bracco International B.V., Amsterdam, Holland). EUS-FNA was performed and a clear liquid with a positive string sign was obtained. Levels of amylase, glucose, CEA, and CA 125 were 14,844 U/L, 3 mg/dL, 3 ng/mL, and 2 U/mL, respectively. Cytology showed few cells, with no evidence of malignancy, and the nodule was confirmed to be mucus. The patient was diagnosed as having a multifocal BD-IPMN and was put under surveillance.

Figure 1 Unilocular cyst in the pancreatic body with a 7 mm nodule confirmed to be mucus

The second case was a 68-year-old woman who was evaluated for bile duct dilatation. Multiple cysts were detected; the largest (15 mm) was located in the head of the pancreas and connected to the pancreatic duct (Fig. 2). No mural nodules or masses were detected. Cytology showed epithelial cells without atypia. The liquid was clear with a positive string sign. Biochemical analysis was not performed because of the elevated viscosity of the sample. The patient was diagnosed as having a multifocal BD-IPMN and was put under surveillance.

Figure 2 Fifteen mm cyst in the uncinate process. Communication with the pancreatic duct, typical of branch-duct intraductal papillary mucinous neoplasm, is observed

A 15-mm cyst was detected in a 59-year-old male. EUS without FNA was performed because of the diagnosis of gastric cancer (stage IV).

Discussion

PCLs are frequently detected incidentally in patients undergoing cross-sectional imaging investigations for other medical indications. Given the malignant potential of some PCLs, incidental detection should be carefully evaluated. PCLs have a broad differential diagnosis, i.e., from non-neoplastic (e.g., pseudocyst, lymphoepithelial cyst, retention cyst) to neoplastic cystic lesions (serous cystic tumor, MCN, IPMN, and solid pseudopapillary neoplasm). Solid tumors can also present necrotic cystic degeneration [4]. Thus, identifying an increasing number of pancreatic cysts in asymptomatic patients has become a clinical challenge.

The majority of incidentally discovered cysts are small and could be BD-IPMNs. Although a cyst with invasive carcinoma is uncommon in patients with an asymptomatic pancreatic cyst, particularly those <10 mm in size, a follow up is still recommended [12]. Therefore, it is essential to know the expected PCL prevalence in asymptomatic individuals. The number of patients with incidentally discovered cysts who subsequently develop a clinically significant lesion, especially pancreatic adenocarcinoma, remains to be identified. Little is known about the natural history of these patients, and the data in the literature are contradictory. Although IPMN and MCN are recognized precursor lesions of pancreatic adenocarcinoma [5,14], a majority of incidentally identified pancreatic cysts are typically small BD-IPMNs that fail to meet the criteria for resection [15].

Some studies showed that small asymptomatic cysts can be safely followed [15]; conversely, other surgical series found a high incidence of neoplastic cysts, even in small pancreatic cysts measuring <1 cm [16,17]. Thus, identifying the prevalence of these asymptomatic cysts in healthy individuals provides an opportunity to reduce pancreatic cancer-associated mortality, which has remained unchanged for the last 20 years [18].

The true prevalence of pancreatic cysts in asymptomatic individuals remains unclear and varies depending on the population studied and the imaging method employed. The age of the equipment used is also a significant factor when interpreting the results, as newer versions of MRI hardware and software are associated with greater numbers of pancreatic cysts detected [19]. Another important consideration is the retrospective nature of the studies, in which only the images of radiological procedures already performed, and not necessarily those with the best protocol scan of the pancreas, are reviewed. Thus, these studies may not clearly reflect the true prevalence of incidental pancreatic cysts. The prevalence identified using MRI is higher than that using CT (2.6-5.4% for CT [20,21] vs. 2.4-19.6% for MRI [8,9,22]).

EUS is an operator-dependent procedure. However, in experienced hands, it allows the pancreas to be completely explored in the vast majority of patients, with a higher resolution than other imaging techniques, especially for small lesions [23,24]. One prospective study performed blinded comparisons of standardized pancreatic protocol CT, secretin-enhanced MRI/magnetic resonance cholangiopancreatography (MRCP), and EUS for screening in high-risk individuals. The results showed that EUS and secretin-enhanced MRI are superior to CT for the detection of small, predominantly cystic, pancreatic lesions. EUS, MRI/MRCP, and CT identified pancreatic lesions in 42.6%, 33.3% and 11% of screened high-risk individuals, respectively [6]. In another study, EUS identified synchronous PCLs unappreciated by initial cross-sectional imaging [23]. Therefore, we could expect that, in subjects without known or suspected pancreatic diseases, the prevalence of EUS-identified pancreatic cysts should be equal to or higher than the prevalence found in studies using MRI. A recent prospective study that included patients with a history of pancreatitis found that only 9.4% of patients had cysts identified with EUS, indicated for conditions unrelated to pancreatic diseases [11]. EUS was indicated in 20% of the cases for chronic abdominal pain. Nevertheless, the proportion of cysts found with EUS was similar to that in other studies that used MRI [8]. In our study, patients with a history of acute pancreatitis and chronic pain of probable pancreatic origin, as well as those with known pancreatic alterations, were excluded, and cysts were found in 21.5% of the remainder. The prevalence of pancreatic cysts in our study was extremely high and similar to that in autopsy studies (i.e., 24.3%) [10].

On major strength of our study is that it was performed in two different hospitals (referral centers for EUS): an 800-bed-capacity tertiary hospital and a 230-bed-capacity community hospital. The proportion of patients in whom cysts were found was similar in both hospitals, with no differences in cyst characteristics. The differences in endoscopy equipment between the two hospitals had no influence on the detection rate, while the prevalence of cysts detected by EUS was strikingly high, which could be attributed to the operators’ expertise. Moreover, the proportion of clinically significant cysts was low and most of the cysts were small (average size 6.3 mm), findings similar to those of previous studies. Only 5.4% were identified as BD-IPMN; the rest of the cysts were non-specific. Therefore, the results of this prospective multicenter study have great value even though surgical treatment would not be indicated in any of the cysts according to guidelines [12]. However, a close follow up will be applied to those cysts classified as BD-IPMN, a premalignant condition, while the other non-specific cysts might be reviewed in order to determine their behavior. A pancreatic MRI/MRCP after a short interval (3-6 months) is now recommended to establish the cysts’ stability. Subsequently, surveillance should be performed according to the size of stratification [25].

The proportion of cysts found during EUS was greater with a linear echoendoscope than with a radial echoendoscope and was much greater when both echoendoscopes were used in the same exploration; nonetheless, the differences were not statistically significant. A previous study also found that linear EUS detects more pancreatic lesions than radial EUS in asymptomatic high-risk individuals for pancreatic cancer undergoing screening EUS. In the same study, a “second-pass effect”, i.e., additional lesions detected with a second EUS examination, was noted [26].

Cyst prevalence increases with age [10,22], with a prevalence of 0.5% in patients <40 years old [8]. Only this variable was significantly associated with the presence of cysts in our study. Cysts in the pancreas are rare (4.2%) in patients <50 years, but their prevalence increases significantly above this age (27%). The mean age of patients with cysts in our study was similar to that in previous studies.

Our study’s limitation was the study population, which was not truly asymptomatic. The patients underwent EUS for a cause unrelated to the pancreas. However, the exclusion criteria ensured that the cysts were incidental findings unrelated to pancreatic pathology and the patients were asymptomatic.

In summary, in our population, we found a high (21.5%) incidental prevalence of pancreatic cysts on EUS, higher than the prevalence estimates in cross-sectional imaging studies.