Elsevier

Gastrointestinal Endoscopy

Volume 62, Issue 3, September 2005, Pages 383-389
Gastrointestinal Endoscopy

Original Article
Cyst fluid analysis in the differential diagnosis of pancreatic cystic lesions: a pooled analysis

https://doi.org/10.1016/S0016-5107(05)01581-6Get rights and content

Background

Pancreatic cystic tumors commonly include serous cystadenoma (SCA), mucinous cystadenoma (MCA), and mucinous cystadenocarcinoma (MCAC). A differential diagnosis with pseudocysts (PC) can be difficult. Radiologic criteria are not reliable. The objective of the study is to investigate the value of cyst fluid analysis in the differential diagnosis of benign (SCA, PC) vs. premalignant or malignant (MCA, MCAC) lesions.

Methods

A search in PubMed was performed with the search terms cyst, pancrea, and fluid. Articles about cyst fluid analysis of pancreatic lesions that contained the individual data of at least 7 patients were included in the study. Data of all individual patients were combined and were plotted in scatter grams. Cutoff levels were determined.

Results

Twelve studies were included, which comprised data of 450 patients. Cysts with an amylase concentration <250 U/L were SCA, MCA, or MCAC (sensitivity 44%, specificity 98%) and, thus, virtually excluded PC. A carcinoembryonic antigen (CEA) <5 ng/mL suggested a SCA or PC (sensitivity 50%, specificity 95%). A CEA >800 ng/mL strongly suggested MCA or MCAC (sensitivity 48%, specificity 98%). A carbohydrate-associated antigen (CA) 19-9 <37 U/mL strongly suggested PC or SCA (sensitivity 19%, specificity 98%). Cytologic examination revealed malignant cells in 48% of MCAC (n = 111).

Discussion

Most pancreatic cystic tumors should be resected without the need for cyst fluid analysis. However, in asymptomatic patients, in patients with an increased surgical risk, and, in patients in whom there is a diagnostic uncertainty about the presence of a PC, cyst fluid analysis helps to determine the optimal therapeutic strategy.

Section snippets

Patients and methods

A search in PubMed was performed in June 2004 with the search terms cyst, pancrea, and fluid. Articles about cyst fluid analysis of pancreatic lesions were included if (a) the individual data of at least 7 patients were present, (b) the diagnosis of a cystadenoma was made by pathologic examination, and (c) the diagnosis of PC was made by pathologic examination or by the presence of a pancreatic cyst in a patient with a history of acute pancreatitis or well-documented chronic pancreatitis in

Results

The PubMed search revealed 12 studies that comprised individual cyst-fluid concentrations of amylase, CA 19-9, or CEA of in total 450 patients (Table 1).1, 2, 3, 7, 9, 10, 11, 12, 13, 14, 15, 16 Two articles from one group partially described the same patient data about the CEA concentrations; duplicate data were excluded.11, 15 In most studies, patients were only included after histologic confirmation of their diagnosis. The diagnosis of nearly all MCA or MCAC was proven by histology or

CEA

The 332 individual CEA concentrations in cyst fluid of PC (n = 125; median, 10 ng/mL), SCA (n = 79; median, 3 ng/mL), MCA (n = 64; median, 400 ng/mL), and MCAC (n = 64; median, 2000 ng/mL) are shown in Figure 2. With a cutoff value of <5 ng/mL (the upper limit of normal serum values of CEA), a clear distinction between MCA/MCAC from SCA/PC was observed (Fig. 2). Cysts with CEA < 5 ng/mL were SCA or PC (sensitivity 50%, specificity 95%, PPV 94%, NPV 55%, accuracy 67%). In one study,16 the CEA

Discussion

Symptomatic resectable lesions should be removed without the need for fluid analysis. However, in asymptomatic patients, in patients with an increased surgical risk, and in patients in whom there is a diagnostic uncertainty about the presence of a PC, it is important to distinguish premalignant or malignant tumors, such as MCA and MCAC from benign tumors, such as PC and SCA. Unfortunately, radiologic criteria have a relatively low accuracy.1, 2, 8 In the present study, we performed a pooled

Cited by (0)

See CME section; p. 414.

View full text