Search tips
Search criteria 


Logo of jnciLink to Publisher's site
J Natl Cancer Inst. 2014 February; 106(2): djt439.
Published online 2014 February 11. doi:  10.1093/jnci/djt439
PMCID: PMC3952201

Proteomic Mucin Profiling for the Identification of Cystic Precursors of Pancreatic Cancer



Pancreatic cystic lesions (PCLs) are increasingly frequent radiological incidentalomas, with a considerable proportion representing precursors of pancreatic cancer. Better diagnostic tools are required for patients to benefit from this development.


To evaluate whether cyst fluid mucin expression could predict malignant potential and/or transformation in PCLs, a proteomic method was devised and prospectively evaluated in consecutive patients referred to our tertiary center for endoscopic ultrasound-guided aspiration of cystic lesions from May 2007 through November 2008 (discovery cohort) and from December 2008 through October 2012 (validation cohort). Cytology and cyst fluid carcinoembryonic antigen (CEA; premalignancy > 192ng/mL, malignancy > 1000ng/mL) were routinely analyzed, and samples were further processed as follows: one-dimensional gel electrophoresis, excision of high-mass areas, tryptic digestion and nano-liquid chromatography–tandem mass spectrometry, with peptide identification by Mascot software and an in-house mucin database. All diagnostic evaluations were blinded to proteomics results. Histology was required to confirm the presence/absence of malignant transformation. All statistical tests were two-sided.


Proteomic mucin profiling proved statistically significantly more accurate (97.5%; 95% confidence interval [CI] = 90.3% to 99.6%) than cytology (71.4%; 95% CI = 59.8% to 80.9%; P < .001) and cyst fluid CEA (78.0%; 95% CI = 65.0% to 87.3%; P < .001) in identifying the 37 (out of 79; 46.8%) lesions with malignant potential (ie, premalignant or malignant tumors). The accuracy of proteomics was nearly identical (96.6% vs 98.0%) between the discovery (n = 29) and validation (n = 50) cohorts. Furthermore, mucin profiling predicted malignant transformation, present in 16 out of 29 (discovery cohort: 9, validation cohort: 20) lesions with available histology, with 89.7% accuracy (95% CI = 71.5% to 97.3%) (for the validation cohort only: 95.0%; 95% CI = 73.1% to 99.7%). This markedly exceeded corresponding results for cytology (51.7%; 95% CI = 32.9% to 70.1%; P = .003) and CEA (57.1%; 95% CI = 34.4% to 77.4%; P = .02).


Proteomic cyst fluid mucin profiling robustly discriminates benign, premalignant, and malignant PCLs. Consequently, it may improve pancreatic cancer prevention and reduce the morbidity burden of unwarranted pancreatic surgery.

Pancreatic cystic lesions (PCLs) are increasingly common unexpected findings on imaging, identifiable on up to 20% of abdominal magnetic resonance imaging examinations (1). Because PCLs almost invariably reflect an underlying inflammatory or neoplastic condition, they rank among the most important incidentalomas to have emerged with radiological advances. Above all, this development has offered a unique opportunity for preventive intervention against pancreatic cancer (ductal adenocarcinoma) because a substantial proportion of cystic tumors can be considered precursor lesions of this devastating disease (2–5).

PCLs fall into two categories: lesions with or without malignant potential. Serous cystic neoplasms are the only pancreatic cystic tumors that can safely be regarded as benign (5,6). PCLs with malignant potential include rare neuroendocrine tumors or ductal adenocarcinomas with cystic degeneration, solid pseudopapillary neoplasms, and the much more prevalent mucinous cystic tumors (5–9). The latter are subclassified as mucinous cystic neoplasms or intraductal papillary mucinous neoplasms (IPMNs). Although generally indolent, they are considered forerunners of pancreatic cancer and may follow a similar disease course once malignant transformation has occurred (10–13).

Presently, even the critical distinction between intrinsically benign, premalignant, and malignant cystic lesions remains problematic. Radiology rarely provides sufficient information for this assessment (14). Endoscopic ultrasound (EUS) with fine-needle aspiration (FNA) may result in cytological diagnosis, but the yield is often scant (15). Cyst fluid carcinoembryonic antigen (CEA) is considered the best indicator of a mucinous tumor, with a diagnostic accuracy of 79% (16,17). However, CEA levels do not correlate with the degree of dysplasia (16). DNA analysis may provide additional information, but its accuracy does not exceed that of CEA (18,19).

The overall aim of this study was to address this unmet clinical need by exploring the potential of cyst fluid mucin expression analysis (mucin profiling) as a diagnostic tool for the evaluation of PCLs.

Mucins are a family of large, membrane-bound or secreted, densely O-glycosylated glycoproteins, which are important for normal epithelial cell barrier function but are also de novo expressed or overexpressed in various cancer types, notably adenocarcinomas (20,21). Postulated roles of membrane-bound mucins in carcinogenesis include the promotion of epithelial growth factor receptor signalling and constitutive activation of cell survival pathways, such as Wnt and NFκB (20–23). Aberrant expression of secreted mucins in (pre)neoplastic lesions may result from epigenetic signalling and has been suggested to provide protection against antitumor immunity (3,20,24,25).

Previous research has demonstrated aberrant expression of both membrane-bound and secreted mucins in ductal adenocarcinomas, pancreatic intraepithelial neoplasia, IPMNs and mucinous cystic neoplasms (3,26–34). Most of these studies have used immunohistochemistry. However, antibody-based detection may lead to under-recognition because the heavy glycosylation of mucins masks their protein identity. Furthermore, the mucin glycosylation pattern is altered in cancer (20,21). To avoid glycosylation-related omissions and obtain complete mucin profiles, we decided to target peptide-protein identification through proteomics.

Proteomics refers to the study of the proteome (ie, the entire set of proteins found in a system in physiological or pathological conditions). Proteomic studies have previously been performed on pancreatic cyst fluid in only a few instances. Two groups used surface-enhanced laser desorption/ionization time-of-flight mass spectrometry to obtain cyst content protein profiles (35,36).This methodology, however, does not permit direct identification of the differentially expressed proteins. Cuoghi et al. assessed aspirates from eight patients by liquid chromatography–tandem mass spectrometry and identified olfactomedin-4 expression as unique to the three mucinous tumors in the study (37). Ke and colleagues applied three different proteomic methods on fluid from 20 lesions and correlated results with histology or CEA levels (38). Several candidate biomarkers were identified, including amylase, mucins, carcinoembryonic antigen-related cell adhesion molecules (CEACAMs), and S100 homologs. The analyses required 8 months of mass spectrometer time.

By contrast, the aim of our study was to specifically evaluate the potential clinical utility of mucins as markers for pre-/cancerous PCLs. Thus, a targeted, potentially high-throughput, proteomic approach was selected. Taking previous findings of MUC6 in serous cystic neoplasms (39,40) and reports on aberrant MUC1 expression in ductal adenocarcinomas (26–29) into account, we formulated two primary hypotheses: 1) proteomic evidence of cyst fluid content of any mucin except MUC6 discriminates PCLs with malignant potential from intrinsically benign lesions; and 2) MUC1 expression indicates malignant transformation (ie, at least high-grade dysplasia).


Study Design and Recruitment of Patients

A prospective study of diagnostic accuracy was designed. Approval was granted by the Vastra-Gotaland Ethics Committee. Study hypotheses and outcome measures were outlined a priori and evaluated at a planned interim analysis after 18 months. The target sample size was also revised at this point. Patients aged 18 years or above referred to Sahlgrenska Hospital, a tertiary center (catchment area = 1.6 million), for EUS-FNA of PCLs between May 2007 and October 2012 were consecutively included. Written informed consent was obtained from all subjects. Pre-established exclusion criteria were 1) solid-pseudopapillary neoplasm and 2) neuroendocrine tumor. These neoplasms are extremely rare (5,6,8,9,41), were deemed unlikely to express mucins, and possess highly characteristic cytological and immunocytochemical features that allow for a relatively straightforward identification (41–44).

EUS Examination, Cytology, and Cyst Fluid CEA Quantification

Lesions were accessed by the transgastric/transduodenal route using a linear echoendoscope (Pentax EG3830UT (Tokyo, Japan)/Olympus GF-UCT140 (Tokyo, Japan)) with a 19, 22, or 25 gauge needle (Wilson-Cook, Limerick, Ireland/Olympus, Aomori, Japan/Boston Scientific, Spencer, IN). A cytopathology technician was present on site. Cytological examination with periodic acid-Schiff diastase staining for mucus and cyst fluid CEA quantification by immunochemoluminescent technology were routinely performed and always prioritized over proteomic analysis. However, a sample volume less than 500 µL precludes CEA quantification by the method used at our hospital. Fluid intended for mucin profiling was stored at −80ºC.

Gel Electrophoresis and Protein Digestion

Samples (approximately 20 µL) were reduced by 100mM dithiothreitol (Sigma-Aldrich, St Louis, MO), heated to 95°C for 30 minutes, subsequently loaded onto a 5% Laemmli polyacrylamide gel with a 3% stacking gel, and separated on a one-dimensional sodium dodecyl sulfate polyacrylamide gel electrophoresis system. Precision protein standards (Bio-Rad, Hercules, CA) were used as molecular mass markers. Gels were stained with Imperial (Thermo Scientific, Rockford, IL). The high molecular mass area (>100kDa) of the separation gel was, together with the stacking gel, excised for each lane. Destaining of the gel pieces, protein digestion by trypsin, and peptide extraction were performed as previously described (45).

Mass Spectrometry and Protein Identification

Peptides were separated on a 50-µm C18 reversed column coupled to a hybrid linear ion trap-Fourier transform ion cyclotron resonance mass spectrometry instrument equipped with a 7-tesla magnet (LTQ-FT; Thermo Electron, Bremen, Germany) and operated in a data-dependent mode to automatically switch between mass spectrometry and tandem mass spectrometry acquisition. Search parameters were set to the following: mass spectrometry accuracy of 5 ppm; tandem mass spectrometry accuracy of 0.5Da; one missed cleavage allowed. For the searches, propionamide (from acrylamide) on cysteine was set as a fixed modification, and oxidation of methionine as a variable modification. Results were evaluated using Mascot software and an in-house mucin database ( Peptides were also screened against the nonredundant protein sequence database of the National Center for Biotechnology Information ( A Mowse score of 45 or greater, including at least one unique peptide at the 99% statistical significance level (ion score >40) or two unique peptides at the 95% significance level (ion score >28), was accepted for protein identification. To test reproducibility, different aspirate portions were prepared and analyzed two to four times for 12 patients (15.4%).


For the assessment of malignant potential, in addition to histology, clinical follow-up was accepted as a reference standard to minimize selection bias and allow inclusion of pseudocysts as negative controls. Case subjects with remaining diagnostic ambiguities after the completion of follow-up (usually including evaluation by a multidisciplinary board) were reviewed by an experienced pancreatic surgeon (A.G. Hyltander). Patients lost to follow-up before a diagnosis could be established were excluded from analysis.

To confirm/exclude malignant transformation, histology was required as an endpoint unless there was evidence of metastasis.

Diagnostic assessments were blinded to proteomics results. Conversely, proteomics results were interpreted (K.S. Jabbar) without knowledge of the diagnostic standard. Histology of surgical specimens was reviewed by the study pathologist (C. Verbeke).

Interim Analysis and Validation

After 18 months an interim analysis was performed. The series then consisted of 29 lesions, denoted the discovery cohort (DC). No alterations of the primary study hypotheses or methodology were deemed necessary.

Interim analysis indicated that for the evaluation of the first study hypothesis, a confidence interval (CI) of maximum ±15% for diagnostic accuracy should preclude overlap between proteomics’ results and the 79% accuracy previously reported for cyst fluid CEA by a large multicenter study (16). At a confidence level of 95%, this equals a validation cohort (VC) of 43 individuals. Regarding the second primary hypothesis, the accuracy of MUC1 considerably exceeded that of conventional methods in the DC. Hence, here a population of 20 was predicted to suffice to detect differences in performance.

Furthermore, based on results from the DC, a further, secondary hypothesis was generated: that a panel of MUC5AC, MUC2, and MUC1 may suffice to stratify benign, premalignant, and malignant PCLs.


The diagnostic accuracy of mucin profiling was compared with corresponding results for standard tests (ie, cyst fluid CEA and cytology). Proteomic evidence of any mucin except MUC6 was regarded as indicative of malignant potential; MUC1 expression was considered predictive of malignancy. For CEA, previously suggested cutoffs of 192ng/mL for premalignancy (16) and 1000ng/mL for malignancy (46) were used. CEA was not quantifiable for 19 lesions, which were excluded from the assessment of its performance. Cytology samples were evaluated by trained cytopathologists as per standard operating protocol. EUS morphology and cyst size [cutoff of 30mm (5,6)] were additionally appraised as malignancy predictors. Sex-specific analyses were not performed.

Fisher exact test and Mann–Whitney U-test were used for the comparison of categorical and continuous data, respectively. All P values are two-tailed, and the statistical significance level was set to less than .05. The Holm–Bonferroni multiple comparisons correction was applied to the primary comparisons of the accuracy of mucin profiling vs conventional methods. Statistical significance levels for other comparisons were adjusted by the Bonferroni procedure whenever appropriate.

Table 1 summarizes how the primary study hypotheses were evaluated.

Table 1.
Study summary*


Study Population

The inclusion/exclusion process that resulted in a series of 78 patients (42 women; median age = 64 years), comprising a DC of 28 patients (n = 29 lesions) and a VC of 50 patients, is outlined in Figure 1. Histology or evidence of metastasis (n = 1 patient) was available for 29 patients (19 women; median age = 64 years; DC = 9 patients, VC = 20 patients). Diagnoses and baseline characteristics of the study population are provided in Table 2. Patients with cystic tumors tended to be older than those with pseudocysts and to have smaller lesions (median size = 20 vs 70mm). The distribution of diagnoses was similar between the DC and the VC, as shown in Figure 2, ,aa and andbb.

Table 2.
Study population; distribution of diagnoses, and baseline characteristics for each diagnosis group*
Figure 1.
Flow chart of patients in the study. Diagnoses provided in italics are tentative. GIST, gastrointestinal stromal tumor; IPMN, intraductal papillary mucinous neoplasm.
Figure 2.
Distribution of diagnoses for the discovery cohort (n = 29) (A) and validation cohort (n = 50) (B). IPMN, intraductal papillary mucinous neoplasm; MCN, mucinous cystic neoplasm.

The endpoints for the evaluation of the first study hypothesis were 1) histology or evidence of metastasis (n = 29; 37.2%); 2) (near-)complete resolution (n = 18; 23.1%); 3) follow-up with imaging (median = 19 months; interquartile range [IQR] = 10–35; n=26; 33.3%); or 4) clinical assessment, in case of diagnostic ambiguity by a multidisciplinary conference (n = 5; 6.4%). Endpoints are summarized in more detail in Supplementary Figure 1 (available online) and are provided for each patient in Supplementary Tables 1 and 2 (available online). Surgery was performed shortly after EUS-FNA (median interval = 2 months; IQR = 1–4).

Cyst Fluid Mucin Profiling

Table 3 summarizes the mucin profiles of the different lesions; individual results are compiled in Supplementary Tables 1 and 2 (available online). Although mucins were not detected in pseudocyst aspirates (with two exceptions), MUC6 was identified in two of five serous cystic tumors, corroborating previous reports (39).

Table 3.
Cyst fluid mucin expression profiles for the different types and subtypes of pancreatic cystic lesions*

Repeated analysis (n = 1–3 times) with up to a 2-year interval of samples from 12 patients (15.4%) showed results to be reproducible (Supplementary Table 3, available online).

Identifying Cystic Lesions With Malignant Potential

As demonstrated in Figure 3, proteomic mucin profiling discriminated PCLs with malignant potential (n = 37 of 79; 46.8%) from benign lesions with 97.5% accuracy (95% CI = 90.3% to 99.6%). This statistically significantly exceeded corresponding values for cytology (71.4%; 95% CI = 59.8% to 80.9%; P < .001), and cyst fluid CEA (78.0%; 95% CI = 65.0% to 87.3%; P < .001). Only one false negative result was recorded, from a grossly hemorrhagic aspirate. The accuracy of mucin profiling was nearly identical between the DC and VC (96.6% vs 98.0%). Statistically significant differences in accuracy compared with conventional methods, were independently observed in the VC. Full performance characteristics for the evaluated analyses, including likelihood ratios, are provided in Supplementary Tables 4 and 5 (available online).

Figure 3.
Identification of pancreatic cystic lesions with malignant potential: performance of proteomics, cytology, and cyst fluid CEA (192ng/mL). The vertically oriented text on the columns representing cytology and CEA refers to two-sided P values for the comparison ...

Of note, cyst fluid requirements for CEA quantification (500 µL) precluded CEA analysis in 19 patients (24.4%), 15 of whom had neoplastic PCLs.

At the point of interim analysis, a diagnostic algorithm for the assessment of PCLs, comprising a panel of three mucins (MUC5AC, MUC2 and MUC1), was tentatively introduced (Figure 4). In the VC, all lesions with malignant potential (n = 23) expressed at least one of these mucins; 19 of 23 (82.6%) contained MUC5AC. Also provided in Figure 4 are the most abundant unique peptides observed for each of these three mucins, information that could potentially be used to facilitate their individual identification. (Basic information on each detected mucin, including a full peptide hit list is compiled in Supplementary Tables 6 and 7, available online.)

Figure 4.
Step-wise approach to the assessment of pancreatic cystic lesions with regard to malignant potential. Columns represent sensitivity values for the identification of pre-/malignant lesions for MUC5AC (left), MUC5AC+MUC2 (middle), and the optimal combination ...

Detecting Manifest Malignancy

For lesions with procurable histology or confirmed metastasis (n = 29), MUC1 expression was separately compared with current state-of-the-art malignancy indicators for PCLs: severe cellular atypia on cytological examination, cyst fluid CEA > 1000ng/mL (46), and EUS morphology suggesting malignant transformation (Figure 5; Supplementary Tables 8 and 9, available online). Malignancy, defined as minimum high-grade dysplasia, was present in 16 of 29 (55.2%) lesions according to the reference standard.

Figure 5.
Detecting manifestly malignant pancreatic cystic lesions: performance of MUC1, cytology, cyst fluid CEA (1000ng/mL), and endoscopic ultrasound (EUS) morphology. The rightmost (light blue) columns refer to a combination of standard analyses, with at least ...

As shown in Figure 5, the accuracy of MUC1 for the identification of malignant PCLs was, at 89.7% (95% CI = 71.5% to 97.3%), markedly higher than corresponding results for cytology, EUS morphology, and cyst fluid CEA (51.7%, 95% CI = 32.9% to 70.1%; 58.6%, 95% CI = 39.1% to 75.9%; and 57.1%, 95% CI = 34.4% to 77.4%; P = .003, .01, and .02, respectively). Conventional methods attained an optimum accuracy of 61.9% (95% CI = 38.7% to 81.0%) when evaluated in combination (P = .04 for the comparison with MUC1). These differences in performance remained statistically significant after Holm–Bonferroni correction for multiple (n = 4) comparisons. In the VC (n = 20), the accuracy of MUC1 was even higher (95.0%; 95% CI = 73.1% to 99.7%).

A size greater than 30mm for a cystic tumor is considered indicative of a substantial risk of malignant transformation (5–7). In our material (excluding three benign lesions that were resected for purely symptomatic reasons), the size criterion demonstrated 37.5% sensitivity and 80.0% specificity for malignancy.

Cyst fluid MUC1 expression was furthermore detected in eight conservatively treated patients who were not included in this analysis. The first three of these subjects died during the course of the study, likely because of metastasized malignancy where no other primary neoplasm has been identified (autopsy not performed). The remaining five were recently sampled and are currently undergoing follow-up; two of them also had CEA levels greater than 1000ng/mL (Supplementary Table 1, available online).


Mucinous cystic tumors, which in the recent past were the topic of case reports, are now known to account for a considerable proportion of PCLs. The prevalence of PCLs, in turn, is presently estimated at 10% to 24% in the elderly population (47,48). At the same time, emerging evidence that mucinous tumors should be considered forerunners of pancreatic cancer has offered a visible target for the prevention of this devastating disease. Unfortunately, the lack of reliable instruments for the differential diagnosis of PCLs could exclude many patients from timely intervention and expose others to the risk of unnecessary pancreatic surgery (49–51)

In this study, we have shown that mucin profiling by a proteomic approach identified cystic forms/precursor lesions of pancreatic cancer with an accuracy of 97.5%. Results from an independent validation cohort were nearly identical. This compares very favorably with cyst fluid CEA (78.0%; P < .001), which is currently considered the most useful diagnostic test for PCLs (16). The performance of conventional methods in this study are in accordance with previous reports (16,17), as is the observation that cyst fluid yield precluded CEA quantification in nearly every fourth (n = 19 of 78) patient and every second (n = 14 of 29) mucinous tumor (52). By contrast, 25 times less material was required for proteomics, allowing for the analysis of all aspirates.

Mucinous tumors are frequent incidental findings, most commonly affecting the elderly (5–7,53,54). Progression to malignancy may take several years (10,11,55). Consequently, although resection is generally life-saving once malignant transformation has occurred, many patients with premalignant tumors would likely benefit greatly from being spared surgery. Currently, however, conventional methods for the prediction of malignancy in PCLs are not sufficiently reliable to safely guide treatment recommendations (16,17). MUC1 expression by contrast, identified malignancy in PCLs with 89.7% accuracy and could thus add a new dimension to the evaluation of these lesions. Further studies are required to validate these promising findings.

Some limitations of the study merit further consideration. First, to avoid bias from underrepresentation of benign lesions, the analysis was not restricted to cases with supportive histology. The follow-up period for conservatively treated lesions was defined by what was considered clinically motivated for each patient and, whenever warranted, supported by review by a multidisciplinary board/senior expert surgeon. Second, the rare solid pseudopapillary neoplasms and cystic neuroendocrine tumors were not considered in this study because they can be readily identified by cytology/immunocytochemistry (41–44). It should therefore be emphasized that mucin profiling is not intended to replace routine cytological assessment. These two analyses can easily be performed in parallel.

Furthermore, for one of the four histological subtypes of IPMNs, the so-called intestinal type, previous reports on MUC1 expression after malignant transformation are conflicting (28,29,56). Evaluation of the performance of MUC1 specifically for the detection of malignancy in this tumor type would require a considerably larger patient series. However, intestinal-type IPMNs typically involve the main pancreatic duct, which is in itself an absolute indication for surgery, making the detection of malignant transformation less critical (5,7,33,34). In addition, intestinal-type IPMNs should be identifiable by mucin profiling given their characteristic predominant MUC2 expression, which could be used to mitigate this possible pitfall (28,29,31,33,34).

The protein and peptide identifications reported in Figure 4 could potentially inform the design of simplified and/or quantitative assays. Possible problems with contamination by gastric [MUC5AC, MUC6 (40,57)] and/or intestinal [MUC2 (57)] epithelium could, for instance, motivate quantitation, although this study did not identify this as a concern. Although antibody-based tests could provide a straightforward solution, the extensive and variable glycosylation of mucins might hamper their performance (20,21). A targeted, quantitative, high-throughput proteomic technique, such as selected reaction monitoring, could be a clinically feasible alternative (58). Meanwhile, mucin profiling, as described here, appears well suited for clinical use. Importantly, results could be obtained for lesions smaller than 1cm and delivered within a working week.

PCLs, although very common, may act as harbingers of one of the deadliest cancer forms known. Thus, accurate diagnostic tools for their assessment are urgently required. We conclude that the proteomic approach evaluated in this study shows great promise in this regard. If further studies can corroborate these results, this methodology should be considered for clinical recruitment.


This work was supported by The Swedish Research Council (no. 7461), the Health and Medical Care Executive Board of the Västra Götaland Region, the Swedish Society of Medicine, the Swedish Cancer Foundation, the Knut and Alice Wallenberg Foundation, the Bengt Ihre Foundation, the Research and Development Council in Södra Älvsborg, the Lindgren′s Stiftelse, the Mr and Mrs Backlund Foundation for Cancer Research, the Syskonen Persson Foundation, the 1012 anno 1964 Foundation, IngaBritt and Arne Lundberg Foundation, Sahlgrenska University Hospital (LUA-ALF), Wilhelm and Martina Lundgren’s Foundation, Torsten and Ragnar Söderberg′s Stiftelser, and the Swedish Foundation for Strategic Research–the Mucus-Bacteria-Colitis Center (MBC) of the Innate Immunity Program.


K.S. Jabbar and R. Sadik had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. K. S. Jabbar, G.C. Hansson, and R. Sadik were responsible for the study concept and design. K. S. Jabbar, C. Verbeke, A.G. Hyltander and R. Sadik were responsible for acquisition of data. K. S. Jabbar, C. Verbeke, A.G. Hyltander, H. Sjövall, G.C. Hansson, and R. Sadik were responsible for analysis and interpretation of data. K.S. Jabbar was responsible for drafting of the manuscript. K. S. Jabbar, C. Verbeke, A.G. Hyltander, H. Sjövall, G.C. Hansson, and R. Sadik were responsible for critical revision of the manuscript for important intellectual content. K.S. Jabbar was responsible for statistical analysis. G.C. Hansson and R. Sadik obtained funding. H. Sjövall, G.C. Hansson, and R. Sadik were responsible for study supervision.

The sponsors had no involvement in the design/conduct of the study, the analysis and interpretation of data, or the preparation, review, and approval of the manuscript.

The authors have no conflicts of interest to declare.

The Proteomics Core Facility and the Bioinformatics Core Facility at the University of Gothenburg are acknowledged for technical help and statistical advice, respectively. Dr Shahid Jabbar is thanked for critical review of the manuscript.


1. Zhang XM, Mitchell DG, Dohke M, et al. Pancreatic cysts: depiction on single-shot fast spinecho MR images. Radiology. 2002;223 (2):547–553 [PubMed]
2. Matthaei H, Schulick RD, Hruban RH, Maitra A. Cystic precursors to invasive pancreatic cancer. Nat Rev Gastroenterol Hepatol. 2011;8(3):141–150 [PMC free article] [PubMed]
3. Strobel O, Rosow DE, Rakhlin EY, et al. Pancreatic duct glands are distinct ductal compartments that react to chronic injury and mediate Shh-induced metaplasia. Gastroenterology. 2010;138(3):1166–1177 [PMC free article] [PubMed]
4. Canto MI, Hruban RH, Fishman EK, et al. Frequent detection of pancreatic lesions in asymptomatic high-risk individuals. Gastroenterology. 2012;142(4):796–804 [PMC free article] [PubMed]
5. Del Chiaro M, Verbeke C, Salvia R, et al. European experts consensus statement on cystic tumours of the pancreas. Dig Liver Dis. 2013;45(9):703–711 [PubMed]
6. Farrell JJ, Férnandez-del Castillo C. Pancreatic cystic neoplasms: management and unanswered questions. Gastroenterology. 2013;144(6):1303–1315 [PubMed]
7. Tanaka M, Férnandez-del Castillo C, Adsay V, et al. International consensus guidelines 2012 for the management of IPMN and MCN of the pancreas. Pancreatology. 2012;12(3):183–197 [PubMed]
8. Klimstra DS, Wenig BM, Heffess CS. Solid-pseudopapillary tumor of the pancreas: a typically cystic carcinoma of low malignant potential. Semin Diagn Pathol. 2000;17(1):66–80 [PubMed]
9. Adsay NV. Cystic lesions of the pancreas. Mod Pathol. 2007;20(Suppl 1):S71–S93 [PubMed]
10. Salvia R, Crippa S, Falconi M, et al. Branch-duct intraductal papillary mucinous neoplasms of the pancreas: to operate or not to operate? Gut. 2007;56(8):1086–1090 [PMC free article] [PubMed]
11. Tanno S, Nakano Y, Nishikawa T, et al. Natural history of branch duct intraductal papillary mucinous neoplasms of the pancreas without mural nodules: long-term follow-up results. Gut. 2008;57 (3):339–343 [PubMed]
12. Maire F, Hammel P, Terris B, et al. Prognosis of malignant intraductal papillary mucinous tumours of the pancreas after surgical resection. Comparison with pancreatic ductal adenocarcinoma. Gut. 2002;51(5):717–722 [PMC free article] [PubMed]
13. Schnelldorfer T, Sarr MG, Nagorney DM, et al. Experience with 208 resections for intraductal papillary mucinous neoplasm of the pancreas. Arch Surg. 2008;143(7):639–646 [PubMed]
14. Visser JC, Yeh BM, Qayyum A, Way LW, McCulloch CE, Coakley FV. Characterization of cystic pancreatic masses: relative accuracy of CT and MRI. Am J Roentgenol. 2007;189(3):648–656 [PubMed]
15. Frossard JL, Amouyal PL, Amouyal G, et al. Performance of endosonography-guided fine needle aspiration and biopsy in the diagnosis of pancreatic cystic lesions. Am J Gastroenterol. 2003;98:1516–1524 [PubMed]
16. Brugge WR, Lewandrowski K, Lee-Lewandrowski E, et al. Diagnosis of pancreatic cystic neoplasms: A Report of the Cooperative Pancreatic Cyst Study. Gastroenterology. 2004;126(5):1330–1336 [PubMed]
17. van der Waaij LA, van Dullemen HM, Porte RJ. Cyst fluid analysis in the differential diagnosis of pancreatic cystic lesions: a pooled analysis. Gastrointest Endosc. 2005;62(3):383–389 [PubMed]
18. Khalid A, Zahid M, Finkelstein SD, et al. Pancreatic cyst fluid DNA analysis in evaluating pancreatic cysts: a report of the PANDA study. Gastrointest Endosc. 2009;69(6):1095–1102 [PubMed]
19. Sawhney MS, Devarajan S, O’Farrel P, et al. Comparison of carcinoembryonic antigen and molecular analysis in pancreatic cyst fluid. Gastrointest Endosc. 2009;69(6):1106–1110 [PubMed]
20. Kufe DW. Mucins in cancer: function, prognosis and therapy. Nat Rev Cancer. 2009;9(12):874–885 [PMC free article] [PubMed]
21. Hollingsworth MA, Swanson BJ. Mucins in cancer: protection and control of the cell surface. Nat Rev Cancer. 2004;4(1):45–60 [PubMed]
22. Kohlgraf KG, Gawron AJ, Higashi M, et al. Contribution of the MUC1 tandem repeat and cytoplasmic tail to invasive and metastatic properties of a pancreatic cancer cell line. Cancer Res. 2003;63 (16):5011–5020 [PubMed]
23. Chaika NV, Gebregiworgis T, Lewallen ME, et al. MUC1 mucin stabilizes and activates hypoxia-inducible factor 1 alpha to regulate metabolism in pancreatic cancer. Proc Natl Acad Sci U S A. 2012;109(34):13787–13922 [PubMed]
24. Vincent A, Perrais M, Desseyn J-L, Aubert J-P, Pigny P, Van Seuningen I. Epigenetic regulation (DNA methylation, histone modifications) of the 11p15 mucin genes (MUC2, MUC5AC, MUC5B, MUC6) in epithelial cancer cells. Oncogene. 2007;26(45):6566–6576 [PubMed]
25. Yamazoe S, Tanaka H, Sawada T, et al. RNA interference suppression of mucin 5AC (MUC5AC) reduces the adhesive and invasive capacity of human pancreatic cancer cells. J Exp Clin Canc Res. 2010;29 (1):53 [PMC free article] [PubMed]
26. Qu CF, Li Y, Song YJ, et al. MUC1 expression in primary and metastatic pancreatic cancer cells for in vitro treatment by (213)Bi-C595 radioimmunoconjugate. Br J Cancer. 2004;91 (12):2086–2093 [PMC free article] [PubMed]
27. Besmer DM, Curry JM, Roy LD, et al. Pancreatic ductal adenocarcinoma mice lacking mucin 1 have a profound defect in tumor growth and metastasis. Cancer Res. 2011;71 (13):4432–4442 [PMC free article] [PubMed]
28. Yonezawa S, Nakamura A, Horinouchi M, Sato E. The expression of several types of mucin is related to the biological behavior of pancreatic neoplasms. J Hepatobiliary Pancreat Surg. 2002;9(3):328–341 [PubMed]
29. Adsay NV, Merati K, Andea A, et al. The dichotomy in the preinvasive neoplasia to invasive carcinoma sequence in the pancreas: differential expression of MUC1 and MUC2 supports the existence of two separate pathways of carcinogenesis. Mod Pathol. 2002;15(10):1087–1095 [PubMed]
30. Kim GE, Bae HI, Park HU, et al. Aberrant expression of MUC5AC and MUC6 gastric mucins and sialyl Tn antigen in intraepithelial neoplasms of the pancreas. Gastroenterology. 2002;123(4):1052–1060 [PubMed]
31. Lüttges J, Feyerabend B, Buchelt T, Pacena M, Klöppel G. The mucin profile of noninvasive and invasive mucinous cystic neoplasms of the pancreas. Am J Surg Pathol. 2002;26(4):466–471 [PubMed]
32. Maker AV, Katabi N, Gonen M, et al. Pancreatic cyst fluid and serum mucin levels predict dysplasia in intraductal papillary mucinous neoplasms of the pancreas. Ann Surg Oncol. 2011;18(1):199–206 [PubMed]
33. Furukawa T, Hatori T, Fujita I, et al. Prognostic relevance of morphological types of intraductal papillary mucinous neoplasms of the pancreas. Gut. 2011;60 (4):509–516 [PubMed]
34. Mino-Kenudson M, Fernández-del Castillo C, Baba Y, et al. Prognosis of invasive intraductal papillary mucinous neoplasm depends on histological and precursor epithelial subtypes. Gut. 2011;60(12):1712–1720 [PMC free article] [PubMed]
35. Scarlett CJ, Samra JS, Xue A, Baxter RC, Smith RC. Classification of pancreatic cystic lesions using SELDI-TOF mass spectrometry. ANZ J Surg. 2007;77(8):648–653 [PubMed]
36. Corcos O, Couvelard A, Dargère D, et al. Proteomic assessment of markers for malignancy in the mucus of intraductal papillary mucinous neoplasms of the pancreas. Pancreas. 2012;41(2):169–174 [PubMed]
37. Cuoghi A, Farina A, Z’graggen K, et al. Role of proteomics to differentiate between benign and potentially malignant pancreatic cysts. J Proteome Res. 2011;10(5):2664–2670 [PubMed]
38. Ke E, Patel BB, Liu T, et al. Proteomic analyses of pancreatic cyst fluids. Pancreas. 2009;38(2):e33–e42 [PMC free article] [PubMed]
39. Kosmahl M, Wagner J, Peters K, Sipos B, Klöppel G. Serous cystic neoplasms of the pancreas: an immunohistochemical analysis revealing alpha-inhibin, neuron-specific enolase, and MUC6 as new markers. Am J Surg Pathol. 2004;28(3):339–346 [PubMed]
40. Bartman AE, Buisine M-P, Aubert J-P, et al. The MUC6 secretory mucin gene is expressed in a wide variety of epithelial tissues. J Pathol. 1998;186(4):398–405 [PubMed]
41. Singhi AD, Chu LC, Tatsas AD, et al. Cystic pancreatic neuroendocrine tumors: a clinicopathologic study. Am J Surg Pathol. 2012;36(11):1666–1673 [PubMed]
42. Yoon WJ, Daglilar ES, Pitman MB, Brugge WR. Cystic pancreatic neuroendocrine tumors: endoscopic ultrasound and fine-needle aspiration characteristics. Endoscopy. 2013;45(3):189–194 [PubMed]
43. Mehta N, Modi L, Patel T, Shah M. Study of cytomorphology of solid pseudopapillary tumor of pancreas and its differentialdiagnosis. J Cytol. 2010;27(4):118–122 [PMC free article] [PubMed]
44. Pettinato G, Di Vizio D, Manivel JC, Pambuccian SE, Somma P, Insabato L. Solid-pseudopapillary tumor of the pancreas: a neoplasm with distinct and highly characteristic cytological features. Diagn Cytopathol. 2002;27(6):325–334 [PubMed]
45. Andersch-Björkman Y, Thomsson KA, Holmén Larsson JM, Ekerhovd E, Hansson GC. Large scale identification of proteins, mucins, and their O-glycosylation in the endocervical mucus during the menstrual cycle. Mol Cell Proteomics. 2007;6(4):708–716 [PubMed]
46. Brugge WR. Editorial: should all pancreatic cystic lesions be resected? Cyst fluid analysis in the differential diagnosis of pancreatic cystic lesions: a metaanalysis. Gastrointest Endosc. 2005;62(3):390–391 [PubMed]
47. De Jong K, Nio CY, Hermans JJ, et al. High prevalence of pancreatic cysts detected by screening magnetic resonance imaging examinations. Clin Gastroenterol Hepatol. 2010;8(9):806–811 [PubMed]
48. Kimura W, Nagai H, Kuroda A, Muto T, Esaki Y. Analysis of small cystic lesions of the pancreas. Int J Pancreatol. 1995;18(3):197–206 [PubMed]
49. Correa-Gallago C, Ferrone C, Thayer SP, Wargo JA, Warshaw AL, Férnandez-del Castillo C. Incidental pancreatic cysts: do we really know what we are watching? Pancreatology. 2010;10(2):144–150 [PMC free article] [PubMed]
50. Pelaez-Luna M, Chari ST. Cyst fluid analysis to diagnose pancreatic cystic lesions: an as yet unfulfilled promise. Gastroenterology. 2006;130(3):1007–1009 [PubMed]
51. Cho CS, Russ AJ, Loeffler AG, et al. Preoperative classification of pancreatic cystic neoplasms: the clinical significance of diagnostic inaccuracy. Ann Surg Oncol. 2013;20(9):3112–3119 [PubMed]
52. De Jong K, Poley JW, van Hooft JE, Visser M, Bruno MJ, Fockens P. Endoscopic ultrasound-guided fine-needle aspiration of pancreatic cystic lesions provides inadequate material for cytology and laboratory analysis: initial results from a prospective study. Endoscopy. 2011;43(7):585–590 [PubMed]
53. Crippa S, Férnandez-del Castillo C, Salvia R, et al. Mucin-producing neoplasms of the pancreas: an analysis of distinguishing clinical and epidemiologic characteristics. Clin Gastroenterol Hepatol. 2010;8(2):213–219 [PMC free article] [PubMed]
54. Rodriguez JR, Salvia R, Crippa S, et al. Branch-duct intraductal papillary mucinous neoplasms: observations in 145 patients who underwent resection. Gastroenterology. 2007;133(1):72–79 [PMC free article] [PubMed]
55. Crippa S, Salvia R, Warshaw A, et al. Mucinous cystic neoplasm of the pancreas is not an aggressive entity: lessons from 163 resected patients. Ann Surg. 2008;247(4):571–579 [PMC free article] [PubMed]
56. Nakamura A, Hourinochi M, Goto M, et al. New classification of pancreatic intraductal papillary–mucinous tumour by mucin expression: its relationship with potential for malignancy. J Pathol. 2002;197(2):201–210 [PubMed]
57. Carrato C, Balague C, de Bolos C, et al. Differential apomucin expression in normal and neoplastic human gastrointestinal tissues. Gastroenterology. 1994;107(1):160–172 [PubMed]
58. Gillette MA, Carr SA. Quantitative analysis of peptides and proteins in biomedicine by targeted mass spectrometry. Nat Methods. 2013;10(1):28–34 [PMC free article] [PubMed]

Articles from JNCI Journal of the National Cancer Institute are provided here courtesy of Oxford University Press