All of the pancreatic specimens were obtained from Peking Union Medical College Hospital, Beijing, China. Between January 2000 and December 2009, a total of 900 pancreatic tumors were resected and confirmed by pathology analysis. A total of two expert pathologists reviewed all slides and classified the cases according to the new World Health Organization (WHO) classification system. A total of 61 patients were diagnosed with IPMN, and 29 exhibited invasive carcinomas associated with IPMN. Of the 61 IPMN cases, there was a failure to extract sufficient genomic DNA to assess KRAS genetic alterations in five cases. All IPMN lesions were associated with the pancreatic duct system, and microscopic observation did not reveal any ‘ovarian-type’ hypercellular stroma surrounding the neoplasms. The present study was approved by the Ethics Committee of Peking Union Medical College Hospital, and informed consent was obtained from all cases.

Histological grading of the IPMN was performed according to the criteria defined by the 2010 edition of the WHO classification system for digestive neoplasms (24). Infiltrative growth accompanied by desmoplastic fibrosis was used as the only criterion to determine infiltration. A small number of samples exhibited evidence of mucin leaking into the interstitial space and forming a simple mucous lake around the pancreatic duct in the absence of floating cancer cells, and this was not regarded as invasion in the present study. Invasive depth was measured from the duct to the deepest infiltrating cancer cells. A cut-off point of 0.5 cm was selected for the depth of invasion and the cases were divided into two groups: Cases with an invasive depth ≤0.5 cm was designated as IPMN-associated minimally invasive carcinoma (IPMN-MI), and invasive depth >0.5 cm was designated as IPMN-associated advanced invasive carcinoma (IPMN-AI). All sections were observed at a magnification of ×40, 100 and 200 by Olympus BX53 microscope (Olympus Corporation, Tokyo, Japan).

Neutral 10% formalin-fixed (at room temperature, overnight) and paraffin-embedded pancreatic neoplasm tissue sections were cut (6 µm), and subsequently microdissected manually to ensure the highest purity and the largest amount of tumor cells in each sample. The resection margins with non-tumorous tissue served as normal controls. Genomic DNA was extracted using the QIAmp DNA Mini kit (Qiagen Inc., Valencia, CA, USA).

A DxS ARMS KRAS mutation test kit (DXS International PLC, Farnham, UK) was used to detect point mutations in codons 12 and 13 of exon 2 of the KRAS gene. This ARMS assay was a quantitative polymerase chain reaction (qPCR) assay that detected a total of eight genetic mutations, including 35G>A, 35G>T, 35G>C, 34G>A, 34G>T, 34G>C, 37G>C and 38G>A. The mutation test was performed according to the manufacturer's protocol. qPCR was performed using an internal Scorpion reporter probe and primers specific for wild and mutant KRAS on the ABI 7500 Real-time polymerase chain reaction system (ABI; Thermo Fisher Scientific, Inc., Waltham, MA, USA). The following thermocycling conditions were used for the qPCR: 50°C for 2 min, initial denaturation at 95°C for 10 min; 40 cycles of 95°C for 15 sec and 60°C for 10 sec. Data analysis was performed using the ABI SDS software v1.4.0.25 (Thermo Fisher Scientific, Inc.). The differences between the cycle threshold (CT) values of the mutant and wild types were calculated. When the difference of CT value between wild and mutant type was less than the cut-off value, we determined that there was a corresponding mutation in the sample (25,26). Subsequently, the mutation of KRAS in codon 61 was detected via direct sequencing of the PCR amplification product with an automated ABI 3730 sequencer (ABI; Thermo Fisher Scientific, Inc.). The primers for codon 61 of KRAS were 5′-TCCCTTCTCAGGATTCCTACA-3′ and 5′-CAAAGAAAGCCCTCCCCAGT-3′. The following thermocycling conditions were used for the PCR amplification of codon 61: Initial denaturation at 94°C for 5 min; 40 cycles of 94°C for 30 sec, 58°C for 30 sec and 72°C for 30 sec; and final extension for 10 min at 72°C. The amplification was performed using TaKaRa Ex Taq polymerase (Takara Bio Inc., Otsu, Japan).

All statistical analyses were performed using SPSS version 12.0 (SPSS Inc., Chicago, IL, USA) software. Data are presented as the mean value. P<0.05 was considered to indicate a statistically significant difference. The comparison between continuous variables was calculated using a Student's t-test. The correlation between categorical variables was acquired by χ² test or Fisher's exact test. The χ² test was used in the IPMN constituent ratio analysis. The Kaplan-Meier method was used to assess survival time, and the log-rank test was applied to determine significance.

Of the 56 cases of IPMN, 33 were in the head of the pancreas, 12 in the tail of the pancreas, and 11 in the whole pancreas. The age of patients ranged between 39 and 79 years, and the average was 61.8 years. The onset manifestation was abdominal discomfort (abdominal pain, abdominal distention and diarrhea) in 34 cases. A total of 13 cases were detected accidentally with no clinical symptoms. The remaining 9 cases represented symptoms with jaundice (n=4), weight loss (n=2), back pain (n=1), a high level of cancer antigen 199 (n=1) and a high creatinine level (n=1). Sex, age and tumor location were not associated with invasion. However, 21/56 (37.5%) patients had a history of acute or chronic pancreatitis with a high amylase level in the blood on at least one occasion prior to the surgery, and 17 patients smoked for >10 years. However, pancreatitis (P=0.408) and smoking (P=0.771) were not associated with KRAS mutations.

IPMN were subclassified into four histological subtypes: gastric-type (15/56; 26.8%), intestinal-type (19/56; 33.9%), pancreatobiliary-type (19/56; 33.9%) and oncocytic-type (3/56; 5.3%). Based on the highest degree of architectural and cytological atypia, the present cohort comprised 12 (21.4%) IPMN cases with low-grade dysplasia, 11 (19.6%) IPMN cases with intermediate-grade dysplasia, 7 (12.5%) IPMN cases with high-grade dysplasia and 26 (46.4%) IPMN-associated invasive carcinomas (Table I). The majority of the gastric-type IPMN cases (14/15) were classified as IPMN with low- and intermediate-grade dysplasia (Fig. 1). Almost all of the pancreatobiliary-type IPMN cases (18/19) were IPMN-associated carcinomas. This suggested that pancreatobiliary-type IPMN are significantly associated with high grade dysplasia and invasive adenocarcinoma, while gastric-type IPMN were associated with low and intermediate grade dysplasia (P=0.004). IPMN-associated carcinomas were primarily PDAs (21/26; 80.8%), colloid carcinomas (4/26; 15.4%) and, rarely, undifferentiated carcinomas (1/26; 3.8%). A total of 11 cases belonged to the IPMN-microinvasive group, including 10 tubular adenocarcinomas and one colloid carcinoma.

A total of five out of 56 patients were lost to follow-up, 35 patients were alive and tumor-free, four patients were alive with IPMN, and 12 patients succumbed to IPMN-associated invasive carcinoma. The mean postoperative follow-up period was 40.8 months (range, 12–112 months) for living patients. The overall 5-year survival rate was 76%. The 5-year survival rate of the patients with non-invasive IPMN was 100%. The 5-year survival rate of the patients with invasive IPMN was not available; instead, the 3-year survival rate of the patients with invasive IPMN was 55%. Of the 12 patients who succumbed, eight presented with pancreatobiliary-type IPMN, three had intestinal-type IPMN and one had oncocytic-type IPMN. A patient diagnosed with colloid carcinoma succumbed due to a complication of Trousseau syndrome 11 months post-surgery. Of the 12 patients who succumbed, all of them presented with IPMN-associated carcinomas and succumbed 7–41 months post-surgery; 11 of these patients succumbed within 2 years. A total of four patients developed recurrent tumors in the remnant pancreas, and these patients presented with gastric-type IPMN with low-grade dysplasia, intestinal-type IPMN with intermediate-grade dysplasia, and intestinal-type and oncocytic-type IPMN with high-grade dysplasia, respectively. Recurrent tumors remained within the pancreatic ducts and exhibited the same histological features as their origin tumors. The first two patients had positive surgical margins in the first segmental pancreatectomy operation.

Disease-specific survival rates were analyzed using the Kaplan-Meier method. Upon comparison of the survival stages of IPMN-MI and IPMN-AI, it was identified that the survival stage of IPMN-AI tumors was significantly shorter compared with that of IPMN-MI tumors (log-rank test; P=0.0006; Fig. 2A). Among the three histological subtypes, gastric-type IPMN had the best prognosis, pancreatobiliary-type IPMN had the worst, and intestinal-type IPMN lay in between (log-rank test; P<0.0001; Fig. 2B). Pancreatobiliary-type IPMN were more likely to develop into infiltrative carcinomas.

KRAS mutations were identified in 28 of the 56 tested patients, with a rate of 50%. No mutation was detected in the matching normal tissues. The frequency of KRAS mutations was 60% (9/15 cases) in gastric-type IPMN, 52.6% (10/19 cases) in intestinal-type, and 47.3% (9/19 cases) in pancreatobiliary-type. KRAS mutations were not present in oncocytic-type IPMN. Except for oncocytic-type IPMN, the frequencies of KRAS mutations in IPMN with low-, intermediate- and high-grade dysplasia, and IPMN-associated carcinomas were 58.3% (7/12 cases), 27.3% (3/11 cases), 80% (4/5 cases) and 56% (14/25 cases), respectively. KRAS mutation rates in low-grade (58.3%) and high-grade IPMNs (80%) were increased compared with intermediate-grade IPMN (27%). There was no association between KRAS mutations and histological subtype (P=0.159), or invasive cancer (P=0.592) (Table I).

There were certain differences in KRAS mutations among the three histological subtypes. Of the 19 intestinal-type IPMN cases, 6/7 invasive intraductal papillary mucinous carcinoma exhibited KRAS mutations, and 2/9 borderline IPMN exhibited KRAS mutations. With the increased degree of dysplasia and invasion of intestinal-type IPMN, KRAS mutation rates increased (P=0.012) (Table II). Compared with intestinal-type IPMN with low- and intermediate-grade dysplasia, gastric-type IPMN with low- and intermediate-grade dysplasia had increased KRAS mutations (57.1 and 22.2%). Intestinal-type IPMN with associated carcinomas had a higher KRAS mutation prevalence compared with pancreatobiliary-type IPMN with associated carcinomas (86.7 and 44.4%). Due to the small number of cases, it was not possible to calculate a P-value.

The present study compared the histopathological features of the invasive group (26 cases) with KRAS mutation occurrence, and demonstrated that the infiltrative depth (P=0.753), differentiation (P=0.909), lymph node metastasis (P=0.177), infiltrative extent (P=0.419) and the type of carcinoma (P=0.468) had no association with KRAS mutation occurrence (Table III). The Kaplan-Meier survival curve demonstrated that survival was not associated with KRAS mutation (log-rank test; P=0.308; Fig. 2C).

The mutated site exhibited single-amino-acid substitutions, including G12D (13 cases; 41%), G12V (12 cases; 38%), G12S (two cases), G12A (one case), G12C (two cases), G12R (one case) and 61 CAA>CGA (one case), and double-amino-acid substitutions, including G12V/G12S (two cases), G12D/G12S (one case) and G12D/G12V (one case). Among the four cases with double mutations, three were invasive carcinomas and the other was a benign tumor. A total of eight out of 13 cases with G12D mutation were IPMN with high-grade dysplasia and/or IPMN-associated carcinomas, and seven out of 12 cases with G12V mutation were IPMN with high-grade dysplasia and/or IPMN-associated carcinomas. The altered amino acids were not associated with IPMN classification. No mutations were identified within codon 13 of exon 2.