Abstract:
Objective:To investigate the clinical features and surgical indications of subtypes of intraductal papillary mucinous neoplasm (IPMN) of the pancreas, and analyze its malignant risk factors.
Methods:The retrospective casecontrol study was conducted. The clinicopathological data of 77 patients with IPMN of the pancreas who were admitted to the First Hospital of Peking University from January 2008 to December 2016 were collected. The subtypes of IPMN of the pancreas detected by preoperative imaging examination included mainduct type (MDIPMN) in 46 patients, branchduct type (BDIPMN) in 12 patients, mixed type (MTIPMN) in 19 patients. The surgical indications were consulted from the Guideline for the diagnosis and treatment of pancreatic cystic lesions composed by the Pancreatic Surgery Group of Surgery Branch of China Medical Association. Surgical procedure was selected according to the location and size of the IPMN. Four to 6 cycles of chemotherapy with S1 and/or Gemcitabine were conducted for patients with malignant IPMN according to the tolerance and baseline characteristics. Observation indicators included: (1) comparison of the clinical features MDIPMN, MTIPMN and BDIPMN; (2) surgical and postoperative conditions; (3) results of postoperative pathological examination and malignant risk factors analysis; (4) accuracy evaluation of Sendai and Fukuoka guidelines for the diagnosis of malignant IPMN of the pancreas; (5) followup results and survival. Patients were followed up by outpatient examination and telephone interview till December 2016. The postoperative adjuvant therapy, tumor recurrence and metastasis of malignant IPMN patients and postoperative survival condition of all the patients were collected. Measurement data with normal distribution were expressed as

±s or average (range), and pairwise comparison was analyzed by t test. Measurement data with skewed distribution were expressed by median (range). Comparison between count data and univariate analysis were done by chisquare test. Multiple factors analysis was done by Logistic regression model. The survival curve was drawn and the survival rate were calculated by KaplanMeier method. The comparison of survival was done by Logrank test.
Results:(1) Comparison of clinical features between MDIPMN, MTIPMN and BDIPMN: The numbers of patients with symptoms, jaundice, those complicated with diabetes and elevated CA199 were 55, 20, 43 and 28 in MDIPMN and MTIPMN, and 6 , 0, 3 and 1 in BDIPMN, with statistically significant difference (x
2=5.421, 3.516, 5.525, 3.834, P<0.05). (2) Surgical and postoperative conditions: the operations for all the 77 patients were successfully done, including pancreaticoduodenectomy with or without preservation of pylorus on 45 patients, resection of head of pancreas with duodenum preservation on 3 patients, distal pancreatectomy on 23 patients, distal pancreatectomy combined with partial resection of spleen and stomach on 2 patients ( with greater curvature of stomach involvement), segmental pancreatectomy on 2 patients, total pancreatectomy on 2 patients. A total of 26 surgeryrelated complications were detected, including pancreatic fistulas (13), delayed gastric emptying (9), wound infection (2), abdominal hemorrhage (2), and all the complications were improved by conservative treatment. There was no perioperative mortality. The mean duration of hospital stay of the 77 patients was 16 days (range, 6-68 days). (3) Results of postoperative pathological examination and malignant risk factor analysis: ① results of postoperative pathological examination: no residual tumor was detected at the resection margin in all the 77 patients, including 47 with benign IPMN (29 with adenoma and 18 with midsevere atypical hyperplasia and without lymph node involvement) and 30 with malignant IPMN (all of them were invasive malignancy, including 17 patients with negative lymph node metastasis and 13 with positive lymph node metastasis). ② Malignant risk factor analysis of IPMN: multivariate analysis showed that age, jaundice, elevated carcinoembryonic antigen (CEA), elevated CA199, tumor diameter, tumor subtypes were associated with malignancy (x
2=6.531, 14.755, 10.243, 12.062, 6.416, 6.143, P<0.05). Multivariate analysis showed that jaundice, elevated CEA, elevated CA199, tumor diameter ≥3.0 cm, MDIPMN were independent risk factors influencing the malignancy of IPMN (OR=9.656, 42.853, 23.243, 34.387, 69.883, 95% confidence interval: 1.392-66.968, 2.088-879.674, 2.991-180.628, 3.313-356.878, 1.247-3 915.467, P<0.05). (4) Accuracy evaluation of the Sendai and Fukuoka guidelines in diagnosis of malignant IPMN. The sensitivity, specificity, positive and negative predictive values were 100.0%(30/30), 14.9%(7/47), 42.9%(30/70) and 100.0%(7/7) for the Sendai guideline and 86.7%(26/30), 48.9%(23/47), 52.0%(26/50) , 85.2%(23/27) for the Fukuoka guideline in diagnosis of malignant IPMN, with no significant difference in the sensitivity between the 2 guidelines (x
2=2.250, P>0.05), while significant difference in the specificity between the 2 guidelines were detected (x
2=12.500, P<0.05). (5) Followup and survival: Seventy of 77 patients were followed up, including 42 with benign IPMN and 28 with malignant IPMN. The median survival time was 35.0 months (range, 6.0-94.0 months). All the malignant IPMN patients received adjuvant therapy. The 1, 3, 5year overall survival rates of 47 patient with benign IPMN were 100.0%, 96.2% and 96.2%, respectively, and 1 patient died of cardiac infarction. The 1, 3, 5year overall survival rates of 30 patients with malignant IPMN were 96.6%, 81.8%, 38.6%, respectively, and 11 patients died of tumor recurrence or metastasis with median time of tumor recurrence or metastasis of 20.5 months (6.0-61.6 months). The 1, 3, 5year overall survival rates of 17 patients with negative lymph node metastasis were 100.0%, 100.0% and 60.0%, respectively, and the 1, 3, 5year overall survival rates of 13 patients with positive lymph node metastasis were 91.7%, 57.1% and 0, respectively. There was statistically significant difference between patients with benign and malignant IPMN (x
2=12.530, P<0.05). There was statistically significant difference between patients with negative lymph node metastasis and those with positive lymph node metastasis (x
2=16.977, P<0.05).
Conclusions:Patients with MDIPMN or MTIPMN are more vulnerable to be complicated with diabetes, jaundice, elevated CA199 and high malignancy, and thus surgery is recommended. Jaundice, elevated CEA and CA199, tumor diameter≥3.0 cm, MDIPMN are the independent risk factors influencing the malignancy of IPMN.