Malignant intraductal papillary mucinous neoplasm: are we doing the right thing?

ABSTRACT

BACKGROUND: Because of the malignant potential, resection has been recommended for some intraductal papillary mucinous neoplasms (IPMN). We hypothesize that a large cancer database could be used to evaluate national resection rates and survival for malignant IPMN. MATERIALS AND METHODS: Using the Surveillance Epidemiology and End Results (SEER) database, 1988-2003, cases of malignant IPMN were identified using histology codes. Age-adjusted incidence rates were calculated; Cochran-Armitage tests evaluated trends over time. Predictors of resection were evaluated using χ(2) and logistic regression. Kaplan-Meier curves and Cox models were constructed to evaluate survival. RESULTS: Of 1834 patients, 209 (11.4%) underwent resection. Annual age-adjusted incidence decreased over the study time-course (P<0.05), while annual proportion of patients presenting with localized lesions and the proportion being resected increased (P<0.05). Predictors of resection on multivariate analysis included localized stage [versus distant, adjusted odds ratio (OR) 31; 95% confidence interval (CI) 17-56], and more recent diagnosis [referent 1988-1991; 2000-2003, OR 3.0 (95%CI 1.7-5.3)]. Median survival for resected patients was 16 mo versus 3 mo without resection (P<0.0001). After adjusting for age, gender, stage, year, and tumor location, surgical resection remained a significant predictor of survival [hazard ratio 0.44 (95% CI 0.36-0.54), P<0.0001]. CONCLUSIONS: In this population-based cohort, detection of malignant IPMNs is decreasing, with an increasing proportion of patients diagnosed at local stages and undergoing resection. Increased awareness of IPMN may be contributing to earlier detection, which might include benign/premalignant lesions, and greater utilization of resection for appropriate candidates; thus, we may be improving survival for this most treatable form of pancreatic cancer.