Background: Classifications of intraductal papillary mucinous neoplasm (IPMN) remain ambiguous, especially for the mixed type. Factors predicting malignancy remain unclear. The aim of this study was to evaluate the usefulness of factors predicting malignancy in the new international consensus guidelines for resection of branch duct-type (BD)-IPMN and to compare them with those in the previous version.
Methods: A prospectively collected database of patients with biopsy-proven BD-IPMN was analysed to compare factors between the first and second consensus guidelines, particularly as predictors of malignancy.
Results: Of 350 patients with BD-IPMN, sensitivity (0.724) and balanced accuracy (0.751) of the second guidelines were superior to those (0.639 and 0.730) in the first version at the expense of slightly reduced specificity (0.779 versus 0.822 for the first version) by random forest models. Multiple logistic regression analysis showed that main pancreatic duct dilatation greater than 5 mm (hazard ratio (HR) 4.54, 95 per cent confidence interval 2.45 to 8.41; P < 0.001), mural nodules (HR 6.27, 3.27 to 12.01; P < 0.001) and carbohydrate antigen 19-9 level above 37 units/ml (HR 4.03, 1.83 to 8.90; P = 0.001) were independent predictors of BD-IPMN malignancy.
Conclusion: The new consensus guidelines provide better sensitivity, performance of factors predicting malignancy, and balanced accuracy in the diagnosis of BD-IPMN malignancy. Size alone was limited in predicting malignancy. Variability in clinical significance of the individual factors associated with a risk of malignancy indicates the need for a tailored approach in the management of patients with BD-IPMN.
© 2014 BJS Society Ltd. Published by John Wiley & Sons Ltd.