Comparison of preoperative conference-based diagnosis with histology of cystic tumors of the pancreas.

Background: Diagnostic errors in the evaluation of cystic neoplasms of the pancreas (PCNs) are quite common. Few data are available regarding the impact of these errors on clinical management. The aim of this study was to determine the accuracy of a pancreatic multidisciplinary conference in diagnosing PCNs, to assess the potential risk of misdiagnosis, and to evaluate the clinical impact of these errors.
Methods: A retrospective consecutive series of patients undergoing surgery for PCNs at Karolinska University Hospital between 2004 and 2012 was analyzed.
Results: During the study period, a total of 141 patients had undergone pancreatic resection for PCN. The overall accuracy of the preoperative diagnosis was 60.9 %. The rate of concordance between preoperative diagnosis and histology was similar for asymptomatic and symptomatic lesions (62.8 vs. 59.1 %; p = NS). The rate of correct diagnosis increased over time (54.5 % in 2004-2006, 61.7 % in 2007-2012, 63.5 % in 2010-2012). Univariate analysis identified the location of the lesion (diffuse pancreatic involvement) and a mucinous nature of the lesion as factors conducive to a correct diagnosis. Reevaluation of the original indication for surgery in light of the exact diagnosis showed that a surgical procedure should not have been performed in 12 patients (8.5 %).
Conclusions: This study confirms that diagnostic errors are fairly common in the preoperative assessment of PCNs, but the errors are clinically relevant in <10 % of patients.