LAUSR

Background. Accurate preoperative diagnosis is critical for the determination of appropriate surgical indications. The aim of this study was to assess the accuracy of preoperative diagnosis and indications for operative therapy for presumed pancreatic neoplasms. Methods. From 2005 to 2013, 851 patients underwent pancreatectomies for presumed pancreatic neoplasms. A formal preoperative diagnosis was established during a multidisciplinary tumor board and compared to the final pathologic examination. The preoperative diagnosis and its accuracy were assessed according to demographics, symptoms, and diagnostic workup. Results. Tumors were benign in 8% of patients (n = 67), premalignant in 43% (n = 370), and malignant in 49% (n = 414). The mean number of preoperative examinations was 3.2; 27% (n = 144) of patients had computed tomography, magnetic resonance imaging, endoscopic ultrasonography, and fine needle examination all performed together. Preoperative diagnosis was confirmed in 89% of patients (n = 754). The morbidity and mortality rates were 65% and 1%, respectively. Of the 97 patients (11%) with a misdiagnosis, operative resection was ultimately relevant (premalignant, malignant tumor, or symptomatic benign tumor) in 51 (6%) but inappropriate in 46 (5%). The rate of misdiagnosis was increased for cystic lesions and in patients under 50 years of age. For lesions <2 cm, diagnostic accuracy was increased when computed tomography, magnetic resonance imaging, endoscopic ultrasonography, and fine needle examination were all performed together. Conclusion. Misdiagnosis can lead to an inappropriate resection in 5% of patients with presumed pancreatic neoplasms. For lesions difficult to characterize, such as small and cystic lesions, association of several modalities of preoperative workup could help to decrease the rate of inappropriate operative care.