Utility of commercial DNA analysis in detecting malignancy within pancreatic cysts.

CONTEXT Pancreatic cysts raise concern because of their malignant potential. Our aims were to assess accuracy of DNA analysis in detecting malignant pancreatic cysts at EUS-FNA and to determine whether DNA analysis added to imaging and cyst fluid studies enhanced International Association of Pancreatology (IAP) guidelines for resection of pancreatic cysts. METHODS This is a retrospective study including pancreatic cysts undergoing EUS-FNA and DNA analysis with k-ras and loss of heterozygosity testing. Diagnostic models of 2006 and 2012 IAP guidelines, DNA analysis alone, and DNA combined with 2012 IAP guidelines were developed, and area under receiver operator characteristic curves (AUC) compared to determine the added value of DNA for detecting malignant cysts at the time of EUS-FNA. RESULTS Two-hundreds and fifty-seven patients were included with 8 (3.1%) malignant cysts. Solid component (P<0.001), main pancreatic duct dilation (P=0.012), suspicious or malignant cytology (P=0.001), and high DNA quantity (P<0.001) were associated with malignancy. Concurrent high amplitude k-ras with loss of heterozygosity mutations was highly specific (98.4%) though insensitive (12.5%) for malignancy. The 2012 IAP guideline (AUC=0.87; 95% CI: 0.82-0.91) was superior to 2006 IAP guideline (AUC=0.54; 95% CI: 0.47-0.60) and DNA analysis alone (AUC=0.60; 95% CI: 0.53-0.66) for detecting malignant cysts (P=0.004 and P=0.002, respectively). Addition of DNA did not improve performance of the 2012 IAP guideline (AUC=0.84; 95% CI: 0.79-0.88). CONCLUSIONS Commercial DNA analysis does not add useful information beyond imaging and cytology for detection of malignant pancreatic cysts. The 2012 IAP guideline better predicted malignant cysts than the 2006 IAP guideline.