Evaluation of Diagnostic Efficacy of Image Guided Fine Needle Aspiration of Pancreatic Lesions. A Retrospective Study of 296 Patients
M Bansal, SE Monaco, WE Khalbuss. University of Pittsburgh, Pittsburgh, PA
Background: Cytology is an important diagnostic tool, used routinely for the diagnosis of pancreatic lesions. In this study we reviewed the fine needle aspiration biopsies (FNAB) of the pancreas and determined it's diagnostic accuracy.
Design: From our institutional data base, 296 pancreatic FNABs from Jan 2006 to Jan 2008 were retrospectively reviewed. The cases were analyzed to determine the nature of the lesion (solid/cystic), the cytological diagnosis (Dx), molecular analysis, and the histological follow-up (F/U). Sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) for the cytology results were calculated using the histological Dx as the gold standard.
Results: The cytological Dx of the 296 pancreatic FNABs were: 126 malignant, 8 suspicious, 60 atypical, 74 negative, and 28 unsatisfactory. There were 78 FNABs (26%)s with histological F/U and the cytological Dx for these cases were: 50 malignant, 6 suspicious, 9 atypical, 8 negative, and 5 unsatisfactory. Histological Dx was malignant in 48 of 50 cases (96%) diagnosed by FNAB as malignant, 5 of 6 cases (83.3%) diagnosed as suspicious, 7 of 9 cases (77.8%) diagnosed as atypical, and in 12 of 13 cases (92.3%) diagnosed as negative or unsatisfactory. Of the 5 false positive cases (FNAB of at least atypical), 2 had no representative material on surgical biopsies, one was a small lesion (7mm), and one had inadequate material on FNAB. Ten of 12( 83.3%) false negative cases were cystic lesions. Correctly diagnosed lesions by FNAB include: 100% of adenocarcinomas, 93% of neuroendocrine neoplasms, 50% of mucinous neoplasms (including intraductal papillary mucinous neoplasms), in addition to schwannoma, solid pseudopapillary tumor, and metastatic renal cell carcinoma. A total of 41 cases (14%) had molecular analysis. Of the these, 33 (81%) were cystic and 16 (48%) of these were positive for a KRAS mutation and 8 (24%) had loss of heterozygosity. Overall the pancreatic FNABs showed sensitivity of 86% and specificity of 55%. The PPV and NPV were 93.5% and 33.3%, respectively.
Conclusions: In our experience, FNAB of pancreatic lesions is a sensitive diagnostic tool with a high PPV, and is particularly helpful for solid lesions. The majority of false negative lesions were cystic mucinous neoplasms, which are extremely difficult due to paucicellular specimens and gastrointestinal contamination, and therefore, other ancillary diagnostic tools (molecular analysis) should be considered if a cystic lesion is suspected radiographically.
Monday, March 9, 2009 9:30 AM
Poster Session I Stowell-Orbison/Autopsy Award # 59, Monday Morning