Fluorescence In Situ Hybridization and K-ras Analyses Improve Diagnostic Yield of Endoscopic UltrasoundYGuided Fine-Needle Aspiration of Solid Pancreatic Masses Sofiya Reicher, MD,* Fatih Z. Boyar, MD,Þ Maher Albitar, MD,Þ Vladimira Sulcova, MS,Þ Sally Agersborg, PhD,Þ Visal Nga, MD,* Ying Zhou, PhD,þ Gang Li, PhD,þ Rose Venegas, MD,§ Samuel W. French, MD,§ David S. Chung, MD,* Bruce E. Stabile, MD,|| Viktor E. Eysselein, MD,*and Arturo Anguiano, MD, FFACMGÞ Objectives: Endoscopic ultrasound (EUS)Yguided fine-needle aspira- tion (FNA) is the main diagnostic modality for pancreatic mass lesions. However, cytology is often indeterminate, leading to repeat FNAs and delay in care. Here, we evaluate whether combining routine cytology with fluorescence in situ hybridization (FISH) and K-ras/p53 analyses improves diagnostic yield of pancreatic EUS-FNA. Methods: Fifty EUS-FNAs of pancreatic masses in 46 patients were retrospectively analyzed. Mean follow-up was 68 months. Thirteen initial cytologic samples (26%) were benign, 23 malignant (46%), and 14 atypical (28%). We performed FISH for p16, p53, LPL, c-Myc, MALT1, topoisomerase 2/human epidermal growth factor receptor 2, and EGFR, as well as K-ras/p53 mutational analyses. Results: On final diagnosis, 11 (79%) of atypical FNAs were malig- nant, and 3 benign (21%). Fluorescence in situ hybridization was neg- ative in all benign and all atypical samples with final benign diagnosis. Fluorescence in situ hybridization plus K-ras analysis correctly identified 60% of atypical FNAs with final malignant diagnosis. Combination of routine cytology with positive FISH and K-ras analyses yielded 87.9% sensitivity, 93.8% specificity, 96.7% positive predictive value, 78.9% negative predictive value, and 89.8% accuracy. Conclusions: Combining routine cytology with FISH and K-ras analyses improves diagnostic yield of EUS-FNA of solid pancreatic masses. We propose to include these ancillary tests in the workup of atypical cytology from pancreatic EUS-FNA. Key Words: pancreatic cancer, endoscopic ultrasound, atypical cytology, diagnostic accuracy Abbreviations: EUS - endoscopic ultrasound, FNA - fine-needle aspiration, FISH - fluorescence in situ hybridization, EGFR - epidermal growth factor receptor, HER2 - human epidermal growth factor receptor 2, TOP2 - topoisomerase 2 (Pancreas 2011;40: 1057Y1062) P ancreatic cancer is the fourth leading cause of cancer-related deaths in the United States, with an overall 5-year survival of less than 5%. 1 Surgery offers the only chance for cure; however, less than 20% of patients are candidates for surgical resection at presentation. Thus, early and expeditious diagnosis of pancreatic malignancy is crucial. In recent years, endoscopic ultrasound (EUS)Yguided fine- needle aspiration (FNA) has become a leading diagnostic mo- dality for pancreatic mass lesions. 2Y5 During this procedure, the pancreatic mass is initially visualized with EUS, and a needle is advanced through the duodenal or stomach wall to aspirate tissue for cytological analysis. The diagnostic yield of EUS-FNA for solid pancreatic masses is 70% to 83%. 2Y5 However, making a diagnosis of pancreatic malignancy can be challenging because of limited cellularity of the specimen and/or inconclusive cyto- logical analysis. The cytological findings are typically reported by the pa- thologist as malignant, benign, or atypical. The malignant cell diagnosis is based on several validated criteria such as nuclear enlargement, variations in nuclear size, nuclear membrane irreg- ularity, and nuclear crowding. 6,7 Atypical cell diagnosis is given when the criteria for malignancy are not met, but aspirated cells are not unequivocally benign. There have been variable reports on the incidence of atypical cell diagnosis after EUS-FNA. 3,8,9 Atypical cell diagnosis creates a difficult management dilemma as this diagnosis is considered insufficient for initiation of chemo- therapy or for surgery. Notably, most atypical cell diagnoses have been determined to be malignant on clinical follow-up or repeat aspirations. 3,10 Repeat FNAs expose the patient to additional risk of complications and ultimately lead to delays in care. One possible approach to enhance the EUS-FNA diagnostic yield is to combine routine cytology with auxiliary diagnostic techniques, for example, tumor marker analysis. In particular, the fluorescence in situ hybridization (FISH) has been recently used in other cancers, including workup of indeterminate biliary strictures. 11 Fluorescence in situ hybridization identifies specific chromosomal changes (gain or loss) using fluorescently labeled DNA probes. 12 Importantly, a number of tumor suppressor and oncogene chromosomal abnormalities typical for pancreatic cancer have been identified in recent years that can be used as targets for FISH. In addition to FISH, mutational analyses for K- ras and p53 are applied for pancreatic cancer detection. 13 In this study, we sought to evaluate whether FISH and K- ras/p53 detection of genomic alterations commonly found in pancreatic cancer can provide additional diagnostic information in cases of indeterminate EUS-FNA cytology and enhance sensitivity. Our results indicate that combining routine cytology with FISH and K-ras analyses improves the diagnostic yield of EUS-guided pancreatic FNA. MATERIALS AND METHODS Patients and Clinical Follow-Up In this study, we retrospectively reviewed 50 EUS-FNAs of solid pancreatic masses in 46 patients treated at Harbor-UCLA ORIGINAL ARTICLE Pancreas & Volume 40, Number 7, October 2011 www.pancreasjournal.com 1057 From the *Division of Gastroenterology, Harbor-UCLA Medical Center, Torrance; †Quest Diagnostics Nichols Institute, San Juan Capistrano; ‡Department of Biostatistics, School of Public Health, UCLA, Los Angeles; and Departments of §Pathology and ||Surgery, Harbor-UCLA Medical Center, Torrance, CA. Received for publication July 13, 2010; accepted April 11, 2011. Reprints: Sofiya Reicher, MD, Division of Gastroenterology, Harbor-UCLA Medical Center, 1124 W Carson St N-21, Torrance, CA 90502 (e-mail: sreicher@sbcglobal.net). The authors declare no conflicts of interest. Copyright * 2011 by Lippincott Williams & Wilkins Copyright © 2011 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.