S52 Results: ‘Suspicious for neoplastic mucinous cyst’ and ‘Intraductal papillary mucinous cyst (or Mucinous cystic neoplasm)’ were preferred and ranked equally. The term ‘Atypical mucinous cyst’ was ranked the lowest. Regarding dysplasia, ‘At least low grade dysplasia (or intermediate grade or high grade dysplasia) is seen’ was ranked highest. Conclusions: Our study shows that clinicians prefer terms in FNA reports that are equivalent or similar to the WHO terminology used in surgical pathology reports for pancreatic neoplastic mucinous tumors. The terminology used in the reports should be mutually understandable and should convey clearly the level of concern by pathologists avoiding ambiguous terminology. 113 The Impact of CEA Fluid Analysis Over Negative and Non-diagnostic Pancreatic Cysts Aspirates: A 5-year Single Institutional Experience Mauro Saieg, MD, PhD, Shanna Colletti, CT(ASCP)CM, Valerie Munson, MS, Larissa Chambers, SCT(ASCP), Aziza Nassar, MD Mayo Clinic, Rochester, Minnesota Introduction: The new proposed Papanicolaou Society of Cytopathology classification for pancreaticobiliary cytology has advocated the use of CEA cyst fluid analysis as an auxiliary diagnostic tool for the determination of mucinous neoplasms in pancreatic cysts aspirates. The objective of the current study was to investigate the impact of the CEA cyst fluid analysis over cases primarily called negative or non-diagnostic. Materials and Methods: We retrospectively collected all pancreatic cyst aspirates from 2011 to 2014. Clinical data such as age, gender, cytological diagnosis and CEA levels were extracted from the patients’ electronic files. Cases primarily classified as non-diagnostic or negative with a cystic CEA level of 192 ng/ml or higher were reclassified as mucinous neoplasms. Statistical analysis was performed using Pearson’s chi-square test. Results: Three hundred and one pancreatic cyst aspirates were retrieved. Median age was 70 years old (range 25-100). Primary diagnoses included 188 (62.5%) negative and 31 (10.3%) non-diagnostic samples. CEA fluid analysis was available for 255 (84.7%) cases. When taking the CEA fluid level into consideration for diagnosis, 113 (37.5%) samples were considered negative and only 16 (5.3%) remained as non-diagnostic, with a significant decrease for both categories (p<0.0001 and pZ0.02, respectively). Conclusions: The use of CEA fluid analysis has significantly decreased the number of samples primarily classified as negative and non-diagnostic, with an important augment in the diagnosis of mucinous cystic neoplasms. As recommended by the new proposed terminology, the analysis of CEA in the cyst aspirate should be a vital part of any pancreatic cysts diagnostic algorithm. 114 The Accuracy of Endoscopic Ultrasound-Guided Fine Needle Aspiration in the Diagnosis of Pancreatic Mucinous Cystic Neoplasms Uma Kundu, MD, Savitri Krishnamurthy, MD MD Anderson Cancer Center, Houston, Texas Introduction: Endoscopic ultrasound-guided fine needle aspiration (EUSFNA) is used for the evaluation of pancreatic cystic lesions. Mucinous cystic neoplasms(MCN) and intraductal papillary mucinous neoplasms(IPMN) are commonly encountered pancreatic cystic neoplasms that require accurate diagnosis for the management/treatment of patients. We evaluated the accuracy of FNA alone and in combination with cyst fluid carcinoembryonic acid (CEA) levels for preoperative diagnosis of MCN and IPMN. Materials and Methods: We searched our pathology database from 20042014 for patients with EUS-FNA of pancreatic cystic lesions subsequently diagnosed as MCN or IPMN on surgical resection. Results of cytopathological examination of procured material by EUS-FNA and the cyst fluid CEA levels were recorded and correlated. Elevated CEA level was considered as >367 ng/ml. The sensitivity of EUS-FNA alone and in combination with cyst fluid CEA levels was determined. Results: 48 patients had pancreatic cystic lesions (41 IPMN and 7 MCN). Four of 7 (57%) patients with MCN and 23 of 41 (56.1%) patients with IPMN were correctly identified by EUS-FNA. The false negative diagnoses in patients
Abstracts diagnosed finally as MCN included mucinous lesion (ML)(2) and mucinous cyst(1). In patients with IPMN, the initial diagnoses other than IPMN in 18 patients included adenocarcinoma (2), MCN/ML(9), mucin (2), and mucinous epithelium (5). Improved sensitivity of 100% was achieved when cytology was combined with elevated CEA levels for the initial diagnosis of MCN. The sensitivity of cytology alone for the initial diagnosis of IPMN of 56% improved to only 59% by combining cytology and CEA levels in cyst fluid. Conclusion: The sensitivity of cytology alone of EUS-FNA of pancreatic cystic lesions for the diagnosis of MCN and IPMN was 57% and 56.1%. While the combination of cytology and CEA levels improved the sensitivity for MCN (100%) significantly, there was no significant improvement in the case of IPMN (59%). 115 Clinical Implications of Accurate Cytologic Diagnosis of Oncocytic Type Intraductal Papillary Mucinous Neoplasm of the Pancreas Christina Stallworth, MD1, Melinda Lewis, MD1, Gizem Akkas, MD1, Olca Basturk, MD2, Volkan Adsay, MD1, Michelle D. Reid, MD1 1 Emory University School of Medicine, Atlanta, Georgia 2 Memorial Sloan Kettering Cancer Center, New York, New York Introduction: Most intraductal papillary mucinous neoplasms (IPMNs) present as innocuous-appearing cystic lesions. A subset classified as oncocytic variant of IPMN frequently presents as complex tumors that are often clinically misinterpreted as cancerous. The cytologic findings in 5 pancreatic oncocytic IPMNs, 4 in head and 1 in body/tail region. Accurate cytologic diagnosis on fine needle aspiration (FNA) was crucial in establishing appropriate treatment and avoiding unnecessary chemo-radiation. Materials and Methods: Five oncocytic IPMNs had endoscopic ultrasoundguided FNA with rapid on-site evaluation (ROSE). All were deemed adequate and 2 were interpreted as suspicious for adenocarcinoma at the time of ROSE. All 5 had follow-up resections that were reviewed. Results: Two males and 3 females, 56e84 years (mean 66) had radiologic diagnoses of “pancreatic ductal adenocarcinoma (PDAC)” (nZ2), IPMN with “invasive cancer” (nZ1), “IPMN vs mucinous cystic neoplasm” (nZ1) and “cystic mass” (nZ1). Cytologic findings included hypercellular smears (80%), composed of oncocytic cells with large nuclei (100%) containing prominent eccentric nucleoli (80%), arranged predominantly in sheets/papillae (100%) with fibrovascular cores (80%), punched out inter-cytoplasmic spaces (80%), threedimensional groups and focal necrosis (60%). Classical IPMN cytology (intracytoplasmic mucin and thick extracellular mucin) was only found after meticulous inspection (80% and 40% respectively). Mean tumor size on resection was 5 cm and invasive carcinoma was seen in 60% (mean size 0.8 cm, 0.1e2.0 cm). Initial cytologic diagnoses included oncocytic IPMN (nZ3), IPMN, NOS (nZ1) and PDAC (nZ1). The latter case proved to be non-invasive oncocytic IPMN on resection. Conclusion: While the cytologic features of oncocytic IPMNs are consistent and classical, and are similar to their histologic counterparts, failure to recognize them on FNA may lead to erroneous diagnosis and overtreatment with neoadjuvant chemoradiation which is typically ineffective in these notoriously indolent neoplasms which (even when invasive) have an overall better prognosis than invasive PDAC. 116 Role of Cytology in Management of Branch Duct Intraductal Papillary Mucinous Neoplasms of the Pancreas: An Institutional Experience Amber Smith, MD, Abha Goyal, MD Cleveland Clinic, Cleveland, Ohio Introduction: The indications for resection of branch duct intraductal papillary mucinous neoplasms (BD-IPMNs) of the pancreas have become more conservative. The 2012 international consensus guidelines recommend resection for “high-risk” (HR) pancreatic cysts and endoscopic ultrasound (EUS) evaluation with cytologic assessment for “worrisome” (WR) cysts. In this study, we evaluate the performance characteristics of cytology in addition to EUS/ imaging and clinical findings in the management of BD-IPMNs at our institution. Materials & Methods: A database search was performed for histologically confirmed BD-IPMNs from 06/2005 to 06/2014. All cases with available
Abstracts cytology (within 6 months prior to the resection), EUS/imaging and clinical findings were included for further analysis. Cytology slides were rereviewed and findings were classified as per the recent Papanicolaou Society of Cytopathology recommendations. Clinico-pathologic information was obtained from the patients’ electronic medical record. Results: Twenty-nine cases met the study criteria: one was HR with solid enhancing component, eleven were with at least one WR feature (9 with cyst size 3 cm, 3 with mural nodules, 2 with thickened wall) and seventeen lacked HR or WR imaging features (cyst size: 0.7-2.7 cm). The HR case was malignant both on histology and cytology. Of the WR cases, 3 (27%) showed high-grade dysplasia (HGD) or worse on histology (one of which was detected on cytology). Of the cases without HR or WR imaging features e 16 were symptomatic, 6 had associated pancreatitis, 3 (18%) showed HGD on histology (one of which was detected on cytology). For cases with diagnostic cytology, its negative predictive value for HGD or worse was 82%. Conclusions: Cytology plays an important role in the pre-operative risk stratification of BD-IPMNs primarily due to its high negative predictive value. In conjunction with EUS and imaging, it also assists in the detection of high-grade dysplasia or worse in WR cases. 117 What Is the Optimal Management for the Atypical Cytologic Diagnostic Category in EUS-FNA of the Pancreas: An Algorithm Based on Regret Decision Curve Analysis Evan Alston, MD, Isam Eltoum, MD, MBA, Kondal Kyanam Kabir Baig, MD, Jayapal Ramesh, MD, FRCP, FASGE University of Alabama Birmingham Hospital, Birmingham, Alabama Introduction: The objective of this study is to apply regret-based decision analysis to identify the best strategy for follow-up of pancreatic lesions with an atypical cytologic diagnosis on EUS-FNA. Materials and Methods: This retrospective study includes all cases of atypical cytologic diagnoses (ADC) of pancreatic EUS-FNAs in the period 2000-2013. Two endoscopists were asked to rate their regret if they selected surgical exploration or watchful clinical follow-up but were proven wrong later. Using the expected regret, we calculated an individualized probability threshold at or below which the endoscopist will choose exploration over clinical follow-up. Six possible follow-up strategies were used to build a regret-based decision curve and the optimum strategy was selected based on least expected regret. Results: 187 (4.9%) of 3832 cases with pancreatic EUS-FNAs were ADC. Of 93 neoplasms (55%) - 61 (36%) carcinomas - 19 (11%) were lost to follow-up. The mean probability threshold for ADC varied from 0.01 to 0.97. Based on minimal regret, we identified three cut-off values of Pt to build an algorithm for individualized triage of ADC (0.1, 0.2, and 0.85). Repeat EUS-FNA or core biopsy was the optimal strategy for a large number of patients and over a large range of probability thresholds. Conversely, watchful clinical follow-up was optimal for only a few cases (all benign). A clinical predictive model performs well at the lower end but not at the higher end of Pt. If a neoplasm was highly suspected and the lesion was resectable, exploration was the optimal strategy. Finally, pancreatic duct brushing was dominated by other strategies. Conclusions: Regret decision curve analysis is a novel, powerful tool for individualized triage of patients with ADC and lends itself to application in other organ systems, for optimal decision-making in the midst of uncertainty, utilizing both intuitive and deliberate processes. 118 Value of Onsite Cytopathologist Evaluation in Endoscopic UltrasoundGuided Fine-Needle Aspiration of the Pancreas Lorene Yoxtheimer, MD, Richard Cantley, MD, Stacy Molnar, BS, SCT(ASCP), Luis De Las Casas, MD University of Toledo, Toledo, Ohio Introduction: Many studies have reported variable results for the value of onsite cytologic evaluation (OSCE) of aspirated material for Endoscopic Ultrasound-Guided Fine-Needle Aspiration (EUSFNA) of the pancreas.
S53 Some studies claim that there is no need for rapid onsite evaluation. We believe that OSCE allows essential exchange of information between cytopathologist and endoscopist for accurate diagnoses and allows adequate and timely triage of aspirated material for ancillary studies. Materials and Methods: A retrospective review of pancreatic EUSFNA in archival pathologic material at our institution from January 2011 to December 2014 was performed. Cytopathologist presence, immediate interpretation, and triage of material were documented in every procedure. Results: 242 pancreatic EUSFNA were diagnosed as neoplastic. Most of the specimens were referrals from outside institutions specifically for EUSFNA. Many patients with malignant pancreatic diagnoses were inoperable and those who were candidates for surgical resection frequently returned to their referring institutions for definitive surgery. 24 pancreatic EUSFNA specimens were found to have surgically excised tissue available for review at our hospital. All 24 cases were confirmed on surgical pathology tissue evaluation (100% positive predictive value), including 11 ductal adenocarcinomas, 4 pancreatic neuroendocrine tumors, 1 solid pseudopapillary tumor, and 8 intraductal papillary mucinous neoplasms. Conclusions: The result of this small pilot study strongly supports that the cytopathologist presence and interaction with the endoscopists during onsite evaluation of pancreatic EUSFNA yields accurate and definitive diagnostic interpretations. 119 Endoscopic Ultrasound Guided Fine Needle Aspiration (EUS-FNA) Diagnosis of Metastatic Neoplasms to the Pancreas: An Institutional Experience Fang Zhou, MD, Dianne Grunes, MD, Melissa Yee-Chang, DO, Gabriel Acosta-Gonzalez, MD, Ronaldo Zamuco, MD, Joan Cangiarella, MD, Xiao-Jun Wei, MD, Aylin Simsir, MD, Yan Shi, MD, PhD New York University Medical Center/School of Medicine, New York, New York Introduction: Metastatic neoplasms (MN) are rare in the pancreas. An accurate diagnosis is challenging because MNs mimic primary pancreatic neoplasms, both clinically and on cytology. However, the distinction is critical for patient management. In this study, we reviewed our experience in diagnosing MNs by EUS-FNA of the pancreas. Material and Methods: We searched our database for pancreatic EUS-FNA specimens with a diagnosis of MN from 1994 to 2014. The clinical history, radiologic findings and follow-up of these cases, if available, were reviewed. Results: There were 17 cases of MNs to the pancreas in 7 males and 10 females, ranging in age from 37 to 85 years (mean Z 62). The primary malignancies included carcinomas of the lung (4), colon (3), breast (2), ovary (1), kidney (1), liver (1), melanoma (3) and sarcoma (2). The pancreatic head and neck were the most common locations (73%).16 cases (94%) had a known prior history of malignancy; the clinical history was not provided in one case. All cases presented as a single mass in the pancreas. The average tumor size was 1.9 cm (range: 0.5 - 4 cm). 12 cases (71%) were poorly-differentiated carcinomas, indistinguishable from a pancreatic adenocarcinoma without immunohistochemical (IHC) studies and/or clinical history. 12 (71%) cases were correctly diagnosed as MN, 3 (18%) cases had indeterminate tumor origin, and 2 (12%) were misdiagnosed as primary pancreatic adenocarcinoma. A correct diagnosis was reached by cytomorphology alone in 3 cases (18%); morphology and immunohistochemical stains in 7 cases (41%); and morphologic comparison to the prior tumors in 2 cases (12%). Conclusions: EUS-FNA is an effective approach to diagnose pancreatic tumors. MNs can be difficult to differentiate from primary pancreatic carcinomas based on cytology alone. Clinical history and adequate cell block for IHC studies are essential to reach an accurate diagnosis. 120 Endoscopic Ultrasound Guided Fine Needle Aspiration Cytology of Metastatic Renal Cell Carcinoma to the Pancreas: A Study of 33 Cases Karyn Hallberg, CT(ASCP)1, Amber Smith, MD2, Aziza Nassar, MD2, Jun Zhang, MD1, Matthew Zarka, MD1, Jordan Reynolds, MD2,