A randomized prospective multicenter trial of pancreaticoduodenectomy with and without routine intraperitoneal drainage.

OBJECTIVE: To test by randomized prospective multicenter trial the hypothesis that pancreaticoduodenectomy (PD) without the use of intraperitoneal drainage does not increase the frequency or severity of complications. BACKGROUND: Some surgeons have abandoned the use of drains placed during pancreas resection. METHODS: We randomized 137 patients to PD with (n = 68, drain group) and without (n = 69, no-drain group) the use of intraperitoneal drainage and compared the safety of this approach and spectrum of complications between the 2 groups. RESULTS: There were no differences between drain and no-drain cohorts in demographics, comorbidities, pathology, pancreatic duct size, pancreas texture, baseline quality of life, or operative technique. PD without intraperitoneal drainage was associated with an increase in the number of complications per patient [1 (0-2) vs 2 (1-4), P = 0.029]; an increase in the number of patients who had at least 1 ≥grade 2 complication [35 (52%) vs 47 (68%), P = 0.047]; and a higher average complication severity [2 (0-2) vs 2 (1-3), P = 0.027]. PD without intraperitoneal drainage was associated with a higher incidence of gastroparesis, intra-abdominal fluid collection, intra-abdominal abscess (10% vs 25%, P = 0.027), severe (≥grade 2) diarrhea, need for a postoperative percutaneous drain, and a prolonged length of stay. The Data Safety Monitoring Board stopped the study early because of an increase in mortality from 3% to 12% in the patients undergoing PD without intraperitoneal drainage. CONCLUSIONS: This study provides level 1 data, suggesting that elimination of intraperitoneal drainage in all cases of PD increases the frequency and severity of complications.

Treatment of bacteriobilia decreases wound infection rates after pancreaticoduodenectomy.

BACKGROUND: Although mortality following pancreaticoduodenectomy is decreasing, postoperative morbidity remains high. It was hypothesized that culture-directed treatment of bacteriobilia would decrease the incidence of infectious complications following pancreaticoduodenectomy. METHODS: In a retrospective study of 197 pancreaticoduodenectomy patients, those in the control group (n = 128, 2005-2009) were given perioperative prophylactic antibiotics, whereas those in the treatment group (n = 69, 2009-2011) were continued on antibiotics until intraoperative bile culture results became available. Patients with bacteriobilia received 10 days of antibiotic treatment, which was otherwise discontinued in patients without bacteriobilia. Various complication rates were compared using Fisher's exact test for categorical variables, Wilcoxon rank sum test for ordinal variables, and a two-sample t-test for continuous variables. RESULTS: Demographics, comorbidities, baseline clinical characteristics, and intraoperative and postoperative variables were similar between the two groups. There were higher incidences of elevated creatinine (19% versus 4%; P = 0.004) and preoperative hyperglycaemia (18% versus 7%; P = 0.053) in the control group. Fewer patients in the control group underwent preoperative biliary stenting (48% versus 67%; P = 0.017) and intraperitoneal drains were placed at the time of resection more frequently in the control group (85% versus 38%; P < 0.001). Bacteriobilia was found in 59% of patients. Treatment of bacteriobilia was associated with a decrease in the rate of postoperative wound infections (12% in the control group versus 3% in the treatment group; P = 0.036) and overall complication severity score (1 in the control group versus 0 in the treatment group; P = 0.027). CONCLUSIONS: Prolonged antibiotic therapy for bacteriobilia may decrease postoperative wound infection rates after pancreaticoduodenectomy. A randomized prospective trial is warranted to provide evidence to further support this practice.

Pancreatic resection without routine intraperitoneal drainage.

BACKGROUND: Most surgeons routinely place intraperitoneal drains at the time of pancreatic resection but this practice has recently been challenged. OBJECTIVE: Evaluate the outcome when pancreatic resection is performed without operatively placed intraperitoneal drains. METHODS: In all, 226 consecutive patients underwent pancreatic resection. In 179 patients drains were routinely placed at the time of surgery and in 47 no drains were placed. Outcomes for these two cohorts were recorded in a prospective database and compared using the χ(2) - /Fisher's exact test for categorical variables, and Wilcoxon's test for continuous variables. RESULTS: Demographic, surgical and pathological details were similar between the two cohorts. Elimination of routine intraperitoneal drainage did not increase the frequency or severity of serious complications. However, when all grades of complications were considered, the number of patients that experienced any complication (65% vs. 47%, P= 0.020) and the median complication severity grade (1 vs. 0, P= 0.027) were increased in the group that had drains placed at the time of surgery. Eliminating intra-operative drains was associated with decreased delayed gastric emptying (24% vs. 9%, P= 0.020) and a trend towards decreased wound infection (12% vs. 2%, P= 0.054). The readmission rate (9% vs. 17% P= 0.007) and number of patients requiring post-operative percutaneous drains (2% vs. 11%, P= 0.001) was higher in patients who did not have operatively placed drains but there was no difference in the re-operation rate (4% vs. 0%, P= 0.210). CONCLUSION: Abandoning the practice of routine intraperitoneal drainage after pancreatic resection may not increase the incidence or severity of severe post-operative complications.

Engineered T cells for pancreatic cancer treatment.

OBJECTIVE: Conventional chemotherapy and radiotherapy produce marginal survival benefits in pancreatic cancer, underscoring the need for novel therapies. The aim of this study is to develop an adoptive T cell transfer approach to target tumours expressing prostate stem cell antigen (PSCA), a tumour-associated antigen that is frequently expressed by pancreatic cancer cells. METHODS: Expression of PSCA on cell lines and primary tumour samples was confirmed by immunohistochemistry. Healthy donor- and patient-derived T cells were isolated, activated in vitro using CD3/CD28, and transduced with a retroviral vector encoding a chimeric antigen receptor (CAR) targeting PSCA. The ability of these cells to kill tumour cells was analysed by chromium-51 (Cr(51)) release. RESULTS: Prostate stem cell antigen was expressed on >70% of the primary tumour samples screened. Activated, CAR-modified T cells could be readily generated in clinically relevant numbers and were specifically able to kill PSCA-expressing pancreatic cancer cell lines with no non-specific killing of PSCA-negative target cells, thus indicating the potential efficacy and safety of this approach. CONCLUSIONS: Prostate stem cell antigen is frequently expressed on pancreatic cancer cells and can be targeted for immune-mediated destruction using CAR-modified, adoptively transferred T cells. The safety and efficacy of this approach indicate that it deserves further study and may represent a promising novel treatment for patients with pancreatic cancer.

Routine nasogastric suction may be unnecessary after a pancreatic resection.

BACKGROUND: Most surgeons routinely place a nasogastric tube at the time of a pancreatic resection. The goal of the present study was to evaluate the outcome when a pancreatic resection is performed without routine post-operative nasogastric suction. METHODS: One hundred consecutive patients underwent a pancreatic resection (64 a pancreaticoduodenectomy, 98% pylorus sparing and 36 a distal pancreatectomy). In the first cohort (50 patients), a nasogastric tube was routinely placed at the time of surgery and in the second cohort (50 patients) the nasogastric was removed in the operating room. Outcomes for these two cohorts were recorded in a prospective database and compared using the χ(2) or Fisher's exact test and Wilcoxon's rank-sum test. RESULTS: Demographical, surgical and pathological details were similar between the two cohorts. A post-operative complication occurred in 22 (44%) in each group (P= 1.000). There were no statistically significant differences in the frequency or severity of complications, or length of stay between groups. The spectrum of complications experienced by the two cohorts was similar including complications that could potentially be related to the use of nasogastric suction such as delayed gastric emptying, anastomotic leak, wound dehiscence and pneumonia. There was no difference between the two groups in the number of patients who required post-operative nasogastric tube placement (or replacement) [2 (4%) vs. 4 (8%), P= 0.678]. CONCLUSION: It may be safe to place a nasogastric tube post-operatively in a minority of patients after a pancreatic resection and spare the majority the discomfort associated with routine post-operative nasogastric suction.

Drain data to predict clinically relevant pancreatic fistula.

BACKGROUND: Post-operative pancreatic fistula (POPF) is a common and potentially devastating complication of pancreas resection. Management of this complication is important to the pancreas surgeon. OBJECTIVE: The aim of the present study was to evaluate whether drain data accurately predicts clinically significant POPF. METHODS: A prospectively maintained database with daily drain amylase concentrations and output volumes from 177 consecutive pancreatic resections was analysed. Drain data, demographic and operative data were correlated with POPF (ISGPF Grade: A--clinically silent, B--clinically evident, C--severe) to determine predictive factors. RESULTS: Twenty-six (46.4%) out of 56 patients who underwent distal pancreatectomy and 52 (43.0%) out of 121 patients who underwent a Whipple procedure developed a POPF (Grade A-C). POPFs were classified as A (24, 42.9%) and C (2, 3.6%) after distal pancreatectomy whereas they were graded as A (35, 28.9%), B (15, 12.4%) and C (2, 1.7%) after Whipple procedures. Drain data analysis was limited to Whipple procedures because only two patients developed a clinically significant leak after distal pancreatectomy. The daily total drain output did not differ between patients with a clinical leak (Grades B/C) and patients without a clinical leak (no leak and Grade A) on post-operative day (POD) 1 to 7. Although the median amylase concentration was significantly higher in patients with a clinical leak on POD 1-6, there was no day that amylase concentration predicted a clinical leak better than simply classifying all patients as 'no leak' (maximum accuracy = 86.1% on POD 1, expected accuracy by chance = 85.6%, kappa = 10.2%). CONCLUSION: Drain amylase data in the early post-operative period are not a sensitive or specific predictor of which patients will develop clinically significant POPF after pancreas resection.

Effect of BioGlue on the incidence of pancreatic fistula following pancreas resection.

BACKGROUND: Despite numerous modifications of surgical technique, pancreatic fistula remains a serious problem and occurs in about 10% of patients following pancreas resection. BioGlue is a new sealant that creates a flexible mechanical seal within minutes independent of the body's clotting mechanism. HYPOTHESIS: Application of BioGlue sealant will reduce the incidence of pancreatic fistula following pancreas resection. METHODS: A retrospective cohort study was performed with 64 patients undergoing pancreas resection. BioGlue sealant was applied to the pancreatic anastomosis (Whipple) or resection margin (distal pancreatectomy) in 32 cases. Factors that could affect the rate of postoperative pancreatic fistula were recorded. Pancreatic fistula was defined as greater than 50 ml of drain output with an amylase content greater than three times normal serum value after postoperative day 10. To improve the sensitivity of our study, we also examined pancreatic fistula with a strict definition of any drain output on or after postoperative day 3 with a high amylase content and graded the fistulas in terms of clinical severity. Grade A leaks were defined as subclinical. Grade B leaks required some response such as making the patient nil per os, parenteral nutrition, octreotide, antibiotics, or a prolonged hospital stay. Grade C leaks were defined as serious and life threatening. They were associated with hemorrhage, sepsis, resulted in deterioration of other organ systems, and mandated intensive care. Comparisons between the two groups were made using the chi-square test or Fisher's exact test for categorical variables and by the Wilcoxon rank-sum test for continuous variables. P values of 0.05 or less were deemed statistically significant. RESULTS: There were no differences between the patients who received BioGlue and the control cohort in terms of comorbid conditions, tumor location, texture of the pancreas, size of the pancreatic duct, or surgical technique. By the common definition, pancreatic fistula occurred in 6% (control) vs. 22% (BioGlue). By the strict definition, a fistula occurred in 41% (control) vs. 60% (BioGlue). In the control group, ten were subclinical (grade A) and two were clinically apparent leaks (grade B). In the BioGlue group, seven were subclinical (grade A), five were clinically apparent (grade B), and three were severe (grade C). There were no statistically significant differences in the incidence or severity grades of postoperative pancreatic fistulas between the two groups. CONCLUSION: Application of BioGlue sealant probably does not reduce the incidence of pancreatic fistula following pancreas resection.

Ampullary Cancers Harbor ELF3 Tumor Suppressor Gene Mutations and Exhibit Frequent WNT Dysregulation.

The ampulla of Vater is a complex cellular environment from which adenocarcinomas arise to form a group of histopathologically heterogenous tumors. To evaluate the molecular features of these tumors, 98 ampullary adenocarcinomas were evaluated and compared to 44 distal bile duct and 18 duodenal adenocarcinomas. Genomic analyses revealed mutations in the WNT signaling pathway among half of the patients and in all three adenocarcinomas irrespective of their origin and histological morphology. These tumors were characterized by a high frequency of inactivating mutations of ELF3, a high rate of microsatellite instability, and common focal deletions and amplifications, suggesting common attributes in the molecular pathogenesis are at play in these tumors. The high frequency of WNT pathway activating mutation, coupled with small-molecule inhibitors of ß-catenin in clinical trials, suggests future treatment decisions for these patients may be guided by genomic analysis.

Antiproliferative effects and mechanisms of liver X receptor ligands in pancreatic ductal adenocarcinoma cells.

Pancreatic ductal adenocarcinoma (PDAC) is difficult to detect early and is often resistant to standard chemotherapeutic options, contributing to extremely poor disease outcomes. Members of the nuclear receptor superfamily carry out essential biological functions such as hormone signaling and are successfully targeted in the treatment of endocrine-related malignancies. Liver X receptors (LXRs) are nuclear receptors that regulate cholesterol homeostasis, lipid metabolism, and inflammation, and LXR agonists have been developed to regulate LXR function in these processes. Intriguingly, these compounds also exhibit antiproliferative activity in diverse types of cancer cells. In this study, LXR agonist treatments disrupted proliferation, cell-cycle progression, and colony-formation of PDAC cells. At the molecular level, treatments downregulated expression of proteins involved in cell cycle progression and growth factor signaling. Microarray experiments further revealed changes in expression profiles of multiple gene networks involved in biological processes and pathways essential for cell growth and proliferation following LXR activation. These results establish the antiproliferative effects of LXR agonists and potential mechanisms of action in PDAC cells and provide evidence for their potential application in the prevention and treatment of PDAC.

An assessment of the necessity of transfusion during pancreatoduodenectomy.

INTRODUCTION: Perioperative transfusion of packed red blood cells (PRBC) has been associated with negative side effects. We hypothesized that a majority of transfusions in our series of patients who underwent pancreaticoduodenectomy (PD) were unnecessary. A retrospective analysis was performed to determine whether transfusions were indicated based on pre-determined criteria, and the impact of perioperative transfusions on postoperative outcomes was assessed. METHODS: Our prospectively maintained database was queried for patients who underwent PD between 2004 and 2011. 200 patients were divided into Cohort 1 (no transfusion) and Cohort 2 (transfusion). Rates of various graded 90-day postoperative complications were compared. Categorical values were compared according to the Common Terminology Criteria for Adverse Events. All cases involving intraoperative blood transfusion were reviewed for associated blood loss, intraoperative vital signs, urine output, hemoglobin values, and presence or absence of EKG changes to determine whether the transfusion was indicated based on these criteria. RESULTS: There were 164 patients (82%) in Cohort 1 (no transfusion) and 36 patients (18%) in Cohort 2 (transfused). Both groups had similar demographics. Patients in Cohort 2 had lesser median preoperative values of hemoglobin (12.3 vs 13.1, P = .002), a greater incidence of vein resection (33% vs. 16%, P = .021), longer operative times (518 vs 440 minutes, P < .0001), a greater estimated blood loss (850 vs. 300 mL, P < .001), and greater intraoperative fluid resuscitation (6,550 vs. 5,300 mL, P = .002). Ninety-day mortality was similar between the 2 groups (3% vs 1%, P = .328). Patients in Cohort 2 (transfused) had increased rates of delayed gastric emptying (36% vs. 20%, P = .031), wound infection (28% vs. 7%, P = .031), pulmonary complications (6% vs. 0%, P = .032), and urinary retention (6% vs. 0%, P = .032). A greater incidence of any complication of grade II severity (67% vs. 35%, P = .0005) or grade III severity (36% vs. 17%, P = .010) was also noted in Cohort 2. Of the 33 intraoperative transfusions, 15 (46%) did not meet any of the predetermined criteria: intraoperative hypotension (<90/60 mmHg), tachycardia (>110 beats per minute), low urine output (<10 mL/hour), decreased oxygen saturation (<95%), excessive blood loss (>1,000 mL), EKG changes, and low hemoglobin (<7.0 g/dL). CONCLUSION: Perioperative transfusions among patients with PD were associated with increased rates of various postoperative complications. A substantive portion (∼46%) of perioperative transfusions in this patient population did not meet predetermined criteria, indicating a potential opportunity for improved blood product use. Further prospective studies are required to determine whether the implementation of these criteria may a positive impact on perioperative outcomes.

A tumorigenic factor interactome connected through tumor suppressor microRNA-198 in human pancreatic cancer.

PURPOSE: The majority of pancreatic cancers overexpress mesothelin (MSLN), which contributes to enhanced proliferation, invasion, and migration. However, the MSLN regulatory network is still unclear. Here, we investigated the regulation of a panel of tumorigenic factors and explored the potential of MSLN-regulated miR-198 treatment in vivo. EXPERIMENTAL DESIGN: The expression and functional regulation of the tumorigenic factors MSLN, NF-κB, and the homeobox transcription factors (TF) POU2F2 (OCT-2), Pre-B-cell leukemia homeobox factor 1 (PBX-1), valosin-containing protein (VCP), and miR-198 were studied in pancreatic cancer cell lines, patient tumor samples, and xenograft pancreatic cancer mouse models. RESULTS: We found that miR-198 is downregulated in pancreatic cancer and is involved in an intricate reciprocal regulatory loop with MSLN, which represses miR-198 through NF-κB-mediated OCT-2 induction. Furthermore, miR-198 repression leads to overexpression of PBX-1 and VCP. The dysregulated PBX-1/VCP axis leads to increased tumorigenicity. Reconstitution of miR-198 in pancreatic cancer cells results in reduced tumor growth, metastasis, and increased survival through direct targeting MSLN, PBX-1, and VCP. Most interestingly, reduced levels of miR-198 in human tissue samples are associated with upregulation of these tumorigenic factors (MSLN, OCT-2, PBX-1, VCP) and predict poor survival. Reduced miR-198 expression links this tumor network signature and prognosticates poor patient outcome. High miR-198 disrupts the network and predicts better prognosis and increased survival. CONCLUSIONS: miR-198 acts as a central tumor suppressor and modulates the molecular makeup of a critical interactome in pancreatic cancer, indicating a potential prognostic marker signature and the therapeutic potential of attacking this tumorigenic network through a central vantage point.

A novel epigenetic CREB-miR-373 axis mediates ZIP4-induced pancreatic cancer growth.

Changes in the intracellular levels of the essential micronutrient zinc have been implicated in multiple diseases including pancreatic cancer; however, the molecular mechanism is poorly understood. Here, we report a novel mechanism where increased zinc mediated by the zinc importer ZIP4 transcriptionally induces miR-373 in pancreatic cancer to promote tumour growth. Reporter, expression and chromatin immunoprecipitation assays demonstrate that ZIP4 activates the zinc-dependent transcription factor CREB and requires this transcription factor to increase miR-373 expression through the regulation of its promoter. miR-373 induction is necessary for efficient ZIP4-dependent enhancement of cell proliferation, invasion, and tumour growth. Further analysis of miR-373 in vivo oncogenic function reveals that it is mediated through its negative regulation of TP53INP1, LATS2 and CD44. These results define a novel ZIP4-CREB-miR-373 signalling axis promoting pancreatic cancer growth, providing mechanistic insights explaining in part how a zinc transporter functions in cancer cells and may have broader implications as inappropriate regulation of intracellular zinc levels plays an important role in many other diseases.

Assessment of the learning curve for pancreaticoduodenectomy.

BACKGROUND: Experience with the Whipple procedure has been associated with improved outcomes, but the learning curve for this complex procedure is not well defined. METHODS: Outcomes with 162 consecutive Whipple procedures during the 1st 11.5 years of practice was documented in a prospective database. A period of low (≤11/y) and high (≥23/y) case volume was compared using the Wilcoxon rank-sum test and Fisher exact test. RESULTS: With low case volume, blood loss was higher (800 vs 400 mL, P = .001), more patients were transfused (44% vs 18%, P = .027), there were more complications (58% vs 46%, P = .0337), and a longer length of stay (10 vs 7 days, P = .006). There was only 1 mortality (.7%). CONCLUSIONS: Frequent repetition of the Whipple procedure is associated with an improvement in quantifiable quality benchmarks, and improvement continues with extensive experience. However, with proper training and the right environment, this procedure can be performed during the learning curve with acceptable outcomes.

Single nucleotide polymorphism in RECQL and survival in resectable pancreatic adenocarcinoma.

BACKGROUND: RECQL is a DNA helicase involved in DNA mismatch repair. The RECQL polymorphism, 3' untranslated region (UTR) A159C, was previously associated with overall survival of patients with resectable pancreatic adenocarcinoma treated with neoadjuvant chemoradiation. In the present study, we examined RECQL for somatic mutations and other polymorphisms and compared these findings with the outcome in patients who received adjuvant or neoadjuvant chemoradiation. We hypothesized that RECQL (i) would be mutated in cancer, (ii) would have polymorphisms linked to the 3'UTR A159C and that either or both events would affect function. We also hypothesized that (iii) these changes would be associated with survival in both cohorts of patients. MATERIAL AND METHODS: We sequenced RECQL's 15 exons and surrounding sequences in paired blood and tumour DNA of 39 patients. The 3'UTR A159C genotype was determined in blood DNA samples from 176 patients with resectable pancreatic adenocarcinoma treated with adjuvant (53) or neoadjuvant (123) chemoradiation. Survival was calculated using the Kaplan-Meier method, with log rank comparisons between groups. The relative impact of genotype on time to overall survival was performed using the Cox proportional hazards model. RESULTS: Somatic mutations were found in UTRs and intronic regions but not in exonic coding regions of the RECQL gene. Two single nucleotide polymorphisms (SNPs), located in introns 2 and 11, were found to be part of the same haplotype block as the RECQL A159C SNP and showed a similar association with overall survival. No short-term difference in survival between treatment strategies was found. We identified a subgroup of patients responsive to neoadjuvant therapy in which the 159 A allele conferred strikingly improved long-term survival. DISCUSSION: The RECQL 3'UTR A159C SNP is not linked with other functional SNPs within RECQL but may function as a site for regulatory molecules. The mechanism of action needs to be clarified further.

Preoperative prediction of complete resection in pancreatic cancer.

BACKGROUND: Accurate preoperative staging is essential in pancreatic cancer to select the 15% of patients who can benefit from surgery and avoid surgery in the 85% with advanced disease. With improvements in computed tomography (CT) scanning, the value of routine laparoscopy for preoperative staging of pancreatic cancer has been questioned because it changes the preoperative plan in less than 20% of unselected cases. METHODS: We retrospectively reviewed our experience with preoperative staging in 88 consecutive patients with pancreatic cancer. All patients had preoperative CT scans, and selective criteria were used to determine which patients would also undergo preoperative staging laparoscopy. Patients were categorized preoperatively as resectable or not resectable (locally advanced or metastatic). Medical records, operative, and pathology reports were reviewed to determine the accuracy of preoperative predictions. RESULTS: Thirty patients were deemed resectable based on CT alone and 27 (90%) were resected (25 R0, 2 R1). Two (7%) had metastatic disease discovered at laparotomy and one (3%) had a R2 resection. Only 19 patients (39%) of 49 patients deemed resectable by CT met our selective criteria for preoperative staging laparoscopy. Laparoscopy changed the treatment plan in 11 (58%) of these patients. Eight were still deemed resectable after staging laparoscopy and 7 (88%) were resected (6 R0, 1 R1). One patient (12%) had metastatic disease diagnosed at laparotomy. If selective staging laparoscopy were eliminated from our algorithm, 49 patients would have been deemed potentially resectable based on CT alone, 34 (69%) would have been found to be resectable at laparotomy (31 R0, 3 R1), and 15 (31%) would have been found to be unresectable at laparotomy (positive predictive value of 69%). The addition of selective staging laparoscopy avoided unnecessary laparotomy in 11 patients and increased the positive predictive value to (34/38) 89%. CONCLUSION: Selective use of laparoscopy increases the positive predictive value of preoperative staging in pancreatic cancer and avoids unnecessary laparoscopy in the majority of patients.

Accuracy of CT in predicting malignant potential of cystic pancreatic neoplasms.

BACKGROUND: Cystic lesions of the pancreas are being identified more frequently. Deciding which asymptomatic lesions can be safely followed with serial imaging and which require resection due to malignant potential is an increasingly common question. Current clinical practice is to rely on characteristics of the lesions on CT scan, and additional information from endoscopic ultrasound with fine-needle aspiration (EUS-FNA) and cyst fluid analysis or endoscopic retrograde pancreatography (ERCP) to assess malignant potential. HYPOTHESIS: The malignant potential of pancreatic cystic lesions cannot be accurately predicted by CT scan. METHODS: CT scans from 48 patients with cystic lesions of the pancreas were stripped of patient identifiers and retrospectively presented to two expert radiologists. The radiologists recorded specific characteristics of the lesions thought to be important in the differential diagnosis and their opinion of the likely diagnosis. Diagnostic accuracy was assessed by comparing the radiologists' diagnoses to the final pathologic diagnosis after resection. To determine if clinical history, EUS-FNA or ERCP findings improved diagnostic accuracy, medical records were retrospectively reviewed and scored as either supporting or not supporting malignant potential of the lesion. RESULTS: Specific diagnoses based on CT findings alone were correct in an average of 39% of the cases. Even when diagnoses were dichotomized as benign (43%) or potentially malignant (57%, papillary mucinous neoplasms, mucinous cystic neoplasms, cancer), determinations based on CT alone were accurate in an average of 61% of cases. Accuracy rates were 60.4 and 62.5% for the two radiologists, although there was only fair agreement between them (Kappa=0.28, 95% CI=(0.01-0.55), p=0.05). When all clinical information available was considered together as a single dichotomous indicator of malignant potential, the indicator was accurate in 90% of the cases (Kappa=0.73, 95% CI=(0.51-0.95, p<0.0001)). CONCLUSION: Specific preoperative diagnosis of pancreatic cystic neoplasms by CT alone is substantially inaccurate. Complementary tests such as EUS-FNA with fluid analysis and ERCP should be recommended to improve diagnosis especially if nonoperative treatment is planned.