Background and study aims: Current guidelines recommend using endoscopic ultrasound (EUS), carcinoembryonic antigen (CEA) testing and cytology to manage incidental pancreatic cystic neoplasms (PCN); however, studies suggest a strategy including integrated molecular pathology (IMP) of cyst fluid may further aid in predicting risk of malignancy. Here, we evaluate several strategies for diagnosing and managing asymptomatic PCN using healthcare economic modeling.Patients and methods: A third-party-payer perspective Markov decision model examined four management strategies in a hypothetical cohort of 1000 asymptomatic patients incidentally found to have a 3cm solitary pancreatic cystic lesion. Strategy I used cross-sectional imaging, recommended surgery only if symptoms or risk factors emerged. Strategy II considered patients for resection without initial EUS.Strategy III (EUS+CEA+Cytology) referred only those with mucinous cysts (CEA >192ng/mL) for resection. Strategy IV implemented IMP; a commercially available panel provided a Benign, Mucinous, or Aggressive classification based on the level of mutational change in cyst fluid. Benign and Mucinous patients were followed with surveillance; Aggressive patients were referred for resection. Quality-adjusted life-years (QALY), relative risk with 95%CI, Number Needed to Treat (NNT), and incremental cost-effectiveness ratios were calculated.Results: Strategy IV provided the greatest increase in QALY at nearly identical cost to the cheapest approach, Strategy I. Relative risk of malignancy compared to the current standard of care and nearest competing strategy, Strategy III, was 0.18 (95%CI 0.06-0.53) with an NNT of 56 (95%CI 34-120).Conclusions: Use of IMP was the most cost-effective strategy, supporting its routine clinical use.

Purpose Progression of Barrett's esophagus (BE) to esophageal adenocarcinoma (EAC) is associated with accumulated genomic instability. Current risk stratification of BE for EAC relies on histological classification and grade of dysplasia. However, histology alone cannot assess the risk of patients with inconsistent or non-dysplastic BE histology. We, therefore, examined the presence and extent of genomic instability in advanced and less advanced BE histology using mutational load (ML). Methods ML summarized the presence and clonality of loss of heterozygosity (LOH) mutations and the emergence of new alleles, manifested as microsatellite instability (MSI) mutations, in ten genomic loci around tumor suppressor genes associated with EAC. The ML of 877 microdissected targets from BE biopsies was correlated to their histology. Histological targets were categorized into three levels: no ML, low ML, and high ML. Results Increasing ML correlated with increasingly severe histology. By contrast, proportions of targets that lacked mutations decreased with increasingly severe histology. A portion of targets with non-dysplastic and low-grade histology shared a similar ML as those with higher risk and EAC disease. The addition of MSI characterization to ML helped to differentiate the ML between advanced and less advanced histology. Conclusions Given that EAC is associated with accumulated genomic instability, high ML in less severe histology may identify BE disease at greater risk of progression to EAC. ML may help to better manage BE in early histological stages and when histology alone provides insufficient information.

Mutations in guanine nucleotide binding protein, alpha stimulating complex locus (GNAS), have been found to have an oncogenic role in intraductal papillary mucinous neoplasms (IPMN).1Wu J. Matthaei H. Maitra A. et al.Recurrent GNAS mutations define an unexpected pathway for pancreatic cyst development.Sci Transl Med. 2011; 3: 92ra66Crossref PubMed Scopus (655) Google Scholar We evaluated whether the analysis of cyst fluid DNA obtained by EUS for GNAS mutations can enhance the preoperative diagnosis of neoplastic pancreatic cysts. We identified 25 patients with pancreatic cysts diagnosed on imaging who underwent EUS-FNA followed by surgical resection. Nine cysts were classified as IPMN according to surgical pathology. The mean cyst size was 2.2 cm. The mean level of cyst fluid carcinoembryonic antigen (CEA) in patients with branch duct IPMNs was 27,822 (44-247,470) ng/mL. The sensitivity, specificity, and diagnostic accuracy of cyst fluid CEA levels >192 ng/mL for the diagnosis of IPMN were 56%, 69%, and 64%, respectively. The presence of KRAS and GNAS mutations was determined in cyst fluid by quantitative polymerase chain reaction analysis. KRAS mutations were present in 67% of IPMNs, 67% of mucinous cystic neoplasms, and 60% of invasive adenocarcinomas. No KRAS mutations were seen in patients with other type of cysts. Four of 9 (44%) IPMN patients carried a mutation in GNAS, all of which harbored degrees of dysplasia. GNAS mutations were not found in patients with other types of cystic lesions or invasive adenocarcinomas (P = .02). The sensitivity, specificity, and diagnostic accuracy of cyst fluid GNAS mutations for the diagnosis of IPMN were 44%, 100%, and 80%, respectively. In aggregate, of the patients with IPMS, 11% harbored both GNAS and KRAS mutations, whereas at least 1 of the 2 genes was mutated in 100% (Table 1). The combination of KRAS and GNAS had a diagnostic accuracy of 80%. The addition of GNAS analysis detected a total of 3 additional IPMN cysts that would otherwise have been missed as a result of these 2 patients having CEA levels <192 ng/mL and no KRAS mutations in cyst fluid. The current study demonstrates that the presence of GNAS in combination with KRAS in pancreatic cystic fluid obtained by EUS-FNA improves the sensitivity for diagnosis of IPMN in comparison with cyst fluid CEA or KRAS alone. Larger prospective trials are needed to further validate the study's findings.TABLE 1Cyst fluid analysis comparing IPMN with non-IPMN pancreatic cystsIPMNOther cystsOdds ratio (95% CI)Statistical significance (P value)Performance characteristics to diagnose IPMNCytology diagnostic for lesion4/98/160.8 (0.2-4.1)1Sensitivity = 44%Specificity = 100%CEA >192ng/mL5/95/162.8 (0.5-14.9).4Sensitivity = 56%Specificity = 69%KRAS mutation6/95/164.4 (0.8-25.1).09Sensitivity = 67%Specificity = 69%GNAS mutation4/90/16Infinity.01⁎Statistically significant.Sensitivity = 44%Specificity = 100%Either KRAS or GNAS mutation9/90/16Infinity<.001⁎Statistically significant.Sensitivity = 100%Specificity = 69%CI, Confidence interval; IPMN, intraductal papillary mucinous neoplasm; CEA; carcinoembryonic antigen; GNAS, alpha stimulating complex locus. Statistically significant. Open table in a new tab CI, Confidence interval; IPMN, intraductal papillary mucinous neoplasm; CEA; carcinoembryonic antigen; GNAS, alpha stimulating complex locus. This study was partially funded by RedPath Integrated Pathology, Inc. The following authors disclosed financial relationships relevant to this publication: E. Ellsworth and S. D. Finkelstin: full-time employees of RedPath Integrated Pathology, Inc. All other authors disclosed no financial relationships relevant to this publication.

Diagnosis of fine-needle aspirations of pancreatic solid masses is complicated by many factors that keep its false-negative rate high. Our novel approach analyzes cell-free cytocentrifugation supernatant, currently a discarded portion of the specimen. Supernatant and cytology slides were collected from 25 patients: 11 cases with confirmed outcome [five positive (adenocarcinoma) and six negative (inflammatory states)], plus 14 without confirmed outcomes. Slides were microdissected, DNA was extracted from microdissections and corresponding supernatants, and all were analyzed for KRAS point mutation and loss of heterozygosity. Notably, higher levels of free DNA were found in supernatants than in corresponding microdissected cells. Supernatants contained sufficient DNA for mutational profiling even when samples contained few to no cells. Mutations were present in 5/5 malignancies and no mutations were present in inflammatory states. In conclusion, these findings support using supernatant for mutational genotyping when diagnostic confirmation is required for pancreatic solid masses.

Purpose: A multi-institutional review of patient medical records was performed in order to determine how the performance of second-line PathfinderTG testing (PFTG, RedPath Integrated Pathology) compared to that of common first-line testing risk features of pancreatic cystic lesions. Methods: Four hundred twenty-two patients previously tested with PFTG had outcomes that met study criteria. Outcomes were classified as benign or malignant based on surgical pathology, unequivocal malignant cytology, evidence of clinical cancer treatment, or at least 2 years of clinical and imaging follow up. PFTG testing was ordered according to standard of care. PFTG diagnoses were classified into four categories: benign, statistically indolent (SI), statistically higher risk (SHR), or aggressive based on predetermined criteria involving clinical, radiological, and molecular characteristics. Results: Cyst size was on average significantly larger in malignant lesions (3.6 cm) than in benign lesions (2.4cm) (p<0.001). CEA was on average significantly higher in malignant lesions (748.1ng/m) than in benign lesions (144.9 ng/mL) (p<0.001). The presence of a mural nodule or associated mass indicated an increase in likelihood of malignancy by 8.3 times (p<0.001). However, no single criterion of the above could reliably predict malignancy in any cyst with high specificity or sensitivity (Figure 1; y-axis). In large lesions (≥3 cm) most outcomes were benign (78/104, 75%), and when CEA levels were elevated (≥192ng/mL) most outcomes were benign (120/146, 82%). In patients with evidence of mural nodule or associated mass, more than half (39/65, 60%) of outcomes were benign. The addition PFTG diagnostic criteria helped to identify benign lesions in most cases with large (≥3 cm) cysts (65/78, 83%), in most cases in which CEA was elevated (≥192 ng/mL) (100/120, 80%), and in most cases with mural nodules or associated masses (29/39, 74%) (Figure 1, Benign and Statistically Indolent categories).Figure: First-line testing cyst characteristics versus PFTG and patient outcome. A) Cyst size (cm) versus PFTG. B) Cyst fluid CEA level (ng/mL) versus PFTG. C) Presence of a mural nodule or associated mass (yes/no) versus PFTG.Conclusion: When first-line CEA, cyst size, or solid component features were worrisome, second-line PFTG testing helped to identify patients with benign disease and distinguish them from those with malignancy. PFTG may help guide the management of cysts with elevated CEA or high risk morphological features. Disclosure - Brendan Corcoran: Employee:RedPath Integrated Pathology; Eric Ellsworth: Employee: RedPath Integrated Pathology; Sara Jackson: Employee:RedPath Integrated Pathology; Sydney Finkelstein, MD: Founder, Board Member:RedPath Integrated Pathology. This research was supported by an industry grant from RedPath Integrated Pathology.

Fine-needle aspiration (FNA) of pancreatic solid masses can be significantly impacted by sampling variation. Molecular analysis of tumor DNA can be an aid for more definitive diagnosis. The aim of this study was to evaluate how molecular analysis of the cell-free cytocentrifugation supernatant DNA can help reduce sampling variability and increase diagnostic yield. Twenty-three FNA smears from pancreatic solid masses were performed. Remaining aspirates were rinsed for preparation of cytocentrifuged slides or cell blocks. DNA was extracted from supernatant fluid and assessed for DNA quantity spectrophotometrically and for amplifiability by quantitative PCR (qPCR). Supernatants with adequate DNA were analyzed for mutations using PCR/capillary electrophoresis for a broad panel of markers (KRAS point mutation by sequencing, microsatellite fragment analysis for loss of heterozygosity (LOH) of 16 markers at 1p, 3p, 5q, 9p, 10q, 17p, 17q, 21q, and 22q). In selected cases, microdissection of stained cytology smears and/or cytocentrifugation cellular slides were analyzed and compared. In all, 5/23 samples cytologically confirmed as adenocarcinoma showed detectable mutations both in the microdissected slide-based cytology cells and in the cytocentrifugation supernatant. While most mutations detected were present in both microdissected slides and supernatant fluid specimens, the latter showed additional mutations supporting greater sensitivity for detecting relevant DNA damage. Clonality for individual marker mutations was higher in the supernatant fluid than in microdissected cells. Cytocentrifugation supernatant fluid contains levels of amplifiable DNA suitable for mutation detection and characterization. The finding of additional detectable mutations at higher clonality indicates that supernatant fluid may be enriched with tumor DNA. Molecular analysis of the supernatant fluid could serve as an adjunct method to reduce sampling variability and increase diagnostic yield, especially in cases with a high clinical suspicion for malignancy and limited number of atypical cells in the smears.

for HMGA2.Results: A total of 32 patients were identified.IPMN types were as noted: branch duct (n = 20), main duct (n = 4), mixed type (n = 8).LGD was present in 5 patients, MD in 17 patients, HGD in 9 patients, and invasive cancer in 1 patient.HMGA2 protein was detected in the cyst fluid of 30/32 (94%) specimens.Mean HMGA2 concentration (ng/ ml): LGD 0.52 ± 0.40, MD 2.56 ± 2.52, HGD 10.5 ±15.22 (p , 0.05).No HMGA2 was detected in the one sample from IPMN with invasive cancer.The mean HMGA2 concentration was significantly higher in the HGD group (10.5 ± 15.22 ng/ml) compared to the concentration in the low-risk IPMN group (2.1 ± 2.38 ng/ml, p = 0.03).The ROC for HMGA2 had an area under the curve of 0.74.Conclusions: HMGA2 protein is present in the cyst fluid of IPMN.Significantly higher concentrations of cyst fluid HMGA2 proteins are found in IPMN with HGD as compared to lesions with LGD or MD.Cyst fluid concentration of HMGA2 may thus serve as a biomarker to differentiate patients with high-risk IPMN lesions from patients with low-risk IPMN lesions.Such a biomarker could help guide clinical decision making regarding which patients may benefit from surgical resection of their IPMN.Further investigation of this cyst fluid biomarker on prospectively obtained samples from EUS-FNA is warranted.

INTRODUCTION: Chronic Atrophic Gastritis (CAG) is the most important independent risk condition for gastric cancer, histology being considered the gold standard for both diagnosis and follow up.The Operative Link for Gastritis Assessment (OLGA) is an innovative histological system proposed for staging atrophic gastritis in a prognostic view.GastroPanel (Biohit, Helsinki, Finland) is a serological kit for a non-invasive diagnosis of chronic atrophic gastritis (CAG), including serum pepsinogen I (PGI), pepsinogen II (PGII), gastrin-17 (G-17) and IgG anti-Helicobacter pylori antibodies (IgG-Hp).Clinical practice is lacking of a non invasive instrument useful to predict and to follow the stage of atrophic gastritis.AIMS& METHODS: To evaluate correlation between GastroPanel and OLGA stage, in a cohort of dyspeptic patients referred from a primary care setting to our G.I. Unit Histological evaluation of gastric biopsies was performed in 125 pts (83 females, mean age 53 yrs).Atrophy was assessed according to the OLGA staging system.OLGA stages III and IV are considered at high risk for gastric cancer development.From each subject, a blood sample was taken for GastroPanel analysis.RESULTS: GastroPanel analysis showed that 79 pts (63,2%) have no gastritis, 22 pts (17,6%) a Hp-related chronic gastritis and 24 (19,2%) atrophic gastritis picture.Comparing GastroPanel results to the histological findings, 48 pts of the first group were confirmed normal at histology, 27 pts have an OLGA stage 0-II, 4 pts have an OLGA stage III, none stage IV.Among the patients with Hp-related gastritis, 1 pt results normal at the histological evaluation, 20 pts have an OLGA stage 0-II, 1 pt has an OLGA stage III; 6 out of the 24 pts with evidence of CAG at GastroPanel have an OLGA stage 0-II, 18 pts have an OLGA stage III-IV.The negative predictive value (NPV) of GastroPanel for CAG is 95,0%, being the positive predictive value (PPV) 75,0%.CONCLUSION: GastroPanel shows an high NPV to predict absence of atrophic gastritis.The relationship beetween GastroPanel and OLGA system seems to confirm the clinical performances of the test Sa1977

Purpose: A multi-institutional review of patient medical records was performed in order to determine the clinical validity (performance characteristics and risk stratification) of PathfinderTG (PFTG, Red-Path Integrated Pathology) when ordered clinically as a second line test. Methods: Four hundred twenty-two patients previously tested with PFTG had outcomes that met study criteria. Patient outcomes were classified into two groups (benign or malignant) based on surgical pathology (n=195), inequivocally diagnostic cytology (n=20), evidence of clinical cancer treatment (n=5) or at least 2 years of clinical and imaging follow up (n=202). All PFTG analyses were ordered according to practitioner standard of care. PFTG diagnoses were classified into four distinct categories: benign, statistically indolent (SI), statistically high risk (SHR), and aggressive based on predetermined molecular, radiographic and clinical criteria. For the purpose of calculating performance characteristics, diagnoses of benign and SI were treated as negative for malignancy (benign), while SHR and aggressive were treated as positive for malignancy. Results: PFTG had a sensitivity and specificity for malignancy of 81% [95% confidence interval [CI] 70-89%] and 91% [95% CI 88-94%], respectively, with an overall accuracy 90% [95% CI 86-92%]. Testing also showed a negative predictive value (NPV) of 96% [95% CI 93-98] and a positive predictive value of 64% [95% CI 53-74%]. Relative to patients with benign diagnoses, patients with SHR and Aggressive diagnoses were 25.95 (p<0.0001) and 51.75 (p<0.0001), respectively more likely to get malignancy. Conclusion: When first line testing results are not definitive, second line PFTG testing can provide clinically valid diagnostic and predictive information that may help to better manage patients. PFTG is able to predict benign and statistically indolent disease in patients with high specificity. Furthermore, PFTG can identify patients with and at higher risk of malignancy (SHR and Aggressive). Disclosure - Brendan Corcoran: Employee:RedPath Integrated Pathology; Eric Ellsworth: Employee: RedPath Integrated Pathology; Sara Jackson: Employee:RedPath Integrated Pathology; Sydney Finkelstein, MD: Founder, Board Member:RedPath Integrated Pathology. This research was supported by an industry grant from RedPath Integrated Pathology.Figure: Kaplan-Meier curve representing the four diagnostic categories of PathFinder TG.

The role of environmental and occupational toxin exposure as a cause of or contributing factor for cancer development and progression is incompletely understood. A unique signature of specific mutational change to discriminate toxin-exposed from sporadic cancer is generally sought but not often encountered. We report an approach to better understand cancer causality based on the measurement of the cumulative DNA damage (via loss of heterozygosity) over a defined genomic region (chromosome 3) that is applicable to archival, fixative-treated tissue and cytology specimens of cancer. Our method was applied to (1) a cohort of 10 brain tumor subjects (9 gliomas, 1 hemangioblastoma) with potential exposure to chlorinated solvents and (2) a control cohort of sporadic brain cancer controls (7 gliomas, 1 hemangioblastoma). We show that brain tumors arising in potentially toxin-exposed subjects bear a significantly higher level of passenger LOH mutations compared to sporadic cancer controls. The methodology utilized tissue microdissection, PCR amplification and capillary electrophoresis (fragment analysis for LOH determination, DNA sequencing for specific point mutations), and examined a panel of 15 microsatellite markers distributed along both arms of chromosome 3 that aimed at capturing passenger mutational change accrued during stages of clonal expansion of neoplastic cells. This proof-of-principle study using mutational profiling for passenger LOH mutational damage provides support for the utility of this approach and further studies in order to differentiate between genotoxin-associated versus sporadic (unexposed) cancer development.

Background: Recent advances in the management of Barrett's Esophagus (BE) have placed greater emphasis on accurate diagnosis of BE as well as better prediction of risk for progression to esophageal adenocarcinoma (EAC). Histological evaluation of BE is particularly challenging with significant inter-observer variability. We explored the presence and extent of genomic instability in BE biopsy specimens as a means to add supplementary information to the histological classification and clinical decision-making related to early disease.Methods: We reviewed histology slides from 271 patients known to have BE. Using histological features as a guide, we microdissected target cell populations with various histological classifications of BE (intestinal metaplasia, "indefinite for dysplasia", low grade dysplasia, or high grade dysplasia). DNA was extracted from microdissected targets and analyzed for loss of heterozygosity (LOH) using a panel of 16 LOH mutational markers associated with tumor suppressor genes at chromosomal loci 1p, 3p, 5q, 9p, 10q, 17p, 17q, 18q, 21q, 22q. The presence or absence of mutations and the clonality of each mutation were determined for each marker.Results: The presence and clonal expansion of LOH mutations was formulated into mutational load (ML) for each microdissected target analyzed. ML correlated with the histological classification of microdissected targets, with increasingly severe histology having higher ML. Three levels of mutation load (no ML, low ML, and high ML) were defined based on the population of microdissected targets histologically classified as intestinal metaplasia. All microdissected targets with dysplasia had mutations, with a high ML consistently present in high grade dysplasia targets. Microdissected targets histologically classified as intestinal metaplasia or "indefinite for dysplasia" spanned a range of no, low, and high ML.Conclusions: The results of this study reinforce the association of genomic instability with disease progression in BE. The presence and extent (clonality) of genomic instability, as assessed by mutational load, may assist histology in defining early stages of BE that are potentially at greater risk for disease progression. Assessment of mutational load using our panel of LOH mutational markers may be a useful adjunct to microscopic inspection of biopsy specimens, and thereby, improve patient management.

Purpose: Long-term surveillance of NDBE has high cost and limited effectiveness in preventing esophageal adenocarcinoma, and EAT for NDBE remains expensive and controversial. Recent studies support a role for molecular markers in defining a subset of patients at greater risk for progression to cancer. Aim: To evaluate cost-effectiveness of EAT of NDBE using risk-stratification with molecular markers. Methods: Using a Markov hybrid decision tree, 4 competing strategies were evaluated in a hypothetical 50-year old cohort of patients with NDBE over their lifetime with societal perspective. In strategy I, all patients with NDBE were followed without specific intervention; in strategy II standard surveillance according to the ACG practice guidelines was performed, with EAT used for patients with high grade dysplasia. In strategy III, EAT was performed in all patients with NDBE. In strategy IV, the PathFinder assay (RedPath Integrated Pathology, Pittsburgh, PA), which assesses mutational load (ML, a semi-quantative measure of genomic instability using loss of heterozygosity mutations in DNA extracted from areas targetted by microdissection from biopsy slides) was used for risk-stratification (Gastroenterol 2012;142 (5):S 749). Patients with no ML (25%) underwent minimal surveillance, patients with low ML (70%) underwent standard surveillance and patients with high ML (5%) were treated selectively by EAT. The model was biased against EAT with conservative estimates of complete response to EAT and continued surveillance afterwards. Transitional probabilities, discounted cost and utility values to estimate quality adjusted life-years (QALY) were obtained from published information and Medicare reimbursement data. Incremental Cost-Effectiveness Ratio (ICER) of different strategies was the main outcome measure. Results: In baseline analysis, strategy IV was clearly dominant (both cheaper and more effective: Table), and sensitivity analyses showed the conclusion to be robust to plausible ranges of parameters. In a second-order Monte Carlo analysis, under strategy IV and III, 45 and 114 esophageal cancers developed over 18,504 and 17,825 person-years of follow-up, respectively [RR 0.39 (95% CI, 0.28 to 0.55); NNT of 14 (95% CI, 11-22)]. Critical determinants of cost-effectiveness of the risk-based strategy were rate of complete response and cost of ablation and, surveillance interval in patients with no ML.Table: [1448] TableConclusion: Use of ML to stratify risk in NDBE patients was the most cost-effective strategy to prevent esophageal adenocarcinoma. Targeting EAT towards patients with high ML presents an opportunity for a paradigm shift in the management of NDBE. Disclosure: Dr. Das: consultant; Mr. Ellsworth, Dr. Smith, Dr. Finklestein: Employee. This research was supported by an industry grant from Research support by RedPath Integrated Pathology, Inc, Pittsburgh, PA.

Cytology diagnosis of pancreatic solid mass fine needle aspiration (FNA) and pancreatic/biliary brushing is limited by sampling variation and low cellularity which together reduce sensitivity for detection of cancer. Two new molecular approaches were used to improve cancer detection sensitivity: 1) extraction of DNA from the cytocentrifugation supernatant fluid which contains free DNA from the extracellular space in the region around the site of sampling and 2) microdissection of stained cytology cells which may lack sufficient microscopic features of cancer. We report our experience with 13 pts negative by first line cytology evaluation but suspected to have aggressive disease.Design: Residual cytology materials (cytocentrifugation supernatant fluid and/or stained cytology slides) were collected from 13 pts (8 pancreatic solid mass, 5 duct brushing) suspected to have pancreaticobiliary cancer but with indeterminate diagnosis or suspected false negative cytology diagnosis (8-insufficient cellularity, 5-benign morphology). DNA was extracted (qiagen) from 1-2 mls of cytocentrifugation supernatant fluid and/or from a single stained cytology slide. DNA mutational analysis used a broad panel of markers (KRAS point mutation by sequencing, loss of heterogeneity at 1p, 3p, 5q, 9p, 10q, 17p, 17q, 21q, 22q). Detectable mutational change was the criteria for molecular detection of cancer. Surgical pathology outcome was available in 5 pts.