Low-risk Cysts Research Articles

IPMN-LEARN: A linear support vector machine learning model for predicting low-grade intraductal papillary mucinous neoplasms.

We aimed to build a machine learning tool to help predict low-grade intraductal papillary mucinous neoplasms (IPMNs) in order to avoid unnecessary surgical resection. IPMNs are precursors to pancreatic cancer. Surgical resection remains the only recognized treatment for IPMNs yet carries some risks of morbidity and potential mortality. Existing clinical guidelines are imperfect in distinguishing low-risk cysts from high-risk cysts that warrant resection. We built a linear support vector machine (SVM) learning model using a prospectively maintained surgical database of patients with resected IPMNs. Input variables included 18 demographic, clinical, and imaging characteristics. The outcome variable was the presence of low-grade or high-grade IPMN based on post-operative pathology results. Data were divided into a training/validation set and a testing set at a ratio of 4:1. Receiver operating characteristics analysis was used to assess classification performance. A total of 575 patients with resected IPMNs were identified. Of them, 53.4% had low-grade disease on final pathology. After classifier training and testing, a linear SVM-based model (IPMN-LEARN) was applied on the validation set. It achieved an accuracy of 77.4%, with a positive predictive value of 83%, a specificity of 72%, and a sensitivity of 83% in predicting low-grade disease in patients with IPMN. The model predicted low-grade lesions with an area under the curve of 0.82. A linear SVM learning model can identify low-grade IPMNs with good sensitivity and specificity. It may be used as a complement to existing guidelines to identify patients who could avoid unnecessary surgical resection.

Updates in diagnosis and management of pancreatic cysts.

Incidental pancreatic cysts are commonly encountered with some cysts having malignant potential. The most common pancreatic cystic neoplasms include serous cystadenoma, mucinous cystic neoplasm and intraductal papillary mucinous neoplasm. Risk stratifying pancreatic cysts is important in deciding whether patients may benefit from endoscopic ultrasound (EUS) or surgical resection. Surgery should be reserved for patients with malignant cysts or cysts at high risk for developing malignancy as suggested by various risk features including solid mass, nodule and dilated main pancreatic duct. EUS may supplement magnetic resonance imaging findings for cysts that remain indeterminate or have concerning features on imaging. Various cyst fluid markers including carcinoembryonic antigen, glucose, amylase, cytology, and DNA markers help distinguish mucinous from nonmucinous cysts. This review will guide the practicing gastroenterologist in how to evaluate incidental pancreatic cysts and when to consider referral for EUS or surgery. For presumed low risk cysts, surveillance strategies will be discussed. Managing pancreatic cysts requires an individualized approach that is directed by the various guidelines.

The faith of non-surveilled pancreatic cysts: a bicentric retrospective study

BackgroundIncidental discovery of pancreatic cystic neoplasms (PCLs) is a common and steadily increasing occurrence. The aim of this study was to investigate a cohort of patients presenting with incidentally detected PCLs which were not included in a surveillance protocol, and to compare their risk of malignant evolution with that of systematically surveilled lesions. Materials and methodsA population of PCLs which did not receive surveillance over a period >10 years (population A) was selected at the Medical University of Vienna. A group of “low risk” branch duct intraductal papillary mucinous neoplasm ≤15 mm in size upon diagnosis undergoing a regular follow-up of at least 5 years at the University of Verona was selected as control (population B). The incidence of pancreatic cancer (PC), cumulative risk of PC and disease-specific survival were compared. ResultsOverall, 376 patients with non-surveilled PCLs were included in study group A and compared to 299 patients in group B. This comparison resulted in similar incidence rates of PC (1.6% vs 1.7%, p = 0.938), a strong similarity in terms of disease-specific mortality rates (1.3% vs 0.3%, p = 0.171) and the 5- and 10-year cumulative risk of PC (≅ 1% and 2%, p = 0.589) and DSS (≅ 100% and 98%, p = 0.050). ConclusionThe “price to pay” for a negligence-based policy in the population of non-surveilled PCLs was reasonable, and the incidence of PC was comparable to that reported for a population of low-risk cysts enrolled to a standardized surveillance protocol.

Progression vs Cyst Stability of Branch-Duct Intraductal Papillary Mucinous Neoplasms After Observation and Surgery

The progression of intraductal papillary mucinous neoplasms (IPMNs) of the pancreas to malignant disease is still poorly understood. Observational and surgical series have failed to provide comprehensive information. To identify dynamic variables associated with the development of malignant neoplasms by combining pathological features with data from preoperative repeated observations. The Crossover Observational Multicentric Study included a retrospective cohort of patients with branch-duct IPMNs (BD IPMNs) enrolled in a surveillance program from January 1, 2000, to December 31, 2019. Patients were enrolled from 5 referral centers: the Pancreas Institute, Verona, Italy; Seoul National University Hospital, Seoul, South Korea; Singapore General Hospital, Singapore; Johns Hopkins School of Medicine, Baltimore, Maryland; and University of Texas MD Anderson Cancer Center, Houston. Patients underwent a minimum of 12 months of preoperative surveillance (median, 37 [interquartile range (IQR), 20-68] months). Dynamic variables associated with malignant disease were explored to estimate the presence of high-grade dysplasia (HGD) and invasive cancer at final pathological examination. A total of 292 patients were included in the analysis (137 women [46.9%] and 155 men [53.1%]; median age, 64 [IQR, 56-71] years). During surveillance, 27 patients (9.2%) developed a worrisome feature after 5 years, and 46 of 276 (16.7%) developed high-risk stigmata (HRS). At final pathological evaluation, 107 patients (36.6%) had HGD or invasive cancer, and 16 (5.5%) had IPMNs with concomitant pancreatic ductal adenocarcinoma. Rates of HGD and invasive cancer at pathological evaluation significantly differed between those without worrisome features and those developing HRS from a previous worrisome feature (9 [27.3%] vs 13 [61.9%]; P < .001). Developing an additional worrisome feature during surveillance (odds ratio [OR], 3.24 [95% CI, 1.38-7.60]; P = .007) or an HRS from a baseline worrisome feature (OR, 2.87 [95% CI, 1.01-8.17]; P = .048) was associated with HGD at final pathological evaluation. Among HRS, development of jaundice on a low-risk cyst was independently associated with invasive cancer (OR, 16.04 [95% CI, 2.94-87.40]; P = .001). These findings suggest that in BD IPMNs under surveillance, harboring a stable worrisome feature carries the lowest risk of malignant disease. Development of additional worrisome features or HRS is associated with the presence of HGD, whereas the occurrence of jaundice is associated with invasive cancer.

A novel microfluidic-based assay to determine high-risk pancreatic cysts for surgery utilizing pancreatic cyst fluid: an international multicenter study

Presenter: Ajay Maker MD | University of California San Francisco Background: Pancreatic cysts are increasingly diagnosed in the general population. We previously reported that the majority of cysts resected in the country are low risk lesions on final pathology, despite multiple established clinical guidelines. Thus, selecting which patients harbor a cyst at high-risk of occult malignancy or malignant degeneration for operative resection remains a clinical dilemma. The aim of the current study is to develop a high-throughout molecular diagnostic platform to identify pancreatic cysts at high risk of malignancy or high-grade dysplasia. Methods: Pancreatic cyst fluid from an international multi-center patient cohort were analyzed on a newly constructed microfluidic TaqMan low density array that was customized with our gene target panel of five miRNAs and seven mRNAs. Signatures were validated utilizing a machine learning random forest model for the optimal accuracy to determine high-risk cysts (high-grade dysplasia or invasive cancer) from low risk cysts for operative intervention. Results: Of 358 pancreatic cyst fluids analyzed, 157 cysts were intraductal papillary mucinous neoplasms (IPMNs), 62 were mucinous cystic neoplasms (MCNs), and the remainder were of various non-mucinous histologies. Of the 219 mucinous cysts, 60 (27%) were found to harbor high-grade dysplasia (HGD) or invasive cancer on final pathological examination. The models were established and then internally validated revealing a maximal accuracy of 87% to determine benign from potentially malignant cysts with an out-of-bag error of 8-12%. Further, the assay performed with an accuracy of 81% to identify high-risk from low-risk mucinous lesions (IPMN+MCN) for operative intervention. Conclusion: A high-throughput, cost-effective microfluidics-based molecular diagnostic assay has been developed that identifies high-risk cystic lesions of the pancreas for surgical evaluation. In this large, international, multi-center study, robust validation sets were performed. Performance of the biosignatures are currently being enhanced with clinical and radiographic features.

Cyst Fluid Carcinoembryonic Antigen Level Difference between Mucinous Cystic Neoplasms and Intraductal Papillary Mucinous Neoplasms.

Background/AimsThe role of cyst fluid carcinoembryonic antigen (CEA) level in differentiating mucinous pancreatic cystic lesions (PCLs) is controversial. We investigated the role of cyst fluid CEA in differentiating low-risk (LR)-intraductal papillary mucinous neoplasms (IPMNs) from high-risk (HR)-IPMNs and LR-mucinous cystic neoplasms (MCNs). MethodsThis was a retrospective study of 466 patients with PCLs who underwent endoscopic ultrasound-guided fine-needleaspiration over a 7-year period. On histology, low-grade dysplasia and intermediate-grade dysplasia were considered LR, whereas high-grade dysplasia and invasive carcinoma were considered HR. ResultsData on cyst fluid CEA levels were available for 50/102 mucinous PCLs with definitive diagnoses. The median CEA (range) levels were significantly higher in HR cysts than in LR cysts (2,624 [0.5–266,510] ng/mL vs. 100 [16.8–53,445]ng/mL, p=0.0012). The area under the receiver operating characteristic curve (AUROC) was 0.930 (95% confidence interval [CI], 0.5–0.8; p<0.001) for differentiating LR-IPMNs from LR-MCNs. The AUROC was 0.921 (95% CI, 0.823–1.000; p<0.001) for differentiating LR-IPMNs from HR-IPMNs. Both had a CEA cutoff level of >100ng/mL, with a negative predictive value (NPV) of 100%. ConclusionsCyst fluid CEA levels significantly vary between LR-IPMNs, LR-MCNs, and HR-IPMNs. A CEA cutoff level of >100ng/mL had a 100% NPV in differentiating LR-IPMNs from LR-MCNs and HR-IPMNs.

Comorbidities Drive the Majority of Overall Mortality in Low-Risk Mucinous Pancreatic Cysts Under Surveillance

Background & AimsThe Charlson Comorbidity Index (CACI) has been suggested as a tool to determine comorbidity burden and guide management for patients with mucinous pancreatic cysts (Intrapapillary Mucinous Neoplasms and Mucinous Cystic Neoplasms), but has not been studied well among “low-risk” mucinous pancreatic cysts i.e. without worrisome features (WF) and high-risk stigmata (HRS). This study sought to determine the comorbidity burden among surveillance population of low-risk pancreatic cysts and provide their follow-up mortality outcomes. MethodsA single center study retrospectively reviewed a prospective pancreatic cyst database and included individuals with low-risk cysts undergoing serial imaging during 2016. Electronic medical records were reviewed to determine their baseline age-adjusted CACI (age-CACI). After 4 years, their progression to WF, disease specific (pancreatic malignancy-related, DSM), extra-pancreatic (EPM), and overall mortalities (OM) were determined using Kaplan-Meir Survival Analysis. Results502 individuals underwent prospective surveillance. The study included 440 individuals with low-risk suspected or presumed mucinous cysts and excluded 50 and 12 individuals with WF and HRS respectively. Over a median follow-up of 56 months, 12 WF progressions, 2 DSMs, 42 EPMs, and 44 OMs were observed. Baseline age-CACI had good predictive capacity for 4-year EPM (Area-Under Curve: 0.87; p< .0001). The median age-CACI of 4 enabled cohort stratification into Low (age-CACI <4) and High CACI (age-CACI ≥4) groups. A significantly higher OM (p< .001) was observed among the High CACI group as compared to the Low CACI group. ConclusionThrough real-time application of CACI to patient outcomes, our analysis supports incorporation of this comorbidity assessment tool in making shared surveillance decisions among low-risk pancreatic cyst population.

S0116 Balancing Risks of Pancreatic Malignancy-Related and Non-Pancreatic-Related Mortality in Surveillance of Low-Risk Presumed Mucinous Pancreatic Cysts

INTRODUCTION: Charlson Comorbidity Index (CACI) has been suggested as a tool to determine comorbidity burden and guide surveillance for patients with presumed mucinous pancreatic cysts (Intrapapillary Mucinous Neoplasms and Mucinous Cystic Neoplasms). This study seeks to define the comorbidity burden among patients with low-risk pancreatic cysts i.e., cysts without worrisome features (WF) and high risk stigmata (HRS) undergoing surveillance, and provide real-world follow-up (including mortality) on this cohort. METHODS: A single center prospective study including individuals with presumed mucinous pancreatic cysts undergoing active surveillance during 2016 was performed. Electronic medical records were reviewed to identify low-risk cysts without WF and HRS, and the CACI for eachindividual was calculated. After 3 years, we performed an interim analysis among these low-risk cysts to determine the disease specific (pancreatic malignancy-related, DSM), extra-pancreatic (EPM), and overall mortality (OM) using Kaplan-Meir model. Ratio of DSM/EPM was calculated for the entire low risk cyst cohort, for different CACI thresholds. RESULTS: The surveillance population of 502 patients included 440 individuals with low-risk cysts, 50 with WF, and 12 with HRS. Among low risk cysts, only one individual (0.23%) had DSM due to metachronous pancreatic malignancy while 36 (8.18%) individuals had EPM over 3 years. Baseline CACI predicted actual mortality at 3 years. Both the CACI cut-offs of 7 and 4 demonstrated significantly higher EPM among groups with “High CACI” vs “Low CACI” groups (log-tank test: p < 0.0001) (Figures 1A and 1B). Upon exclusion of individuals with CACI ≥7 (N = 378) and CACI ≥4 (N = 221), DSM was observed in 1 patient each (0.26% & 0.45%) but the EPM was 14 (3.17%) and 1 (0.45%) respectively. The ratio of DSM/EPM for CACI <4, CACI <7, and the entire cohort was 100%, 7.14%, and 2.78% respectively (Figure 2). CONCLUSION: Our interim analysis demonstrates no drastic alteration in DSM but reduction in EPM as we excluded individuals from surveillance cohorts based on their CACI. Though the EPM is inevitable, due to lack of survival benefit, surveillance cessation at CACI ≥7 and 4 would prevent unnecessary procedures, anxiety and health resource allocation among 62 and 219 individuals respectively. Our study supports incorporation of CACI in shared decisions for surveillance cessation among high CACI individuals with low risk cysts.Figure 1.: Kaplan Meier Survival curves for EPM among “High CACI” vs “Low CACI” low-risk cysts with 1A: CACI cut-off of 7 (χ2 = 91.4, p < 0.0001) and 1B: CACI cut-off of 4 (χ2 = 34.54, p < 0.0001).Figure 2.: Distribution of EPM, DSM and the ratios among entire cohort (N = 440), Cohort with CACI <7 (N = 378), and those with CACI <4 (N = 221).

S0001 Using Large Sample Real World Data to Study the Progression of Low Risk Pancreatic Cysts: New Onset Diabetes as a Potential Biomarker of Malignant Transformation

INTRODUCTION: Although the risk of the transformation of pancreatic cystic neoplasms to pancreatic cancer (PC) has been extensively investigated, prior study has been significantly limited by small sample sizes and selection bias, yielding highly variable progression results. Furthermore, due to these study limitations, it has been difficult to reliably identify novel risk factors and biomarkers for malignant transformation, such as patient metabolic status. METHODS: Patients were sourced from the IBM Marketscan insurance claims databases, covering over 180 million patients from 2008 to 2017. ICD-9/10 codes were used to identify patients with low-risk pancreatic cysts. Patients with a history of worrisome features or prior pancreatic pathology were excluded. Patients were evaluated for at least two years following cyst diagnosis. Patients who developed PC or had surgery within 6 months of diagnosis were excluded. Kaplan-Meier analysis and Cox proportional hazard modeling were used to assess the risk of progression to PC and the association of progression with metabolic status. Results were standardized by age and insurance coverage. RESULTS: From the 137,970 patients with a diagnosis of pancreatic cyst, 14,279 low risk patients fulfilled study criteria. At 5 years, the adjusted cumulative progression rate to PC was 2.2% (95% CI: 1.8%–2.6%) (Figure 1Figure 1.: Cumulative Risk of Progression of Low Risk Pancreatic Cysts to Pancreatic Cancer (N = 14,249).). The cumulative risk of progression increased at a nearly linear rate over 7 years (R = 0.979; P < 0.001). A prior history of diabetes (HR = 2.01; 95% CI: 1.89–2.14) and the development of new onset diabetes (NODM) following cyst diagnosis (HR = 3.24; 95% CI: 3.03–3.46) were both associated with a greater risk of progression. The 5 year adjusted progression rates to PC for patients without diabetes, a prior history of diabetes, and NODM were respectively: 1.5% (95% CI: 1.1%–1.9%); 3.1% (95% CI: 2.1%–4.1%); 4.6% (95% CI: 3.1%–6.1%) (Figure 2Figure 2.: Cumulative Risk of Progression of Low Risk Cysts to Pancreatic Cancer by Metabolic Status.). CONCLUSION: This is the first large scale, real world study of the malignant progression of low risk pancreatic cysts. Using significantly larger sample sizes than prior studies, results suggest that the risk of progression is small among patients with low risk cystic lesions, but increases linearly with time. The finding that new onset diabetes is associated with a significantly higher risk of progression, provides strong support for the inclusion of HbA1C, fasting glucose, and other markers of diabetes mellitus in cyst surveillance algorithms.

Cyst Fluid Biosignature to Predict Intraductal Papillary Mucinous Neoplasms of the Pancreas with High Malignant Potential

Current standard-of-care technologies, such as imaging and cyst fluid analysis, are unable to consistently distinguish intraductal papillary mucinous neoplasms (IPMNs) of the pancreas at high risk of pancreatic cancer from low-risk IPMNs. The objective was to create a single-platform assay to identify IPMNs that are at high risk for malignant progression. Building on the Verona International Consensus Conference branch duct IPMN biomarker review, additional protein, cytokine, mucin, DNA, and microRNA cyst fluid targets were identified for creation of a quantitative polymerase chain reaction-based assay. This included messenger RNA markers: ERBB2, GNAS, interleukin 1β, KRAS, MUCs1, 2, 4, 5AC, 7, prostaglandin E2R, PTGER2, prostaglandin E synthase 2, prostaglandin E synthase 1, TP63; microRNA targets: miRs 101, 106b, 10a, 142, 155, 17, 18a, 21, 217, 24, 30a, 342, 532, 92a, and 99b; and GNAS and KRAS mutational analysis. A multi-institutional international collaborative contributed IPMN cyst fluid samples to validate this platform. Cyst fluid gene expression levels were normalized, z-transformed, and used in classification and regression analysis by a support vector machine training algorithm. From cyst fluids of 59 IPMN patients, principal component analysis confirmed no institutional bias/clustering. Lasso (least absolute shrinkage and selection operator)-penalized logistic regression with binary classification and 5-fold cross-validation used area under the curve as the evaluation criterion to create the optimal signature to discriminate IPMNs as low risk (low/moderate dysplasia) or high risk (high-grade dysplasia/invasive cancer). The most predictive signature was achieved with interleukin 1β, MUC4, and prostaglandin E synthase 2 to accurately discriminate high-risk cysts from low-risk cysts with an area under the curve of up to 0.86 (p= 0.002). We have identified a single-platform polymerase chain reaction-based assay of cyst fluid to accurately predict IPMNs with high malignant potential for additional studies.

Long-term investigation of decompression as a definitive treatment for mandibular cysts associated with impacted third molars

BackgroundClinicians sometimes use decompression for secondary, low-risk cyst enucleation. The authors explored whether the decompression of dentigerous cysts associated with third molars is a reliable, long-term, definitive treatment option. MethodsThe authors monitored 25 mandibular cysts associated with impacted third molars in adults after surgical decompression without the extraction of the related tooth for a mean (standard deviation) time of 37 (15) months (range, 12-71 months). The authors carefully evaluated the postoperative clinical situation and the extent of radiographic shrinkage. ResultsA minimal epithelial slit remained patent in all patients. All lacked clinical problems, had no need for further intervention, and exhibited persistent impaction of the teeth. The cyst reduction rate calculated on panoramic radiographs ranged from 63.4% to 98.8% (mean [standard deviation] 87.5% [10.6%]) and was statistically significant (P < .05). In 13 patients, the reduction rate was greater than 90%. ConclusionsDecompression triggered marked radiographic reductions of cysts when epithelial communication persisted. The situation stabilized after the first 6 through 8 months, and no further intervention was required. Practical ImplicationsIt is risky to enucleate cysts associated with impacted third molars and extract the molars. Clinicians can solve the problem in dental practice by using surgical decompression.

Predicting the Grade of Dysplasia of Pancreatic Cystic Neoplasms Using Cyst Fluid DNA Methylation Markers

Purpose: Pancreatic cysts are common and pose diagnostic and management challenges. Pancreatic cyst fluid markers have the potential to aid in the management of cysts with concerning imaging findings. Our aim was to evaluate cyst fluid methylated DNA markers for their accuracy for predicting the histologic grade of neoplastic pancreatic cysts.Experimental Design: Pancreatic cyst fluid samples from 183 patients (29 discovery and 154 validation) aspirated after surgical resection were analyzed for methylated DNA at selected genes (SOX17, BNIP3, FOXE1, PTCHD2, SLIT2, EYA4, and SFRP1) using methylation-specific droplet-digital PCR (dd-QMSP). Methylated DNA levels were evaluated for their accuracy at predicting the grade of dysplasia of the pancreatic cyst.Results: All six markers evaluated in the validation set could accurately distinguish high-risk cystic neoplasms (with high-grade dysplasia and/or associated invasive cancer) from low-risk cysts (lower grades of dysplasia) with accuracies from 79.8% to 83.6%. Methylated SOX17 had the highest overall accuracy as a single marker (sensitivity, 78.4%; specificity, 85.6%; accuracy 83.6%, cutoff; 25 methylated DNA molecules/μL cyst fluid). The best four-gene combination had 84.3% sensitivity, 89.4% specificity, and 88.0% accuracy at distinguishing cysts with high-grade dysplasia and/or invasive cancer from those without. All six markers were independent predictors of having invasive cancer/high-grade dysplasia after adjusting for clinical/imaging factors known to be associated with grade of dysplasia. The combination of methylated SOX17 with cytology better predicted neoplastic grade than cytology alone.Conclusions: A panel of methylated gene markers quantified by dd-QMSP can be used to predict the grade of dysplasia of pancreatic cysts. Clin Cancer Res; 23(14); 3935-44. ©2017 AACR.

Competing Risks for Mortality in Patients With Asymptomatic Pancreatic Cystic Neoplasms: Implications for Clinical Management.

Pancreatic cystic neoplasms (PCNs) are being detected with increased frequency. Current clinical practice guidelines emphasize management based on cyst-related features. We aimed to evaluate the impact of comorbidity on mortality in PCN patients via competing risk analysis. We analyzed a retrospective cohort of patients diagnosed between 2005-2010, with follow-up through 2015, for overall and cause-specific mortality. Comorbidities were classified by the Charlson comorbidity index. We used Cox proportional hazards regression to evaluate the independent effect of cyst features, age, gender, and comorbidities on cause-specific mortality. Subgroup analysis was performed to determine the cause-specific mortality based on four a priori clinical profiles-healthy patients with low- or high-risk cysts, and high-comorbidity patients with low- or high-risk cysts. A total of 1,800 patients with PCNs comprised the study cohort (median follow-up 5.7 years). A total of 402 deaths (22.3%) occurred during the study period: 43 pancreatic cancer and 359 non-pancreatic cancer deaths. Compared to healthy patients without any high-risk cyst features (reference group), patients with high comorbidity as well as high-risk cyst features had an increased risk of overall mortality (Cox hazard ratio 6.30, 95% confidence interval (CI) 4.71, 8.42, P<0.01), pancreatic cancer mortality (subdistribution hazard ratio (SHR) 51.13, 95% CI 6.35, 411.29, P<0.01), as well as non-pancreatic cancer mortality (SHR 5.24, 95% CI 3.85, 7.12, P<0.01). Meanwhile, low-risk patients with a high-risk cyst were more likely to experience pancreatic cancer mortality (SHR 68.14, 95% CI 9.27, 501.01, P<0.01) rather than non-pancreatic cancer mortality (SHR 1.22, 95% CI 0.88, 1.71, P=0.23), compared to the reference group. Similarly, compared to the reference group, high-risk patients with a low-risk cyst were more likely to experience non-pancreatic cancer mortality (SHR 3.96, 95% CI 2.98, 5.26, P<0.01) rather than pancreatic cancer mortality (SHR 2.35, 95% CI 0.14, 38.82, P=0.55). Most of the deaths in the study were unrelated to pancreatic cancer. This has implications for clinical management. By applying patient-related factors in conjunction with cyst features, we defined commonly encountered patient profiles to help guide PCN clinical management.

A single-platform cyst fluid assay to predict IPMN with high-malignant potential for surgical resection: a multi-institutional international report

Objective: Current standard-of-care technologies are unable to distinguish IPMN at high-risk of malignancy from low-risk lesions. The objective of this study was to create a single-platform assay to identify IPMN that are at high-risk for malignant progression. Methods: Building on the Verona consensus conference BD-IPMN biomarker study; specific protein, cytokine, mucin, and miRNA cyst fluid targets were identified for creation of a q-PCR based assay. A multi-institutional international IPMN cyst fluid collaborative contributed patient samples to validate this platform. Cyst fluid gene expression levels were processed to obtain RQ values that were normalized, z-transformed, and utilized in classification and regression analysis by a support vector machine (SVM) training algorithm. Results: From 59 cyst fluid samples, principal component analysis confirmed no institutional bias/clustering. Sixty percent of eligible samples were randomized to a training set, followed by SVM model optimization with 10-fold cross-validation, and then applied to a test set. The model was repeated 100 times and performance determined by ROC analysis. Machine learning methods classified samples into low-risk (low/moderate dysplasia) or high risk (high-grade dysplasia/invasive cancer). The assay accurately discriminated high from low-risk cysts with a c-statistic (AUC) of 0.83 (figure). Conclusion: We have identified a single-platform PCR-based assay using multiple targets to predict IPMN with high-malignant potential. The creation of this test may allow patients with low-risk IPMN to avoid pancreatic surgery, while identifying patients with high-risk lesions so that they may undergo surgery before the development of invasive disease.

Preoperative characteristics and cytological features of 136 histologically confirmed pancreatic mucinous cystic neoplasms.

Mucinous cystic neoplasms (MCNs) of the pancreas present a management conundrum. The majority are benign but all are resected due to their malignant potential. Recent studies have recommended nonsurgical management. In the current study, the authors analyzed the preoperative imaging, cytology, and cyst fluid characteristics of 136 histologically confirmed MCNs to assess predictors of a high-risk (HR) cyst for surgical triage. MCNs resected at the Massachusetts General Hospital between 1990 and 2014 formed the study cohort. Patient demographics, cyst size, and mural nodules (MNs) by endoscopic ultrasound, cytology, and cyst fluid carcinoembryonic antigen and amylase levels were correlated with histological grade. A HR cyst was defined as high-grade dysplasia or invasive carcinoma on histology. Performance characteristics were assessed for each parameter, with a cyst size ≥3 cm or a MN on imaging and malignant cytology considered to be "true-positive" results for predicting malignancy. Only 15 of the 136 cysts had HR histology (11%). On average, patients with HR cysts were older than those with low-risk cysts (55 years vs 49 years, respectively). High-grade cytology was the most accurate predictor of malignancy (95%) followed by MN and cyst size together (88%) and MN alone (83%). The average carcinoembryonic antigen level (in ng/mL) increased with the grade of dysplasia but the ranges overlapped between low risk and HR cysts. To the authors' knowledge, the current study is the largest series to date analyzing the cytological features of histologically confirmed MCN. Cytology is insensitive but very specific for detecting a HR MCN and outperformed imaging for the detection of HR MCN. Endoscopic ultrasound-guided fine-needle aspiration and cytology should be performed on any clinically suspected MCN that is being considered for conservative management. Cancer Cytopathol 2017;125:169-177. © 2016 American Cancer Society.

Determining the natural history of pancreatic cystic neoplasms: a Manitoban cohort study

BackgroundMost pancreatic cystic neoplasms (PCN) are thought to harbor a low malignant potential. This historical cohort study attempts to describe the natural history of these lesions in a provincial cohort, to assess the safety of non-surgical management. Pathological diagnosis of malignancy was the primary outcome measure of interest. MethodsAll adult patients (age 18+) with PCN seen between 2000 and 2012 by the two main institutions in Manitoba were included in this study. PCN were graded as high and low risk, which dictated initial treatment plan (surgery or observation). Predictors of initial surgical treatment, delayed surgery in the observation group and the clinical/radiological predictors of malignancy were determined. Results497 patients were included in this study. 43 (8.7%) high-risk lesions underwent initial surgery, with 13 (30.2%) cases of malignancy discovered. 450 (90.5%) low-risk cysts were observed for a median of 17.3 months (range: 0.00–142.3). 29 (6.4%) cases of delayed surgery occurred, with malignancy discovered in five (17.2%). ConclusionsThis study supports current selection criteria for management of PCNs. Due to the low incidence of malignancy in low-risk PCN, it appears that long-term observation is safe and should be the treatment modality of choice in the absence of high-risk features.