Utilisation of artificial intelligence for the development of an EUS-convolutional neural network model trained to enhance the diagnosis of autoimmune pancreatitis.

OBJECTIVE: The diagnosis of autoimmune pancreatitis (AIP) is challenging. Sonographic and cross-sectional imaging findings of AIP closely mimic pancreatic ductal adenocarcinoma (PDAC) and techniques for tissue sampling of AIP are suboptimal. These limitations often result in delayed or failed diagnosis, which negatively impact patient management and outcomes. This study aimed to create an endoscopic ultrasound (EUS)-based convolutional neural network (CNN) model trained to differentiate AIP from PDAC, chronic pancreatitis (CP) and normal pancreas (NP), with sufficient performance to analyse EUS video in real time. DESIGN: A database of still image and video data obtained from EUS examinations of cases of AIP, PDAC, CP and NP was used to develop a CNN. Occlusion heatmap analysis was used to identify sonographic features the CNN valued when differentiating AIP from PDAC. RESULTS: From 583 patients (146 AIP, 292 PDAC, 72 CP and 73 NP), a total of 1 174 461 unique EUS images were extracted. For video data, the CNN processed 955 EUS frames per second and was: 99% sensitive, 98% specific for distinguishing AIP from NP; 94% sensitive, 71% specific for distinguishing AIP from CP; 90% sensitive, 93% specific for distinguishing AIP from PDAC; and 90% sensitive, 85% specific for distinguishing AIP from all studied conditions (ie, PDAC, CP and NP). CONCLUSION: The developed EUS-CNN model accurately differentiated AIP from PDAC and benign pancreatic conditions, thereby offering the capability of earlier and more accurate diagnosis. Use of this model offers the potential for more timely and appropriate patient care and improved outcome.

Determining age and sex-specific distribution of pancreatic whole-gland CT attenuation using artificial intelligence aided image segmentation: Associations with body composition and pancreatic cancer risk.

BACKGROUND & AIMS: Increased intrapancreatic fat is associated with pancreatic diseases; however, there are no established objective diagnostic criteria for fatty pancreas. On non-contrast computed tomography (CT), adipose tissue shows negative Hounsfield Unit (HU) attenuations (-150 to -30 HU). Using whole organ segmentation on non-contrast CT, we aimed to describe whole gland pancreatic attenuation and establish 5th and 10th percentile thresholds across a spectrum of age and sex. Subsequently, we aimed to evaluate the association between low pancreatic HU and risk of pancreatic ductal adenocarcinoma (PDAC). METHODS: The whole pancreas was segmented in 19,456 images from 469 non-contrast CT scans. A convolutional neural network was trained to assist pancreas segmentation. Mean pancreatic HU, volume, and body composition metrics were calculated. The lower 5th and 10th percentile for mean pancreatic HU were identified, examining the association with age and sex. Pre-diagnostic CT scans from patients who later developed PDAC were compared to cancer-free controls. RESULTS: Less than 5th percentile mean pancreatic HU was significantly associated with increase in BMI (OR 1.07; 1.03-1.11), visceral fat (OR 1.37; 1.15-1.64), total abdominal fat (OR 1.12; 1.03-1.22), and diabetes mellitus type 1 (OR 6.76; 1.68-27.28). Compared to controls, pre-diagnostic scans in PDAC cases had lower mean whole gland pancreatic HU (-0.2 vs 7.8, p = 0.026). CONCLUSION: In this study, we report age and sex-specific distribution of pancreatic whole-gland CT attenuation. Compared to controls, mean whole gland pancreatic HU is significantly lower in the pre-diagnostic phase of PDAC.

Combined Celiac Ganglia and Plexus Neurolysis Shortens Survival, Without Benefit, vs Plexus Neurolysis Alone.

BACKGROUND & AIMS: Pancreatic cancer produces debilitating pain that opioids often ineffectively manage. The suboptimal efficacy of celiac plexus neurolysis (CPN) might result from brief contact of the injectate with celiac ganglia. We compared the effects of endoscopic ultrasound-guided celiac ganglia neurolysis (CGN) vs the effects of CPN on pain, quality of life (QOL), and survival. METHODS: We performed a randomized, double-blind trial of patients with unresectable pancreatic ductal adenocarcinoma and abdominal pain; 60 patients (age 66.4±11.6 years; male 66%) received CPN and 50 patients (age 66.8±10.0 years; male 56%) received CGN. Primary outcomes included pain control and QOL at week 12 and survival (overall median and 12 months). Secondary outcomes included morphine response, performance status, secondary neurolytic effects, and adverse events. RESULTS: Rates of pain response at 12 weeks were 46.2% for CGN and 40.4% for CPN (P = .84). There was no significant difference in improvement of QOL between the techniques. The median survival time was significantly shorter for patients receiving CGN (5.59 months) compared to (10.46 months) (hazard ratio for CGN, 1.49; 95% CI, 1.02-2.19; P = .042), particularly for patients with non-metastatic disease (hazard ratio for CGN, 2.95; 95% CI, 1.61-5.45; P < .001). Rates of survival at 12 months were 42% for patients who underwent CPN vs 26% for patients who underwent CGN. The number of adverse events did not differ between techniques. CONCLUSION: In a prospective study of patients with unresectable pancreatic ductal adenocarcinoma and abdominal pain, we found CGN to reduce median survival time without improving pain, QOL, or adverse events, compared to CPN. The role of CGN must be therefore be reassessed. Clinicaltrials.gov no: NCT01615653.

Safety, Diagnostic Accuracy, and Effects of Endoscopic Ultrasound Fine-Needle Aspiration on Detection of Extravascular Migratory Metastases.

BACKGROUND & AIMS: Tumor cells can migrate via diminutive perivascular cuffing to distant sites along blood vessels to form extravascular migratory metastases (EVMM). These metastases usually are identified during surgery or autopsies. We aimed to evaluate the feasibility and safety of endoscopic ultrasound fine-needle aspiration (EUS-FNA) of perivascular soft-tissue cuffs to detect EVMM. We compared findings from EUS with those from noninvasive cross-sectional imaging (reference standard) of patients who underwent EUS-FNA to assess suspected EVMM and studied the effects on pancreatic tumor staging and determination of resectability. METHODS: We performed a retrospective analysis of 253 patients (mean age, 62 ± 12 y) who underwent EUS-FNA of 267 vessels for evaluation of suspected EVMM, from April 2001 through May 2018. EUS findings were compared with those from computed tomography (CT) and magnetic resonance imaging (MRI) as the reference standard. Lesions were considered to be malignant based on cytology analysis of FNA samples, histology analyses of surgical or biopsy specimens, or vascular abnormalities detected by CT or MRI that clearly indicate EVMM. RESULTS: Thirty patients were found to have benign lesions. The remaining 223 patients who had malignancies (166 with pancreatic ductal adenocarcinomas [PDACs]), underwent further analyses. A median of 4 FNAs (range, 1-20 FNAs) were obtained from 4-mm perivascular soft-tissue cuffs (range, 2-20 mm). FNA and cytology analysis showed malignant cells in 163 vessels (69.4%) from 157 patients (70.4%). CT or MRI did not detect EVMM in 44 patients (28%) with malignancies, including 24 patients (24%) with newly diagnosed PDAC. Detection of EVMM by EUS-FNA resulted in upstaging of 15 patients and conversion of 14 patients with PDAC from resectable (based on CT or MRI) to unresectable. No adverse events were reported during a follow-up period of 3.9 months (range, 0-117 mo). CONCLUSIONS: EUS-FNA and cytologic analysis of perivascular soft-tissue cuffs can detect EVMM that were not found in 28% of patients by CT or MRI. Detection of EVMM affects tumor staging and determination of tumor resectability.

High-Grade Dysplasia in Resected Main-Duct Intraductal Papillary Mucinous Neoplasm (MD-IPMN) is Associated with an Increased Risk of Subsequent Pancreatic Cancer.

BACKGROUND: There is lack of consensus on post-operative surveillance for resected non-invasive intraductal papillary neoplasms (IPMNs). In this study we explored risk factors for subsequent PC in patients with MD-IPMN undergoing partial pancreatectomy. METHODS: We searched the Mayo Clinic surgical pathology database for all cases of resected MD-IPMN between 1997 and 2014. Cases with histologically confirmed main pancreatic duct involvement either isolated or in a mixed pattern with branch-duct involvement were included. Outcomes of PC in the remnant pancreas, and death related to MD-IPMN were assessed with survival analyses (Kaplan-Meier and Cox regression). RESULTS: Among the 179 patients with resected MD-IPMN the incidence of concomitant PC and high-grade dysplasia (HGD) in the resected specimen was 23 and 14%, respectively. The mean duration of follow-up was 4.31 years (range 0.12-13.5 years). Excluding 28 subjects who either underwent initial total pancreatectomy or partial pancreatectomy with surgical margins positive for PC/HGD, the 5-year incidence of subsequent PC was 12%, including 60.6% and 15.6% in those with initial PC and HGD, respectively. The 10-year incidence of PC was 21.2% overall, 60.6% for PC, 38.3% for HGD, and 3.0% for LGD. Risk of subsequent PC was significantly higher for those with initial PC compared with HGD (HR = 4.95, 95% CI: 1.63-15.03, p = 0.005 and for HGD compared with LGD (HR = 11.30, 95% CI: 1.55-82.26, p = 0.017). CONCLUSIONS: Patients with MD-IPMN with PC or HGD undergoing segmental pancreatectomy are at higher risk of subsequent PC and may benefit from post-operative surveillance. The post-operative surveillance intervals in resected MD-IPMNs need to be tailored based on dysplasia grade.

Classic chronic pancreatitis is associated with prior acute pancreatitis in only 50% of patients in a large single-institution study.

BACKGROUND: The sentinel acute pancreatitis event (SAPE) hypothesis for pathogenesis of chronic pancreatitis (CP) postulates that acute pancreatitis (AP), especially recurrent AP (RAP), precedes development of CP. However, in a recent population-based study, 52/89 (58.4%) of CP had no prior episodes of AP. In a large clinic-based CP cohort, we aimed to determine the incidence and timing of prior AP in patients diagnosed with CP. METHODS: We retrospectively identified 499 consecutive patients with classic CP diagnosed at our institution from January 2013 through December 2015. We abstracted their demographic and clinical data, especially regarding prior AP. RESULTS: We identified 3 cohorts: 1) CP with no AP (n = 231 [46.3%]), 2) AP before CP (n = 250 [50.1%]), and 3) AP after CP (n = 18 [3.6%]). At CP diagnosis, 249 patients (49.9%) had no prior AP. Compared with the "CP preceded by AP" cohort, the "CP without AP"' cohort was older (59.2 ±â€¯13.9 vs 48.6 ±â€¯15.7 years; P < .001), had a higher prevalence of diabetes mellitus (30.3% vs 12.4%; P < .001), reported less pain (52.8% vs 87.6%; P < .001), and had a lower rate of endoscopic interventions (16.0% vs 39.2%; P < .001). In the "CP preceded by AP" cohort, 117 (46.8%) had a single episode of AP and 133 (53.2%) had RAP. CONCLUSION: Nearly half the patients with classic CP did not have prior AP. Only a quarter of patients had CP that could potentially have evolved from prior RAP. Development of CP may be attributable to an altogether different pathogenesis (a non-SAPE pathway) for a considerable proportion of patients.

Rituximab Maintenance Therapy Reduces Rate of Relapse of Pancreaticobiliary Immunoglobulin G4-related Disease.

BACKGROUND & AIMS: IgG4-related disease (IgG4-RD), a multi-organ fibroinflammatory syndrome, typically responds to steroids. However, some cases are steroid resistant, and pancreaticobiliary IgG4-RD commonly relapses after steroid withdrawal. Rituximab induces remission of IgG4-RD, but the need for and safety of maintenance rituximab treatment are unknown. We compared outcomes of patients with pancreaticobiliary IgG4-RD treated with or without maintenance rituximab therapy. METHODS: We performed a retrospective study of patients with pancreaticobiliary IgG4-RD treated with rituximab at the Mayo Clinic in Rochester, Minnesota, from January 2005 through December 2015. The cohort was divided into patients who received only rituximab induction therapy (group 1, n = 14) and patients who received rituximab induction followed by maintenance therapy (group 2, n = 29). We collected data on recurrence of IgG4-RD symptoms and findings, as well as information on evaluations, treatment, and adverse events. RESULTS: Median follow-up times were similar between group 1 (34 mo) and group 2 (27 mo) (P = .99). Thirty-seven patients (86%) were in steroid-free remission 6 months after rituximab initiation. A higher proportion of patients in group 1 had disease relapse (3-year event rate, 45%) than in group 2 (3-year event rate, 11%) (P = .034). Younger age, higher IgG4 responder index score after induction therapy, and increased serum levels of alkaline phosphatase at baseline or after rituximab induction were associated with relapse. Infections developed in 6 of 43 patients, all in group 2 (P = .067 vs group 1); all but 1 occurred during maintenance therapy. CONCLUSIONS: In a retrospective study of patients with pancreaticobiliary IgG4-RD, we found rituximab maintenance therapy prolongs remission. Relapses are uncommon among patients receiving maintenance therapy, but maintenance therapy may increase risk of infection. Patients with factors that predict relapse could be candidates for rituximab maintenance therapy.

Risk of Pancreatic Cancer in Patients With Pancreatic Cysts and Family History of Pancreatic Cancer.

BACKGROUND & AIMS: A diagnosis of pancreatic cancer in a first-degree relative increases an individuals' risk of this cancer. However, it is not clear whether this cancer risk increases in individuals with pancreatic cystic lesions who have a first-degree relative with pancreatic cancer. The Fukuoka criteria are used to estimate risk of pancreatic cancer for patients with pancreatic cystic lesions: individuals with cysts with high risk or worrisome features (Fukuoka positive) have a higher risk of pancreatic cancer than individuals without these features (Fukuoka negative). We aimed to compare the risk of pancreatic cancer and surgery based on presence or absence of pancreatic cystic lesions and a first-degree relative with pancreatic cancer. METHODS: We performed a retrospective study of patients seen at the Mayo Clinic in Rochester, Minnesota, from January 1, 2000, through December 31, 2012. We identified individuals with: pancreatic cystic lesions and first-degree relative with pancreatic cancer (group 1, n = 269), individuals with pancreatic cystic lesions but no first-degree relative with pancreatic cancer (group 2, n = 1195), and individuals without pancreatic cystic lesions but with a first-degree relative with pancreatic cancer (group 3, n = 720). We compared, among groups, as well among patients with cysts classified according to Fukuoka criteria, proportions of individuals who developed pancreatic cancer or underwent pancreatic surgery within a 5-year period. RESULTS: A significantly higher proportion of individuals in group 1 developed pancreatic cancer during the 5-year period than in group 3 (6.64% vs 1.69%; P = .03); there was no significant difference between the percentage of individuals in group 1 vs group 2 who developed pancreatic cancer (6.64% vs 4.05%; P = .41). There was no significant difference in pancreatic cancer development among individuals with Fukuoka-positive cysts with vs without a family history of pancreatic cancer (P = .39). There was no significant difference in the proportion of patients in group 1 vs group 2 who underwent pancreatic surgery for their pancreatic cyst over the 5-year period (14.37% vs 11.80%; P = .59). Among patients with Fukuoka-negative cysts, a significantly higher proportion underwent surgery in group 1 than in group 2 (10.90% vs 5.90%; P = .03). However, among patients with Fukuoka-positive cysts, there was no difference in proportions of patients who underwent surgery between groups 1 and 2 (P = .66). CONCLUSIONS: In a retrospective study of patients with pancreatic cysts and/or cancer, we found that a family history of pancreatic cancer does not affect 5-year risk of pancreatic cancer in patients with pancreatic cystic lesions. Despite this, among patients with Fukuoka-negative cysts, a higher proportion of those with a family history of pancreatic cancer undergo surgery than patients without family history of pancreatic cancer.

Large-caliber metal stents versus plastic stents for the management of pancreatic walled-off necrosis.

BACKGROUND AND AIMS: Symptomatic pancreatic walled-off necrosis (WON) may be managed by endoscopic transmural drainage and endoscopic transmural necrosectomy, with stent placement at endoscopic drainage sites. The optimal stent choice is yet to be determined. We compared outcomes after endoscopic management of WON using either large-caliber fully covered self-expandable metal stents (LC-SEMSs) or double-pigtail plastic stents (DPPSs). METHODS: We performed a retrospective comparison of outcomes among patients who received LC-SEMSs or DPPSs before endoscopic transmural necrosectomy for WON. RESULTS: Among 94 patients included, WON resolution rates did not differ between the DPPS (36 patients) and LC-SEMS (58 patients) groups, whether concomitant percutaneous drainage was considered a failure (75% vs 82.8%; P = .36) or not (91.7% vs 94.8%; P = .55). Of 75 patients (80%) successfully treated without percutaneous drainage, 37 (49%) underwent endoscopic transmural drainage without subsequent endoscopic transmural necrosectomy. WON was more likely to resolve without subsequent endoscopic transmural necrosectomy in the LC-SEMS group than the DPPS group (60.4% vs 30.8%; P = .01). WON resolution without subsequent endoscopic transmural necrosectomy remained more likely with LC-SEMSs (odds ratio, 4.5 [95% confidence interval, 1.5-15.5]) after adjusting for patient age and size and location of WON. Rates of adverse events were similar except for clinically significant bleeding requiring endoscopic intervention, which was higher with DPPSs than LC-SEMSs (14% vs 2%; P = .02). CONCLUSION: Management of pancreatic WON with LC-SEMSs appears to decrease both the need for repeated necrosectomy procedures and the risk of intervention-related hemorrhage.

Metal stents versus plastic stents for the management of pancreatic walled-off necrosis: a systematic review and meta-analysis.

BACKGROUND AND AIMS: Endoscopic transluminal drainage of symptomatic walled-off necrosis (WON) is a good management option, although the optimal choice of drainage site stent is unclear. We performed a systematic review and meta-analysis to compare metal stents (MSs) and plastic stents (PSs) in terms of WON resolution, likelihood of resolution after 1 procedure, and adverse events. METHODS: An expert librarian queried several databases to identify studies that assessed WON management, and selection was according to a priori criteria. Publication bias, heterogeneity, and study quality were evaluated with the appropriate tools. We performed single and 2-arm meta-analyses for noncomparative and comparative studies using event rate random-effects model and odds ratio (OR)/difference in means, respectively. RESULTS: We included 41 studies involving 2213 patients. In 2-arm study meta-analysis, WON resolution was more likely with MSs compared with PSs (OR, 2.8; 95% confidence interval, 1.7-4.6; P < .001). Resolution with a single endoscopic procedure was similar between stents (47% vs 44%), although for those cases requiring more than 1 intervention, the MS group had fewer interventions, favored by a mean difference of -.9 procedures (95% CI, -1.283 to -.561). In single-arm study meta-analysis, when compared with PSs, MS use was associated with lower bleeding (5.6% vs 12.6%; P = .02), a trend toward lower perforation and stent occlusion (2.8% vs 4.3%, P = .2, and 9.5% vs 17.4%, P = .07), although with higher migration (8.1% vs 5.1%; P = .1). CONCLUSION: Evidence suggests that MSs are superior for WON resolution, with fewer bleeding events, trend toward less occlusion and perforation rate, but increased migration rate compared with PSs.

Endoscopic Ultrasound Fine-Needle Aspiration Diagnosis of Synchronous Primary Pancreatic Adenocarcinoma and Effects on Staging and Resectability.

Synchronous primary pancreatic adenocarcinoma, defined as the simultaneous presence of 2 or more newly identified and anatomically separate primary adenocarcinomas within the pancreas, is reported rarely. We compared endoscopic ultrasound (EUS) and computed tomography (CT) and magnetic resonance imaging (MRI) findings from patients with synchronous primary pancreatic adenocarcinoma and their effects on cancer staging and treatment. We performed a retrospective analysis of the EUS database at the Mayo Clinic, from September 2008 through May 2016, to collect EUS, CT, MRI, and clinical data from patients with synchronous primary pancreatic adenocarcinoma. EUS and separate fine-needle aspiration of both tumors detected synchronous primary pancreatic adenocarcinoma in 11 patients (70.9 ± 10.4 y; 64% men). Of the 22 cancers, CT (n = 9) and MRI (n = 2) detected 9 (41%) cancers; in only 2 patients did CT detect both cancers. EUS increased cancer stage for 7 of the 11 (64%) patients and changed the status from resectable to unresectable for 3 of the 9 (33%) patients, compared with CT or MRI. EUS findings altered the likely extent of surgical resection for 3 patients. Synchronous primary pancreatic adenocarcinoma is reported rarely and may be undetected by CT or MRI; this could account for the false presumption of early tumor recurrence, rather than actual residual second tumor, leading to incomplete resection.

Remote malignant intravascular thrombi: EUS-guided FNA diagnosis and impact on cancer staging.

BACKGROUND AND AIMS: Malignant vascular invasion usually results from gross direct infiltration from a primary tumor and impacts cancer staging, prognosis, and therapy. However, patients may also develop a remote malignant thrombi (RMT), defined as a malignant intravascular thrombus located remote and noncontiguous to the primary tumor. Our aim was to compare EUS, CT, and magnetic resonance imaging (MRI) findings of RMT and to explore the potential impact on cancer staging. METHODS: Patients with RMT were identified from a prospectively maintained EUS database. Retrospective chart review was performed to obtain EUS, CT/MRI, clinical, and outcome data. RESULTS: A median of 3 FNAs (range, 1-8) was obtained from RMT in 17 patients (60 ± 14.1 years, 56% men) between April 2003 and August 2016, with the finding of malignant cytology in 12 patients (70.6%; 10 positive, 2 suspicious). CT/MRI detected the RMT in 5 patients (29.4%), 4 of whom had positive or suspicious EUS-FNA cytology. Among the 8 newly diagnosed pancreatic adenocarcinoma (PaC) patients, CT did not detect the RMT in 5 (63%), of whom 3 patients had positive or suspicious intravascular EUS-FNA cytology. For newly diagnosed PaC patients (n = 8), the EUS-FNA diagnosis of a biopsy specimen-proven RMT upstaged 3 patients (37.5%) and converted 2 patients (25%) from CT resectable to unresectable disease. No adverse events were reported. The mean follow-up was 18.9 ± 27.7 months. CONCLUSIONS: Our study demonstrates the ability and potential safety of intravascular FNA to detect radiographically occult RMT, which substantially impacts cancer staging and resectability.

Clinical profiles and outcomes in idiopathic duct-centric chronic pancreatitis (type 2 autoimmune pancreatitis): the Mayo Clinic experience.

OBJECTIVE: Idiopathic duct-centric chronic pancreatitis (IDCP), also known as type 2 autoimmune pancreatitis (AIP), is an uncommon subtype of AIP. International Consensus Diagnostic Criteria for IDCP propose that the diagnosis requires pancreatic histology and/or concurrent IBD. We examined our experience with IDCP (type 2 AIP) to assess the appropriateness of these criteria, and identify unique characteristics in patients presenting with acute pancreatitis. DESIGN: We reviewed the Mayo Clinic AIP database through May 2014 to identify subjects with either definitive (n=31) or probable (n=12) IDCP. We compared demographic and clinical factors based on strength of diagnostic confidence (definitive versus probable), presence of IBD, and acute pancreatitis as the presenting manifestation. Relapse-free survival was determined using the Kaplan-Meier method. RESULTS: The clinical profiles were similar irrespective of the diagnostic criteria fulfilled. Common clinical presentations included acute pancreatitis (n=25, 58.1%, 12 of whom (27.9%) had recurrent pancreatitis) and pancreatic mass/obstructive jaundice (n=15, 34.9%). The cumulative relapse rate was 10.6% at 3 years (median follow-up 2.9 years). Relapse-free survival was similar for the different diagnostic categories, but was decreased in those initially presenting with acute pancreatitis (p=0.047) or treated with steroids (vs surgery, p=0.049). CONCLUSIONS: The current diagnostic classification of probable IDCP and the inclusion of IBD as a supportive criterion appear valid, because patients have similar clinical profiles and disease-related outcomes to those with definitive IDCP. Concurrent IBD, especially in young patients, may suggest when IDCP is the underlying cause of recurrent acute pancreatitis, but additional studies are needed for validation.

Obstructive jaundice in autoimmune pancreatitis can be safely treated with corticosteroids alone without biliary stenting.

OBJECTIVE: Autoimmune pancreatitis (AIP) responds dramatically to corticosteroids treatment. We reviewed our experience to determine the safety and effectiveness of treating obstructive jaundice in definitive AIP with corticosteroids alone without biliary stenting. METHODS: From our AIP database, we retrospectively identified type 1 AIP subjects whose jaundice was treated with corticosteroids alone without biliary stenting. Their medical records were reviewed and clinical data were evaluated to determine the outcomes. RESULTS: Fifteen AIP subjects (87% male, mean age 68.4 years) were treated with corticosteroids at initial presentation (n = 8), first (n = 5) or subsequent (n = 2) relapse. Mean values (upper limit of normal, ULN) of liver tests prior to corticosteroids were aspartate aminotransferase (AST) 203.5u/l (4 × ULN), alanine aminotransferase (ALT) 325.8u/l (6 × ULN), alkaline phosphatase (ALP) 567.4u/l (5 × ULN), and total bilirubin (TB) 5.9 mg/dl (5.9 × ULN). At first follow-up (mean 4 days) the decrease was 54.9% for AST, 51.6% for ALT, 33% for ALP and 47.2% for TB (all p < 0.05). After 15-45 days, all patients had normal AST, 3/15 had ALT > 1.5 × ULN, 1/15 had ALP > 1.5 × ULN, 1/15 had TB > 1.5 × ULN. No patient required biliary stent placement, or developed cholangitis or other infectious complications during steroid treatment. CONCLUSION: Under the supervision of an experienced pancreatologist and with close monitoring of patients, obstructive jaundice secondary to definitive AIP can be safely and effectively managed with corticosteroids alone, without the need for biliary stenting.

EUS-guided ethanol lavage does not reliably ablate pancreatic cystic neoplasms (with video).

BACKGROUND: The true efficacy of EUS-guided ethanol lavage (EEL) of pancreatic cystic neoplasms is unclear. This study aimed to assess long-term outcomes and adverse events of EEL by using a standardized protocol. METHODS: Single-center, prospective, pilot study in which participants with suspected mucinous cyst neoplasms or branch duct intraductal papillary mucinous neoplasms ≥1 cm in maximum diameter underwent EEL with 80% ethanol. Follow-up cross-sectional imaging was obtained to assess for changes in cyst volume. RESULTS: Twenty-three patients underwent EEL (57% male, mean age 70 years). Mean duration of follow-up was 40 months (range 9-82 months). Mean calculated final concentration of ethanol achieved in treated cysts was 50% (range 0%-79%). Complete resolution of pancreatic cystic neoplasms occurred in 2 participants (9%). When stratified into those participants who achieved ≥80% versus <80% reduction in cyst volume, no statistically significant differences were seen with regard to patient demographics, cyst characteristics, or final concentration of ethanol achieved in the treated cyst. Greater decreases in cyst volume were seen in presumed nonmucinous cysts compared with presumed mucinous cysts (P = .006). Two early adverse events occurred. Five participants died during the study follow-up period (4 from nonpancreatic causes), including 1 participant who was diagnosed with pancreatic adenocarcinoma thought to have arisen from the treated branch duct intraductal papillary mucinous neoplasm 41 months after undergoing EEL. CONCLUSIONS: As performed in this study, EEL therapy does not appear to be a promising method for prevention of malignancy in pancreatic cysts. Endoscopic methods that effectively and completely ablate pancreatic cystic neoplasms are needed. ( CLINICAL TRIAL REGISTRATION NUMBER: NCT02158039.).

Pancreatic cyst epithelial denudation: a natural phenomenon in the absence of treatment.

BACKGROUND AND AIMS: The presence and significance of epithelial denudation among treatment-naïve pancreatic cystic lesions (PCLs) remain undetermined. The aims of this study were to determine the prevalence, extent, and predictors of epithelial denudation in treatment-naïve PCLs. METHODS: Single-center retrospective study including patients who underwent EUS preceded by cross-sectional imaging and who subsequently underwent surgical resection of treatment-naïve PCLs. Surgically resected PCLs were reviewed by a pathologist in a fashion that allowed evaluation from evenly distributed regions of the cyst. RESULTS: A total of 140 patients were identified (60% female, mean age 63 years). Eighty-five cysts (60.7%) were classified as intraductal papillary mucinous neoplasms (IPMNs), 33 (23.5%) as main duct IPMNs (m-IPMNs), 11 (7.9%) as serous cystadenomas (SCAs), and 11 (7.9%) were composed of other cyst subtypes. A greater extent of epithelial denudation was seen in mucinous cystic neoplasm (MCN) compared with IPMN and SCA (mean percentage of denuded epithelium 45.1%, 10.8%, and 22.4%, respectively [P < .0001]). An association existed between the extent of denuded epithelium and degree of cyst epithelial dysplasia for IPMN and MCN combined (mean percentage of denuded epithelium for low-, moderate-, and high-grade dysplasia being 23.3%, 4.5%, and 1.2%, respectively; P = .02). PCLs resected from the neck and/or body and/or tail of the pancreas were associated with a greater extent of mean percentage of denuded epithelium than PCLs resected from the head and/or uncinate of the pancreas (23.9% vs 13.4%; P = .035). CONCLUSIONS: The presence and extent of cyst epithelial denudation of treatment-naïve PCLs vary with cyst histology and other factors. The observation of denudation after intracystic ablative therapy may not provide an adequate metric of successful intervention. Further studies are needed to validate these findings.

Impact of celiac neurolysis on survival in patients with pancreatic cancer.

BACKGROUND: Pancreatic cancer (PC) often produces pain that is difficult to control. Celiac neurolysis (CN) is performed with the goal of improving pain control and quality of life while reducing opioid-related side effects. OBJECTIVE: We aimed to evaluate whether CN provides a survival advantage for PC patients. DESIGN: Retrospective case-control study. SETTING: Single tertiary-care referral center. PATIENTS: Review of a prospectively maintained database identified patients with unresectable PC who underwent CN over a 12-year period. Each patient was matched to 2 control patients with unresectable PC. INTERVENTION: CN, which included both celiac plexus neurolysis (CPN) and celiac ganglia neurolysis (CGN). MAIN OUTCOME MEASUREMENTS: Median survival in Kaplan-Meier curves and hazard ratios. RESULTS: A total of 417 patients underwent CN and were compared with 840 controls with PC. Baseline characteristics were similar except the CN group had greater weight loss and pain requiring opioids. A mean of 16.6 ± 5.8 mL of alcohol was administered. For patients who underwent CN, the median survival from the time of presentation was shorter compared with controls (193 vs 246 days; hazard ratio 1.32; 95% confidence interval, 1.13-1.54). There was no difference in survival with unilateral or bilateral injection. However, EUS-guided CN was associated with longer survival compared with non-EUS approaches, and those who received CPN had longer survival compared with CGN. LIMITATIONS: Single center, retrospective. CONCLUSION: Our study suggests that CN is an independent predictor of shortened survival in PC patients. A prospective study is needed to verify the findings and determine whether shortened survival results from CN or from other features such as performance status and tumor-related characteristics. It is also imperative to verify our finding that EUS-guided CN provides a survival advantage over other approaches and whether CPN prolongs survival compared with CGN.

The string sign for diagnosis of mucinous pancreatic cysts.

BACKGROUND AND STUDY AIMS: Pancreas cyst fluid analysis does not provide optimal discrimination between mucinous and nonmucinous cysts. The aim of this study was to assess the performance characteristics of the "string sign" - a test performed at the time of endoscopic ultrasound-guided fine-needle aspiration (EUS-FNA), for the diagnosis of mucinous pancreatic cysts (branch duct intraductal papillary mucinous neoplasms [bIPMN] and mucinous cystic neoplasms). PATIENTS AND METHODS: Patients undergoing EUS-FNA of pancreatic cystic lesions at one referral center between 2003 and 2012 were included. The string sign was performed prospectively, and was considered positive if ≥ 1 cm string formed in cyst fluid and lasted for ≥ 1 second. Performance characteristics of the string sign and a sequential cyst fluid test interpretation model were assessed. RESULTS: For 98 histologically proven cases, the sensitivity, specificity, positive predictive value, and negative predictive value of the string sign for diagnosis of mucinous cysts were 58 % (95 % confidence interval [CI] 44 % - 70 %), 95 % (83 % - 99 %), 94 % (81 % - 99 %), and 60 % (46 % - 72 %), respectively. When string sign results and carcinoembryonic antigen (CEA) concentration (≥ 200 ng/mL) were combined, diagnostic accuracy improved from 74 % and 83 %, respectively, to 89 % (P ≤ 0.03). Among bIPMN, a positive string sign was associated with gastric and intestinal epithelial subtypes. The sequential cyst fluid test interpretation model (including cytology, mucin stain, CEA, and string sign) yielded an overall sensitivity for mucinous lesions of 96 %, with a specificity of 90 %. CONCLUSIONS: The string sign is highly specific for diagnosis of mucinous pancreatic cysts, and improves overall diagnostic accuracy of pancreatic cyst fluid analysis. Sequential cyst fluid test interpretation yields high diagnostic sensitivity and specificity for mucinous cysts.

Treatment of relapsing autoimmune pancreatitis with immunomodulators and rituximab: the Mayo Clinic experience.

BACKGROUND: There is a paucity of data on long-term management of type 1 autoimmune pancreatitis (AIP), a relapsing steroid-responsive disorder. OBJECTIVE: We describe our experience with treatment of relapses and maintenance of remission using steroid-sparing immunomodulators (IMs) and induction of remission using rituximab (RTX). METHODS: We obtained details of disease relapse and treatment in 116 type 1 AIP patients from clinic visits, medical records and telephone interviews. We compared relapse free survival in those treated with IMs versus those treated with steroids alone, assessed patients' response to RTX, and identified treatment-related complications. RESULTS: During a median follow-up of 47 months, 52/116 AIP patients experienced 76 relapse episodes. The first relapse was treated with another course of steroids in 24 patients, and with steroids plus IM in another 27 patients; subsequent relapse-free survival until a second relapse was similar in the two groups (p=0.23). 38 patients received an IM for >2 months; failure or intolerance of IM therapy occurred in 17 (45%). 12 patients with steroid or IM intolerance/resistance were treated with RTX, an antiCD20 antibody; 10 (83%) experienced complete remission and had no relapses while on maintenance therapy. Treatment-limiting side effects related to RTX were uncommon. CONCLUSIONS: In type 1 AIP relapses are common. Relapse-free survival is similar in those treated with steroids plus IM compared to those treated with steroids alone. Nearly half the patients on IMs will relapse during treatment. RTX is effective in the treatment of both IM resistant and steroid intolerant patients.