Prediction of thirty-day morbidity and mortality after duodenal switch using an artificial neural network

BACKGROUND

Understanding factors that increase risk of both mortality and specific measures of morbidity after duodenal switch (DS) is important in deciding to offer this weight loss operation. Artificial neural networks (ANN) are computational deep learning approaches that model complex interactions among input factors to optimally predict an outcome. Here, a comprehensive national database is examined for patient factors associated with poor outcomes, while comparing the performance of multivariate logistic regression and ANN models in predicting these outcomes.

METHODS

2907 DS patients from the 2019 Metabolic and Bariatric Surgery Accreditation and Quality Improvement Program database were assessed for patient factors associated with the previously validated composite endpoint of 30-day postoperative reintervention, reoperation, readmission, or mortality using bivariate analysis. Variables associated (P ≤ 0.05) with the endpoint were imputed in a multivariate logistic regression model and a three-node ANN with 20% holdback for validation. Goodness-of-fit was assessed using area under receiver operating curves (AUROC).

RESULTS

There were 229 DS patients with the composite endpoint (7.9%), and 12 mortalities (0.4%). Associated patient factors on bivariate analysis included advanced age, non-white race, cardiac history, hypertension requiring 3 + medications (HTN), previous foregut/obesity surgery, obstructive sleep apnea (OSA), and higher creatinine (P ≤ 0.05). Upon multivariate analysis, independently associated factors were non-white race (odds ratio 1.40; P = 0.075), HTN (1.55; P = 0.038), previous foregut/bariatric surgery (1.43; P = 0.041), and OSA (1.46; P = 0.018). The nominal logistic regression multivariate analysis (n = 2330; R² = 0.02, P < 0.001) and ANN (R² = 0.06; n = 1863 [training set], n = 467 [validation]) models generated AUROCs of 0.619, 0.656 (training set) and 0.685 (validation set), respectively.

CONCLUSION

Readily obtainable patient factors were identified that confer increased risk of the 30-day composite endpoint after DS. Moreover, use of an ANN to model these factors may optimize prediction of this outcome. This information provides useful guidance to bariatricians and surgical candidates alike.

Rendezvous transgastric and percutaneous sinus tract endoscopy (STE) for debridement of necrotic collections with deep retroperitoneal extension: a case series (with video): Meeting presentations: Digestive Disease Week 2018

Background and study aims  Infected necrotic collections extending deep into the retroperitoneum may not be amenable solely to endoscopic necrosectomy. Rendezvous transgastric and percutaneous sinus tract endoscopic necrosectomy was conceived to maximize debridement and obviate the need for open necrosectomy or video-assisted retroperitoneal debridement. Patients and methods  Patients who underwent simultaneous rendezvous transgastric and sinus tract endoscopic transluminal necrosectomy at the same session were identified from a prospectively maintained database. Demographic data, clinical outcomes, immediate and late complications were evaluated. Results  Among 415 patients with necrotizing pancreatitis, four patients (three males, median age 47 years) underwent this intervention after a median 29.5 days following placement of percutaneous drain. Intra-procedural bleeding following dilation of percutaneous tract required placement of esophageal stent for tamponade. No patients required rescue open necrosectomy or video assisted retroperitoneal debridement. Complete removal of percutaneous drains was accomplished in all patients after a median of 78.5 days. Conclusions  Our novel approach is safe and effective and can expand the available armamentarium for management of large necrotic collection with deep retroperitoneal extension.

Evaluation of Rosemont criteria for non-calcific chronic pancreatitis (NCCP) based on histopathology - A retrospective study

BACKGROUND

Rosemont classification for chronic pancreatitis has not been evaluated specifically in non-calcific chronic pancreatitis (NCCP) patients and to this date, it has not been correlated with the gold standard namely histopathology.

OBJECTIVE

To assess the correlation of EUS Rosemont criteria for NCCP with histopathology from surgical specimens and evaluate the impact of age, sex, BMI, smoking and alcohol on Rosemont classification.

METHODS

Adult patients undergoing TPIAT for NCCP between July 2009 and January 2013 were identified from our institutional database. The presence or absence of standard and Rosemont (major and minor) criteria were determined by expert endosonographers using linear endosonography. Patients were categorized into normal, indeterminate and suggestive with CP based on Rosemont classification. Histology was obtained at time of TPIAT from the resected pancreas by wedge biopsy of head, body and tail. All histopathology were re-reviewed by a GI pathologist blinded to endosonographic features and clinical outcomes. Available pancreatic tissue was graded for severity of intralobular and perilobular pancreatic fibrosis by the Ammann classification system.

RESULTS

50 patients with NCCP (42 females, mean age± SD = 37.9 ± 10.8) underwent TPIAT with preoperative EUS during the study period. Univariate analysis of features such as age, sex, BMI, smoking and alcohol history showed no significant difference between patients identified as normal and those identified as indeterminate/suggestive (p > 0.05). Rosemont "Normal" was poor in excluding CP as 5/9 patients (55.5%) had CP on histopathology. 25/26 patients (96.2%) with features "suggestive" of CP had evidence of CP on histopathology. 12/15 patients (80.0%) with "indeterminate" features had CP on histopathology.

CONCLUSIONS

Rosemont classification can be used independent of patient characteristics (age, sex and BMI) and environmental factors (smoking and alcohol exposure). In our cohort, Rosemont classification was strongly predictive of CP in patients with features "suggestive" of CP. However, "normal" Rosemont classification had poor correlation in this study. This is maybe due to lack of true comparator "normal" pancreas which cannot be obtained reasonably. The strength of agreement for diagnosis of CP was substantial between the standard and Rosemont criteria.

Interobserver agreement among cytopathologists in the evaluation of pancreatic endoscopic ultrasound-guided fine needle aspiration cytology specimens

BACKGROUND AND AIMS

Endoscopic ultrasound with fine needle aspiration (EUS-FNA) has become the standard of care in the evaluation of solid pancreatic lesions. Limited data exist on interobserver agreement (IOA) among cytopathologists in assessing solid pancreatic EUS-FNA specimens. This study aimed to evaluate IOA among cytopathologists in assessing EUS-FNA cytology specimens of solid pancreatic lesions using a novel standardized scoring system and to assess individual clinical and cytologic predictors of IOA.

METHODS

Consecutive patients who underwent EUS-FNA of solid pancreatic lesions at a tertiary care referral center were included. EUS-FNA slides were evaluated by four blinded cytopathologists using a standardized scoring system that assessed final cytologic diagnosis and quantitative (number of nucleated/diagnostic cells) and qualitative (bloodiness, inflammation/necrosis, contamination, artifact) cytologic parameters. Final clinical diagnosis was based on final cytology, surgical pathology, or 1-year clinical follow-up. IOA was calculated using multi-rater kappa (κ) statistics. Bivariate analyses were performed comparing cases with and without uniform agreement among the cytopathologists followed by logistic regression with backward elimination to model likelihood of uniform agreement.

RESULTS

Ninety-nine patients were included (49 % males, mean age 64 years, mean lesion size 26 mm). IOA for final diagnosis was moderate (κ = 0.45, 95 % confidence interval (CI) 0.4 - 0.49) with minimal improvement when combining suspicious and malignant diagnoses (κ = 0.54, 95 %CI 0.49 - 0.6). The weighted kappa value for overall diagnosis was 0.65 (95 %CI 0.54 - 0.76). IOA was slight to fair (κ = 0.04 - 0.32) for individual cytologic parameters. A final clinical diagnosis of malignancy was the most significant predictor of agreement [OR 3.99 (CI 1.52 - 10.49)].

CONCLUSIONS

Interobserver agreement among cytopathologists for pancreatic EUS-FNA specimens is moderate-substantial for the final cytologic diagnosis. The final clinical diagnosis of malignancy was the strongest predictor of agreement. These results have significant implications for patient management and need to be validated in future trials.

Human bile contains microRNA-laden extracellular vesicles that can be used for cholangiocarcinoma diagnosis

Cholangiocarcinoma (CCA) presents significant diagnostic challenges, resulting in late patient diagnosis and poor survival rates. Primary sclerosing cholangitis (PSC) patients pose a particularly difficult clinical dilemma because they harbor chronic biliary strictures that are difficult to distinguish from CCA. MicroRNAs (miRs) have recently emerged as a valuable class of diagnostic markers; however, thus far, neither extracellular vesicles (EVs) nor miRs within EVs have been investigated in human bile. We aimed to comprehensively characterize human biliary EVs, including their miR content. We have established the presence of extracellular vesicles in human bile. In addition, we have demonstrated that human biliary EVs contain abundant miR species, which are stable and therefore amenable to the development of disease marker panels. Furthermore, we have characterized the protein content, size, numbers, and size distribution of human biliary EVs. Utilizing multivariate organization of combinatorial alterations (MOCA), we defined a novel biliary vesicle miR-based panel for CCA diagnosis that demonstrated a sensitivity of 67% and specificity of 96%. Importantly, our control group contained 13 PSC patients, 16 with biliary obstruction of varying etiologies (including benign biliary stricture, papillary stenosis, choledocholithiasis, extrinsic compression from pancreatic cysts, and cholangitis), and 3 with bile leak syndromes. Clinically, these types of patients present with a biliary obstructive clinical picture that could be confused with CCA.

CONCLUSION

These findings establish the importance of using extracellular vesicles, rather than whole bile, for developing miR-based disease markers in bile. Finally, we report on the development of a novel bile-based CCA diagnostic panel that is stable, reproducible, and has potential clinical utility.

Enteral stenting versus gastrojejunostomy for palliation of malignant gastric outlet obstruction

BACKGROUND AND AIMS

Endoscopic placement of enteral self-expandable metallic stents is an alternative to surgical gastrojejunostomy (GJ) for palliation of malignant gastric outlet obstruction (GOO). Factors associated with clinical outcomes are not known. The aims of this study are to compare the overall complication rate and effectiveness (duration of oral intake) between endoscopic stenting (ES) and GJ in patients with GOO and identify predictors of clinical outcomes.

PATIENTS AND METHODS

This was a retrospective cohort study at a single tertiary academic center. Patients who underwent ES or GJ for treatment of GOO between 1/2001 and 12/2010 were identified using an institutional claims database. The electronic medical records for each patient were reviewed. Univariate and multivariate logistic regression analyses were performed to study the association of treatment outcomes with patient factors and cancer therapy.

RESULTS

120 patients had ES while 227 had GJ. Technical success was higher for GJ (99 vs. 96 %, p = 0.004). Complication rates were higher in the GJ group (22.10 vs. 11.66 %, p = 0.02). Reintervention was more common with ES [adjusted odds ratio (OR) 9.18, p < 0.0001]. Mean length of hospital stay (LOHS) was shorter (adjusted p = 0.005) in the ES compared with the GJ group. However, mean hospital charges, including reinterventions, were greater in the ES group (US $34,250 vs. US $27,599, p = 0.03). ES and GJ had comparable reintervention-free time in patients who had reintervention (88 vs. 106 days, respectively, p = 0.79). Chemotherapy [adjusted hazard ratio (HR) 3 > 0.57, p = 0.04] and radiation therapy (adjusted HR 0.35, p = 0.03) were associated with significantly longer duration of oral intake after ES or GJ.

CONCLUSION

ES is associated with fewer complications, shorter LOHS, but higher reintervention rates and overall charges.

Tumor size and location correlate with behavior of pancreatic serous cystic neoplasms

OBJECTIVES

The majority of pancreatic serous cystic neoplasms (SCNs) are benign. However, these neoplasms can cause symptoms and rarely can be aggressive. Identification of factors associated with symptomatic or aggressive SCNs may aid management decisions. The aim of this study was to identify variables that predict aggressive SCNs.

METHODS

Prospective pathology database was queried for SCNs that were surgically resected at Johns Hopkins Hospital. Tumors were considered aggressive if they invaded surrounding structures and/or vessels or if they metastasized to lymph nodes or distant organs. The associations of gender, tumor size, and tumor location, with the presence or absence of symptoms and tumor behavior were examined using Fisher's exact test, logistic regression, and multivariate analyses.

RESULTS

A total of 257 patients with SCNs underwent surgical resection. Mean tumor diameter was 4.9 cm. Tumor location in the head of pancreas (HOP) was associated with symptoms (odds ratio (OR) 1.87, 95% confidence interval (CI) 1.1-3.3). Computed tomography (CT) predicted the diagnosis of SCN in approximately a quarter of patients. Thirteen tumors (mean 10.5 cm) were considered aggressive. Multivariate analysis showed that tumor diameter (OR 1.53, 95% CI 1.24-1.89) and location of tumor in pancreatic head (OR 10.44, 95% CI 1.73-63.04) were independently associated with aggressive behavior.

CONCLUSIONS

We describe the largest case series of patients with pathologically proven SCNs. CT performed poorly in preoperative diagnosis of SCNs. Large tumor size and head location predicted aggressive behavior. These factors should be considered in the management of patients with SCN.

EUS is still superior to multidetector computerized tomography for detection of pancreatic neuroendocrine tumors

BACKGROUND

The role of EUS for detection of pancreatic neuroendocrine tumors (PNETs) is not clearly defined in institutions that use multidetector CT for pancreatic imaging.

OBJECTIVE

The aims of this study were to (1) compare the detection rates of EUS and CT by type and size of PNET and calculate the incremental benefit of EUS over CT, (2) evaluate the CT detection rate for PNETs adjusted for improved CT technology over time, and (3) determine the factors associated with CT-negative PNETs.

DESIGN

Retrospective single-center cohort study.

SETTING

Johns Hopkins Hospital.

PATIENTS

Patients with pathologically proven PNETs with preoperative CT. Incidentally found PNETs (resection specimens) and those without Johns Hopkins Hospital CT imaging were excluded.

MAIN OUTCOME MEASUREMENT

Detection rates of CT and EUS were compared by using pathology as the reference standard.

RESULTS

In 217 patients (with 231 PNETs) studied, CT detected 84% of tumors (54.3% of insulinomas). The sensitivity of CT for the detection of PNETs significantly increased with improvement in CT technology (P = .02; χ(2) for trend). CT was more likely to miss lesions <2 cm (P = .005) and insulinomas (P < .0001). In 56 patients who had both CT and EUS, the sensitivity of EUS was greater than CT (91.7% vs 63.3%; P = .0002), particularly for insulinomas (84.2% vs 31.6%; P = .001). EUS detected 20 of 22 CT-negative tumors (91%).

LIMITATIONS

Retrospective nonrandomized design and referral bias.

CONCLUSIONS

The detection rate of CT has significantly improved over time. CT-negative tumors are small and more likely to be insulinomas. A sequential approach of CT followed by EUS can detect most PNETs. EUS is a more sensitive initial test for the detection of suspected insulinomas.