A multi-stage machine learning model for diagnosis of esophageal manometry.

High-resolution manometry (HRM) is the primary procedure used to diagnose esophageal motility disorders. Its manual interpretation and classification, including evaluation of swallow-level outcomes and then derivation of a study-level diagnosis based on Chicago Classification (CC), may be limited by inter-rater variability and inaccuracy of an individual interpreter. We hypothesized that an automatic diagnosis platform using machine learning and artificial intelligence approaches could be developed to accurately identify esophageal motility diagnoses. Further, a multi-stage modeling framework, akin to the step-wise approach of the CC, was utilized to leverage advantages of a combination of machine learning approaches including deep-learning models and feature-based models. Models were trained and tested using a dataset comprised of 1741 patients' HRM studies with CC diagnoses assigned by expert physician raters. In the swallow-level stage, three models based on convolutional neural networks (CNNs) were developed to predict swallow type and swallow pressurization (test accuracies of 0.88 and 0.93, respectively), and integrated relaxation pressure (IRP)(regression model with test error of 4.49 mmHg). At the study-level stage, model selection from families of the expert-knowledge-based rule models, xgboost models and artificial neural network(ANN) models were conducted. A simple model-agnostic strategy of model balancing motivated by Bayesian principles was utilized, which gave rise to model averaging weighted by precision scores. The averaged (blended) models and individual models were compared and evaluated, of which the best performance on test dataset is 0.81 in top-1 prediction, 0.92 in top-2 predictions. This is the first artificial-intelligence style model to automatically predict esophageal motility (CC) diagnoses from HRM studies using raw multi-swallow data and it achieved high accuracy. Thus, this proposed modeling framework could be broadly applied to assist with HRM interpretation in a clinical setting.

Validation of Clinically Relevant Thresholds of Esophagogastric Junction Obstruction Using FLIP Panometry.

BACKGROUND & AIMS: This study aimed to assess the accuracy of functional luminal imaging probe (FLIP) panometry to detect esophagogastric junction (EGJ) obstruction assigned by high-resolution manometry (HRM) and the Chicago Classification version 4.0 (CCv4.0). METHODS: Six hundred eighty-seven adult patients who completed FLIP and HRM for primary esophageal motility evaluation and 35 asymptomatic volunteers (controls) were included. EGJ opening was evaluated with 16-cm FLIP during sedated endoscopy via EGJ-distensibility index (DI) and maximum EGJ diameter. HRM was classified according to CCv4.0 and focused on studies with a conclusive disorder of EGJ outflow (ie, achalasia subtypes I, II, or III; or EGJ outflow obstruction with abnormal timed barium esophagogram) or normal EGJ outflow. RESULTS: All 35 controls had EGJ-DI >3.0 mm2/mmHg and maximum EGJ diameter >16 mm. Per HRM and CCv4.0, 245 patients had a conclusive disorder of EGJ outflow, and 314 patients had normal EGJ outflow. Among the 241 patients with reduced EGJ opening (EGJ-DI <2.0 mm2/mmHg and maximum EGJ diameter <12 mm) on FLIP panometry, 86% had a conclusive disorder of EGJ outflow per CCv4.0. Among the 203 patients with normal EGJ opening (EGJ-DI ≥2.0 mm2/mmHg and maximum EGJ diameter ≥16 mm) on FLIP panometry, 99% had normal EGJ outflow per CCv4.0. CONCLUSIONS: FLIP panometry accurately identified clinically relevant conclusive EGJ obstruction as defined by CCv4.0 in patients evaluated for esophageal motor disorders. Thus, FLIP panometry is a valuable tool for both independent and complementary evaluation of esophageal motility.

Validation of secondary peristalsis classification using FLIP panometry in 741 subjects undergoing manometry.

BACKGROUND AND AIMS: This study aimed to systematically evaluate a classification scheme of secondary peristalsis using functional luminal imaging probe (FLIP) panometry through comparison with primary peristalsis on high-resolution manometry (HRM). METHODS: 706 adult patients that completed FLIP and HRM for primary esophageal motility evaluation and 35 asymptomatic volunteers ("controls") were included. Secondary peristalsis, that is, contractile responses (CRs), was classified on FLIP panometry by the presence and pattern of contractility as normal (NCR), borderline (BCR), impaired/disordered (IDCR), absent (ACR), or spastic-reactive (SRCR). Primary peristalsis on HRM was assessed according to the Chicago Classification. RESULTS: All 35 of the controls had antegrade contractions on FLIP panometry with either NCR (89%) or BCR (11%). The average percentages of normal swallows on HRM varied across contractile response patterns from 84% in NCR, 68% in BCR, 39% in IDCR, to 11% in ACR, as did the percentage of failed swallows on HRM: 4% in NCR, 12% in BCR, 36% in IDCR, and 79% in ACR. SRCR on FLIP panometry was observed in 18/57 (32%) patients with type III achalasia, 4/15 (27%) with distal esophageal spasm, and 7/15 (47%) with hypercontractile esophagus on HRM. CONCLUSIONS: The FLIP panometry contractile response patterns reflect a pathophysiologic transition from normal to abnormal esophageal peristaltic function with shared features with primary peristaltic function/dysfunction on HRM. Thus, these patterns of the contractile response to distension can facilitate the evaluation of esophageal motility using FLIP panometry.

Classifying Esophageal Motility by FLIP Panometry: A Study of 722 Subjects With Manometry.

INTRODUCTION: Functional luminal imaging probe (FLIP) panometry can evaluate esophageal motility in response to sustained esophageal distension at the time of sedated endoscopy. This study aimed to describe a classification of esophageal motility using FLIP panometry and evaluate it against high-resolution manometry (HRM) and Chicago Classification v4.0 (CCv4.0). METHODS: Five hundred thirty-nine adult patients who completed FLIP and HRM with a conclusive CCv4.0 diagnosis were included in the primary analysis. Thirty-five asymptomatic volunteers ("controls") and 148 patients with an inconclusive CCv4.0 diagnosis or systemic sclerosis were also described. Esophagogastric junction (EGJ) opening and the contractile response (CR) to distension (i.e., secondary peristalsis) were evaluated with a 16-cm FLIP during sedated endoscopy and analyzed using a customized software program. HRM was classified according to CCv4.0. RESULTS: In the primary analysis, 156 patients (29%) had normal motility on FLIP panometry, defined by normal EGJ opening and a normal or borderline CR; 95% of these patients had normal motility or ineffective esophageal motility on HRM. Two hundred two patients (37%) had obstruction with weak CR, defined as reduced EGJ opening and absent CR or impaired/disordered CR, on FLIP panometry; 92% of these patients had a disorder of EGJ outflow per CCv4.0. DISCUSSION: Classifying esophageal motility in response to sustained distension with FLIP panometry parallels the swallow-associated motility evaluation provided with HRM and CCv4.0. Thus, FLIP panometry serves as a well-tolerated method that can complement, or in some cases be an alternative to HRM, for evaluating esophageal motility disorders.

Prediction of Esophageal Retention: A Study Comparing High-Resolution Manometry and Functional Luminal Imaging Probe Panometry.

INTRODUCTION: High-resolution manometry (HRM) is generally considered the primary method to evaluate esophageal motility; functional luminal imaging probe (FLIP) panometry represents a novel method to do so and is completed during sedated endoscopy. This study aimed to compare HRM and FLIP panometry in predicting esophageal retention on timed barium esophagram (TBE). METHODS: A total of 329 adult patients who completed FLIP, HRM, and TBE for primary esophageal motility evaluation were included. An abnormal TBE was defined by a 1-minute column height >5 cm or impaction of a 12.5-mm barium tablet. The integrated relaxation pressure (IRP) on HRM was assessed in the supine and upright patient positions. Esophagogastric junction (EGJ) opening was evaluated with 16-cm FLIP performed during sedated endoscopy through EGJ-distensibility index and maximum EGJ diameter. RESULTS: Receiver operating characteristic curves to identify an abnormal TBE demonstrated AUC (95% confidence interval) of 0.79 (0.75-0.84) for supine IRP, 0.79 (0.76-0.86) for upright IRP, 0.84 (0.79-0.88) for EGJ-distensibility index, and 0.88 (0.85-0.92) for maximum EGJ diameter. Logistic regression to predict abnormal TBE showed odds ratios (95% confidence interval) of 1.8 (0.84-3.7) for consistent IRP elevation and 39.7 (16.4-96.2) for reduced EGJ opening on FLIP panometry. Of 40 patients with HRM-FLIP panometry discordance, HRM-IRP was consistent with TBE in 23% while FLIP panometry was consistent with TBE in 78%. DISCUSSION: FLIP panometry provided superior detection of esophageal retention over IRP on HRM. However, application of a complementary evaluation involving FLIP panometry, HRM, and TBE may be necessary to accurately diagnose esophageal motility disorders.

Evaluating esophageal motility beyond primary peristalsis: Assessing esophagogastric junction opening mechanics and secondary peristalsis in patients with normal manometry.

BACKGROUND: Functional luminal imaging probe (FLIP) Panometry assesses the esophageal response to distention and may complement the assessment of primary peristalsis on high-resolution manometry (HRM). We aimed to investigate whether FLIP Panometry provides complementary information in patients with normal esophageal motility on HRM. METHODS: Adult patients that completed FLIP and had an HRM classification of normal motility were retrospectively identified for inclusion. 16-cm FLIP studies performed during endoscopy were evaluated to assess EGJ distensibility, secondary peristalsis, and identify an abnormal response to distention involving sustained LES contraction (sLESC). Clinical characteristics and esophagram were assessed when available. KEY RESULTS: Of 164 patients included (mean(SD) age 48(16) years, 75% female), 111 (68%) had normal Panometry with EGJ-distensibility index (DI) ≥2.0 mm2 /mmHg, maximum EGJ diameter ≥16mm and antegrade contractions. Abnormal EGJ distensibility was observed in 44/164 (27%), and 38/164 (23%) had an abnormal contractile response to distension. sLESC was observed in 11/164 (7%). Among 68 patients that completed esophagram, abnormal EGJ distensibility was more frequently observed with an abnormal esophagram than normal EGJ opening: 14/23 (61%) vs 10/45 (22%); P=0.003. Epiphrenic diverticula were present in 3/164 patients: 2/3 had sLESC. CONCLUSIONS & INFERENCES: Symptomatic patients with normal esophageal motility on HRM predominantly have normal FLIP Panometry; however, abnormal FLIP findings can be observed. While abnormal Panometry findings appear clinically relevant via an association with abnormal bolus retention, complementary tests, such as provocative maneuvers with HRM and timed barium esophagram, are useful to determine clinical context.

A deep-learning-based unsupervised model on esophageal manometry using variational autoencoder.

High-resolution manometry (HRM) is the primary method for diagnosing esophageal motility disorders and its interpretation and classification are based on variables (features) from data of each swallow. Modeling and learning the semantics directly from raw swallow data could not only help automate the feature extraction, but also alleviate the bias from pre-defined features. With more than 32-thousand raw swallow data, a generative model using the approach of variational auto-encoder (VAE) was developed, which, to our knowledge, is the first deep-learning-based unsupervised model on raw esophageal manometry data. The VAE model was reformulated to include different types of loss motivated by domain knowledge and tuned with different hyper-parameters. Training of the VAE model was found sensitive on the learning rate and hence the evidence lower bound objective (ELBO) was further scaled by the data dimension. Case studies showed that the dimensionality of latent space have a big impact on the learned semantics. In particular, cases with 4-dimensional latent variables were found to encode various physiologically meaningful contraction patterns, including strength, propagation pattern as well as sphincter relaxation. Cases with so-called hybrid L2 loss seemed to better capture the coherence of contraction/relaxation transition. Discriminating capability was further evaluated using simple linear discriminative analysis (LDA) on predicting swallow type and swallow pressurization, which yields clustering patterns consistent with clinical impression. The current work on modeling and understanding swallow-level data will guide the development of study-level models for automatic diagnosis as the next stage.

Multiple rapid swallows and rapid drink challenge in patients with esophagogastric junction outflow obstruction on high-resolution manometry.

BACKGROUND: Esophagogastric junction outflow obstruction (EGJOO) as defined by Chicago Classification of esophageal motility disorders (CCv3.0) encompasses a broad range of diagnoses, thus posing clinical challenges. Our aims were to evaluate multiple rapid swallow (MRS) and rapid drink challenge (RDC) during high-resolution manometry (HRM) to aid identifying clinically relevant EGJOO. METHODS: Patients with a HRM diagnosis of EGJOO based on CCv3.0 that also completed MRS and RDC during HRM and barium esophagram were retrospectively identified. Radiographic EGJOO (RAD-EGJOO) was defined by either liquid barium retention or delayed passage of a barium tablet on barium esophagram. Thirty healthy asymptomatic controls that completed HRM were also included. MRS involved drinking 2 mL for 5 successive swallows. RDC involved rapid drinking of 200 mL liquid. Integrated relaxation pressure (IRP) and presence of panesophageal pressurization (PEP) during MRS and RDC were assessed. KEY RESULTS: One hundred one patients, mean (SD) age 56 (16) years, were included; 32% had RAD-EGJOO, 68% did not. RAD-EGJOO patients more frequently had elevated (>12 mmHg) upright IRP (100%), MRS-IRP (56%), RDC-IRP (53%), and PEP during RDC (66%) than both controls [17%; 0%; 7%; 3%] and patients without RAD-EGJOO [83%; 35%; 39%; 41%] Having IRP >12 mmHg during both MRS and RDC was twice as likely to be associated with RAD-EGJOO (19%) than those without RAD-EGJOO (9%) among patients with upright IRP >12 mmHg. CONCLUSIONS AND INFERENCES: Adjunctive HRM maneuvers MRS and RDC appear to help identify clinically significant EGJOO. While future outcome studies are needed, comprehensive multimodal evaluation helps clarify relevance of EGJOO on HRM.

Esophagogastric Junction Opening Parameters Are Consistently Abnormal in Untreated Achalasia.

Achalasia is a disorder of impaired lower esophageal sphincter (LES) relaxation and failed peristalsis traditionally characterized by manometry.1 As impaired LES relaxation is a mechanism of reduced esophagogastric junction (EGJ) opening, abnormally reduced EGJ distensibility assessed with functional luminal imaging probe (FLIP) was reported among patients with untreated achalasia.2-5 Therefore, we aimed to describe the performance characteristics of EGJ opening parameters on FLIP panometry among a large cohort of treatment-naïve achalasia patients.

Normal Functional Luminal Imaging Probe Panometry Findings Associate With Lack of Major Esophageal Motility Disorder on High-Resolution Manometry.

BACKGROUND & AIMS: A normal esophageal response to distension on functional luminal imaging probe (FLIP) panometry during endoscopy might indicate normal esophageal motor function. We aimed to investigate the correlation of normal FLIP panometry findings with esophageal high-resolution manometry (HRM) and outcomes of discrepant patients. METHODS: We performed a retrospective study using data from a registry of patients who completed FLIP during sedated endoscopy. We identified 111 patients with normal FLIP panometry findings (mean age, 42 y; 69% female) and corresponding HRM data. A normal FLIP panometry was defined as an esophagogastric junction (EGJ) distensibility index greater than 3.0 mm2/mm Hg, an absence of repetitive retrograde contractions, and a repetitive antegrade contraction pattern that met the Rule-of-6s: ≥6 consecutive antegrade contractions of ≥6-cm in length, at a rate of 6 ± 3 contractions per minute. HRM findings were classified by the Chicago classification system version 3.0. RESULTS: HRM results were classified as normal motility in 78 patients (70%), ineffective esophageal motility in 10 patients (9%), EGJ outflow obstruction in 20 patients (18%), and 3 patients (3%) as other. In patients with EGJ outflow obstruction based on HRM, the integrated relaxation pressure normalized on adjunctive swallows in 16 of 20 patients (80%), and in 8 of 9 patients (88%) who completed a barium esophagram and had normal barium clearance. Thus, although 23 of 111 patients (21%) with normal FLIP panometry had abnormal HRM findings, these HRMs often were considered to be false-positive or equivocal results. All patients with an abnormal result from HRM were treated conservatively. CONCLUSIONS: In a retrospective cohort study, we found that patients with normal FLIP panometry results did not have a clinical impression of a major esophageal motor disorder. Normal FLIP panometry results can exclude esophageal motility disorders at the time of endoscopy, possibly negating the need for HRM in select patients.

Achalasia subtypes can be identified with functional luminal imaging probe (FLIP) panometry using a supervised machine learning process.

BACKGROUND: Achalasia subtypes on high-resolution manometry (HRM) prognosticate treatment response and help direct management plan. We aimed to utilize parameters of distension-induced contractility and pressurization on functional luminal imaging probe (FLIP) panometry and machine learning to predict HRM achalasia subtypes. METHODS: One hundred eighty adult patients with treatment-naïve achalasia defined by HRM per Chicago Classification (40 type I, 99 type II, 41 type III achalasia) who underwent FLIP panometry were included: 140 patients were used as the training cohort and 40 patients as the test cohort. FLIP panometry studies performed with 16-cm FLIP assemblies were retrospectively analyzed to assess distensive pressure and distension-induced esophageal contractility. Correlation analysis, single tree, and random forest were adopted to develop classification trees to identify achalasia subtypes. KEY RESULTS: Intra-balloon pressure at 60 mL fill volume, and proportions of patients with absent contractile response, repetitive retrograde contractile pattern, occluding contractions, sustained occluding contractions (SOC), contraction-associated pressure changes >10 mm Hg all differed between HRM achalasia subtypes and were used to build the decision tree-based classification model. The model identified spastic (type III) vs non-spastic (types I and II) achalasia with 90% and 78% accuracy in the train and test cohorts, respectively. Achalasia subtypes I, II, and III were identified with 71% and 55% accuracy in the train and test cohorts, respectively. CONCLUSIONS AND INFERENCES: Using a supervised machine learning process, a preliminary model was developed that distinguished type III achalasia from non-spastic achalasia with FLIP panometry. Further refinement of the measurements and more experience (data) may improve its ability for clinically relevant application.

Repetitive Antegrade Contractions: A novel response to sustained esophageal distension is modulated by cholinergic influence.

BACKGROUND & AIMS: A unique motor response to sustained esophageal distension, repetitive antegrade contractions (RACs), is observed using functional luminal imaging probe (FLIP) panometry. However, physiologic mechanisms related to this response are unexplored. This study aimed to evaluate the impact of cholinergic inhibition with atropine on the esophageal contractile response to sustained distention, including RACs, among healthy volunteers. METHODS: 8 asymptomatic volunteers (ages 22-45) were evaluated in a crossover study design with 16-cm FLIP positioned across the esophagogastric junction and distal esophagus during sedated upper endoscopy. The FLIP study involving stepwise volumetric distension was performed twice in each subject, at baseline and again after atropine (15 mcg/kg) was administered intravenously. FLIP panometry was analyzed to assess the contractile response to distension. RESULTS: Antegrade contractions, lumen-occluding contractions, and a RAC pattern were observed in 8/8, 8/8, and 7/8(88%) subjects, respectively, at baseline and in 5/8 (63%), 2/8 (25%) and 2/8 (25%) subjects after atropine. The rate of contractions in the RAC pattern was similar (6-7 contractions per minute) before and after atropine. Compared with the baseline study, distension-induced contractility was triggered at higher fill volumes after atropine. FLIP pressures were lower in response to volumetric filling after atropine than at baseline. CONCLUSIONS: The vigor and triggering of the esophageal contractile response to distension is reduced by cholinergic inhibition in asymptomatic controls. The observation that the rate of contractions did not change when patients developed repetitive contractile responses suggests that this rate is not modified by cholinergic inhibition once contractility is triggered.

EsoFLIP for esophageal dilation: proposed advantages.

PURPOSE OF REVIEW: The EsoFLIP integrates impedance planimetry technology into a dilator balloon capable of dilating from diameters between 10 and 30 mm via controlled volumetric distension while providing real-time visualization and objective measurement during the dilation procedure, potentially negating the need for fluoroscopy. This review aims to describe the use and application of EsoFLIP and the potential advantages this novel technology may afford. RECENT FINDINGS: Small pilot and retrospective studies demonstrate EsoFLIP feasibility and safety, but larger studies are needed to understand its impact on clinical outcomes. SUMMARY: EsoFLIP is an appealing new technology that offers advantages in therapeutic dilation of the lower esophageal sphincter in achalasia or of esophageal strictures because of previous surgery, gastroesophageal reflux disease, or radiation.

Endoscopic Dilation of Bariatric RNY Anastomotic Strictures: a Systematic Review and Meta-analysis.

Gastrojejunostomy anastomotic strictures are a complication of Roux-en-Y gastric bypass surgery without an established treatment guideline. A systematic review and meta-analysis were performed to determine the safety and efficacy of endoscopic dilation in their management. PubMed, Web of Science, and Cochrane Central (1994-2017) were searched. Data was analyzed with random effects meta-analysis and mixed effects meta-regression. Twenty-one observational studies (896 patients) were included. The stricture rate for laparoscopic patients was 6% (95% CI, 5-9%). Only 38% (95% CI, 30-47%) required greater than one dilation. Symptom improvement occurred in 97% (95% CI, 94-98%). The complication rate was 4% (95% CI, 3-6%). Endoscopic dilation of GJA strictures is safe, effective, and sustaining. This study can guide endoscopists in the treatment of a common bariatric surgical complication.

Severe Thrombotic Complication of Eltrombopag in a Cirrhotic Patient.

We present a patient with hepatitis C virus (HCV) and cirrhosis who was treated with eltrombopag for idiopathic thrombocytopenic purpura and was incidentally found to have a right atrial thrombus with extension into the left internal jugular vein. Eltrombopag was discontinued and the patient was treated with thrombectomy and anticoagulation. Given the proposed use of eltrombopag in HCV-associated thrombocytopenia, we advise caution when treating cirrhotics who are at higher intrinsic risk of thrombosis.

Benefit of Early Palliative Care Intervention in End-Stage Liver Disease Patients Awaiting Liver Transplantation.

BACKGROUND: Patients with end-stage liver disease have a predictable and progressive decline in their quality of life because of physical symptoms and psychological distress. Early palliative care intervention (EPCI) correlates with better symptom control and mood. We aimed to improve symptomatology and mood in liver transplant candidates by implementing a longitudinal multidisciplinary EPCI. MEASURES: Depression level and symptom burden were assessed with Center for Epidemiological Studies Depression Scale and a modified liver-specific Edmonton Symptom Assessment System scale. INTERVENTION: All patients referred for liver transplant evaluation between July 2013 and May 2014 were scheduled for EPCI. OUTCOMES: After EPCI, 50% of moderate-to-severe symptoms improved (P < 0.05), and 43% of patients showed improvement in clinically significant depressive symptoms (P = 0.003). Notably, patients with more symptoms showed a greater improvement in Center for Epidemiological Studies Depression Scale scores (P = 0.001). CONCLUSIONS/LESSONS LEARNED: Implementation of EPCI improved symptom burden and mood in end-stage liver disease patients awaiting transplant.