Risk prediction of progression of reflux esophagitis based on contrast-enhanced ultrasound technology and machine learning

Risk prediction of progression of reflux esophagitis based on contrast-enhanced ultrasound technology and machine learning | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Risk prediction of progression of reflux esophagitis based on contrast-enhanced ultrasound technology and machine learning Ying Tu, Jing-Yu Wang, Wen-Ying Wang, Zhen-Yu Song, Chun-Yan Xie, and 2 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8494694/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 12 You are reading this latest preprint version Abstract BACKGROUND The incidence of reflux esophagitis (RE) is increasing worldwide. RE can cause heartburn, acid reflux, and swallowing discomfort, thereby diminishing quality of life, with potential progression to Barrett's esophagus or esophageal adenocarcinoma. Gastroscopy for the diagnosis and grading of RE can be uncomfortable, and other modalities lack predictive accuracy. Thus, highly non-invasive diagnostic methods for RE are needed. AIM To explore the value of contrast-enhanced ultrasound (CEUS) examination for predicting the progression of RE. METHODS We retrospectively selected 296 patients with RE who were hospitalized at West China Longquanyi Hospital of Sichuan University from February 2019 to January 2023. Basic clinical characteristics such as age and gender were collected. CEUS was used to measure the diameter of the esophageal hiatus, frequency of reflux within 5 min, duration of reflux within 5 min, and length of the abdominal esophagus. According to the Los Angeles (LA) classification, the patients were divided into three groups: 92 patients with “LA-A” grade (Group A), 85 with “LA-B” grade (Group B), and 20 with “LA-C/D” grade (Group C). Group differences in clinical characteristics were analyzed. Key factors influencing the progression of RE were analyzed using ordered logistic regression, random forest, and support vector machine (SVM). Statistical methods included paired the t -test, Mann–Whitney U test, Chi-square test, Kruskal–Wallis H test, and one-way analysis of variance. RESULTS Comparisons among the three groups showed statistically significant differences in the diameter of the esophageal hiatus and frequency and duration of reflux within 5 min. Pairwise comparisons revealed that the frequency of reflux within 5 min in group C was significantly higher than in groups A and B. Ordered logistic regression, random forest, and SVM analyses all indicated that the frequency and duration of reflux within 5 min were key factors influencing the progression of RE. The SVM analysis showed that the diameter of the esophageal hiatus was a key factor influencing the progression of RE. CONCLUSION CEUS can better predict RE progression. The number of reflux events within 5 min, duration of reflux within 5 min, and diameter of the esophageal hiatus are key factors influencing RE progression. Reflux esophagitis Contrast-enhanced ultrasound Machine learning Figures Figure 1 Figure 2 Figure 3 Core Tip Reflux esophagitis (RE) poses a growing global health burden, with potential progression to severe conditions like Barrett's esophagus and esophageal adenocarcinoma. Conventional diagnostic methods such as gastroscopy are invasive, while other non-invasive approaches lack sufficient predictive accuracy. This study used contrast-enhanced ultrasound (CEUS) and 3 models (ordered logistic regression, random forest, SVM) to explore RE progression predictors. For 197 LA-classified RE patients, it found 5-min reflux frequency/duration (all models) and esophageal hiatus diameter (SVM) key, supporting CEUS in non-invasive RE prediction. INTRODUCTION Reflux esophagitis (RE) is the most common subtype of gastroesophageal reflux disease, and its incidence is increasing year by year worldwide. It is characterized by obvious gastroesophageal reflux, and patients often suffer from symptoms such as heartburn, acid reflux, and swallowing discomfort. Long-term and repeated attacks not only seriously affect quality of life but may also progress to Barrett's esophagus or even esophageal adenocarcinoma, posing significant challenges for clinical diagnosis and treatment[ 1 ]. Currently, gastroscopy remains the "gold standard" for the diagnosis and grading of RE, but its invasive operation can easily cause discomfort to patients, while traditional non-invasive examinations such as esophageal pH monitoring and symptom scoring are often insufficient in predictive accuracy due to individual differences and operational norms[ 2 ]. Therefore, exploring non-invasive and highly accurate prediction methods for RE has become an important direction in clinical research. The development of contrast-enhanced ultrasound (CEUS) technology provides a new solution to meet this need[ 3 ]. As a non-invasive imaging technique, CEUS can capture the blood flow perfusion characteristics, microcirculation status, and structural changes of esophageal mucosa and surrounding tissues by oral or intravenously injected contrast agents. These data not only reflect the hemodynamic abnormalities in the lesion area but also indirectly indicate the degree of mucosal inflammation and the extent of damage. CEUS has the advantages of convenient operation, high repeatability, and no radiation risk, providing quantitative data for the non-invasive assessment of RE[ 4 ]. In recent years, machine learning algorithms have made significant progress in the medical field, and their advantages in processing high-dimensional data and mining hidden patterns have shown significant potential in the identification of risk factors[ 5 ]. Among them, logistic regression, as a classic statistical learning method, with its strong model interpretability and easy quantification of results, has long been the benchmark tool for identifying linearly associated risk factors[ 6 ]. Support vector machines (SVMs), through kernel function mapping, transform nonlinear problems into linear classification problems in high-dimensional space, exhibit excellent generalization ability in small samples and high-dimensional data, and are suitable for capturing complex interactions between variables[ 7 ]. Combining multiple decision trees for prediction, random forests, representative of ensemble learning, can effectively reduce overfitting risks and intuitively quantify the contribution of different factors through variable importance scoring, and are especially suitable for handling datasets with multicollinearity or nonlinear relationships[ 8 ]. Based on the above, this study intends to use CEUS data as a basis, combined with a random forest algorithm, SVM algorithm, and ordered logistic regression, to identify factors related to the occurrence and progression of RE. The aim is to achieve non-invasive and efficient prediction of RE through the mining of contrast imaging data, thereby providing objective auxiliary diagnostic evidence for clinicians, reducing unnecessary invasive examinations, improving the efficiency of early intervention in the disease, and ultimately providing a new technical path for the precise diagnosis and treatment of RE. MATERIALS AND METHODS Study population This study was approved by the Ethics Committee of the First People's Hospital of Longquanyi District, Chengdu City, Sichuan Province (Approval Number: af - key − 20101) and has been registered in the Chinese Clinical Trial Registry (No: ChiCTR1900021934, registration date: March 16, 2019). In total, 296 patients with RE hospitalized at West China Longquanyi Hospital of Sichuan University from February 2019 to January 2023 were retrospectively selected. All patients underwent upper gastrointestinal endoscopy, and the diagnostic criteria for RE were based on the Los Angeles (LA) classification[ 9 ]. The inclusion criteria were: (1) age ranging from 18 to 80 years; and (2) clear diagnosis of RE by endoscopy. The exclusion criteria were: (1) esophagitis caused by other etiologies; (2) suspected or confirmed esophageal squamous cell carcinoma or other malignant tumors at the time of admission; (3) severe diseases in other systems such as the heart, lung, and kidney; and (4) incomplete data or loss to follow-up. Patients were divided into three groups by experienced endoscopy physicians: “LA-A” (Group A), “LA-B (Group B), and LA-C/D” (Group C) according to the LA classification. Data collection All patients underwent CEUS, and the diameter of the esophageal hiatus, frequency of reflux within 5 min, duration of reflux within 5 min, length of the abdominal esophagus, etc., were measured. Basic clinical characteristics of the patients, such as age and gender, were also collected. The frequency of reflux within 5 min was recorded as 0 for zero times, 1 for one time, 2 for twice, 3 for three times, and 4 for four or more for four times. The duration of reflux within 5 min was recorded as 0 for 0 seconds, 1 for 1 second, 2 for 2 seconds, 3 for 3 seconds, and 4 for 4 or more for 4 seconds or more. Statistical analysis Statistical analysis was performed using SPSS 26.0 software (IBM Corp., Armonk, NY, USA). Measurement data that followed a normal distribution were expressed as (mean ± standard deviation). Comparisons between two groups were conducted using the paired t-test, while comparisons among multiple groups were carried out using one-way analysis of variance. Measurement data that deviated from a normal distribution were represented as median (quartile 1, quartile 3). Comparisons among multiple groups were conducted using the Kruskal–Wallis H test, and comparisons between two groups were conducted using the Mann–Whitney U test. Count data were expressed as the number of cases and percentages, and comparisons between groups were conducted using the chi-square test. We used multivariate ordered logistic regression to analyze the factors related to the grading of RE patients. A P value < 0.05 was considered statistically significant. Relevant variables were selected based on the random forest algorithm through the "VSURF", "Boruta", and "varSelRF" data packages in R version 4.2.2 (R Foundation for Statistical Computing, Vienna, Austria)。This algorithm integrates the prediction results of multiple decision trees to reduce the overfitting risk of a single model and enhance the stability of variable selection. The core parameters of the model are optimized through 5-fold cross-validation. The specific settings are as follows: (1) Number of decision trees (n_estimators) = 1000; (2) Maximum tree depth (max_depth) = 20; (3) Minimum sample number for node splitting (min_samples_split) = 5; (4) Minimum sample number for leaf nodes (min_samples_leaf) = 2; (5) Maximum number of features for each split (max_features) = "sqrt"; (6) Random seed (random_state) = 42 (to ensure reproducibility of results). The variable selection combines three functions: "VSURF" for stepwise selection, "Boruta" for verifying the validity of shadow features, and "varSelRF" for optimizing based on importance scores and cross-validation errors. Finally, the core variables are determined. Relevant variables were selected based on the SVM algorithm through the "caret", "e1071", and "doParallel" data packages in R version 4.2.2 (R Foundation for Statistical Computing, Vienna, Austria). This algorithm achieves effective classification and feature screening of high-dimensional data by seeking the optimal separation hyperplane. Among them, the "doParallel" package is used to enable parallel computing and improve the efficiency of model training; The "caret" package is used for hyperparameter optimization and cross-validation of the model. The "e1071" package is used for building the core model. The core parameters of the model are set as follows: (1) Kernel Function type = Radial Basis Function (RBF), this kernel function has good nonlinear mapping ability and can adapt to the complex relationship between features and outcome variables; (2) The penalty parameter = 1.0 is used to balance the training error and generalization error of the model. The larger the C value, the higher the degree of fit of the model to the training set, and the more likely overfitting occurs. Cross-validation determines that the model performance is optimal when C = 1.0. (3) Kernel function parameter = 0.1, controlling the width of the radial basis kernel function. The larger the gamma value, the stronger the local fitting ability of the model. It is determined by grid search that the classification accuracy of the model is the highest when gamma = 0.1. (4) Cross-validation strategy = 5-fold cross-validation. The dataset is randomly divided into 5 equal parts, with 4 parts used as the training set and 1 part as the validation set in sequence. After repeating this 5 times, the average prediction error is calculated to evaluate the generalization ability of the model. (5) Data preprocessing = Standardizing all input features (z-score standardization, that is, converting to a standard normal distribution with a mean of 0 and a standard deviation of 1) to eliminate the influence of dimensional differences of different features on the model. Variable screening is achieved through the Recursive Feature Elimination (RFE) method in the "caret" package, gradually eliminating the features that contribute less to the model prediction, and finally retaining the feature subset that minimizes the cross-validation error of the model. Ethical considerations This study received ethical approval from our hospital's Medical Ethics Committee (AF-KY- 201901) and was registered in the Chinese Clinical Trial Registry (registration number: ChiCTR1900021934). The study was conducted in accordance with the Helsinki Declaration and Good Clinical Practice guidelines. Before the study started, we obtained informed consent from each participant. If informed consent was not obtained, participation in the study was not allowed. RESULTS Clinical characteristics of the study population A total of 296 patients were initially included in this study. Among them, 13 were excluded due to being over 80 years old, 15 were excluded because they were subsequently diagnosed with esophageal cancer and other malignant tumors, 11 were excluded due to having major diseases of the heart, lungs, kidneys and other organs, and 60 were excluded due to missing data. A total of 197 patients with RE who underwent CEUS were finally included in this study, including 92 patients of LA-A grade (Group A), 85 of LA-B grade (Group B), and 20 of LA-C/D grade (Group C). The median age of the patients in Group A was 54.5 years, with 54.35% being male. The median age of the patients in Group B was 56 years, with 64.71% being male. The median age of the patients in Group C was 60.5 years, with 65% being male. The reflux frequency within 5 min, reflux time within 5 min (s), and diameter of the esophageal hiatus (mm) were statistically significantly different among multiple groups. Further pairwise comparisons showed that the reflux frequency within 5 min in Groups A and B was significantly lower than that in Group C ( P < 0.05) (Table 1 ). Table 1 Analysis of baseline clinical characteristics Group A ( n = 92) Group B ( n = 85) Group C ( n = 20) P value P P a P b P C Age (years) 54.50 (44.00, 65.00) 56.00 (40.50, 67.00) 60.50 (45.25, 68.00) 0.485 0.916 0.234 0.278 Male ( n ) 50 (54.35%) 55 (64.71%) 13 (65.00%) 0.330 0.161 0.384 0.980 Reflux frequency within 5 minutes 0.25 ± 0.74 0.19 ± 0.63 3.30 ± 1.03 < 0.001 0.298 0.006 < 0.001 Reflux time within 5 minutes, (s) 0.29 ± 0.83 0.27 ± 0.93 3.55 ± 1.23 < 0.001 0.853 0.152 0.182 Abdominal esophageal length, (mm) 32.00 (25.00, 38.00) 32.00 (26.00, 37.50) 28.00 (21.25, 35.75) 0.302 0.989 0.143 0.139 Esophageal hiatus (mm) 13.27 ± 3.12 13.20 ± 3.56 15.45 ± 3.12 0.020 0.429 0.947 0.611 a Comparative analysis of Groups A and B; b Comparative analysis of Groups A and C; C Comparative analysis of Groups B and C. Ordered logistic regression analysis screening for risk factors influencing the grading of RE We used a multivariate ordered logistic regression model to determine the influence of several independent variables on the grading of RE, including age, gender, reflux frequency within 5 minutes, reflux time within 5 min, abdominal esophageal length, and esophageal hiatus. RE was classified into Groups A (LA-A), B (LA-B), and C (LA-C/D), with Group C as the reference category. Our analysis indicated that patients with a shorter reflux time within 5 min progress to a lower grade of RE compared to those with a reflux time of 4 seconds or more within 5 min. Compared with patients who had four or more reflux episodes within 5 min, those with fewer reflux episodes within 5 min progressed to a lower grade of RE, and the difference was statistically significant ( P < 0.05) (Table 2 ). Table 2 Ordered logistic regression analysis of factors predicting reflux esophagitis grade B Std error Wald Exp (B) Sig 95% CI for Exp (B) Lower Upper Threshold LA-A –8.538 2.204 15.002 < 0.001 0.000 –12.859 –4.218 LA-B –4.333 2.032 4.549 0.013 0.033 –8.314 –0.351 Esophageal hiatus 0.021 0.050 0.174 1.021 0.676 –0.077 0.118 Abdominal esophageal length 0.004 0.019 0.054 1.004 0.816 –0.033 0.041 Male 0.449 0.314 2.037 1.567 0.153 –0.167 1.065 Age 0.001 0.010 0.009 1.001 0.924 –0.019 0.021 Reflux time within 5 minutes = 0 –3.651 1.804 4.095 0.026 0.043 –7.188 –0.115 Reflux time within 5 minutes = 1 –4.889 1.537 10.126 0.008 0.001 –7.901 –1.878 Reflux time within 5 minutes = 2 –8.012 1.942 17.020 < 0.001 < 0.001 –11.818 –4.206 Reflux time within 5 minutes = 3 –4.246 1.648 6.641 0.014 0.010 –7.475 –1.017 Reflux time within 5 minutes = 4 0 a Reflux frequency within 5 minutes = 0 –5.673 1.581 12.882 0.003 < 0.001 –8.771 –2.575 Reflux frequency within 5 minutes = 1 –5.673 1.626 12.178 0.003 < 0.001 –8.860 –2.487 Reflux frequency within 5 minutes = 2 –5.618 1.893 8.807 0.004 0.003 –9.328 -1.907 Reflux frequency within 5 minutes = 3 –3.433 1.720 3.983 0.032 0.046 –6.804 –0.061 Reflux frequency within 5 minutes = 4 0 a 0 Random forest algorithm screening for risk factors influencing the grading of RE The influence of each variable on the grading of RE was analyzed using the "VSURF" and "varSelRF" packages in R. Variable importance scores were calculated; the top two importance scores among all the variables included in the analysis were those of reflux frequency and reflux time, and these scores were significantly higher than those of the other factors (Fig. 1 ). The Boruta algorithm under the random forest framework was used to screen for factors the influencing the grading of RE. The importance scores for reflux frequency and reflux time, which were ultimately classified as "confirmed variables", were significantly higher than the highest importance score of the shadow features (Fig. 2 ). SVM algorithm screening for risk factors influencing the grading of RE The influence of each variable on the grading of RE was analyzed based on the SVM algorithm using the "caret" package in R. The performance of the SVM model under different quantities of variables was evaluated through fivefold cross-validation, thereby selecting the optimal feature subset and calculating the importance scores of each variable. Reflux frequency, reflux count, and esophageal hiatus had the highest scores among all the variables included in the analysis, and their scores were significantly higher than those of the other factors (Fig. 3 ). DISCUSSION The diagnosis and prognostic prediction of RE have always been key challenges in clinical practice. The introduction of contrast-enhanced ultrasound technology has led to breakthrough progress in this field. By CEUS, clinical data (reflux time, reflux frequency, etc.) can be precisely quantified, and predictive models can be constructed[ 10 ]. For example, by analyzing the association between relevant clinical data and the grade of RE through the random forest algorithm, patients can be more accurately classified by severity. By using SVMs to handle nonlinear features in the data, individuals at high risk of progressing to advanced RE can be identified in advance[ 11 ]. This "radiomics + clinical data" integration model not only improves the sensitivity and specificity of predictive models but also reveals the key role of CEUS technology in disease progression through variable importance analysis, providing decision support for individualized treatment plans[ 12 ]. This study revealed that ordered logistic regression, random forest, and SVM analyses all indicated that reflux time (in seconds) within 5 minutes and reflux frequency on enhanced ultrasound contrast imaging were key factors predicting the severity of RE. Reflux time directly reflects the duration of contact between gastric acid, pepsin, and the esophageal mucosa. A longer reflux duration increases the chance of the mucosal tissue being exposed to an acidic environment, leading to an increased risk of squamous epithelial cell damage, release of inflammatory factors, and disruption of the mucosal barrier[ 13 , 14 ]. Reflux frequency reflects the repetitiveness of reflux events, where frequent reflux will further compound the effects of mucosal damage, causing the inflammation to progress from mild congestion and edema to erosion, ulceration, and even stenosis[ 15 , 16 ]. CEUS can dynamically monitor the reflux process and accurately capture these two core parameters, which precisely align with the pathological feature of "damage – accumulation" of RE[ 17 ]. Therefore, it can effectively distinguish lesions by severity. Additionally, our comparison between groups showed that the reflux frequency in group C was significantly higher than that in groups A and B, but the reflux time did not show a similar result. We believe that the reason for this might be the smaller amount of data collected, with fewer patients having a reflux time of 4 seconds or longer, and where the reflux time in group C was significantly longer than that in groups A and B. The above results all suggest that reflux time and frequency enable more accurate prediction of high-grade RE. Frequent and prolonged gastroesophageal reflux is likely to be a key factor leading to higher-grade esophagitis. It is worth noting that the SVM algorithm additionally identified the diameter of the esophageal hiatus as an important factor influencing the grading of RE. This discovery has provided a new dimension for the disease grading mechanism. The esophageal hiatus is the anatomical structure where the esophagus connects to the stomach, and as such its functional integrity (including the size and tension of the hiatus) directly affects the anti-reflux barrier function: a relaxed or enlarged hiatus may lead to a greater likelihood of gastroesophageal reflux and make it more difficult to clear reflux substances in time, causing a "synergistic damage effect" according to the frequency of reflux[ 18 , 19 ]. Compared to logistic regression, the SVM, as an algorithm better suited for handling nonlinear relationships[ 20 ], may have more accurately captured the interaction between the diameter of the esophageal hiatus and the intensity of reflux, revealing complex correlations that traditional linear models have difficulty identifying. This study has certain limitations. First, it is a retrospective study with a small amount of data. In the future, the sample size should be expanded, and prospective verification should be conducted. Second, due to the influence of the retrospective data collection, the number of confounding factors was relatively small. In the future, the number of confounding factors should be increased. In summary, based on the CEUS data combined with ordered logistic regression, random forest, and SVM analyses, this study identified the key factors affecting the progression of RE, providing a new quantitative tool for monitoring RE. CONCLUSION CEUS can better predict the progression of RE. The number of reflux events within 5 minutes, duration of reflux within 5 min, and diameter of the esophageal hiatus are the key factors influencing the progression of RE. Declarations Ethics approval and consent to participate This study was approved by the Ethics Committee of the First People's Hospital of Longquanyi District, Chengdu City, Sichuan Province (Approval Number: af - key - 20101) and has been registered in the Chinese Clinical Trial Registry (No: ChiCTR1900021934, registration date: March 16, 2019). Consent for publication Not Applicable. Availability of data and materials All authors declare that the data in this study is true and reliable. If there is any information that needs further verification or if you need to obtain the collated data set, please feel free to communicate with the authors at any time. Competing interests All authors declare that during the research process and the writing of the manuscript, there were no competing interests that might affect the objectivity and fairness of the research. Funding This research is not currently supported by any funds. Authors' contributions Y.T., J.Y.W. and S.C.Z. designed the project; Y.T., W.Y.W. and J.C. conducted the statistical analysis; Y.T. and C.Y.X drew the charts; Y.T., Z.Y.S. and S.C.Z. wrote the main manuscript text. All authors reviewed the manuscript. Acknowledgements First of all, I would like to express my sincere gratitude to all the subjects and their families who participated in this study for their trust and cooperation. Thanks to the Ethics Committee of the First People's Hospital of Longquanyi District, Chengdu City Guidance and review of the ethical work of this study by Sichuan Province and the Chinese Clinical Trial Registry; At the same time, we would like to express our gratitude to the medical staff and workers of the Gastroenterology Department of the First People's Hospital of Longquanyi District, Chengdu City, Sichuan Province for their technical support and platform guarantee during the research process. Finally, we would like to express our sincere gratitude to all the individuals and institutions that have provided assistance for the successful completion of this research and the publication of the manuscript. References Shibli F, Mari A, Fass R. Drug treatment strategies for erosive esophagitis in adults: a narrative review. Transl Gastroenterol Hepatol. 2025-01-01; 10 54. 10.21037/tgh-24-168 . PMID: 40755734. Davis TA, Gyawali CP. Refractory Gastroesophageal Reflux Disease: Diagnosis and Management. J NEUROGASTROENTEROL. 2024-01-30; 30 (1): 17–28. 10.5056/jnm23145 . PMID: 38173155. Emanuel AL, Meijer RI, van Poelgeest E, Spoor P, Serné EH, Eringa EC. Contrast-enhanced ultrasound for quantification of tissue perfusion in humans. MICROCIRCULATION. 2020-01-01; 27 (1): e12588. 10.1111/micc.12588 . PMID: 31465606. Wang JY, Luo Y, Wang WY, Zheng SC, He L, Xie CY, Peng L. Contrast-enhanced ultrasound using SonoVue mixed with oral gastrointestinal contrast agent to evaluate esophageal hiatal hernia: Report of three cases and a literature review. World J Clin Cases. 2021-04-16; 9 (11): 2679–2687. 10.12998/wjcc.v9.i11.2679 . PMID: 33889636. Daidone M, Ferrantelli S, Tuttolomondo A. Machine learning applications in stroke medicine: advancements, challenges, and future prospectives. NEURAL REGEN RES. 2024-04-01; 19 (4): 769–773. 10.4103/1673-5374.382228 . PMID: 37843210. Xi Y, Wang H, Sun N. Machine learning outperforms traditional logistic regression and offers new possibilities for cardiovascular risk prediction: A study involving 143,043 Chinese patients with hypertension. Front Cardiovasc Med. 2022-01-01; 9 1025705. 10.3389/fcvm.2022.1025705 . PMID: 36451926. Silva RHLE, Machado AMC. Automatic measurement of pressure ulcers using Support Vector Machines and GrabCut. COMPUT METH PROG BIO. 2021-03-01; 200 105867. 10.1016/j.cmpb.2020.105867 . PMID: 33261945. Alakus C, Larocque D, Labbe A. Covariance regression with random forests. BMC Bioinformatics. 2023-06-17; 24 (1): 258. 10.1186/s12859-023-05377-y . PMID: 37330468. Yadlapati R, Gyawali CP, Pandolfino JEAGA. Clinical Practice Update on the Personalized Approach to the Evaluation and Management of GERD: Expert Review. CLIN GASTROENTEROL H. 2022-05-01; 20 (5): 984–994.e1. 10.1016/j.cgh.2022.01.025 . PMID: 35123084. Sofia C, Solazzo A, Cattafi A, Chimenz R, Cicero G, Marino MA, D'angelo T, Manti L, Condorelli E, Ceravolo G, Mazziotti S, Ascenti G. Contrast-enhanced voiding urosonography in the assessment of vesical-ureteral reflux: the time has come. RADIOL MED. 2021-07-01; 126 (7): 901–909. 10.1007/s11547-021-01360-w . PMID: 33954899. Liu SW, Li P, Li XQ et al. Recent advances in machine learning for precision diagnosis and treatment of esophageal disorders. World J Gastroenterol. 2025; 31 (23): 105076. 10.3748/wjg.v31.i23.105076 . PMID: 40575335. Pan G, Pan Z, Chen W et al. Integration of MRI radiomics and clinical data for preoperative prediction of vascular invasion in breast cancer: A deep learning approach. Magn Reson Imaging. 2025; 118 110339. 10.1016/j.mri.2025.110339 . PMID: 39880177. Prakash C, Clouse RE. Value of extended recording time with wireless pH monitoring in evaluating gastroesophageal reflux disease. CLIN GASTROENTEROL H. 2005-04-01; 3 (4): 329 – 34. 10.1016/s1542-3565(05)00021-2 . PMID: 15822037. Souza RF, Huo X, Mittal V, Schuler CM, Carmack SW, Zhang HY, Zhang X, Yu C, Hormi-Carver K, Genta RM, Spechler SJ. Gastroesophageal reflux might cause esophagitis through a cytokine-mediated mechanism rather than caustic acid injury. Gastroenterology. 2009-11-01; 137 (5): 1776-84. 10.1053/j.gastro.2009.07.055 . PMID: 19660463. Morozov S, Sentsova T, 1084:, LOCAL EXPRESSION OF PROINFLAMMATORY CYTOKINES GENES IN PATIENTS WITH GASTROESOPHAGEAL REFLUX DISEASE AND ITS ASSOCIATION WITH TYPES AND NUMBER OF GASTROESOPHAGEAL REFLUXES GASTROENTEROLOGY. 2022-05-01; 162 (7): S–252. 10.1016/s0016-5085(22)60600-0 Kruse-Andersen S, Wallin L, Madsen T. Reflux patterns and related oesophageal motor activity in gastro-oesophageal reflux disease. GUT. 1990-06-01; 31 (6): 633-8. 10.1136/gut.31.6.633 . PMID: 2379866. Farina R, Pennisi F, La Rosa M, Puglisi C, Mazzone G, Riva G, Foti PV, Ettorre GC. Contrast-enhanced colour-Doppler sonography versus pH-metry in the diagnosis of gastro-oesophageal reflux in children. RADIOL MED. 2008-06-01; 113 (4): 591-8. 10.1007/s11547-008-0267-4 . PMID: 18478190. Wang JY, Luo Y, Wang WY, Zheng SC, He L, Xie CY, Peng L. Contrast-enhanced ultrasound using SonoVue mixed with oral gastrointestinal contrast agent to evaluate esophageal hiatal hernia: Report of three cases and a literature review. World J Clin Cases. 2021-04-16; 9 (11): 2679–2687. 10.12998/wjcc.v9.i11.2679 . PMID: 33889636. Franzén T, Tibbling L. Is the severity of gastroesophageal reflux dependent on hiatus hernia size? WORLD J GASTROENTERO. 2014-02-14; 20 (6): 1582-4. 10.3748/wjg.v20.i6.1582 . PMID: 24587634. Piccialli V, Sciandrone M. Nonlinear optimization and support vector machines ANN OPER RES. 2022-04-14; 10.1007/s10479-022-04655-x Additional Declarations No competing interests reported. Cite Share Download PDF Status: Under Review Version 1 posted Reviews received at journal 22 Feb, 2026 Reviewers agreed at journal 22 Feb, 2026 Reviews received at journal 18 Feb, 2026 Reviewers agreed at journal 17 Feb, 2026 Reviews received at journal 13 Feb, 2026 Reviewers agreed at journal 11 Feb, 2026 Reviewers agreed at journal 11 Feb, 2026 Reviewers invited by journal 11 Feb, 2026 Editor assigned by journal 09 Feb, 2026 Editor invited by journal 21 Jan, 2026 Submission checks completed at journal 20 Jan, 2026 First submitted to journal 20 Jan, 2026 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-8494694","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":591227498,"identity":"9e7707f6-4dbe-4881-a1e9-193885d8c612","order_by":0,"name":"Ying Tu","email":"","orcid":"","institution":"Sichuan University/First People's Hospital of Longquanyi District","correspondingAuthor":false,"prefix":"","firstName":"Ying","middleName":"","lastName":"Tu","suffix":""},{"id":591227499,"identity":"70bf8388-444f-4cc6-9ddf-85937ffb1229","order_by":1,"name":"Jing-Yu Wang","email":"","orcid":"","institution":"Sichuan University/First People's Hospital of Longquanyi District","correspondingAuthor":false,"prefix":"","firstName":"Jing-Yu","middleName":"","lastName":"Wang","suffix":""},{"id":591227500,"identity":"65555174-6714-4335-916d-4182657e33ac","order_by":2,"name":"Wen-Ying Wang","email":"","orcid":"","institution":"Sichuan University/First People's Hospital of Longquanyi District","correspondingAuthor":false,"prefix":"","firstName":"Wen-Ying","middleName":"","lastName":"Wang","suffix":""},{"id":591227501,"identity":"9667e47d-dd5c-47b3-ab4c-516344b94260","order_by":3,"name":"Zhen-Yu Song","email":"","orcid":"","institution":"Sichuan University/First People's Hospital of Longquanyi District","correspondingAuthor":false,"prefix":"","firstName":"Zhen-Yu","middleName":"","lastName":"Song","suffix":""},{"id":591227502,"identity":"5af58883-7d74-406d-9403-214e716c4c97","order_by":4,"name":"Chun-Yan Xie","email":"","orcid":"","institution":"Sichuan University/First People's Hospital of Longquanyi District","correspondingAuthor":false,"prefix":"","firstName":"Chun-Yan","middleName":"","lastName":"Xie","suffix":""},{"id":591227503,"identity":"301ab2b1-e2a6-4055-abe7-6ee84da77f10","order_by":5,"name":"Jie Chen","email":"","orcid":"","institution":"Sichuan University/First People's Hospital of Longquanyi District","correspondingAuthor":false,"prefix":"","firstName":"Jie","middleName":"","lastName":"Chen","suffix":""},{"id":591227504,"identity":"216f8c73-31fe-47f8-aa15-db5f49c88861","order_by":6,"name":"Shi-Cheng Zheng","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAy0lEQVRIiWNgGAWjYHAC5ocffjDwsMk/PnAAyCAKsBlL9jDI8TOkJR6c2UOkNRI8bAzGkg05xoc52IhQLt9/+IGBBA9D4oYDZz4cZuBhkOcXO4BfC+OMNIMHBRZALQd7NxwGMgxnzk7Ar4UZaAPQlv+JGw7zbjg8g4chweA2AS1s/GfAfknccIznwWEgg7AWHoYcqPd7eBiI0yIhkWYGCWQJNgNgIEsQ9gswxB5DolKC+fGHDz9s5PmlCWjBsJU05aNgFIyCUTAKsAMAtBNAFxOFFl8AAAAASUVORK5CYII=","orcid":"","institution":"Sichuan University/First People's Hospital of Longquanyi District","correspondingAuthor":true,"prefix":"","firstName":"Shi-Cheng","middleName":"","lastName":"Zheng","suffix":""}],"badges":[],"createdAt":"2026-01-01 10:53:39","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8494694/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8494694/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":102828861,"identity":"00ac4f91-ee93-40d9-894c-626b98842727","added_by":"auto","created_at":"2026-02-17 09:26:35","extension":"jpeg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":22826,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eRandom forest analysis of factors predicting reflux esophagitis grade.\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"floatimage1.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-8494694/v1/9d94a7f816bf85554576b789.jpeg"},{"id":102828903,"identity":"c289dd51-6b37-43bd-a733-6fd0ab8e0cb7","added_by":"auto","created_at":"2026-02-17 09:26:40","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":8798,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eBoruta algorithm analysis of factors predicting reflux esophagitis grade.\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-8494694/v1/0beafb773fd51d45cc5ce413.png"},{"id":102828899,"identity":"6f1d877b-f9c8-4e33-ad04-979f424e2ef3","added_by":"auto","created_at":"2026-02-17 09:26:36","extension":"jpeg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":25277,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eSupport vector machine analysis of factors predicting reflux esophagitis grade.\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"floatimage3.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-8494694/v1/8ef06a1a23334e24fbe8f3ad.jpeg"},{"id":102829047,"identity":"a62211b8-5100-42cd-a60a-894f8076d8eb","added_by":"auto","created_at":"2026-02-17 09:27:12","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":879713,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8494694/v1/2319058a-02d1-45f7-81f2-6f44f1c15962.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Risk prediction of progression of reflux esophagitis based on contrast-enhanced ultrasound technology and machine learning","fulltext":[{"header":"Core Tip","content":"\u003cp\u003eReflux esophagitis (RE) poses a growing global health burden, with potential progression to severe conditions like Barrett's esophagus and esophageal adenocarcinoma. Conventional diagnostic methods such as gastroscopy are invasive, while other non-invasive approaches lack sufficient predictive accuracy. This study used contrast-enhanced ultrasound (CEUS) and 3 models (ordered logistic regression, random forest, SVM) to explore RE progression predictors. For 197 LA-classified RE patients, it found 5-min reflux frequency/duration (all models) and esophageal hiatus diameter (SVM) key, supporting CEUS in non-invasive RE prediction.\u003c/p\u003e\n\u003cp\u003e\u003cbr\u003e\u003c/p\u003e"},{"header":"INTRODUCTION","content":"\u003cp\u003eReflux esophagitis (RE) is the most common subtype of gastroesophageal reflux disease, and its incidence is increasing year by year worldwide. It is characterized by obvious gastroesophageal reflux, and patients often suffer from symptoms such as heartburn, acid reflux, and swallowing discomfort. Long-term and repeated attacks not only seriously affect quality of life but may also progress to Barrett's esophagus or even esophageal adenocarcinoma, posing significant challenges for clinical diagnosis and treatment[\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e]. Currently, gastroscopy remains the \"gold standard\" for the diagnosis and grading of RE, but its invasive operation can easily cause discomfort to patients, while traditional non-invasive examinations such as esophageal pH monitoring and symptom scoring are often insufficient in predictive accuracy due to individual differences and operational norms[\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e]. Therefore, exploring non-invasive and highly accurate prediction methods for RE has become an important direction in clinical research.\u003c/p\u003e \u003cp\u003eThe development of contrast-enhanced ultrasound (CEUS) technology provides a new solution to meet this need[\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e]. As a non-invasive imaging technique, CEUS can capture the blood flow perfusion characteristics, microcirculation status, and structural changes of esophageal mucosa and surrounding tissues by oral or intravenously injected contrast agents. These data not only reflect the hemodynamic abnormalities in the lesion area but also indirectly indicate the degree of mucosal inflammation and the extent of damage. CEUS has the advantages of convenient operation, high repeatability, and no radiation risk, providing quantitative data for the non-invasive assessment of RE[\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eIn recent years, machine learning algorithms have made significant progress in the medical field, and their advantages in processing high-dimensional data and mining hidden patterns have shown significant potential in the identification of risk factors[\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e]. Among them, logistic regression, as a classic statistical learning method, with its strong model interpretability and easy quantification of results, has long been the benchmark tool for identifying linearly associated risk factors[\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e]. Support vector machines (SVMs), through kernel function mapping, transform nonlinear problems into linear classification problems in high-dimensional space, exhibit excellent generalization ability in small samples and high-dimensional data, and are suitable for capturing complex interactions between variables[\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e]. Combining multiple decision trees for prediction, random forests, representative of ensemble learning, can effectively reduce overfitting risks and intuitively quantify the contribution of different factors through variable importance scoring, and are especially suitable for handling datasets with multicollinearity or nonlinear relationships[\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eBased on the above, this study intends to use CEUS data as a basis, combined with a random forest algorithm, SVM algorithm, and ordered logistic regression, to identify factors related to the occurrence and progression of RE. The aim is to achieve non-invasive and efficient prediction of RE through the mining of contrast imaging data, thereby providing objective auxiliary diagnostic evidence for clinicians, reducing unnecessary invasive examinations, improving the efficiency of early intervention in the disease, and ultimately providing a new technical path for the precise diagnosis and treatment of RE.\u003c/p\u003e"},{"header":"MATERIALS AND METHODS","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eStudy population\u003c/h2\u003e \u003cp\u003e This study was approved by the Ethics Committee of the First People's Hospital of Longquanyi District, Chengdu City, Sichuan Province (Approval Number: af - key \u0026minus;\u0026thinsp;20101) and has been registered in the Chinese Clinical Trial Registry (No: ChiCTR1900021934, registration date: March 16, 2019). In total, 296 patients with RE hospitalized at West China Longquanyi Hospital of Sichuan University from February 2019 to January 2023 were retrospectively selected. All patients underwent upper gastrointestinal endoscopy, and the diagnostic criteria for RE were based on the Los Angeles (LA) classification[\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e]. The inclusion criteria were: (1) age ranging from 18 to 80 years; and (2) clear diagnosis of RE by endoscopy. The exclusion criteria were: (1) esophagitis caused by other etiologies; (2) suspected or confirmed esophageal squamous cell carcinoma or other malignant tumors at the time of admission; (3) severe diseases in other systems such as the heart, lung, and kidney; and (4) incomplete data or loss to follow-up.\u003c/p\u003e \u003cp\u003ePatients were divided into three groups by experienced endoscopy physicians: \u0026ldquo;LA-A\u0026rdquo; (Group A), \u0026ldquo;LA-B (Group B), and LA-C/D\u0026rdquo; (Group C) according to the LA classification.\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eData collection\u003c/h3\u003e\n\u003cp\u003eAll patients underwent CEUS, and the diameter of the esophageal hiatus, frequency of reflux within 5 min, duration of reflux within 5 min, length of the abdominal esophagus, etc., were measured. Basic clinical characteristics of the patients, such as age and gender, were also collected. The frequency of reflux within 5 min was recorded as 0 for zero times, 1 for one time, 2 for twice, 3 for three times, and 4 for four or more for four times. The duration of reflux within 5 min was recorded as 0 for 0 seconds, 1 for 1 second, 2 for 2 seconds, 3 for 3 seconds, and 4 for 4 or more for 4 seconds or more.\u003c/p\u003e \u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003eStatistical analysis\u003c/h2\u003e \u003cp\u003eStatistical analysis was performed using SPSS 26.0 software (IBM Corp., Armonk, NY, USA). Measurement data that followed a normal distribution were expressed as (mean\u0026thinsp;\u0026plusmn;\u0026thinsp;standard deviation). Comparisons between two groups were conducted using the paired t-test, while comparisons among multiple groups were carried out using one-way analysis of variance. Measurement data that deviated from a normal distribution were represented as median (quartile 1, quartile 3). Comparisons among multiple groups were conducted using the Kruskal\u0026ndash;Wallis H test, and comparisons between two groups were conducted using the Mann\u0026ndash;Whitney U test. Count data were expressed as the number of cases and percentages, and comparisons between groups were conducted using the chi-square test. We used multivariate ordered logistic regression to analyze the factors related to the grading of RE patients. A P value\u0026thinsp;\u0026lt;\u0026thinsp;0.05 was considered statistically significant.\u003c/p\u003e \u003cp\u003eRelevant variables were selected based on the random forest algorithm through the \"VSURF\", \"Boruta\", and \"varSelRF\" data packages in R version 4.2.2 (R Foundation for Statistical Computing, Vienna, Austria)。This algorithm integrates the prediction results of multiple decision trees to reduce the overfitting risk of a single model and enhance the stability of variable selection. The core parameters of the model are optimized through 5-fold cross-validation. The specific settings are as follows: (1) Number of decision trees (n_estimators)\u0026thinsp;=\u0026thinsp;1000; (2) Maximum tree depth (max_depth)\u0026thinsp;=\u0026thinsp;20; (3) Minimum sample number for node splitting (min_samples_split)\u0026thinsp;=\u0026thinsp;5; (4) Minimum sample number for leaf nodes (min_samples_leaf)\u0026thinsp;=\u0026thinsp;2; (5) Maximum number of features for each split (max_features) = \"sqrt\"; (6) Random seed (random_state)\u0026thinsp;=\u0026thinsp;42 (to ensure reproducibility of results). The variable selection combines three functions: \"VSURF\" for stepwise selection, \"Boruta\" for verifying the validity of shadow features, and \"varSelRF\" for optimizing based on importance scores and cross-validation errors. Finally, the core variables are determined.\u003c/p\u003e \u003cp\u003eRelevant variables were selected based on the SVM algorithm through the \"caret\", \"e1071\", and \"doParallel\" data packages in R version 4.2.2 (R Foundation for Statistical Computing, Vienna, Austria). This algorithm achieves effective classification and feature screening of high-dimensional data by seeking the optimal separation hyperplane. Among them, the \"doParallel\" package is used to enable parallel computing and improve the efficiency of model training; The \"caret\" package is used for hyperparameter optimization and cross-validation of the model. The \"e1071\" package is used for building the core model. The core parameters of the model are set as follows: (1) Kernel Function type\u0026thinsp;=\u0026thinsp;Radial Basis Function (RBF), this kernel function has good nonlinear mapping ability and can adapt to the complex relationship between features and outcome variables; (2) The penalty parameter\u0026thinsp;=\u0026thinsp;1.0 is used to balance the training error and generalization error of the model. The larger the C value, the higher the degree of fit of the model to the training set, and the more likely overfitting occurs. Cross-validation determines that the model performance is optimal when C\u0026thinsp;=\u0026thinsp;1.0. (3) Kernel function parameter\u0026thinsp;=\u0026thinsp;0.1, controlling the width of the radial basis kernel function. The larger the gamma value, the stronger the local fitting ability of the model. It is determined by grid search that the classification accuracy of the model is the highest when gamma\u0026thinsp;=\u0026thinsp;0.1. (4) Cross-validation strategy\u0026thinsp;=\u0026thinsp;5-fold cross-validation. The dataset is randomly divided into 5 equal parts, with 4 parts used as the training set and 1 part as the validation set in sequence. After repeating this 5 times, the average prediction error is calculated to evaluate the generalization ability of the model. (5) Data preprocessing\u0026thinsp;=\u0026thinsp;Standardizing all input features (z-score standardization, that is, converting to a standard normal distribution with a mean of 0 and a standard deviation of 1) to eliminate the influence of dimensional differences of different features on the model. Variable screening is achieved through the Recursive Feature Elimination (RFE) method in the \"caret\" package, gradually eliminating the features that contribute less to the model prediction, and finally retaining the feature subset that minimizes the cross-validation error of the model.\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eEthical considerations\u003c/h3\u003e\n\u003cp\u003e This study received ethical approval from our hospital's Medical Ethics Committee (AF-KY- 201901) and was registered in the Chinese Clinical Trial Registry (registration number: ChiCTR1900021934). The study was conducted in accordance with the Helsinki Declaration and Good Clinical Practice guidelines. Before the study started, we obtained informed consent from each participant. If informed consent was not obtained, participation in the study was not allowed.\u003c/p\u003e"},{"header":"RESULTS","content":"\u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003eClinical characteristics of the study population\u003c/h2\u003e \u003cp\u003eA total of 296 patients were initially included in this study. Among them, 13 were excluded due to being over 80 years old, 15 were excluded because they were subsequently diagnosed with esophageal cancer and other malignant tumors, 11 were excluded due to having major diseases of the heart, lungs, kidneys and other organs, and 60 were excluded due to missing data. A total of 197 patients with RE who underwent CEUS were finally included in this study, including 92 patients of LA-A grade (Group A), 85 of LA-B grade (Group B), and 20 of LA-C/D grade (Group C). The median age of the patients in Group A was 54.5 years, with 54.35% being male. The median age of the patients in Group B was 56 years, with 64.71% being male. The median age of the patients in Group C was 60.5 years, with 65% being male. The reflux frequency within 5 min, reflux time within 5 min (s), and diameter of the esophageal hiatus (mm) were statistically significantly different among multiple groups. Further pairwise comparisons showed that the reflux frequency within 5 min in Groups A and B was significantly lower than that in Group C (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05) (Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eAnalysis of baseline clinical characteristics\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"8\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c8\" colnum=\"8\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c2\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eGroup A (\u003cem\u003en\u003c/em\u003e\u0026thinsp;=\u0026thinsp;92)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eGroup B (\u003cem\u003en\u003c/em\u003e\u0026thinsp;=\u0026thinsp;85)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eGroup C (\u003cem\u003en\u003c/em\u003e\u0026thinsp;=\u0026thinsp;20)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"4\" nameend=\"c8\" namest=\"c5\"\u003e \u003cp\u003e\u003cem\u003eP\u003c/em\u003e value\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003e\u003cem\u003eP\u003c/em\u003e\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u003cem\u003eP\u003c/em\u003e \u003csup\u003ea\u003c/sup\u003e\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u003cem\u003eP\u003c/em\u003e \u003csup\u003eb\u003c/sup\u003e\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c8\"\u003e \u003cp\u003e\u003cem\u003eP\u003c/em\u003e \u003csup\u003eC\u003c/sup\u003e\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAge (years)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e54.50 (44.00, 65.00)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e56.00 (40.50, 67.00)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e60.50 (45.25, 68.00)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.485\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.916\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.234\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.278\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMale (\u003cem\u003en\u003c/em\u003e)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e50 (54.35%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e55 (64.71%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e13 (65.00%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.330\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.161\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.384\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.980\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eReflux frequency within 5 minutes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.25\u0026thinsp;\u0026plusmn;\u0026thinsp;0.74\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.19\u0026thinsp;\u0026plusmn;\u0026thinsp;0.63\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e3.30\u0026thinsp;\u0026plusmn;\u0026thinsp;1.03\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.298\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.006\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eReflux time within 5 minutes, (s)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.29\u0026thinsp;\u0026plusmn;\u0026thinsp;0.83\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.27\u0026thinsp;\u0026plusmn;\u0026thinsp;0.93\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e3.55\u0026thinsp;\u0026plusmn;\u0026thinsp;1.23\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.853\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.152\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.182\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAbdominal\u0026nbsp;esophageal\u0026nbsp;length, (mm)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e32.00 (25.00, 38.00)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e32.00 (26.00, 37.50)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e28.00 (21.25, 35.75)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.302\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.989\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.143\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.139\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eEsophageal hiatus (mm)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e13.27\u0026thinsp;\u0026plusmn;\u0026thinsp;3.12\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e13.20\u0026thinsp;\u0026plusmn;\u0026thinsp;3.56\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e15.45\u0026thinsp;\u0026plusmn;\u0026thinsp;3.12\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.020\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.429\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.947\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.611\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"8\"\u003e\u003csup\u003ea\u003c/sup\u003eComparative analysis of Groups A and B; \u003csup\u003eb\u003c/sup\u003eComparative analysis of Groups A and C; \u003csup\u003eC\u003c/sup\u003e Comparative analysis of Groups B and C.\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eOrdered logistic regression analysis screening for risk factors influencing the grading of RE\u003c/h3\u003e\n\u003cp\u003eWe used a multivariate ordered logistic regression model to determine the influence of several independent variables on the grading of RE, including age, gender, reflux frequency within 5 minutes, reflux time within 5 min, abdominal esophageal length, and esophageal hiatus. RE was classified into Groups A (LA-A), B (LA-B), and C (LA-C/D), with Group C as the reference category. Our analysis indicated that patients with a shorter reflux time within 5 min progress to a lower grade of RE compared to those with a reflux time of 4 seconds or more within 5 min. Compared with patients who had four or more reflux episodes within 5 min, those with fewer reflux episodes within 5 min progressed to a lower grade of RE, and the difference was statistically significant (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05) (Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eOrdered logistic regression analysis of factors predicting reflux esophagitis grade\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"9\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c8\" colnum=\"8\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c9\" colnum=\"9\"\u003e\u003c/div\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eB\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eStd error\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eWald\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eExp (B)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eSig\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c9\" namest=\"c8\"\u003e \u003cp\u003e95% CI for Exp (B)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003eLower\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003eUpper\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eThreshold\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eLA-A\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u0026ndash;8.538\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2.204\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e15.002\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e\u0026ndash;12.859\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e\u0026ndash;4.218\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eLA-B\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u0026ndash;4.333\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2.032\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e4.549\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.013\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.033\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e\u0026ndash;8.314\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e\u0026ndash;0.351\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"13\" rowspan=\"14\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eEsophageal hiatus\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.021\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.050\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.174\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e1.021\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.676\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e\u0026ndash;0.077\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.118\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAbdominal\u0026nbsp;esophageal\u0026nbsp;length\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.004\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.019\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.054\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e1.004\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.816\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e\u0026ndash;0.033\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.041\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMale\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.449\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.314\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e2.037\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e1.567\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.153\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e\u0026ndash;0.167\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e1.065\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAge\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.010\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.009\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e1.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.924\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e\u0026ndash;0.019\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.021\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReflux time within 5 minutes\u0026thinsp;=\u0026thinsp;0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u0026ndash;3.651\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1.804\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e4.095\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.026\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.043\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e\u0026ndash;7.188\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e\u0026ndash;0.115\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReflux time within 5 minutes\u0026thinsp;=\u0026thinsp;1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u0026ndash;4.889\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1.537\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e10.126\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.008\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e\u0026ndash;7.901\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e\u0026ndash;1.878\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReflux time within 5 minutes\u0026thinsp;=\u0026thinsp;2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u0026ndash;8.012\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1.942\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e17.020\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e\u0026ndash;11.818\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e\u0026ndash;4.206\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReflux time within 5 minutes\u0026thinsp;=\u0026thinsp;3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u0026ndash;4.246\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1.648\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e6.641\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.014\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.010\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e\u0026ndash;7.475\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e\u0026ndash;1.017\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReflux time within 5 minutes\u0026thinsp;=\u0026thinsp;4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0\u003csup\u003ea\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReflux frequency within 5 minutes\u0026thinsp;=\u0026thinsp;0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u0026ndash;5.673\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1.581\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e12.882\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.003\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e\u0026ndash;8.771\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e\u0026ndash;2.575\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReflux frequency within 5 minutes\u0026thinsp;=\u0026thinsp;1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u0026ndash;5.673\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1.626\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e12.178\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.003\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e\u0026ndash;8.860\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e\u0026ndash;2.487\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReflux frequency within 5 minutes\u0026thinsp;=\u0026thinsp;2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u0026ndash;5.618\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1.893\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e8.807\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.004\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.003\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e\u0026ndash;9.328\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e-1.907\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReflux frequency within 5 minutes\u0026thinsp;=\u0026thinsp;3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u0026ndash;3.433\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1.720\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e3.983\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.032\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.046\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e\u0026ndash;6.804\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e\u0026ndash;0.061\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReflux frequency within 5 minutes\u0026thinsp;=\u0026thinsp;4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0\u003csup\u003ea\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e\n\u003ch3\u003eRandom forest algorithm screening for risk factors influencing the grading of RE\u003c/h3\u003e\n\u003cp\u003eThe influence of each variable on the grading of RE was analyzed using the \"VSURF\" and \"varSelRF\" packages in R. Variable importance scores were calculated; the top two importance scores among all the variables included in the analysis were those of reflux frequency and reflux time, and these scores were significantly higher than those of the other factors (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). The Boruta algorithm under the random forest framework was used to screen for factors the influencing the grading of RE. The importance scores for reflux frequency and reflux time, which were ultimately classified as \"confirmed variables\", were significantly higher than the highest importance score of the shadow features (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003eSVM algorithm screening for risk factors influencing the grading of RE\u003c/h2\u003e \u003cp\u003eThe influence of each variable on the grading of RE was analyzed based on the SVM algorithm using the \"caret\" package in R. The performance of the SVM model under different quantities of variables was evaluated through fivefold cross-validation, thereby selecting the optimal feature subset and calculating the importance scores of each variable. Reflux frequency, reflux count, and esophageal hiatus had the highest scores among all the variables included in the analysis, and their scores were significantly higher than those of the other factors (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e"},{"header":"DISCUSSION","content":"\u003cp\u003eThe diagnosis and prognostic prediction of RE have always been key challenges in clinical practice. The introduction of contrast-enhanced ultrasound technology has led to breakthrough progress in this field. By CEUS, clinical data (reflux time, reflux frequency, etc.) can be precisely quantified, and predictive models can be constructed[\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e]. For example, by analyzing the association between relevant clinical data and the grade of RE through the random forest algorithm, patients can be more accurately classified by severity. By using SVMs to handle nonlinear features in the data, individuals at high risk of progressing to advanced RE can be identified in advance[\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e]. This \"radiomics\u0026thinsp;+\u0026thinsp;clinical data\" integration model not only improves the sensitivity and specificity of predictive models but also reveals the key role of CEUS technology in disease progression through variable importance analysis, providing decision support for individualized treatment plans[\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eThis study revealed that ordered logistic regression, random forest, and SVM analyses all indicated that reflux time (in seconds) within 5 minutes and reflux frequency on enhanced ultrasound contrast imaging were key factors predicting the severity of RE. Reflux time directly reflects the duration of contact between gastric acid, pepsin, and the esophageal mucosa. A longer reflux duration increases the chance of the mucosal tissue being exposed to an acidic environment, leading to an increased risk of squamous epithelial cell damage, release of inflammatory factors, and disruption of the mucosal barrier[\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e, \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e]. Reflux frequency reflects the repetitiveness of reflux events, where frequent reflux will further compound the effects of mucosal damage, causing the inflammation to progress from mild congestion and edema to erosion, ulceration, and even stenosis[\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e, \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e]. CEUS can dynamically monitor the reflux process and accurately capture these two core parameters, which precisely align with the pathological feature of \"damage \u0026ndash; accumulation\" of RE[\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e]. Therefore, it can effectively distinguish lesions by severity. Additionally, our comparison between groups showed that the reflux frequency in group C was significantly higher than that in groups A and B, but the reflux time did not show a similar result. We believe that the reason for this might be the smaller amount of data collected, with fewer patients having a reflux time of 4 seconds or longer, and where the reflux time in group C was significantly longer than that in groups A and B. The above results all suggest that reflux time and frequency enable more accurate prediction of high-grade RE. Frequent and prolonged gastroesophageal reflux is likely to be a key factor leading to higher-grade esophagitis.\u003c/p\u003e \u003cp\u003eIt is worth noting that the SVM algorithm additionally identified the diameter of the esophageal hiatus as an important factor influencing the grading of RE. This discovery has provided a new dimension for the disease grading mechanism. The esophageal hiatus is the anatomical structure where the esophagus connects to the stomach, and as such its functional integrity (including the size and tension of the hiatus) directly affects the anti-reflux barrier function: a relaxed or enlarged hiatus may lead to a greater likelihood of gastroesophageal reflux and make it more difficult to clear reflux substances in time, causing a \"synergistic damage effect\" according to the frequency of reflux[\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e, \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e]. Compared to logistic regression, the SVM, as an algorithm better suited for handling nonlinear relationships[\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e], may have more accurately captured the interaction between the diameter of the esophageal hiatus and the intensity of reflux, revealing complex correlations that traditional linear models have difficulty identifying.\u003c/p\u003e \u003cp\u003eThis study has certain limitations. First, it is a retrospective study with a small amount of data. In the future, the sample size should be expanded, and prospective verification should be conducted. Second, due to the influence of the retrospective data collection, the number of confounding factors was relatively small. In the future, the number of confounding factors should be increased. In summary, based on the CEUS data combined with ordered logistic regression, random forest, and SVM analyses, this study identified the key factors affecting the progression of RE, providing a new quantitative tool for monitoring RE.\u003c/p\u003e"},{"header":"CONCLUSION","content":"\u003cp\u003eCEUS can better predict the progression of RE. The number of reflux events within 5 minutes, duration of reflux within 5 min, and diameter of the esophageal hiatus are the key factors influencing the progression of RE.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003e\u003cem\u003eEthics approval and consent to participate\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study was approved by the Ethics Committee of the First People\u0026apos;s Hospital of Longquanyi District, Chengdu City, Sichuan Province (Approval Number: af - key - 20101) and has been registered in the Chinese Clinical Trial Registry (No: ChiCTR1900021934, registration date: March 16, 2019).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003eConsent for publication\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot Applicable.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003e\u0026nbsp;Availability of data and materials\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll authors declare that the data in this study is true and reliable. If there is any information that needs further verification or if you need to obtain the collated data set, please feel free to communicate with the authors at any time.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003eCompeting interests\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll authors declare that during the research process and the writing of the manuscript, there were no competing interests that might affect the objectivity and fairness of the research.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003eFunding\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis research is not currently supported by any funds.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003e\u0026nbsp;Authors\u0026apos; contributions\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eY.T., J.Y.W. and S.C.Z. designed the project; Y.T., W.Y.W. and J.C. conducted the statistical analysis; Y.T. and C.Y.X drew the charts; Y.T., Z.Y.S. and S.C.Z. wrote the main manuscript text. All authors reviewed the manuscript.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003e\u0026nbsp;Acknowledgements\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eFirst of all, I would like to express my sincere gratitude to all the subjects and their families who participated in this study for their trust and cooperation. Thanks to the Ethics Committee of the First People\u0026apos;s Hospital of Longquanyi District, Chengdu City Guidance and review of the ethical work of this study by Sichuan Province and the Chinese Clinical Trial Registry; At the same time, we would like to express our gratitude to the medical staff and workers of the Gastroenterology Department of the First People\u0026apos;s Hospital of Longquanyi District, Chengdu City, Sichuan Province for their technical support and platform guarantee during the research process. Finally, we would like to express our sincere gratitude to all the individuals and institutions that have provided assistance for the successful completion of this research and the publication of the manuscript.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eShibli F, Mari A, Fass R. Drug treatment strategies for erosive esophagitis in adults: a narrative review. Transl Gastroenterol Hepatol. 2025-01-01; 10 54. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.21037/tgh-24-168\u003c/span\u003e\u003cspan address=\"10.21037/tgh-24-168\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. PMID: 40755734.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDavis TA, Gyawali CP. Refractory Gastroesophageal Reflux Disease: Diagnosis and Management. J NEUROGASTROENTEROL. 2024-01-30; 30 (1): 17\u0026ndash;28. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.5056/jnm23145\u003c/span\u003e\u003cspan address=\"10.5056/jnm23145\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. PMID: 38173155.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eEmanuel AL, Meijer RI, van Poelgeest E, Spoor P, Sern\u0026eacute; EH, Eringa EC. Contrast-enhanced ultrasound for quantification of tissue perfusion in humans. MICROCIRCULATION. 2020-01-01; 27 (1): e12588. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1111/micc.12588\u003c/span\u003e\u003cspan address=\"10.1111/micc.12588\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. PMID: 31465606.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang JY, Luo Y, Wang WY, Zheng SC, He L, Xie CY, Peng L. Contrast-enhanced ultrasound using SonoVue mixed with oral gastrointestinal contrast agent to evaluate esophageal hiatal hernia: Report of three cases and a literature review. World J Clin Cases. 2021-04-16; 9 (11): 2679\u0026ndash;2687. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.12998/wjcc.v9.i11.2679\u003c/span\u003e\u003cspan address=\"10.12998/wjcc.v9.i11.2679\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. PMID: 33889636.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDaidone M, Ferrantelli S, Tuttolomondo A. Machine learning applications in stroke medicine: advancements, challenges, and future prospectives. NEURAL REGEN RES. 2024-04-01; 19 (4): 769\u0026ndash;773. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.4103/1673-5374.382228\u003c/span\u003e\u003cspan address=\"10.4103/1673-5374.382228\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. PMID: 37843210.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eXi Y, Wang H, Sun N. Machine learning outperforms traditional logistic regression and offers new possibilities for cardiovascular risk prediction: A study involving 143,043 Chinese patients with hypertension. Front Cardiovasc Med. 2022-01-01; 9 1025705. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3389/fcvm.2022.1025705\u003c/span\u003e\u003cspan address=\"10.3389/fcvm.2022.1025705\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. PMID: 36451926.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSilva RHLE, Machado AMC. Automatic measurement of pressure ulcers using Support Vector Machines and GrabCut. COMPUT METH PROG BIO. 2021-03-01; 200 105867. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.cmpb.2020.105867\u003c/span\u003e\u003cspan address=\"10.1016/j.cmpb.2020.105867\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. PMID: 33261945.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAlakus C, Larocque D, Labbe A. Covariance regression with random forests. BMC Bioinformatics. 2023-06-17; 24 (1): 258. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1186/s12859-023-05377-y\u003c/span\u003e\u003cspan address=\"10.1186/s12859-023-05377-y\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. PMID: 37330468.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYadlapati R, Gyawali CP, Pandolfino JEAGA. Clinical Practice Update on the Personalized Approach to the Evaluation and Management of GERD: Expert Review. CLIN GASTROENTEROL H. 2022-05-01; 20 (5): 984\u0026ndash;994.e1. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.cgh.2022.01.025\u003c/span\u003e\u003cspan address=\"10.1016/j.cgh.2022.01.025\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. PMID: 35123084.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSofia C, Solazzo A, Cattafi A, Chimenz R, Cicero G, Marino MA, D'angelo T, Manti L, Condorelli E, Ceravolo G, Mazziotti S, Ascenti G. Contrast-enhanced voiding urosonography in the assessment of vesical-ureteral reflux: the time has come. RADIOL MED. 2021-07-01; 126 (7): 901\u0026ndash;909. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s11547-021-01360-w\u003c/span\u003e\u003cspan address=\"10.1007/s11547-021-01360-w\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. PMID: 33954899.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiu SW, Li P, Li XQ et al. Recent advances in machine learning for precision diagnosis and treatment of esophageal disorders. World J Gastroenterol. 2025; 31 (23): 105076. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3748/wjg.v31.i23.105076\u003c/span\u003e\u003cspan address=\"10.3748/wjg.v31.i23.105076\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. PMID: 40575335.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePan G, Pan Z, Chen W et al. Integration of MRI radiomics and clinical data for preoperative prediction of vascular invasion in breast cancer: A deep learning approach. Magn Reson Imaging. 2025; 118 110339. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.mri.2025.110339\u003c/span\u003e\u003cspan address=\"10.1016/j.mri.2025.110339\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. PMID: 39880177.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePrakash C, Clouse RE. Value of extended recording time with wireless pH monitoring in evaluating gastroesophageal reflux disease. CLIN GASTROENTEROL H. 2005-04-01; 3 (4): 329\u0026thinsp;\u0026ndash;\u0026thinsp;34. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/s1542-3565(05)00021-2\u003c/span\u003e\u003cspan address=\"10.1016/s1542-3565(05)00021-2\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. PMID: 15822037.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSouza RF, Huo X, Mittal V, Schuler CM, Carmack SW, Zhang HY, Zhang X, Yu C, Hormi-Carver K, Genta RM, Spechler SJ. Gastroesophageal reflux might cause esophagitis through a cytokine-mediated mechanism rather than caustic acid injury. Gastroenterology. 2009-11-01; 137 (5): 1776-84. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1053/j.gastro.2009.07.055\u003c/span\u003e\u003cspan address=\"10.1053/j.gastro.2009.07.055\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. PMID: 19660463.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMorozov S, Sentsova T, 1084:, LOCAL EXPRESSION OF PROINFLAMMATORY CYTOKINES GENES IN PATIENTS WITH GASTROESOPHAGEAL REFLUX DISEASE AND ITS ASSOCIATION WITH TYPES AND NUMBER OF GASTROESOPHAGEAL REFLUXES GASTROENTEROLOGY. 2022-05-01; 162 (7): S\u0026ndash;252. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/s0016-5085(22)60600-0\u003c/span\u003e\u003cspan address=\"10.1016/s0016-5085(22)60600-0\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKruse-Andersen S, Wallin L, Madsen T. Reflux patterns and related oesophageal motor activity in gastro-oesophageal reflux disease. GUT. 1990-06-01; 31 (6): 633-8. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1136/gut.31.6.633\u003c/span\u003e\u003cspan address=\"10.1136/gut.31.6.633\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. PMID: 2379866.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFarina R, Pennisi F, La Rosa M, Puglisi C, Mazzone G, Riva G, Foti PV, Ettorre GC. Contrast-enhanced colour-Doppler sonography versus pH-metry in the diagnosis of gastro-oesophageal reflux in children. RADIOL MED. 2008-06-01; 113 (4): 591-8. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s11547-008-0267-4\u003c/span\u003e\u003cspan address=\"10.1007/s11547-008-0267-4\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. PMID: 18478190.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang JY, Luo Y, Wang WY, Zheng SC, He L, Xie CY, Peng L. Contrast-enhanced ultrasound using SonoVue mixed with oral gastrointestinal contrast agent to evaluate esophageal hiatal hernia: Report of three cases and a literature review. World J Clin Cases. 2021-04-16; 9 (11): 2679\u0026ndash;2687. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.12998/wjcc.v9.i11.2679\u003c/span\u003e\u003cspan address=\"10.12998/wjcc.v9.i11.2679\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. PMID: 33889636.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFranz\u0026eacute;n T, Tibbling L. Is the severity of gastroesophageal reflux dependent on hiatus hernia size? WORLD J GASTROENTERO. 2014-02-14; 20 (6): 1582-4. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3748/wjg.v20.i6.1582\u003c/span\u003e\u003cspan address=\"10.3748/wjg.v20.i6.1582\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. PMID: 24587634.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePiccialli V, Sciandrone M. Nonlinear optimization and support vector machines ANN OPER RES. 2022-04-14; \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s10479-022-04655-x\u003c/span\u003e\u003cspan address=\"10.1007/s10479-022-04655-x\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"bmc-gastroenterology","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"bmge","sideBox":"Learn more about [BMC Gastroenterology](http://bmcgastroenterol.biomedcentral.com/)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/bmge/default.aspx","title":"BMC Gastroenterology","twitterHandle":"BMC_series","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"em","reportingPortfolio":"BMC Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"Reflux esophagitis, Contrast-enhanced ultrasound, Machine learning","lastPublishedDoi":"10.21203/rs.3.rs-8494694/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8494694/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eBACKGROUND\u003c/h2\u003e \u003cp\u003eThe incidence of reflux esophagitis (RE) is increasing worldwide. RE can cause heartburn, acid reflux, and swallowing discomfort, thereby diminishing quality of life, with potential progression to Barrett's esophagus or esophageal adenocarcinoma. Gastroscopy for the diagnosis and grading of RE can be uncomfortable, and other modalities lack predictive accuracy. Thus, highly non-invasive diagnostic methods for RE are needed.\u003c/p\u003e\u003ch2\u003eAIM\u003c/h2\u003e \u003cp\u003eTo explore the value of contrast-enhanced ultrasound (CEUS) examination for predicting the progression of RE.\u003c/p\u003e\u003ch2\u003eMETHODS\u003c/h2\u003e \u003cp\u003eWe retrospectively selected 296 patients with RE who were hospitalized at West China Longquanyi Hospital of Sichuan University from February 2019 to January 2023. Basic clinical characteristics such as age and gender were collected. CEUS was used to measure the diameter of the esophageal hiatus, frequency of reflux within 5 min, duration of reflux within 5 min, and length of the abdominal esophagus. According to the Los Angeles (LA) classification, the patients were divided into three groups: 92 patients with \u0026ldquo;LA-A\u0026rdquo; grade (Group A), 85 with \u0026ldquo;LA-B\u0026rdquo; grade (Group B), and 20 with \u0026ldquo;LA-C/D\u0026rdquo; grade (Group C). Group differences in clinical characteristics were analyzed. Key factors influencing the progression of RE were analyzed using ordered logistic regression, random forest, and support vector machine (SVM). Statistical methods included paired the \u003cem\u003et\u003c/em\u003e-test, Mann\u0026ndash;Whitney U test, Chi-square test, Kruskal\u0026ndash;Wallis H test, and one-way analysis of variance.\u003c/p\u003e\u003ch2\u003eRESULTS\u003c/h2\u003e \u003cp\u003eComparisons among the three groups showed statistically significant differences in the diameter of the esophageal hiatus and frequency and duration of reflux within 5 min. Pairwise comparisons revealed that the frequency of reflux within 5 min in group C was significantly higher than in groups A and B. Ordered logistic regression, random forest, and SVM analyses all indicated that the frequency and duration of reflux within 5 min were key factors influencing the progression of RE. The SVM analysis showed that the diameter of the esophageal hiatus was a key factor influencing the progression of RE.\u003c/p\u003e\u003ch2\u003eCONCLUSION\u003c/h2\u003e \u003cp\u003eCEUS can better predict RE progression. The number of reflux events within 5 min, duration of reflux within 5 min, and diameter of the esophageal hiatus are key factors influencing RE progression.\u003c/p\u003e","manuscriptTitle":"Risk prediction of progression of reflux esophagitis based on contrast-enhanced ultrasound technology and machine learning","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-02-17 09:23:35","doi":"10.21203/rs.3.rs-8494694/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"editorInvitedReview","content":"","date":"2026-02-22T17:33:05+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"289011203264198096931810078894438423836","date":"2026-02-22T17:21:42+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-02-18T12:34:49+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"328043763305807548090982398264258424349","date":"2026-02-17T17:53:18+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-02-14T00:48:58+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"178714095673259957920304499742571405115","date":"2026-02-11T22:28:11+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"231891430231441213852513799302007462468","date":"2026-02-11T09:07:49+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2026-02-11T08:42:02+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-02-10T04:41:07+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2026-01-21T07:02:47+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2026-01-21T00:07:45+00:00","index":"","fulltext":""},{"type":"submitted","content":"BMC Gastroenterology","date":"2026-01-21T00:02:24+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"bmc-gastroenterology","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"bmge","sideBox":"Learn more about [BMC Gastroenterology](http://bmcgastroenterol.biomedcentral.com/)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/bmge/default.aspx","title":"BMC Gastroenterology","twitterHandle":"BMC_series","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"em","reportingPortfolio":"BMC Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"d085de45-0584-4880-a185-4d992c322206","owner":[],"postedDate":"February 17th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[],"tags":[],"updatedAt":"2026-02-17T09:23:39+00:00","versionOfRecord":[],"versionCreatedAt":"2026-02-17 09:23:35","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8494694","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8494694","identity":"rs-8494694","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}