Triglyceride and glucose (TyG) index is an effective biomarker to identify severe acute pancreatitis

Application Value of the Automated Machine Learning Model Based on Modified Computed Tomography Severity Index Combined With Serological Indicators in the Early Prediction of Severe Acute Pancreatitis

BACKGROUND AND AIMS

Machine learning (ML) algorithms are widely applied in building models of medicine due to their powerful studying and generalizing ability. To assess the value of the Modified Computed Tomography Severity Index (MCTSI) combined with serological indicators for early prediction of severe acute pancreatitis (SAP) by automated ML (AutoML).

PATIENTS AND METHODS

The clinical data, of the patients with acute pancreatitis (AP) hospitalized in Hospital 1 and hospital 2 from January 2017 to December 2021, were retrospectively analyzed. Serological indicators within 24 hours of admission were collected. MCTSI score was completed by noncontrast computed tomography within 24 hours of admission. Data from the hospital 1 were adopted for training, and data from the hospital 2 were adopted for external validation. The diagnosis of AP and SAP was based on the 2012 revised Atlanta classification of AP. Models were built using traditional logistic regression and AutoML analysis with 4 types of algorithms. The performance of models was evaluated by the receiver operating characteristic curve, the calibration curve, and the decision curve analysis based on logistic regression and decision curve analysis, feature importance, SHapley Additive exPlanation Plot, and Local Interpretable Model Agnostic Explanation based on AutoML.

RESULTS

A total of 499 patients were used to develop the models in the training data set. An independent data set of 201 patients was used to test the models. The model developed by the Deep Neural Net (DL) outperformed other models with an area under the receiver operating characteristic curve (areas under the curve) of 0.907 in the test set. Furthermore, among these AutoML models, the DL and gradient boosting machine models achieved the highest sensitivity values, both exceeding 0.800.

CONCLUSION

The AutoML model based on the MCTSI score combined with serological indicators has good predictive value for SAP in the early stage.

A highly sensitive optical fiber sensor enables rapid triglycerides-specific detection and measurement at different temperatures using convolutional neural networks

For specific recognition and sensitive detection of triglycerides (TGs), an optical fiber sensor (OFS) based on an enhanced core diameter mismatch was proposed. The sensitivity of the sensor is significantly increased due to the repetitive excitation of the higher-order cladding modes. A technique for immobilizing lipase using covalent binding technology was presented and demonstrated by Fourier transform infrared (FTIR) spectroscopy and scanning electron microscopy. The interference dip of the sensor was shifted due to TGs being hydrolyzed in the presence of lipase. The sensor shows an optimal response within 3 min and exhibits a high sensitivity of 0.9933 nm/(mg/ml) and a limit of detection of 0.0822 mg/ml in the concentration range 0-8 mg/ml at a temperature of 37 °C and a pH of 7.4. The response of the sensor to TGs concentration at different temperatures and pH was investigated. The reproducibility, reusability, and stability of the proposed sensor were tested and verified experimentally. The biosensor is highly specific for TGs and unaffected by many other interfering substances. Further, the measurement of TGs concentration at different temperatures was realized. This method provides a new way to detect TGs rapidly and reliably and has potential applications in medical research and clinical diagnosis.

Risk factors for post‑retrograde cholangiopancreatography pancreatitis in patients with common bile duct stones: A meta‑analysis

Endoscopic retrograde cholangiopancreatography (ERCP) has become a common treatment method for common bile duct stones. However, ERCP is also associated with a high risk of post-endoscopic retrograde cholangiopancreatography pancreatitis (PEP). Identification of risk factors is essential for reducing the incidence of PEP. The present study aimed to summarize the risk factors for PEP by performing a meta-analysis. Therefore, studies published between 2000 and 2022 were screened in PubMed, Cochrane Library, Embase, Web of Science, China National Knowledge Infrastructure, Wanfang Digital Periodicals and the Weipu Database, with no language restrictions. Newcastle-Ottawa Scale was used to assess the quality of the included studies. Stata 17.0 software was utilized for the meta-analysis of 14 possible risk factors. Overall, 15 high-quality studies were included into the present meta-analysis. The results showed that female [odds ratio (OR), 1.42; 95% CI, 1.23-1.64), age <60 years (OR, 1.53; 95% CI, 1.06-2.21), difficult intubation (OR, 4.87; 95% CI, 2.73-8.68), ≥3 cannulation attempts (OR, 9.64; 95% CI, 4.16-22.35), cannulation time ≥10 min (OR, 2.37; 95% CI, 1.67-3.35), history of pancreatitis (OR, 2.95; 95% CI, 1.06-5.51), pancreatic duct visualization (OR, 3.63; 95% CI, 2.47-5.34) and sphincter of Oddi dysfunction (OR, 5.72; 95% CI, 1.80-18.24) are potential risk factors for PEP (P<0.05). In conclusion, the present meta-analysis suggests that PEP can be affected by several risk factors, particularly the technique-related factors such as the frequency and time of cannulation. Therefore, effective precautions should be taken as early as possible to reduce the incidence of PEP.

A New Risk Score to Predict Intensive Care Unit Admission for Patients with Acute Pancreatitis 48 Hours After Admission: Multicenter Study

AIMS

The objective of this study was to develop and validate an easy-to-use risk score (APRS) to predict which patients with acute pancreatitis (AP) will need intensive care unit (ICU) treatment within 48 h post-hospitalization on the basis of the ubiquitously available clinical records.

METHODS

Patients with acute pancreatitis were retrospectively included from three independent institutions (RM cohort, 5280; TJ cohort, 262; SN cohort, 196), with 56 candidate variables collected within 48 h post-hospitalization. The RM cohort was randomly divided into a training set (N = 4220) and a test set (N = 1060). The most predictive features were extracted by LASSO from the RM cohort and entered into multivariate analysis. APRS was constructed using the coefficients of the statistically significant variables weighted by the multivariable logistic regression model. The APRS was validated by RM, TJ, and SN cohorts. The C-statistic was employed to evaluate the APRS's discrimination. DeLong test was used to compare area under the receiver operating characteristic curve (AUC) differences.

RESULTS

A total of 5738 patients with AP were enrolled. Eleven variables were selected by LASSO and entered into multivariate analysis. APRS was inferred using the above five factors (pleural effusion, ALT/AST, ALB/GLB, urea, and glucose) weighted by their regression coefficients in the multivariable logistic regression model. The C-statistics of APRS were 0.905 (95% CI 0.82-0.98) and 0.889 (95% CI 0.81-0.96) in RM and TJ validation. An online APRS web-based calculator was constructed to assist the clinician to earlier assess the clinical outcomes of patients with AP.

CONCLUSION

APRS could effectively stratify patients with AP into high and low risk of ICU admission within 48 h post-hospitalization, offering clinical value in directing management and personalize therapeutic selection for patients with AP.

A population-based cohort study on risk factors for acute pancreatitis: A comparison by age group

BACKGROUND

/objectives: Acute pancreatitis (AP) is an acute inflammatory disorder that can occur in all age groups. The risk of AP has been shown to increase with age. However, no study has compared risk factors for AP according to age group yet. Thus, the aim of this study was to perform such comparison.

METHODS

Clinical data from individuals 20 years of age and older who received a health examination arranged by the Korean national health insurance program in 2009 (n = 4,238,822) were used. First-attack AP was identified using claims data from baseline to December 2018. Incidence and risk factors of AP were analyzed for young (20-39 years old), middle-aged (40-64 years old), and old (over 65 years old) groups.

RESULTS

Incidences of AP in young, middle-aged, and old groups were 16.30, 27.85, and 57.19 per 100,000 person-years, respectively. Smoking, alcohol drinking, diabetes, gallstone, and chronic pancreatitis were associated with increased risk of AP in all age groups. Meanwhile, male, older age, and higher waist circumference were associated with increased risk of AP in middle-aged and old groups. In young and middle-aged groups, risk of AP was increased in the presence of hypertension and dyslipidemia. However, high income was associated with decreased risk of AP in these groups.

CONCLUSIONS

In this population-based cohort study, incidences and risk factors for AP differed according to age group. Thus, a tailored strategy might be needed to prevent AP according to age group.

DIFFERENTIAL DIAGNOSIS BETWEEN BILIARY AND NONBILIARY ACUTE PANCREATITIS: WHAT IS THE IMPORTANCE OF LABORATORY TESTS?

BACKGROUND

The differential diagnosis of the causal factors of acute pancreatitis is fundamental for its clinical follow-up, becoming relevant to establishing laboratory criteria that elucidate the difference between biliary and nonbiliary causes.

AIM

The aim of this study was to establish criteria based on laboratory tests for the differential diagnosis between acute pancreatitis of biliary and nonbiliary causes and to identify laboratory tests with sufficient sensitivity to propose the creation of an algorithm for differential diagnosis between the causes.

METHODS

The research consisted of observational analysis, with a cross-sectional design of laboratory tests of two groups of patients with acute pancreatitis: group A: nonbiliary cause and group B: biliary cause. Hematocrit, white blood cell count, lactate dehydrogenase, glucose, lipase, amylase, total bilirubin, oxalacetic transaminase, pyruvic transaminase, gamma-glutamyltransferase, and alkaline phosphatase were investigated. Data were submitted to nonparametric tests and receiver operating characteristics.

RESULTS

Hematocrit values, number of leukocytes, lactate dehydrogenase, and glucose showed no significant difference between the groups (p>0.1). Lipase, amylase, total bilirubin, oxalacetic transaminase, pyruvic transaminase, gamma-glutamyltransferase, and alkaline phosphatase values showed a significant difference between groups (p<0.05). The oxalacetic transaminase, pyruvic transaminase, and alkaline phosphatase tests were most sensitive in determining the biliary cause, allowing the establishment of a cutoff point by the receiver operating characteristic test: pyruvic transaminase: 123.0 U/L (sensitivity: 69.2%; specificity: 81.5%), oxalacetic transaminase: 123.5 U/L (sensitivity: 57.3%; specificity: 78.8%), and alkaline phosphatase: 126.5 U/L (sensitivity: 66.1%; specificity: 69.4%), from which the probability of a correct answer increases.

CONCLUSION

It was possible to establish criteria based on laboratory tests for the differential diagnosis between acute pancreatitis of biliary and nonbiliary origin; however, the tests did not show enough sensitivity to propose the creation of an algorithm for differential diagnosis between the same causes.

Automated Machine Learning for the Early Prediction of the Severity of Acute Pancreatitis in Hospitals

Background

Machine learning (ML) algorithms are widely applied in building models of medicine due to their powerful studying and generalizing ability. This study aims to explore different ML models for early identification of severe acute pancreatitis (SAP) among patients hospitalized for acute pancreatitis.

Methods

This retrospective study enrolled patients with acute pancreatitis (AP) from multiple centers. Data from the First Affiliated Hospital and Changshu No. 1 Hospital of Soochow University were adopted for training and internal validation, and data from the Second Affiliated Hospital of Soochow University were adopted for external validation from January 2017 to December 2021. The diagnosis of AP and SAP was based on the 2012 revised Atlanta classification of acute pancreatitis. Models were built using traditional logistic regression (LR) and automated machine learning (AutoML) analysis with five types of algorithms. The performance of models was evaluated by the receiver operating characteristic (ROC) curve, the calibration curve, and the decision curve analysis (DCA) based on LR and feature importance, SHapley Additive exPlanation (SHAP) Plot, and Local Interpretable Model Agnostic Explanation (LIME) based on AutoML.

Results

A total of 1,012 patients were included in this study to develop the AutoML models in the training/validation dataset. An independent dataset of 212 patients was used to test the models. The model developed by the gradient boost machine (GBM) outperformed other models with an area under the ROC curve (AUC) of 0.937 in the validation set and an AUC of 0.945 in the test set. Furthermore, the GBM model achieved the highest sensitivity value of 0.583 among these AutoML models. The model developed by eXtreme Gradient Boosting (XGBoost) achieved the highest specificity value of 0.980 and the highest accuracy of 0.958 in the test set.

Conclusions

The AutoML model based on the GBM algorithm for early prediction of SAP showed evident clinical practicability.

A Machine Learning-Modified Novel Nomogram to Predict Perioperative Blood Transfusion of Total Gastrectomy for Gastric Cancer

Background

Perioperative blood transfusion reserves are limited, and the outcome of blood transfusion remains unclear. Therefore, it is important to prepare plans for perioperative blood transfusions. This study aimed to establish a risk assessment model to guide clinical patient management.

Methods

This retrospective comparative study involving 513 patients who had total gastrectomy (TG) between January 2018 and January 2021 was conducted using propensity score matching (PSM). The influencing factors were explored by logistic regression, correlation analysis, and machine learning; then, a nomogram was established.

Results

After assessment of the importance of factors through machine learning, blood loss, preoperative controlling nutritional status (CONUT), hemoglobin (Hb), and the triglyceride–glucose (TyG) index were considered as the modified transfusion-related factors. The modified model was not considered to be different from the original model in terms of performance, but is simpler. A nomogram was created, with a C-index of 0.834, and the decision curve analysis (DCA) demonstrated good clinical benefit.

Conclusions

A nomogram was established and modified with machine learning, which suggests the importance of the patient’s integral condition. This emphasizes that caution should be exercised regarding transfusions, and, if necessary, preoperative nutritional interventions or delayed surgery should be implemented for safety.

Hypertriglyceridemia is a Risk Factor for Fever in Early Acute Non-biliary Pancreatitis

Background/Aims

Fever is a common symptom of acute pancreatitis (AP). This study examined the factors associated with fever due to pancreatic inflammation in the early stages of non-biliary AP.

Methods

This study analyzed the AP database from Kangwon National University Hospital from January 2018 until April 2021 and identified patients who developed fever within 1 week of hospitalization. Patients with gallstone pancreatitis, pseudocyst, walled-off necrosis, chronic pancreatitis, bacteremia, and other site infections were excluded. The febrile group was compared with the afebrile group.

Results

One hundred and fifty-two patients were analyzed, and fever was diagnosed in 79 patients (52.0%). Based on multivariate analysis, fever was positively correlated with hypertriglyceridemia-induced AP (OR 16.8, 95% CI 4.0-70.7, p<0.01) and computed tomography severity index (OR 1.7, 95% CI 1.2-2.6, p<0.01). Recurrent AP was negatively associated with fever (OR 0.3, 95% CI 0.1-0.8, p=0.01). Fever was more frequent in patients with higher initial serum triglyceride (TG) levels than those with lower levels (TG <200 mg/dL; 35.1%, 200≤TG<500 mg/dL; 42.3%, TG ≥500 mg/dL; 88.6%, p<0.01).

Conclusions

Hypertriglyceridemia and CT severity index are associated with fever in early non-biliary AP.

Triglyceride-Glucose Index Is Not Associated With Lung Cancer Risk: A Prospective Cohort Study in the UK Biobank

BACKGROUND

The triglyceride-glucose (TyG) index is a practical substitute measure for insulin resistance (IR). The relationship between IR and lung cancer has been examined in previous studies; however, the findings have been controversial. In addition, previous studies had small sample sizes. Thus, we systematically examined the association between IR and lung cancer risk based on the UK Biobank with IR measured by the TyG index and further examined the interactions and joint effects for lung cancer.

METHODS

A total of 324,334 individuals free from any type of cancer at recruitment from the UK Biobank prospective cohort were included. The participants were predominantly between 40 and 70 years old. After adjusting for relevant confounders, multivariable Cox regression models were constructed to examine the relationship between the TyG index and the risk of lung cancer. We also checked the interactions and joint effects using a polygenic risk score (PRS) for lung cancer.

RESULTS

During a median follow-up of 9 years, 1,593 individuals were diagnosed with lung cancer. No association was found between the TyG index and lung cancer risk after multivariate Cox regression analysis adjusted for risk factors (hazard ratio: 0.91; 95% confidence interval: 0.64-1.18). No interaction or joint effects for genetic risk and the TyG index were observed.

CONCLUSION

The TyG index was not associated with the risk of lung cancer. Our results provide limited evidence that IR is not correlated with the risk of lung cancer.

Dual-fiber optic bioprobe system for triglyceride detection using surface plasmon resonance sensing and lipase-immobilized magnetic bead hydrolysis

The rapid and accurate detection of triglyceride (TG) plays a valuable role in the prevention and control of dyslipidemia. In this paper, a novel method for TG detection using a dual-fiber optic bioprobe system, which can accurately detect different levels of TG concentration in serum, is proposed. The system employs disposable microprobe-type fiber optic surface plasmon resonance (SPR) biosensors for signal acquisition, providing high stability and portability while avoiding cross-contamination caused by repeated use. The proposed biosensor with a high sensitivity of 1.25 nm/(mg/mL) for TG detection in serum and a tiny diameter of 125 μm, was fabricated using a novel multimode fiber-single-mode fiber-reflector (MSR) structure, which has been scarcely ever reported to the best of our knowledge. In the process of TG detection, lipase-immobilized magnetic beads were introduced to specifically hydrolyze TG, and the relationship between the TG content and the SPR differential signal was obtained from dual-fiber optic bioprobe measurements of the TG sample before and after hydrolysis. The proposed method achieved TG detection in the concentration range of 0-8 mg/mL (including healthy and unhealthy levels of TG concentration in the human body). Additionally, the miniaturized fiber optic biosensors used in this work have the advantages of low sample consumption, high sensitivity, simple operation, label-free measurement, high selectivity, and low cost. This method provides a new pathway for rapid and reliable TG detection and has potential applications in medical research and clinical diagnosis.

Metabolomic-based clinical studies and murine models for acute pancreatitis disease: A review

Acute pancreatitis (AP) is one of the most common gastroenterological disorders requiring hospitalization and is associated with substantial morbidity and mortality. Metabolomics nowadays not only help us to understand cellular metabolism to a degree that was not previously obtainable, but also to reveal the importance of the metabolites in physiological control, disease onset and development. An in-depth understanding of metabolic phenotyping would be therefore crucial for accurate diagnosis, prognosis and precise treatment of AP. In this review, we summarized and addressed the metabolomics design and workflow in AP studies, as well as the results and analysis of the in-depth of research. Based on the metabolic profiling work in both clinical populations and experimental AP models, we described the metabolites with potential utility as biomarkers and the correlation between the altered metabolites and AP status. Moreover, the disturbed metabolic pathways correlated with biological function were discussed in the end. A practical understanding of current and emerging metabolomic approaches applicable to AP and use of the metabolite information presented will aid in designing robust metabolomics and biological experiments that result in identification of unique biomarkers and mechanisms, and ultimately enhanced clinical decision-making.