Utility of leucocytes, inflammatory markers and pancreatic enzymes as indicators of gangrenous cholecystitis

An explainable predictive machine learning model of gangrenous cholecystitis based on clinical data: a retrospective single center study

BACKGROUND

Gangrenous cholecystitis (GC) is a serious clinical condition associated with high morbidity and mortality rates. Machine learning (ML) has significant potential in addressing the diverse characteristics of real data. We aim to develop an explainable and cost-effective predictive model for GC utilizing ML and Shapley Additive explanation (SHAP) algorithm.

RESULTS

This study included a total of 1006 patients with 26 clinical features. Through 5-fold CV, the best performing integrated learning model, XGBoost, was identified. The model was interpreted using SHAP to derive the feature subsets WBC, NLR, D-dimer, Gallbladder width, Fibrinogen, Gallbladder wallness, Hypokalemia or hyponatremia, these subsets comprised the final diagnostic prediction model.

CONCLUSIONS

The study developed a explainable predictive tool for GC at an early stage. This could assist doctors to make quick surgical intervention decisions and perform surgery on patients with GC as soon as possible.

Predictive Factors for the Development of Gallbladder Necrosis

Introduction Acute cholecystitis is a common complication of gallstone disease. Likewise, gallbladder necrosis is a complication of cholecystitis associated with higher risks of morbidity and mortality. Identification of risk factors which portend to gallbladder necrosis is key in prioritizing the management of higher-risk patients. This study aimed to identify such factors that predict the development of gallbladder necrosis. Method A retrospective review of all patients undergoing emergency cholecystectomy in a tertiary hospital over a two-year period was performed. Gallbladder necrosis was diagnosed on histopathological examination of operative specimens. Multivariable logistic regression was performed to determine risk factors for gallbladder necrosis. Results A total of 163 patients underwent acute cholecystectomy and 43 (26%) had proven gallbladder necrosis. Multivariable analysis demonstrated that elevated white cell count (WCC) (OR 1.122, 95%CI 1.031-1.221, p=0.007), elevated C-reactive protein (CRP) (OR 1.004, 95%CI 1.001-1.008, p=0.022) and positive smoking status (OR 5.724, 95%CI 1.323-24.754, p=0.020) were independently predictive of gallbladder necrosis. Notably, advancing age, elevated BMI, diabetes mellitus or American Society of Anesthesiologists (ASA) grade were not found to be associated with developing necrosis. Conclusion Patients at risk of gallbladder necrosis include those with higher WCC, CRP, and active smokers. Given the increased potential complications, these risk factors should be identified early in the management of those admitted with gallstone disease to ensure such patients receive aggressive medical therapy alongside timely and guided surgical intervention.

Value of monocyte distribution width for predicting severe cholecystitis: a retrospective cohort study

OBJECTIVES

Acute cholecystitis is a gallbladder inflammation, and the Tokyo Guidelines 2018 (TG18) can be used to predict its presence and severity with high sensitivity and specificity. However, TG18 grading require the collection of excessive parameters. Monocyte distribution width (MDW) is a parameter used to detect sepsis early. Therefore, we investigated the correlation between MDW and cholecystitis severity.

METHODS

We conducted a retrospective study of patients with cholecystitis admitted to our hospital from November 1, 2020, to August 31, 2021. The primary outcome was severe cholecystitis analyzed as a composite of intensive care unit (ICU) admission and mortality. The secondary outcomes were length of hospital stay, ICU stay, and TG18 grade.

RESULTS

A total of 331 patients with cholecystitis were enrolled in this study. The average MDWs for TG18 grades 1, 2, and 3 were 20.21 ± 3.99, 20.34 ± 3.68, and 25.77 ± 6.61, respectively. For patients with severe cholecystitis, the average MDW was 25.42 ± 6.83. Using the Youden J statistic, we set a cutoff MDW of 21.6. Multivariate logistic regression revealed that patients with an MDW≥21.6 had a higher risk of severe cholecystitis (odds ratio=4.94; 95 % CI, 1.71-14.21; p=0.003). The Cox model revealed that patients with an MDW≥21.6 were more likely to have a prolonged hospital stay.

CONCLUSIONS

MDW is a reliable indicator of severe cholecystitis and prolonged length of stay. Additional MDW testing and a complete blood count may provide simple information for predicting severe cholecystitis early.

The Use of Systemic Inflammatory Response Syndrome (SIRS) and Elevated Liver Enzymes as Predictive Factors of Gangrenous Cholecystitis: A Case Report

Gangrenous cholecystitis is a severe complication of acute cholecystitis. It is often found incidentally during laparoscopic cholecystectomy or during conversion to open surgery and diagnosed with subsequent pathological analysis. While intraoperative diagnosis is typically through direct visualization of the gallbladder, specific diagnostic modalities may guide physicians toward an earlier diagnosis. Surgical intervention and a more aggressive approach are often needed to prevent the advancement of the disease and its catastrophic complications. This case report illustrates the distinct risk factors predisposing a patient to develop gangrenous cholecystitis. Comorbidities such as hypertension, coronary artery disease, age, the relevance of the SIRS criteria, and elevated liver enzymes are explored as predictive factors in a patient with gangrenous cholecystitis.

Prognostic clinical indexes for prediction of acute gangrenous cholecystitis and acute purulent cholecystitis

Background

Preoperative prediction of severe cholecystitis (SC), including acute gangrenous cholecystitis (AGC) and acute purulent cholecystitis (APC), as opposed to acute exacerbation of chronic cholecystitis (ACC), is of great significance, as SC is associated with high mortality rate.

Methods

In this study, we retrospectively investigated medical records of 114 cholecystitis patients, treated in Shanghai No. 6 People’s Hospital from February 2009 to July 2020. Gallbladder wall thickness (GBWT), indexes of blood routine examination, including white blood cell (WBC), alkaline phosphatase (ALP), the percentage of neutrophil, alanine transaminase (ALT), aspartate aminotransferase (AST), fibrinogen (FIB), gamma-glutamyl transferase, prothrombin time and total bilirubin were evaluated. One-way analysis of variance (ANOVA) was used to evaluate significant differences between a certain kind of SC and ACC to select a prediction index for each kind of SC. Receiver operating characteristic (ROC) curve analysis was conducted to identify the prediction effectiveness of these indexes and their optimal cut-off values.

Results

Higher WBC and lower ALP were associated with AGC diagnosis (P < 0.05). Higher percentage of neutrophils was indicative of APC and AGC, while higher GBWT was significantly associated with APC diagnosis (P < 0.05) The optimal cut-off values for these indexes were established at 11.1*109/L (OR: 5.333, 95% CI 2.576–10.68, P < 0.0001, sensitivity: 72.73%, specificity: 66.67%), 79.75% (OR: 5.735, 95% CI 2.749–12.05, P < 0.0001, sensitivity: 77.92%, specificity: 61.9%) and 5.5 mm (OR: 22, 95% CI 4.757–83.42, P < 0.0001, sensitivity: 78.57%, specificity: 85.71%), respectively.

Conclusion

We established a predictive model for the differentiations of APC and AGC from ACC using clinical indexes, such as GBWT, the percentage of neutrophil and WBC, and determined cut-off values for these indexes based on ROC curves. Index values exceeding these cut-off values will allow to diagnose patients as APC and AGC, as opposed to a diagnosis of ACC.

The predictive significance of neutrophil-to-lymphocyte ratio in cholecystitis: a systematic review and meta-analysis

PURPOSE

The aim of this review was to examine whether neutrophil-to-lymphocyte ratio (NLR) can predict the presence of cholecystitis and distinguish between simple and severe cholecystitis.

METHODS

A systematic literature search was performed. Risk of bias was assessed using the Newcastle-Ottawa Scale. Random effects model was used to calculate mean difference (MD) in two situations: (a) no cholecystitis versus cholecystitis and (b) simple versus severe cholecystitis. Receiver operating characteristic (ROC) curve analysis was performed to determine cut-off values of NLR for the above situations.

RESULTS

Ten retrospective studies comprising of 2827 patients were included. Three hundred twenty-seven had no cholecystitis, 2100 had simple cholecystitis and the remaining 400 had severe cholecystitis. NLR was significantly higher in acute cholecystitis compared to "no cholecystitis" (MD = 8.05 (95% CI 7.71-8.38), p < 0.01) and in severe cholecystitis when compared with simple cholecystitis (MD = 3.14 (95% CI 1.26-5.02), p < 0.01). For patients with cholecystitis compared to those without cholecystitis, an NLR cut-off value of 2.98 was identified (AUC = 0.90). Logistic regression analysis confirmed an NLR > 2.9 was an independent predictor of cholecystitis (OR 36.0, p = 0.006). In simple versus severe cholecystitis, an NLR cut-off value of 8.5 was identified (AUC = 0.73). Binary logistic regression analysis suggested an NLR > 8.5 was not an independent predictor of severe cholecystitis (OR 6.5 p = 0.090).

CONCLUSION

NLR is significantly higher in patients with cholecystitis of any severity compared to patients without cholecystitis. Moreover, NLR can predict acute cholecystitis. However, NLR cannot predict the severity of disease due to inadequately powered studies. Future research is required.

Plasma Procalcitonin Is Useful for Predicting the Severity of Acute Cholecystitis

Background

Acute cholecystitis is a common complication of cholelithiasis. Delayed diagnosis may constitute morbidity and mortality; therefore, early diagnosis and determining the severity of acute cholecystitis are crucial.

Objectives

This study aimed to determine the validity of blood procalcitonin (PCT) levels in assessing the severity of acute cholecystitis.

Methods

The Emergency Department (ED) patients diagnosed as acute cholecystitis were included in the study. Patients were allocated into three severity grades according to the Tokyo Guidelines 2013. PCT level was measured after the clinical and radiological diagnosis of acute cholecystitis in the ED.

Results

Ninety-five patients diagnosed with acute cholecystitis, among them 48 of were male. Forty patients (42.1%) were allocated to grade 1, 19 (20%) to grade 2, and 36 (37.9%) to grade 3. The median values of PCT were 0.104 ng/ml, 0.353 ng/ml, and 1.466 ng/ml for grade 1, 2, and 3 patients, respectively (p < 0.001).

Conclusion

Blood procalcitonin levels can be used to determine the severity of acute cholecystitis effectively.