Predictive modeling for early detection of biliary atresia in infants with cholestasis: Insights from a machine
Xuting Chen1, Dongying Zhao1, Haochen Ji2
1Department of Neonatology, Xinhua Hospital, Shanghai JiaoTong University School of Medicine, Shanghai, China.
Insights
This study developed a new AI model to help diagnose biliary atresia (BA) in infants with cholestasis. The model accurately identifies BA, aiding early detection and improving infant health outcomes.
Area of Science:
- Pediatric Gastroenterology
- Medical Informatics
- Machine Learning in Healthcare
Background:
- Cholestasis in neonates and infants presents diagnostic challenges, particularly differentiating biliary atresia (BA) from other causes due to similar clinical signs.
- Early and accurate diagnosis of BA is crucial for timely intervention and preventing severe liver dysfunction.
Purpose of the Study:
- To develop and validate a screening model for prognosticating outcomes in cases of biliary atresia (BA) in infants.
- To introduce a novel wrapper feature selection model, bWFMVO-SVM-FS, for improved diagnostic accuracy.
Main Methods:
- A wrapper feature selection model, bWFMVO-SVM-FS, was developed, combining the water flow-based multi-verse optimizer (WFMVO) and support vector machine (SVM).
- The WFMVO algorithm was benchmarked against eleven other algorithms on IEEE CEC 2017 and 2022 datasets.
- The bWFMVO-SVM-FS model was applied to a dataset of 870 infants with cholestasis (BA or non-BA) from two major hospitals.
Main Results:
- The bWFMVO-SVM-FS model achieved a high predictive accuracy of 92.639% and specificity of 88.865%.
- Key features identified for early BA diagnosis include gamma-glutamyl transferase levels, triangular cord sign, weight, gallbladder appearance, and stool color.
- These identified features enhance the interpretability of the model for clinicians.
Conclusions:
- The developed bWFMVO-SVM-FS model demonstrates significant potential for accurately screening and diagnosing biliary atresia in infants.
- The identified clinical and biochemical markers provide valuable insights for early detection and clinical decision-making in pediatric cholestasis.
Abstract:
Cholestasis, characterized by the obstruction of bile flow, poses a significant concern in neonates and infants. It can result in jaundice, inadequate weight gain, and liver dysfunction. However, distinguishing between biliary atresia (BA) and non-biliary atresia in these young patients presenting with cholestasis poses a formidable challenge, given the similarity in their clinical manifestations. To this end, our study endeavors to construct a screening model aimed at prognosticating outcomes in cases of BA. Within this study, we introduce a wrapper feature selection model denoted as bWFMVO-SVM-FS, which amalgamates the water flow-based multi-verse optimizer (WFMVO) and support vector machine (SVM) technology. Initially, WFMVO is benchmarked against eleven state-of-the-art algorithms, with its efficiency in searching for optimized feature subsets within the model validated on IEEE CEC 2017 and IEEE CEC 2022 benchmark functions. Subsequently, the developed bWFMVO-SVM-FS model is employed to analyze a cohort of 870 consecutively registered cases of neonates and infants with cholestasis (diagnosed as either BA or non-BA) from Xinhua Hospital and Shanghai Children's Hospital, both affiliated with Shanghai Jiao Tong University. The results underscore the remarkable predictive capacity of the model, achieving an accuracy of 92.639 % and specificity of 88.865 %. Gamma-glutamyl transferase, triangular cord sign, weight, abnormal gallbladder, and stool color emerge as highly correlated with early symptoms in BA infants. Furthermore, leveraging these five significant features enhances the interpretability of the machine learning model's performance outcomes for medical professionals, thereby facilitating more effective clinical decision-making.
