1
|
Kimura N, Takahashi K, Setsu T, Horibata Y, Kaneko Y, Miyazaki H, Ogawa K, Kawata Y, Sakai N, Watanabe Y, Abe H, Kamimura H, Sakamaki A, Yokoo T, Kamimura K, Tsuchiya A, Terai S. Development and validation of machine learning model for predicting treatment responders in patients with primary biliary cholangitis. Hepatol Res 2024; 54:67-77. [PMID: 37691006 DOI: 10.1111/hepr.13966] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Submit a Manuscript] [Subscribe] [Scholar Register] [Received: 05/27/2023] [Revised: 08/24/2023] [Accepted: 08/31/2023] [Indexed: 09/12/2023]
Abstract
AIMS Ursodeoxycholic acid is the first-line treatment for primary biliary cholangitis, and treatment response is one of the factors predicting the outcome. To prescribe alternative therapies, clinicians might need additional information before deciphering the treatment response to ursodeoxycholic acid, contributing to a better patient prognosis. In this study, we developed and validated machine learning (ML) algorithms to predict treatment responses using pretreatment data. METHODS This multicenter cohort study included collecting datasets from two data samples. Data 1 included 245 patients from 18 hospitals for ML development, and was divided into (i) training and (ii) development sets. Data 2 (iii: test set) included 51 patients from our hospital for validation. An extreme gradient boosted tree predicted the treatment response in the ML model. The area under the curve was used to evaluate the efficacy of the algorithm. RESULTS Data 1 showed that patients complying with the Paris II treatment response had significantly lower serum alkaline phosphatase and total bilirubin levels than those who did not respond. Three factors, total bilirubin, total protein, and alanine aminotransferase levels were selected as essential variables for prediction. Data 2 showed that patients complying with the Paris II criteria had significantly high prothrombin time and low total bilirubin levels. The area under the curve of extreme gradient boosted tree was good for (ii) (0.811) and (iii) (0.856). CONCLUSIONS We demonstrated the efficacy of ML in predicting the treatment response for patients with primary biliary cholangitis. Early identification of cases requiring additional treatment with our novel ML model may improve prognosis.
Collapse
Affiliation(s)
- Naruhiro Kimura
- Division of Gastroenterology and Hepatology, Niigata University Graduate School of Medical and Dental Sciences, Niigata, Japan
| | - Kazuya Takahashi
- Division of Gastroenterology and Hepatology, Niigata University Graduate School of Medical and Dental Sciences, Niigata, Japan
| | - Toru Setsu
- Division of Gastroenterology and Hepatology, Niigata University Graduate School of Medical and Dental Sciences, Niigata, Japan
| | - Yusuke Horibata
- Division of Gastroenterology and Hepatology, Niigata University Graduate School of Medical and Dental Sciences, Niigata, Japan
| | - Yusuke Kaneko
- Division of Gastroenterology and Hepatology, Niigata University Graduate School of Medical and Dental Sciences, Niigata, Japan
| | - Haruka Miyazaki
- Division of Gastroenterology and Hepatology, Niigata University Graduate School of Medical and Dental Sciences, Niigata, Japan
| | - Kohei Ogawa
- Division of Gastroenterology and Hepatology, Niigata University Graduate School of Medical and Dental Sciences, Niigata, Japan
| | - Yuzo Kawata
- Division of Gastroenterology and Hepatology, Niigata University Graduate School of Medical and Dental Sciences, Niigata, Japan
| | - Norihiro Sakai
- Division of Gastroenterology and Hepatology, Niigata University Graduate School of Medical and Dental Sciences, Niigata, Japan
| | - Yusuke Watanabe
- Division of Gastroenterology and Hepatology, Niigata University Graduate School of Medical and Dental Sciences, Niigata, Japan
| | - Hiroyuki Abe
- Division of Gastroenterology and Hepatology, Niigata University Graduate School of Medical and Dental Sciences, Niigata, Japan
| | - Hiroteru Kamimura
- Division of Gastroenterology and Hepatology, Niigata University Graduate School of Medical and Dental Sciences, Niigata, Japan
| | - Akira Sakamaki
- Division of Gastroenterology and Hepatology, Niigata University Graduate School of Medical and Dental Sciences, Niigata, Japan
| | - Takeshi Yokoo
- Division of Gastroenterology and Hepatology, Niigata University Graduate School of Medical and Dental Sciences, Niigata, Japan
| | - Kenya Kamimura
- Division of Gastroenterology and Hepatology, Niigata University Graduate School of Medical and Dental Sciences, Niigata, Japan
| | - Atsunori Tsuchiya
- Division of Gastroenterology and Hepatology, Niigata University Graduate School of Medical and Dental Sciences, Niigata, Japan
| | - Shuji Terai
- Division of Gastroenterology and Hepatology, Niigata University Graduate School of Medical and Dental Sciences, Niigata, Japan
| |
Collapse
|
2
|
Kimura N, Takahashi K, Setsu T, Goto S, Miida S, Takeda N, Kojima Y, Arao Y, Hayashi K, Sakai N, Watanabe Y, Abe H, Kamimura H, Sakamaki A, Yokoo T, Kamimura K, Tsuchiya A, Terai S. Machine learning prediction model for treatment responders in patients with primary biliary cholangitis. JGH Open 2023; 7:431-438. [PMID: 37359114 PMCID: PMC10290270 DOI: 10.1002/jgh3.12915] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 01/09/2023] [Revised: 03/27/2023] [Accepted: 05/11/2023] [Indexed: 06/28/2023]
Abstract
Background and Aim Treatment response to ursodeoxycholic acid may predict the prognosis of patients with primary biliary cholangitis (PBC). Recent studies have suggested the benefits of using machine learning (ML) to forecast complex medical predictions. We aimed to predict treatment response in patients with PBC using ML and pretreatment data. Methods We conducted a single-center retrospective study and collected data from 194 patients with PBC who were followed up for at least 12 months after treatment initiation. Patient data were analyzed with five ML models, namely random forest, extreme gradient boosting (XGB), decision tree, naïve Bayes, or logistic regression, to predict treatment response using the Paris II criteria. The established models were assessed using an out-of-sample validation. The area under the curve (AUC) was used to evaluate the efficacy of each algorithm. Overall survival and liver-related deaths were analyzed using Kaplan-Meier analysis. Results Compared to logistic regression (AUC = 0.595, P = 0.0219, 0.031 models), ML analyses showed significantly high AUC in the random forest (AUC = 0.84) and XGB (AUC = 0.83) models; however, the AUC was not significantly high for decision tree (AUC = 0.633) or naïve Bayes (AUC = 0.584) models. Kaplan-Meier analysis showed significantly improved prognoses in patients predicted to achieve the Paris II criteria by XGB (log-rank = 0.005 and 0.007). Conclusion ML algorithms could improve treatment response prediction using pretreatment data, which could lead to better prognoses. In addition, the ML model using XGB could predict the prognosis of patients before treatment initiation.
Collapse
Affiliation(s)
- Naruhiro Kimura
- Division of Gastroenterology and HepatologyNiigata University Graduate School of Medical and Dental SciencesNiigataJapan
| | - Kazuya Takahashi
- Division of Gastroenterology and HepatologyNiigata University Graduate School of Medical and Dental SciencesNiigataJapan
| | - Toru Setsu
- Division of Gastroenterology and HepatologyNiigata University Graduate School of Medical and Dental SciencesNiigataJapan
| | - Shu Goto
- Division of Gastroenterology and HepatologyNiigata University Graduate School of Medical and Dental SciencesNiigataJapan
| | - Suguru Miida
- Division of Gastroenterology and HepatologyNiigata University Graduate School of Medical and Dental SciencesNiigataJapan
| | - Nobutaka Takeda
- Division of Gastroenterology and HepatologyNiigata University Graduate School of Medical and Dental SciencesNiigataJapan
| | - Yuichi Kojima
- Division of Gastroenterology and HepatologyNiigata University Graduate School of Medical and Dental SciencesNiigataJapan
| | - Yoshihisa Arao
- Division of Gastroenterology and HepatologyNiigata University Graduate School of Medical and Dental SciencesNiigataJapan
| | - Kazunao Hayashi
- Division of Gastroenterology and HepatologyNiigata University Graduate School of Medical and Dental SciencesNiigataJapan
| | - Norihiro Sakai
- Division of Gastroenterology and HepatologyNiigata University Graduate School of Medical and Dental SciencesNiigataJapan
| | - Yusuke Watanabe
- Division of Gastroenterology and HepatologyNiigata University Graduate School of Medical and Dental SciencesNiigataJapan
| | - Hiroyuki Abe
- Division of Gastroenterology and HepatologyNiigata University Graduate School of Medical and Dental SciencesNiigataJapan
| | - Hiroteru Kamimura
- Division of Gastroenterology and HepatologyNiigata University Graduate School of Medical and Dental SciencesNiigataJapan
| | - Akira Sakamaki
- Division of Gastroenterology and HepatologyNiigata University Graduate School of Medical and Dental SciencesNiigataJapan
| | - Takeshi Yokoo
- Division of Gastroenterology and HepatologyNiigata University Graduate School of Medical and Dental SciencesNiigataJapan
| | - Kenya Kamimura
- Division of Gastroenterology and HepatologyNiigata University Graduate School of Medical and Dental SciencesNiigataJapan
| | - Atsunori Tsuchiya
- Division of Gastroenterology and HepatologyNiigata University Graduate School of Medical and Dental SciencesNiigataJapan
| | - Shuji Terai
- Division of Gastroenterology and HepatologyNiigata University Graduate School of Medical and Dental SciencesNiigataJapan
| |
Collapse
|