โก Quick Summary
This study developed a machine learning-based prediction model for portal vein system thrombosis (PVST) occurring within 3 months after splenectomy in patients with cirrhosis. The AdaBoost model demonstrated a mean area under the receiver operating characteristic curve (AUROC) of 0.72, indicating its potential for early risk assessment.
๐ Key Details
- ๐ Dataset: 392 patients with cirrhosis who underwent splenectomy
- ๐งฉ Features used: 37 candidate predictors including demographic, clinical, and laboratory data
- โ๏ธ Technology: Five machine learning algorithms, with a focus on AdaBoost
- ๐ Performance: AdaBoost model achieved AUROC of 0.72 (95% CI 0.60 to 0.84)
๐ Key Takeaways
- ๐ PVST is a significant complication after splenectomy in cirrhosis patients.
- ๐ก Early prediction of PVST is crucial for timely intervention.
- ๐ค Machine learning can enhance risk assessment in clinical settings.
- ๐ฅ The AdaBoost model showed the best performance among tested algorithms.
- ๐ SHAP analysis provided insights into key predictors of PVST risk.
- ๐ Important predictors included albumin levels, platelet count, and D-dimer levels.
- ๐ Study conducted at the Second Affiliated Hospital of Xi’an Jiaotong University.
- ๐๏ธ Follow-up period: 3 months post-splenectomy.
๐ Background
Portal vein system thrombosis (PVST) is a common and potentially life-threatening complication that can arise after splenectomy, particularly in patients suffering from cirrhosis and portal hypertension. The ability to predict PVST early can significantly improve patient outcomes by allowing for timely interventions. This study aims to leverage machine learning techniques to create a predictive model that can assess the risk of PVST within a critical timeframe.
๐๏ธ Study
Conducted at the Second Affiliated Hospital of Xi’an Jiaotong University, this study enrolled 392 patients who underwent splenectomy between July 2016 and December 2022. The researchers collected a comprehensive dataset comprising 37 candidate predictors derived from clinical data, including demographic characteristics, disease features, imaging results, and laboratory values. After thorough analysis, eight key predictors were selected for the model construction.
๐ Results
Over the 3-month follow-up period, 144 patients (approximately 36.73%) developed PVST. The AdaBoost model emerged as the most effective, achieving a mean AUROC of 0.72, indicating moderate discriminative ability. Key predictors identified through SHAP analysis included albumin levels, platelet count, diameter of the portal vein, ฮณ-glutamyl transferase levels, length of hospital stay, activated partial thromboplastin time, D-dimer levels, and a history of preoperative gastrointestinal bleeding.
๐ Impact and Implications
The findings from this study highlight the potential of machine learning models in clinical practice, particularly for predicting PVST in patients with cirrhosis. By providing a transparent and explainable risk assessment, healthcare providers can implement targeted preventive measures, ultimately reducing the incidence of PVST and improving patient care. This approach could pave the way for more personalized medicine strategies in managing complications associated with cirrhosis.
๐ฎ Conclusion
This study underscores the transformative potential of machine learning in predicting PVST risk post-splenectomy. The AdaBoost model, with its moderate discriminative ability, offers a promising tool for clinicians to identify high-risk patients and tailor interventions accordingly. As we continue to explore the integration of AI in healthcare, further research is essential to refine these models and enhance their applicability in diverse clinical settings.
๐ฌ Your comments
What are your thoughts on the use of machine learning for predicting complications in cirrhosis patients? We would love to hear your insights! ๐ฌ Join the conversation in the comments below or connect with us on social media:
Development of an explainable prediction model for portal vein system thrombosis post-splenectomy in patients with cirrhosis.
Abstract
BACKGROUND: Portal vein system thrombosis (PVST) is a common and potentially life-threatening complication following splenectomy plus pericardial devascularisation (SPDV) in patients with cirrhosis and portal hypertension. Early prediction of PVST is critical for timely intervention. This study aimed to develop a machine learning-based prediction model for PVST occurrence within 3 months after splenectomy.
METHODS: 392 patients with cirrhosis who underwent splenectomy at the Second Affiliated Hospital of Xi’an Jiaotong University between 1 July 2016 and 31 December 2022 were enrolled in this study and followed up for 3 months. The predictive model integrated 37 candidate predictors based on accessible clinical data, including demographic characteristics, disease features, imaging results, laboratory values, perioperative details and postoperative prophylactic therapies, and finally, eight predictors were selected for model construction. The five machine learning algorithms (logistic regression, Gaussian Naive Bayes, decision tree, random forest and AdaBoost) were employed to train the predictive models for assessing risks of PVST, which were validated using five fold cross-validation. Model discrimination and calibration were estimated using receiver operating characteristic curves(ROC), accuracy, sensitivity, specificity, positive predictive value, negative predictive value and Brier scores. The outcome of the predictive model was interpreted using SHapley Additive exPlanations (SHAP), which provided insights into the factors influencing PVST risk prediction.
RESULTS: During the 3-month follow-up, a total of 144 (36.73%) patients developed PVST. The AdaBoost model demonstrated the highest discriminative ability, with a mean area under the receiver operating characteristic curve (AUROC) of 0.72 (95% CI 0.60 to 0.84). Important features for predicting PVST included albumin, platelet addition, the diameter of the portal vein, ฮณ-glutamyl transferase, length of stay, activated partial thromboplastin time, D-dimer level and history of preoperative gastrointestinal bleeding, as revealed by SHAP analysis.
CONCLUSIONS: The machine learning-based prediction models can provide an initial assessment of 3-month PVST risk after SPDV in patients with cirrhosis and portal hypertension. The AdaBoost model demonstrates moderate discriminative ability in distinguishing between high-risk and low-risk patients, with an AUROC of 0.72 (95% CI 0.60 to 0.84). By incorporating SHAP analysis, the model can offer transparent explanations for personalised risk predictions, facilitating targeted preventive interventions and reducing excessive interventions across the entire patient population.
Author: [‘Qu D’, ‘Dai D’, ‘Li G’, ‘Zhou R’, ‘Dong C’, ‘Zhao J’, ‘An L’, ‘Song X’, ‘Zhu J’, ‘Li ZF’]
Journal: BMJ Health Care Inform
Citation: Qu D, et al. Development of an explainable prediction model for portal vein system thrombosis post-splenectomy in patients with cirrhosis. Development of an explainable prediction model for portal vein system thrombosis post-splenectomy in patients with cirrhosis. 2025; 32:(unknown pages). doi: 10.1136/bmjhci-2024-101319