Use of artificial neural network to predict warfarin individualized dosage regime in Chinese patients receiving low-intensity anticoagulation after heart valve replacement

Machine learning guided prediction of warfarin blood levels for personalized medicine based on clinical longitudinal data from cardiac surgery patients: a prospective observational study

BACKGROUND

Warfarin is a common oral anticoagulant, and its effects vary widely among individuals. Numerous dose-prediction algorithms have been reported based on cross-sectional data generated via multiple linear regression or machine learning. This study aimed to construct an information fusion perturbation theory and machine learning prediction model of warfarin blood levels based on clinical longitudinal data from cardiac surgery patients.

METHODS AND MATERIAL

The data of 246 patients were obtained from electronic medical records. Continuous variables were processed by calculating the distance of the raw data with the moving average (MA ∆vki(sj)), and categorical variables in different attribute groups were processed using Euclidean distance (ED ǁ∆vk(sj)ǁ). Regression and classification analyses were performed on the raw data, MA ∆vki(sj), and ED ǁ∆vk(sj)ǁ. Different machine-learning algorithms were chosen for the STATISTICA and WEKA software.

RESULTS

The random forest (RF) algorithm was the best for predicting continuous outputs using the raw data. The correlation coefficients of the RF algorithm were 0.978 and 0.595 for the training and validation sets, respectively, and the mean absolute errors were 0.135 and 0.362 for the training and validation sets, respectively. The proportion of ideal predictions of the RF algorithm was 59.0%. General discriminant analysis (GDA) was the best algorithm for predicting the categorical outputs using the MA ∆vki(sj) data. The GDA algorithm's total true positive rate (TPR) was 95.4% and 95.6% for the training and validation sets, respectively, with MA ∆vki(sj) data.

CONCLUSIONS

An information fusion perturbation theory and machine learning model for predicting warfarin blood levels was established. A model based on the RF algorithm could be used to predict the target international normalized ratio (INR), and a model based on the GDA algorithm could be used to predict the probability of being within the target INR range under different clinical scenarios.

Warfarin-A natural anticoagulant: A review of research trends for precision medication

BACKGROUND

Warfarin is a widely prescribed anticoagulant in the clinic. It has a more considerable individual variability, and many factors affect its variability. Mathematical models can quantify the quantitative impact of these factors on individual variability.

PURPOSE

The aim is to comprehensively analyze the advanced warfarin dosing algorithm based on pharmacometrics and machine learning models of personalized warfarin dosage.

METHODS

A bibliometric analysis of the literature retrieved from PubMed and Scopus was performed using VOSviewer. The relevant literature that reported the precise dosage of warfarin calculation was retrieved from the database. The multiple linear regression (MLR) algorithm was excluded because a recent systematic review that mainly reviewed this algorithm has been reported. The following terms of quantitative systems pharmacology, mechanistic model, physiologically based pharmacokinetic model, artificial intelligence, machine learning, pharmacokinetic, pharmacodynamic, pharmacokinetics, pharmacodynamics, and warfarin were added as MeSH Terms or appearing in Title/Abstract into query box of PubMed, then humans and English as filter were added to retrieve the literature.

RESULTS

Bibliometric analysis revealed important co-occuring MeShH and index keywords. Further, the United States, China, and the United Kingdom were among the top countries contributing in this domain. Some studies have established personalized warfarin dosage models using pharmacometrics and machine learning-based algorithms. There were 54 related studies, including 14 pharmacometric models, 31 artificial intelligence models, and 9 model evaluations. Each model has its advantages and disadvantages. The pharmacometric model contains biological or pharmacological mechanisms in structure. The process of pharmacometric model development is very time- and labor-intensive. Machine learning is a purely data-driven approach; its parameters are more mathematical and have less biological interpretation. However, it is faster, more efficient, and less time-consuming. Most published models of machine learning algorithms were established based on cross-sectional data sourced from the database.

CONCLUSION

Future research on personalized warfarin medication should focus on combining the advantages of machine learning and pharmacometrics algorithms to establish a more robust warfarin dosage algorithm. Randomized controlled trials should be performed to evaluate the established algorithm of warfarin dosage. Moreover, a more user-friendly and accessible warfarin precision medicine platform should be developed.

Concomitant administration of warfarin and toremifene: A case report

WHAT IS KNOWN AND OBJECTIVE

Antiestrogen agents have been reported to enhance the anticoagulant activity of warfarin. The use of tamoxifen with warfarin has been contraindicated. However, warfarin in combination with toremifene has not been reported. We report a case in which warfarin was combined with toremifene and applied warfarin dose prediction models to predict the dose of warfarin.

CASE SUMMARY

We report the case of a 50-year-old woman with a history of breast cancer, who underwent long-term toremifene therapy after mastectomy. The patient was treated with warfarin after prosthetic valve replacement and had a fluctuating international normalized ratio (INR) following the concomitant administration of toremifene. We applied the warfarin dose prediction model to adjust the warfarin dose during treatment. Finally, her INR stabilized with a lower dose of warfarin, and there was no serious bleeding during the 1-year follow-up.

WHAT IS NEW AND CONCLUSION

Warfarin does not have a serious interaction with toremifene in this case, but it needed about 37.5% dose reduction which was comparable to the interaction of some common antibiotics with warfarin.

Nonlinear Machine Learning in Warfarin Dose Prediction: Insights from Contemporary Modelling Studies

Objective: This study aimed to systematically assess the characteristics and risk of bias of previous studies that have investigated nonlinear machine learning algorithms for warfarin dose prediction. Methods: We systematically searched PubMed, Embase, Cochrane Library, Chinese National Knowledge Infrastructure (CNKI), China Biology Medicine (CBM), China Science and Technology Journal Database (VIP), and Wanfang Database up to March 2022. We assessed the general characteristics of the included studies with respect to the participants, predictors, model development, and model evaluation. The methodological quality of the studies was determined, and the risk of bias was evaluated using the Prediction model Risk of Bias Assessment Tool (PROBAST). Results: From a total of 8996 studies, 23 were assessed in this study, of which 23 (100%) were retrospective, and 11 studies focused on the Asian population. The most common demographic and clinical predictors were age (21/23, 91%), weight (17/23, 74%), height (12/23, 52%), and amiodarone combination (11/23, 48%), while CYP2C9 (14/23, 61%), VKORC1 (14/23, 61%), and CYP4F2 (5/23, 22%) were the most common genetic predictors. Of the included studies, the MAE ranged from 1.47 to 10.86 mg/week in model development studies, from 2.42 to 5.18 mg/week in model development with external validation (same data) studies, from 12.07 to 17.59 mg/week in model development with external validation (another data) studies, and from 4.40 to 4.84 mg/week in model external validation studies. All studies were evaluated as having a high risk of bias. Factors contributing to the risk of bias include inappropriate exclusion of participants (10/23, 43%), small sample size (15/23, 65%), poor handling of missing data (20/23, 87%), and incorrect method of selecting predictors (8/23, 35%). Conclusions: Most studies on nonlinear-machine-learning-based warfarin prediction models show poor methodological quality and have a high risk of bias. The analysis domain is the major contributor to the overall high risk of bias. External validity and model reproducibility are lacking in most studies. Future studies should focus on external validity, diminish risk of bias, and enhance real-world clinical relevance.

Warfarin maintenance dose prediction for Chinese after heart valve replacement by a feedforward neural network with equal stratified sampling

Patients requiring low-dose warfarin are more likely to suffer bleeding due to overdose. The goal of this work is to improve the feedforward neural network model's precision in predicting the low maintenance dose for Chinese in the aspect of training data construction. We built the model from a resampled dataset created by equal stratified sampling (maintaining the same sample number in three dose-groups with a total of 3639) and performed internal and external validations. Comparing to the model trained from the raw dataset of 19,060 eligible cases, we improved the low-dose group's ideal prediction percentage from 0.7 to 9.6% and maintained the overall performance (76.4% vs. 75.6%) in external validation. We further built neural network models on single-dose subsets to invest whether the subsets samples were sufficient and whether the selected factors were appropriate. The training set sizes were 1340 and 1478 for the low and high dose subsets; the corresponding ideal prediction percentages were 70.2% and 75.1%. The training set size for the intermediate dose varied and was 1553, 6214, and 12,429; the corresponding ideal prediction percentages were 95.6, 95.1%, and 95.3%. Our conclusion is that equal stratified sampling can be a considerable alternative approach in training data construction to build drug dosing models in the clinic.

A post-hoc interpretable ensemble model to feature effect analysis in warfarin dose prediction for Chinese patients

To interprete the importance of clinical features and genotypes for warfarin daily dose prediction, we developed a post-hoc interpretable framework based on an ensemble predictive model. This framework includes permutation importance for global interpretation and local interpretable model-agnostic explanation (LIME) and shapley additive explanations (SHAP) for local explanation. The permutation importance globally ranks the importance of features on the whole data set. This can guide us to build a predictive model with less variables and the complexity of final predictive model can be reduced. LIME and SHAP together explain how the predictive model give the predicted dosage for specific samples. This help clinicians prescribe accurate doses to patients using more effective clinical variables. Results showed that both the permutation importance and SHAP demonstrated that VKORC1, age, serum creatinine (SCr), left atrium (LA) size, CYP2C9 and weight were the most important features on the whole data set. In specific samples, both SHAP and LIME discovered that in Chinese patients, wild-type VKORC1-AA, mutant-type CYP2C9*3, age over 60, abnormal LA size, SCr within the normal range, and using amiodarone definitely required dosage reduction, whereas mutant-type VKORC1-AG/GG, small age, SCr out of normal range, normal LA size, diabetes and heavy weight required dosage enhancement.

An Adapted Neural-Fuzzy Inference System Model Using Preprocessed Balance Data to Improve the Predictive Accuracy of Warfarin Maintenance Dosing in Patients After Heart Valve Replacement

BACKGROUND

Tailoring warfarin use poses a challenge for physicians and pharmacists due to its narrow therapeutic window and substantial inter-individual variability. This study aimed to create an adapted neural-fuzzy inference system (ANFIS) model using preprocessed balance data to improve the predictive accuracy of warfarin maintenance dosing in Chinese patients undergoing heart valve replacement (HVR).

METHODS

This retrospective study enrolled patients who underwent HVR between June 1, 2012, and June 1, 2016, from 35 centers in China. The primary outcomes were the mean difference between predicted warfarin dose by ANFIS models and actual dose and the models' predictive accuracy, including the ideal predicted percentage, the mean absolute error (MAE), and the mean squared error (MSE). The eligible cases were divided into training, internal validation, and external validation groups. We explored input variables by univariate analysis of a general linear model and created two ANFIS models using imbalanced and balanced training sets. We finally compared the primary outcomes between the imbalanced and balanced ANFIS models in both internal and external validation sets. Stratified analyses were conducted across warfarin doses (low, medium, and high doses).

RESULTS

A total of 15,108 patients were included and grouped as follows: 12,086 in the imbalanced training set; 2820 in the balanced training set; 1511 in the internal validation set; and 1511 in the external validation set. Eight variables were explored as predictors related to warfarin maintenance doses, and imbalanced and balanced ANFIS models with multi-fuzzy rules were developed. The results showed a low mean difference between predicted and actual doses (< 0.3 mg/d for each model) and an accurate prediction property in both the imbalanced model (ideal prediction percentage, 74.39-78.16%; MAE, 0.37 mg/daily; MSE, 0.39 mg/daily) and the balanced model (ideal prediction percentage, 73.46-75.31%; MAE, 0.42 mg/daily; MSE, 0.43 mg/daily). Compared to the imbalanced model, the balanced model had a significantly higher prediction accuracy in the low-dose (14.46% vs. 3.01%; P < 0.001) and the high-dose warfarin groups (34.71% vs. 23.14%; P = 0.047). The results from the external validation cohort confirmed this finding.

CONCLUSIONS

The ANFIS model can accurately predict the warfarin maintenance dose in patients after HVR. Through data preprocessing, the balanced model contributed to improved prediction ability in the low- and high-dose warfarin groups.

An ensemble learning based framework to estimate warfarin maintenance dose with cross-over variables exploration on incomplete data set

MOTIVATION

Warfarin is a widely used oral anticoagulant, but it is challenging to select the optimal maintenance dose due to its narrow therapeutic window and complex individual factor relationships. In recent years, machine learning techniques have been widely applied for warfarin dose prediction. However, the model performance always meets the upper limit due to the ignoration of exploring the variable interactions sufficiently. More importantly, there is no efficient way to resolve missing values when predicting the optimal warfarin maintenance dose.

METHODS

Using an observational cohort from the Xinhua Hospital affiliated to Shanghai Jiaotong University School of Medicine, we propose a novel method for warfarin maintenance dose prediction, which is capable of assessing variable interactions and dealing with missing values naturally. Specifically, we examine single variables by univariate analysis initially, and only statistically significant variables are included. We then propose a novel feature engineering method on them to generate the cross-over variables automatically. Their impacts are evaluated by stepwise regression, and only the significant ones are selected. Lastly, we implement an ensemble learning based approach, LightGBM, to learn from incomplete data directly on the selected single and cross-over variables for dosing prediction.

RESULTS

377 unique patients with eligible and time-independent 1173 warfarin order events are included in this study. Through the comprehensive experimental results in 5-fold cross-validation, our proposed method demonstrates the efficiency of exploring the variable interactions and modeling on incomplete data. The R² can achieve 75.0% on average. Moreover, the subgroup analysis results reveal that our method performs much better than other baseline methods, especially in the medium-dose and high-dose subgroups. Lastly, the IWPC dosing prediction model is used for further comparison, and our approach outperforms it by a significant margin.

CONCLUSION

In summary, our proposed method is capable of exploring the variable interactions and learning from incomplete data directly for warfarin maintenance dose prediction, which has a great premise and is worthy of further research.

Evaluating warfarin dosing models on multiple datasets with a novel software framework and evolutionary optimisation

Warfarin is an effective preventative treatment for arterial and venous thromboembolism, but requires individualised dosing due to its narrow therapeutic range and high individual variation. Many machine learning techniques have been demonstrated in this domain. This study evaluated the accuracy of the most promising algorithms on the International Warfarin Pharmacogenetics Consortium dataset and a novel clinical dataset of South African patients. Support vectors and linear regression were amongst the top performers in both datasets and performed comparably to recent stacked ensemble approaches, whilst neural networks were one of the worst performers in both datasets. We also introduced genetic programming to automatically optimise model architectures and hyperparameters without human guidance. Remarkably, the generated models were found to match the performance of the best models hand-crafted by human experts. Finally, we present a novel software framework (Warfit-learn) for warfarin dosing research. It leverages the most successful techniques in preprocessing, imputation, and parallel evaluation, with the goal of accelerating research and making results in this domain more reproducible.

Warfarin dosing algorithms: A systematic review

Aims

Numerous algorithms have been developed to guide warfarin dosing and improve clinical outcomes. We reviewed the algorithms available for various populations and the covariates, performances and risk of bias of these algorithms.

Methods

We systematically searched MEDLINE up to 20 May 2020 and selected studies describing the development, external validation or clinical utility of a multivariable warfarin dosing algorithm. Two investigators conducted data extraction and quality assessment.

Results

Of 10 035 screened records, 266 articles were included in the review, describing the development of 433 dosing algorithms, 481 external validations and 52 clinical utility assessments. Most developed algorithms were for dose initiation (86%), developed by multiple linear regression (65%) and mostly applicable to Asians (49%) or Whites (43%). The most common demographic/clinical/environmental covariates were age (included in 401 algorithms), concomitant medications (270 algorithms) and weight (229 algorithms) while CYP2C9 (329 algorithms), VKORC1 (319 algorithms) and CYP4F2 (92 algorithms) variants were the most common genetic covariates. Only 26% and 7% algorithms were externally validated and evaluated for clinical utility, respectively, with <2% of algorithm developments and external validations being rated as having a low risk of bias.

Conclusion

Most warfarin dosing algorithms have been developed in Asians and Whites and may not be applicable to under‐served populations. Few algorithms have been externally validated, assessed for clinical utility, and/or have a low risk of bias which makes them unreliable for clinical use. Algorithm development and assessment should follow current methodological recommendations to improve reliability and applicability, and under‐represented populations should be prioritized.

Warfarin maintenance dose Prediction for Patients undergoing heart valve replacement- a hybrid model with genetic algorithm and Back-Propagation neural network

Warfarin is the most recommended anticoagulant drug for patients undergoing heart valve replacement. However, due to the narrow therapeutic window and individual dose, the use of warfarin needs more advanced technology. We used the data collected from a multi-central registered clinical system all over China about the patients who have undergone heart valve replacement, subsequently divided into three groups (training group: 10673 cases; internal validation group: 3558 cases; external validation group: 1463 cases) in order to construct a hybrid model with genetic algorithm and Back-Propagation neural network (BP-GA), For testing the model's prediction accuracy, we used Mean absolute error (MAE), Root mean squared error (RMSE) and the ideal predicted percentage of total and dose subgroups. In results, whether in internal or in external validation group, the total ideal predicted percentage was over 58% while the intermediate dose subgroup manifested the best. Moreover, it showed higher prediction accuracy, lower MAE value and lower RMSE value in the external validation group than that in the internal validation group (p < 0.05). In conclusion, BP-GA model is promising to predict warfarin maintenance dose.

A prediction study of warfarin individual stable dose after mechanical heart valve replacement: adaptive neural-fuzzy inference system prediction

Background

It’s difficult but urgent to achieve the individualized rational medication of the warfarin, we aim to predict the individualized warfarin stable dose though the artificial intelligent Adaptive neural-fuzzy inference system (ANFIS).

Methods

Our retrospective analysis based on a clinical database, involving 21,863 patients from 15 Chinese provinces who receive oral warfarin after the heart valve replacement. They were allocated into four groups: the external validation group (A group), the internal validation group (B group), training group (C group) and stratified training group (D group). We used a univariate analysis of general linear models(GLM-univariate) to select the input variables and construct two prediction models by the ANFIS with the training and stratified training group, and then verify models with two validation groups by the mean squared error(MSE), mean absolute error(MAE) and the ideal predicted percentage.

Results

A total of 13,639 eligible patients were selected, including 1639 in A group, 3000 in B group, 9000 in C group, and 3192 in D group. Nine input variables were selected out and two five-layered ANFIS models were built. ANFIS model achieved the highest total ideal predicted percentage 63.7%. In the dose subgroups, all the models performed best in the intermediate-dose group with the ideal predicted percentage 82.4~ 86.4%, and the use of the stratified training group slightly increased the prediction accuracy in low-dose group by 8.8 and 5.2%, respectively.

Conclusion

As a preliminary attempt, ANFIS model predicted the warfarin stable dose properly after heart valve surgery among Chinese, and also proved that Chinese need lower anticoagulation intensity INR (1.5–2.5) to warfarin by reference to the recommended INR (2.5–3.5) in the developed countries.

Modeling of glucose release from native and modified wheat starch gels during in vitro gastrointestinal digestion using artificial intelligence methods

Estimation of the amounts of glucose release (AGR) during gastrointestinal digestion can be useful to identify food of potential use in the diet of individuals with diabetes. In this work, adaptive neuro-fuzzy inference system (ANFIS), genetic algorithm-artificial neural network (GA-ANN) and group method of data handling (GMDH) models were applied to estimate the AGR from native (NWS), cross-linked (CLWS) and hydroxypropylated wheat starch (HPWS) gels during digestion under simulated gastrointestinal conditions. The GA-ANN and ANFIS were fed with 3 inputs of digestion time (1-120min), gel volume (7.5 and 15ml) and concentration (8 and 12%, w/w) for prediction of the AGR. The developed ANFIS predictions were close to the experimental data (r=0.977-0.996 and RMSE=0.225-0.619). The optimized GA-ANN, which included 6-7 hidden neurons, predicted the AGR with a good precision (r=0.984-0.993 and RMSE=0.338-0.588). Also, a three layers GMDH model with 3 neurons accurately predicted the AGR (r=0.979-0.986 and RMSE=0.339-0.443). Sensitivity analysis data demonstrated that the gel concentration was the most sensitive factor for prediction of the AGR. The results dedicated that the AGR will be accurately predictable through such soft computing methods providing less computational cost and time.

Development of neuro-fuzzy model to explore gene-nutrient interactions modulating warfarin dose requirement

AIM

To investigate the influence of alterations in vitamin K (K1, K2 and K3) in modulating warfarin dose requirement.

PATIENTS & METHODS

Reverse phase HPLC to determine the plasma vitamin K; PCR-RFLP to detect polymorphisms; and the neuro-fuzzy model to predict warfarin dose were used.

RESULTS

The developed neuro-fuzzy model showed a mean absolute error of 0.000024 mg/week. CYP2C9*2 and CYP2C9*3 mediated warfarin sensitivity was observed when vitamin K is in high and low tertiles, respectively. VKORC1-1639G>A exhibited warfarin sensitivity in all combinations. Higher vitamin K1 was observed in CYP4F2 V433M polymorphism. The requirement of warfarin is low in GGCX 8016 GG genotype compared with GA and AA genotypes.

CONCLUSION

Vitamin K profile along with genetic testing ensures precision in warfarin dose optimization.

Artificial neural network-based pharmacogenomic algorithm for warfarin dose optimization

AIM

To develop more precise pharmacogenomic algorithm for prediction of safe and effective dose of warfarin.

MATERIALS & METHODS

An artificial neural network (ANN) algorithm was developed by using age, gender, BMI, plasma vitamin K levels, thyroid status and ten genetic variables as the inputs and therapeutic warfarin dose as the output. Hyperbolic tangent function was used to build an ANN architecture.

RESULTS

This model explained 93.5% variability in warfarin dosing and predicted warfarin dose accurately in 74.5% patients whose international normalized ratio (INR) was less than 2.0 and in 83.3% patients whose INR was more than 3.5. This algorithm reduced the out-of-range INRs (odds ratio [OR]: 0.49; 95% CI: 0.30-0.79; p = 0.003), the rate of adverse drug reactions (OR: 0.00; 95% CI: 0.00-1.21; p = 0.06) and time to reach first therapeutic INR (OR: 6.73; 95% CI: 2.17-22.31; p < 0.0001). This algorithm was found to be applicable in both euthyroid and hypothyroid status. S-warfarin/7-hydroxywarfarin ratio was found to increase in subjects with CYP2C9*2 and CYP2C9*3 justifying the warfarin sensitivity attributed to these variants.

CONCLUSION

An application of ANN for warfarin dosing improves predictability and provides safe and effective dosing.