Predicting Nanoparticle Delivery to Tumors Using Machine Learning and Artificial Intelligence Approaches

Lin, Zhoumeng; Chou, Wei-Chun; Cheng, Yi-Hsien; He, Chunla; Monteiro‐Riviere, Nancy A.; Riviere, Jim E.

doi:10.2147/ijn.s344208

Cited by 57 publications

(33 citation statements)

References 54 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Previous studies have suggested that the tumor size is related to the EPR magnitude of tumors, which is revealed in our models as the positive relationship between delivery efficacy and tumor weight. [13] The tumor type was reported as another important determinant in predicting delivery efficacy, [10,13] which is also confirmed by the importance of genomic profile in our models. Because there is strong multicollinearity between tumor types/lineages and the genomic profiles, we excluded the tumor types/lineages from the model construction to increase the reliability of the statistical inferences.…”

Section: Discussionsupporting

confidence: 81%

“…The delivery efficacy dataset was obtained from the previously published meta-analysis, [10] containing the delivery efficacy at 24 h (DE24), 168 h (DE168), and the maximum delivery efficacy (DEmax) after intravenous injection of NPs. The original dataset contains 330 samples.…”

Section: Methodsmentioning

confidence: 99%

“…[6][7][8][9] Very recently, several reports applied machine learning (ML) and literature data mining techniques to predict the biodistribution of NPs in vivo from the combination of their properties. [10,11] These ML models reflect the complex relationship between the delivery efficacy and the properties of NPs, which is not possible for a single study based on a small number of NPs. [12] Nonetheless, it remains challenging to use these ML models to explain the difference in biodistribution among different tumor models and how much an NP property influences the final delivery efficacy in a certain tumor.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Interpretable XGBoost-SHAP model predicts the nanoparticles delivery and reveals its interaction with tumor genomic profiles

Wang

Tang

2022

Preprint

View full text Add to dashboard Cite

Understanding the complex interaction between nanoparticles (NPs) and tumors in vivo and how it dominates the delivery efficacy of NPs is critical for the translation of nanomedicine. Herein, we proposed an interpretable XGBoost-SHAP model by integrating the information of NPs physicochemical properties and tumor genomic profile to predict the delivery efficacy. The correlation coefficients were > 0.99 for all training sets, and 0.830, 0.839, and 0.741 for the prediction of maximum delivery efficacy (DEmax), delivery efficacy at 24 h (DE24), and delivery efficacy at 168 (DE168) for test sets. The analysis of the feature importance revealed that the tumor genomic mutations and their interaction with NPs properties played an important role in the delivery of NPs. The functional profile of the NP-delivery-related genes was further explored through gene ontology enrichment analysis. Our work provides a method to accurately predict the delivery efficacy of NPs to heterogeneous tumors and highlights the power of simultaneously using omics data and interpretable machine learning algorithms for discovering the interaction between NPs and tumors, which is important for the development of precision nanomedicine.

show abstract

Section: Discussionsupporting

confidence: 81%

Section: Methodsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Interpretable XGBoost-SHAP model predicts the nanoparticles delivery and reveals its interaction with tumor genomic profiles

Wang

Tang

2022

Preprint

View full text Add to dashboard Cite

show abstract

“…Seventh, the present model is in healthy rats, and thus cannot be used to predict delivery efficiency of AuNPs to tumor. However, the present model provides a basis to be extrapolated to tumor-bearing animals by adding a tumor compartment to help address low tumor delivery efficiency issue [4,52]. Finally, the model does not include the lymphatic system due to the lack of pharmacokinetic data of AuNPs in lymph nodes.…”

Section: Discussionmentioning

confidence: 99%

Development of a multi-route physiologically based pharmacokinetic (PBPK) model for nanomaterials: a comparison between a traditional versus a new route-specific approach using gold nanoparticles in rats

et al. 2022

Self Cite

View full text Add to dashboard Cite

Background Physiologically based pharmacokinetic (PBPK) modeling is an important tool in predicting target organ dosimetry and risk assessment of nanoparticles (NPs). The methodology of building a multi-route PBPK model for NPs has not been established, nor systematically evaluated. In this study, we hypothesized that the traditional route-to-route extrapolation approach of PBPK modeling that is typically used for small molecules may not be appropriate for NPs. To test this hypothesis, the objective of this study was to develop a multi-route PBPK model for different sizes (1.4–200 nm) of gold nanoparticles (AuNPs) in adult rats following different routes of administration (i.e., intravenous (IV), oral gavage, intratracheal instillation, and endotracheal inhalation) using two approaches: a traditional route-to-route extrapolation approach for small molecules and a new approach that is based on route-specific data that we propose to be applied generally to NPs. Results We found that the PBPK model using this new approach had superior performance than the traditional approach. The final PBPK model was optimized rigorously using a Bayesian hierarchical approach with Markov chain Monte Carlo simulations, and then converted to a web-based interface using R Shiny. In addition, quantitative structure–activity relationships (QSAR) based multivariate linear regressions were established to predict the route-specific key biodistribution parameters (e.g., maximum uptake rate) based on the physicochemical properties of AuNPs (e.g., size, surface area, dose, Zeta potential, and NP numbers). These results showed the size and surface area of AuNPs were the main determinants for endocytic/phagocytic uptake rates regardless of the route of administration, while Zeta potential was an important parameter for the estimation of the exocytic release rates following IV administration. Conclusions This study suggests that traditional route-to-route extrapolation approaches for PBPK modeling of small molecules are not applicable to NPs. Therefore, multi-route PBPK models for NPs should be developed using route-specific data. This novel PBPK-based web interface serves as a foundation for extrapolating to other NPs and to humans to facilitate biodistribution estimation, safety, and risk assessment of NPs.

show abstract

“…Since the rapid progress of computational facilities, artificial intelligence using machine learning (ML) has developed rapidly and has been used in some of the research areas in the medical field, including cancer, cardiovascular disease, neurological disease, emergency medicine and even in the pharmacological field, etc. [ 17 , 18 , 19 , 20 , 21 ]. The definition of ML is the study of computer algorithms that can improve automatically through experience and by the use of data [ 22 ].…”

Section: Introductionmentioning

confidence: 99%

Comparison between Machine Learning and Multiple Linear Regression to Identify Abnormal Thallium Myocardial Perfusion Scan in Chinese Type 2 Diabetes

et al. 2022

View full text Add to dashboard Cite

Type 2 diabetes mellitus (T2DM) patients have a high risk of coronary artery disease (CAD). Thallium-201 myocardial perfusion scan (Th-201 scan) is a non-invasive and extensively used tool in recognizing CAD in clinical settings. In this study, we attempted to compare the predictive accuracy of evaluating abnormal Th-201 scans using traditional multiple linear regression (MLR) with four machine learning (ML) methods. From the study, we can determine whether ML surpasses traditional MLR and rank the clinical variables and compare them with previous reports.In total, 796 T2DM, including 368 men and 528 women, were enrolled. In addition to traditional MLR, classification and regression tree (CART), random forest (RF), stochastic gradient boosting (SGB) and eXtreme gradient boosting (XGBoost) were also used to analyze abnormal Th-201 scans. Stress sum score was used as the endpoint (dependent variable). Our findings show that all four root mean square errors of ML are smaller than with MLR, which implies that ML is more precise than MLR in determining abnormal Th-201 scans by using clinical parameters. The first seven factors, from the most important to the least are:body mass index, hemoglobin, age, glycated hemoglobin, Creatinine, systolic and diastolic blood pressure. In conclusion, ML is not inferior to traditional MLR in predicting abnormal Th-201 scans, and the most important factors are body mass index, hemoglobin, age, glycated hemoglobin, creatinine, systolic and diastolic blood pressure. ML methods are superior in these kinds of studies.

show abstract

Predicting Nanoparticle Delivery to Tumors Using Machine Learning and Artificial Intelligence Approaches

Cited by 57 publications

References 54 publications

Interpretable XGBoost-SHAP model predicts the nanoparticles delivery and reveals its interaction with tumor genomic profiles

Interpretable XGBoost-SHAP model predicts the nanoparticles delivery and reveals its interaction with tumor genomic profiles

Development of a multi-route physiologically based pharmacokinetic (PBPK) model for nanomaterials: a comparison between a traditional versus a new route-specific approach using gold nanoparticles in rats

Comparison between Machine Learning and Multiple Linear Regression to Identify Abnormal Thallium Myocardial Perfusion Scan in Chinese Type 2 Diabetes

Contact Info

Product

Resources

About