Abstract
Introduction
Isoniazid is a first-line antituberculosis agent with high variability, which would profit from individualized dosing. Concentrations of isoniazid at 2 h (C2h), as an indicator of safety and efficacy, are important for optimizing therapy.
Objective
The objective of this study was to establish machine learning (ML) models to predict the C2h, that can be used for establishing an individualized dosing regimen in clinical practice.
Methods
Published population pharmacokinetic (PopPK) models for adults were searched based on PubMed and ultimately four reliable models were selected for simulating individual C2h datasets under different conditions (demographics, genotype, ethnicity, etc.). Machine learning models were trained on simulated C2h obtained from the four PopPK models. Five different algorithms were used for ML model building to predict C2h. Real-world data were used for predictive performance evaluations. Virtual trials were used to compare ML-optimized doses with PopPK model-optimized doses.
Results
Categorical boosting (CatBoost) exhibited the highest prediction ability. Target C2h can be predicted using the ML model combined with the dosing regimen and three covariates (N-acetyltransferase 2 [NAT2] genotypes, weight and race [Asians and Africans]). Real-world data validation results showed that the ML model can achieve an overall prediction accuracy of 93.4%. Using the final ML model, the mean absolute prediction error value decreased by 45.7% relative to the average of PopPK models. Using the ML-optimized dosing regimen, the probability of target attainment increased by 43.7% relative to the PopPK model-optimized dosing regimens.
Conclusion
Machine learning models were developed with great predictive performance, which can be used to determine the individualized initial dose of isoniazid in adult patients.
![](http://media.springernature.com/m312/springer-static/image/art%3A10.1007%2Fs40262-024-01400-4/MediaObjects/40262_2024_1400_Fig1_HTML.png)
![](http://media.springernature.com/m312/springer-static/image/art%3A10.1007%2Fs40262-024-01400-4/MediaObjects/40262_2024_1400_Fig2_HTML.png)
![](http://media.springernature.com/m312/springer-static/image/art%3A10.1007%2Fs40262-024-01400-4/MediaObjects/40262_2024_1400_Fig3_HTML.png)
![](http://media.springernature.com/m312/springer-static/image/art%3A10.1007%2Fs40262-024-01400-4/MediaObjects/40262_2024_1400_Fig4_HTML.png)
![](http://media.springernature.com/m312/springer-static/image/art%3A10.1007%2Fs40262-024-01400-4/MediaObjects/40262_2024_1400_Fig5_HTML.png)
References
Pai M, Behr MA, Dowdy D, Dheda K, Divangahi M, Boehme CC, et al. Tuberculosis. Nat Rev Dis Primers. 2016. https://doi.org/10.1038/nrdp.2016.76.
Bagcchi S. WHO’s global tuberculosis report 2022. Lancet Microbe. 2023;4(1): e20.
Soedarsono S, Jayanti RP, Mertaniasih NM, Kusmiati T, Permatasari A, Indrawanto DW, et al. Development of population pharmacokinetics model of isoniazid in Indonesian patients with tuberculosis. Int J Infect Dis. 2022;117:8–14.
Rogers Z, Hiruy H, Pasipanodya JG, Mbowane C, Adamson J, Ngotho L, et al. The non-linear child: ontogeny, isoniazid concentration, and NAT2 genotype modulate enzyme reaction kinetics and metabolism. EBioMedicine. 2016;11:118–26.
Kinzig-Schippers M, Tomalik-Scharte D, Jetter A, Scheidel B, Jakob V, Rodamer M, et al. Should we use N-acetyltransferase type 2 genoty** to personalize isoniazid doses? Antimicrob Agents Chemother. 2005;49(5):1733–8.
Chen B, Shi H-Q, Feng MR, Wang X-H, Cao X-M, Cai W-M. Population pharmacokinetics and pharmacodynamics of isoniazid and its metabolite acetylisoniazid in Chinese population. Front Pharmacol. 2022. https://doi.org/10.3389/fphar.2022.932686.
Pasipanodya JG, Gumbo T. A new evolutionary and pharmacokinetic–pharmacodynamic scenario for rapid emergence of resistance to single and multiple anti-tuberculosis drugs. Curr Opin Pharmacol. 2011;11(5):457–63.
Bekker A, Schaaf HS, Draper HR, van der Laan L, Murray S, Wiesner L, et al. Pharmacokinetics of rifampin, isoniazid, pyrazinamide, and ethambutol in infants dosed according to revised WHO-recommended treatment guidelines. Antimicrob Agents Chemother. 2016;60(4):2171–9.
Alsultan A, Peloquin CA. Therapeutic drug monitoring in the treatment of tuberculosis: an update. Drugs. 2014;74(8):839–54.
Anderson G, Vinnard C. Diagnostic accuracy of therapeutic drug monitoring during tuberculosis treatment. J Clin Pharmacol. 2022;62(10):1206–14.
Peloquin CA. Therapeutic drug monitoring in the treatment of tuberculosis. Drugs. 2002;62(15):2169–83.
Prahl JB, Johansen IS, Cohen AS, Frimodt-Møller N, Andersen ÅB. Clinical significance of 2 h plasma concentrations of first-line anti-tuberculosis drugs: a prospective observational study--authors' response. J Antimicrob Chemother. 2015;70(1):321–2.
Eckardt JN, Wendt K, Bornhäuser M, Middeke JM. Reinforcement learning for precision oncology. Cancers. 2021;13(18):4624.
Bräm DS, Nahum U, Schropp J, Pfister M, Koch G. Low-dimensional neural ODEs and their application in pharmacokinetics. J Pharmacokinet Pharmacodyn. 2024;51(2):123–40.
Bräm DS, Koch G, Allegaert K, van den Anker J, Pfister M. Applying neural ODEs to derive a mechanism-based model for characterizing maturation-related serum creatinine dynamics in preterm newborns. J Clin Pharmacol. 2024. https://doi.org/10.1002/jcph.2460.
Lu J, Deng KW, Zhang XY, Liu GB, Guan YF. Neural-ODE for pharmacokinetics modeling and its advantage to alternative machine learning models in predicting new dosing regimens. Iscience. 2021;24(7):102804.
Woillard JB, Labriffe M, Aurélie P, Marquet P. Estimation of drug exposure by machine learning based on simulations from published pharmacokinetic models: the example of tacrolimus. Pharmacol Res. 2021;167:105578.
Cho YS, Jang TW, Kim HJ, Oh JY, Lee HK, Park HK, et al. Isoniazid population pharmacokinetics and dose recommendation for Korean patients with tuberculosis based on target attainment analysis. J Clin Pharmacol. 2021;61(12):1567–78.
Denti P, Jeremiah K, Chigutsa E, Faurholt-Jepsen D, PrayGod G, Range N, et al. Pharmacokinetics of isoniazid, pyrazinamide, and ethambutol in newly diagnosed pulmonary TB patients in Tanzania. PLoS One. 2015;10(10): e0141002.
Gao YZ, Forsman LD, Ren WH, Zheng XB, Bao ZW, Hu Y, et al. Drug exposure of first-line anti-tuberculosis drugs in China: A prospective pharmacological cohort study. Br J Clin Pharmacol. 2021;87(3):1347–58.
Naidoo A, Chirehwa M, Ramsuran V, McIlleron H, Naidoo K, Yende-Zuma N, et al. Effects of genetic variability on rifampicin and isoniazid pharmacokinetics in South African patients with recurrent tuberculosis. Pharmacogenomics. 2019;20(4):225–40.
Ben Fredj N, Ben Romdhane H, Woillard JB, Chickaid M, Ben Fadhel N, Chadly Z, et al. Population pharmacokinetic model of isoniazid in patients with tuberculosis in Tunisia. Int J Infect Dis. 2021;104:562–7.
Ogami C, Tsuji Y, Seki H, Kawano H, To H, Matsumoto Y, et al. An artificial neural network-pharmacokinetic model and its interpretation using Shapley additive explanations. CPT Pharmacometr Syst Pharmacol. 2021;10(7):760–8.
**g W, Zong ZJ, Tang BH, Wang J, Zhang TT, Wen S, et al. Population pharmacokinetic analysis of isoniazid among pulmonary tuberculosis patients from China. Antimicrob Agents Chemother. 2020;64(3):e01736-19. https://doi.org/10.1128/AAC.01736-19.
Donald PR, Parkin DP, Seifart HI, Schaaf HS, van Helden PD, Werely CJ, et al. The influence of dose and N-acetyltransferase-2 (NAT2) genotype and phenotype on the pharmacokinetics and pharmacodynamics of isoniazid. Eur J Clin Pharmacol. 2007;63(7):633–9.
Treatment of Tuberculosis: Guidelines. 4th ed. Geneva: World Health Organization. 2010.
Payam N, Dorman SE, Narges A, Barry PM, Brozek JL, Adithya C, et al. Official American Thoracic Society/Centers for Disease Control and Prevention/Infectious Diseases Society of America Clinical Practice Guidelines: treatment of drug-susceptible tuberculosis. Clin Infect Dis. 2016;63(7): e147.
Lin S-Y, Law K-M, Yeh Y-C, Wu K-C, Lai J-H, Lin C-H, et al. Applying machine learning to carotid sonographic features for recurrent stroke in patients with acute stroke. Front Cardiovasc Med. 2022;9: 804410.
Mohammadi MR, Hadavimoghaddam F, Pourmahdi M, Atashrouz S, Munir MT, Hemmati-Sarapardeh A, et al. Modeling hydrogen solubility in hydrocarbons using extreme gradient boosting and equations of state. Sci Rep. 2021. https://doi.org/10.1038/s41598-021-97131-8.
Li QY, Tang BH, Wu YE, Yao BF, Zhang W, Zheng Y, et al. Machine learning: a new approach for dose individualization. Clin Pharmacol Ther. 2023;115(4):727–44.
Muscat JE, Pittman B, Kleinman W, Lazarus P, Stellman SD, Richie JP. Comparison of CYP1A2 and NAT2 phenotypes between black and white smokers. Biochem Pharmacol. 2008;76(7):929–37.
Liang HY, Tsui BY, Ni H, Valentim CCS, Baxter SL, Liu G, et al. Evaluation and accurate diagnoses of pediatric diseases using artificial intelligence. Nat Med. 2019;25(3):433–8.
Motwani M, Dey D, Berman DS, Germano G, Achenbach S, Al-Mallah MH, et al. Machine learning for prediction of all-cause mortality in patients with suspected coronary artery disease: a 5-year multicentre prospective registry analysis. Eur Heart J. 2017;38(7):500–7.
Keutzer L, You H, Farnoud A, Nyberg J, Wicha SG, Maher-Edwards G, et al. Machine learning and pharmacometrics for prediction of pharmacokinetic data: differences, similarities and challenges illustrated with rifampicin. Pharmaceutics. 2022;14(8):1530.
Woillard J-B, Labriffe M, Debord J, Marquet P. Mycophenolic acid exposure prediction using machine learning. Clin Pharmacol Ther. 2021;110(2):370–9.
Woillard J-B, Labriffe M, Debord J, Marquet P. Tacrolimus exposure prediction using machine learning. Clin Pharmacol Ther. 2021;110(2):361–9.
Miljkovic F, Martinsson A, Obrezanova O, Williamson B, Johnson M, Sykes A, et al. Machine learning models for human in vivo pharmacokinetic parameters with in-house validation. Mol Pharm. 2021;18(12):4520–30.
Ota R, Yamashita F. Application of machine learning techniques to the analysis and prediction of drug pharmacokinetics. J Control Release. 2022;352:961–9.
Acknowledgments
We thank all of the patients who participated in this study and all of the participants and research staff. All authors approved the final version of the manuscript.
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Funding
This paper is funded by the National Key R&D Program of China (2023YFC2706100, 2022YFC0868600), National Natural Science Foundation of China (82173897), Distinguished Young and Middle-aged Scholar of Shandong University, Innovation and Development Joint Fund of Natural Science Foundation of Shandong Province (ZR2022LSW007), Natural Science Foundation of Shandong Province (ZR2022QH004), the Capital’s Funds for Health Improvement and Research (2020-2-2161) and Bei**g High-Level Public Health Talent Program (G2003-2-002). We declare that we have no conflicts of interest relevant to this article.
Conflicts of Interest
John van den Anker is an Editorial Board member of Clinical Pharmacokinetics. John van den Anker was not involved in the selection of peer reviewers for the manuscript nor any of the subsequent editorial decisions. The authors declared no competing interests for this work.
Ethics Approval
All the data were obtained from previous studies. These studies were approved by the institutional ethics committee.
Consent to Participate
All the data were obtained from previous studies. These studies were approved by the institutional ethics committee.
Consent for Publication
All participants received written informed consent in the previous studies.
Availability of Data and Material
The data that support the findings of this study are available on reasonable request from the corresponding authors.
Code Availability
Research codes are available in the supplements.
Authors’ Contributions
B-H.T., X-F.Z., G-X.H. and J.V.D.A. wrote the manuscript; B-H.T., X-F.Z., S-M.F., B-F.Y., Y-E.W., Y.Z., Y.Z., H-R.H., G-X.H.and W.Z. designed the research; B-H.T., X-F.Z., H-R.H., G-X.H.and W.Z. performed the research; B-H.T. and X-F.Z. analyzed the data.
Supplementary Information
Below is the link to the electronic supplementary material.
Rights and permissions
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
About this article
Cite this article
Tang, BH., Zhang, XF., Fu, SM. et al. Machine Learning Approach in Dosage Individualization of Isoniazid for Tuberculosis. Clin Pharmacokinet (2024). https://doi.org/10.1007/s40262-024-01400-4
Accepted:
Published:
DOI: https://doi.org/10.1007/s40262-024-01400-4