{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"institution":[{"name":"bioRxiv"}],"indexed":{"date-parts":[[2026,7,30]],"date-time":"2026-07-30T09:17:14Z","timestamp":1785403034366,"version":"3.56.0"},"posted":{"date-parts":[[2025,7,18]]},"group-title":"Pharmacology and Toxicology","reference-count":28,"publisher":"openRxiv","license":[{"start":{"date-parts":[[2025,7,18]],"date-time":"2025-07-18T00:00:00Z","timestamp":1752796800000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/creativecommons.org\/licenses\/by-nc\/4.0\/"}],"funder":[{"award":["DK132408"],"award-info":[{"award-number":["DK132408"]}],"id":[{"id":"https:\/\/ror.org\/00adh9b73","id-type":"ROR","asserted-by":"publisher"}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"accepted":{"date-parts":[[2025,7,18]]},"abstract":"<jats:title>ABSTRACT<\/jats:title>\n                <jats:p>Model-informed precision dosing (MIPD) utilizes pharmacokinetic\/pharmacodynamic (PK\/PD) models to optimize drug therapy. However, conventional MIPD often requires manual simulation and regimen selection, which are time-consuming and demand specialized expertise. Reinforcement learning (RL), in which an agent learns optimal decisions through iterative interactions with an environment, offers a scalable and automated alternative. In this study, we developed a model-informed Deep Q-Network (DQN) to personalize infliximab dosing for patients with Crohn\u2019s disease. The DQN was trained in a simulation environment incorporating a population PK model, inter-individual variability, and assay error. Virtual patients with randomly sampled covariates were used to explore dosing strategies at infusions 1, 3, and 4. Doses ranged from 1 to 10 mg\/kg at infusion 1 and from 1 to 20 mg\/kg thereafter, with intervals of 4 to 12 weeks. The reward function prioritized achieving trough concentrations of 18\u201326 \u00b5g\/mL before infusion 3 and 5\u201310 \u00b5g\/mL before infusions 4 and 5, while penalizing overtreatment and additional infusions. The DQN policy converged after 80,000 episodes, yielding target attainment probabilities (PTAs) of 92.9% and 98.4% at infusions 4 and 5, respectively, in 1,000 virtual patients. High doses (11\u201320 mg\/kg) were selected in only 0.2% of cases. At infusion 4, 66.8% of patients received an 8-week interval, and 57.3% at infusion 5. Retrospective real-world validation showed that patients whose actual doses matched DQN recommendations had trough levels significantly closer to target ranges. These findings support the feasibility of using DQN-based agents to enhance and automate infliximab individualized dosing in pediatric populations.<\/jats:p>","DOI":"10.1101\/2025.07.14.664804","type":"posted-content","created":{"date-parts":[[2025,7,19]],"date-time":"2025-07-19T00:40:27Z","timestamp":1752885627000},"source":"Crossref","is-referenced-by-count":1,"title":["Model-informed Deep Q-Networks to Guide Infliximab Dosing in Pediatric Crohn\u2019s Disease"],"prefix":"10.64898","author":[{"given":"Kei","family":"Irie","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Phillip","family":"Minar","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Jack","family":"Reifenberg","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Brendan M","family":"Boyle","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Joshua D","family":"Noe","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"S Jeffrey","family":"Hyams","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Tomoyuki","family":"Mizuno","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"54368","reference":[{"key":"2025072216100912000_2025.07.14.664804v1.1","doi-asserted-by":"crossref","unstructured":"van Rheenen, P.F. et al. The Medical Management of Paediatric Crohn\u2019s Disease: an ECCO-ESPGHAN Guideline Update. J Crohns Colitis, (2020).","DOI":"10.1093\/ecco-jcc\/jjaa161"},{"key":"2025072216100912000_2025.07.14.664804v1.2","doi-asserted-by":"publisher","DOI":"10.1007\/s11894-023-00895-4"},{"key":"2025072216100912000_2025.07.14.664804v1.3","doi-asserted-by":"crossref","first-page":"68","DOI":"10.1097\/MPG.0000000000002304","article-title":"Development of Infliximab Target Concentrations During Induction in Pediatric Crohn Disease Patients","volume":"69","year":"2019","journal-title":"J Pediatr Gastroenterol Nutr"},{"key":"2025072216100912000_2025.07.14.664804v1.4","doi-asserted-by":"publisher","DOI":"10.1002\/cpt.2148"},{"key":"2025072216100912000_2025.07.14.664804v1.5","doi-asserted-by":"publisher","DOI":"10.1097\/FTD.0000000000001204"},{"key":"2025072216100912000_2025.07.14.664804v1.6","doi-asserted-by":"publisher","DOI":"10.1093\/ibd\/izac237"},{"key":"2025072216100912000_2025.07.14.664804v1.7","doi-asserted-by":"crossref","first-page":"853","DOI":"10.1002\/cpt.1777","article-title":"Model-Informed Artificial Intelligence: Reinforcement Learning for Precision Dosing","volume":"107","year":"2020","journal-title":"Clin Pharmacol Ther"},{"key":"2025072216100912000_2025.07.14.664804v1.8","doi-asserted-by":"crossref","first-page":"241","DOI":"10.1002\/psp4.12588","article-title":"Reinforcement learning and Bayesian data assimilation for model-informed precision dosing in oncology","volume":"10","year":"2021","journal-title":"CPT Pharmacometrics Syst Pharmacol"},{"key":"2025072216100912000_2025.07.14.664804v1.9","doi-asserted-by":"crossref","first-page":"1497","DOI":"10.1002\/psp4.12858","article-title":"Model enhanced reinforcement learning to enable precision dosing: A theoretical case study with dosing of propofol","volume":"11","year":"2022","journal-title":"CPT Pharmacometrics Syst Pharmacol"},{"key":"2025072216100912000_2025.07.14.664804v1.10","doi-asserted-by":"crossref","first-page":"619","DOI":"10.1002\/cpt.3356","article-title":"Model-Informed Reinforcement Learning for Enabling Precision Dosing Via Adaptive Dosing","volume":"116","year":"2024","journal-title":"Clin Pharmacol Ther"},{"key":"2025072216100912000_2025.07.14.664804v1.11","doi-asserted-by":"crossref","first-page":"17882","DOI":"10.1038\/s41598-021-97028-6","article-title":"Reinforcement learning derived chemotherapeutic schedules for robust patient-specific therapy","volume":"11","year":"2021","journal-title":"Sci Rep"},{"key":"2025072216100912000_2025.07.14.664804v1.12","doi-asserted-by":"crossref","first-page":"104267","DOI":"10.1016\/j.jbi.2022.104267","article-title":"Optimizing warfarin dosing using deep reinforcement learning","volume":"137","year":"2023","journal-title":"J Biomed Inform"},{"key":"2025072216100912000_2025.07.14.664804v1.13","doi-asserted-by":"crossref","first-page":"1270443","DOI":"10.3389\/fphar.2023.1270443","article-title":"Simulating clinical trials for model-informed precision dosing: using warfarin treatment as a use case","volume":"14","year":"2023","journal-title":"Front Pharmacol"},{"key":"2025072216100912000_2025.07.14.664804v1.14","doi-asserted-by":"publisher","DOI":"10.1111\/cts.70086"},{"key":"2025072216100912000_2025.07.14.664804v1.15","doi-asserted-by":"crossref","unstructured":"Garcia, K.S. et al. Efficacy of Early Optimization of Infliximab Guided by Therapeutic Drug Monitoring during Induction-A Prospective Trial. Biomedicines 11, (2023).","DOI":"10.3390\/biomedicines11061757"},{"key":"2025072216100912000_2025.07.14.664804v1.16","doi-asserted-by":"crossref","first-page":"542","DOI":"10.1093\/ecco-jcc\/jjz162","article-title":"Therapeutic Drug Monitoring of Biologics During Induction to Prevent Primary Non-Response","volume":"14","year":"2020","journal-title":"J Crohns Colitis"},{"key":"2025072216100912000_2025.07.14.664804v1.17","doi-asserted-by":"publisher","DOI":"10.1038\/nature14236"},{"key":"2025072216100912000_2025.07.14.664804v1.18","unstructured":"Wang, Z. , Schaul, T. , Hessel, M. , Hasselt, H.V. , Lanctot, M. & Freitas, N.D . Dueling network architectures for deep reinforcement learning. In Proceedings of the 33rd International Conference on International Conference on Machine Learning - Volume 48 1995\u20132003."},{"key":"2025072216100912000_2025.07.14.664804v1.19","doi-asserted-by":"crossref","unstructured":"van Hasselt, H. , Guez, A. & Silver, D . Deep Reinforcement Learning with Double Q-Learning. Proceedings of the AAAI Conference on Artificial Intelligence 30, (2016).","DOI":"10.1609\/aaai.v30i1.10295"},{"key":"2025072216100912000_2025.07.14.664804v1.20","unstructured":"Schaul, T. , Quan, J. , Antonoglou, I. & Silver, D. Prioritized experience replay. arXiv preprint arXiv:151105952, (2015)."},{"key":"2025072216100912000_2025.07.14.664804v1.21","first-page":"508","article-title":"Comparison of three commercially available ELISA assays for anti-infliximab antibodies","volume":"53","year":"2021","journal-title":"Pathology"},{"key":"2025072216100912000_2025.07.14.664804v1.22","doi-asserted-by":"crossref","first-page":"825","DOI":"10.1002\/cpt.3176","article-title":"Reinforcement Learning and PK-PD Models Integration to Personalize the Adaptive Dosing Protocol of Erdafitinib in Patients with Metastatic Urothelial Carcinoma","volume":"115","year":"2024","journal-title":"Clin Pharmacol Ther"},{"key":"2025072216100912000_2025.07.14.664804v1.23","doi-asserted-by":"crossref","unstructured":"Del Valle-Moreno, P. et al. Model-Informed Precision Dosing Software Tools for Dosage Regimen Individualization: A Scoping Review. Pharmaceutics 15, (2023).","DOI":"10.3390\/pharmaceutics15071859"},{"key":"2025072216100912000_2025.07.14.664804v1.24","doi-asserted-by":"crossref","first-page":"115421","DOI":"10.1016\/j.addr.2024.115421","article-title":"Model-informed precision dosing: State of the art and future perspectives","volume":"215","year":"2024","journal-title":"Adv Drug Deliv Rev"},{"key":"2025072216100912000_2025.07.14.664804v1.25","doi-asserted-by":"crossref","first-page":"1418","DOI":"10.1111\/bcp.14426","article-title":"Clinical implementation of pharmacogenetics and model-informed precision dosing to improve patient care","volume":"88","year":"2022","journal-title":"Br J Clin Pharmacol"},{"key":"2025072216100912000_2025.07.14.664804v1.26","doi-asserted-by":"crossref","unstructured":"De Carlo, A. , Tosca, E.M. & Magni, P . Precision Dosing in Presence of Multiobjective Therapies by Integrating Reinforcement Learning and PK-PD Models: Application to Givinostat Treatment of Polycythemia Vera. CPT Pharmacometrics Syst Pharmacol, (2025).","DOI":"10.1002\/psp4.70012"},{"key":"2025072216100912000_2025.07.14.664804v1.27","doi-asserted-by":"crossref","unstructured":"Tu, Z. et al. Deep reinforcement learning for multi-targets propofol dosing. J Clin Monit Comput, (2025).","DOI":"10.1007\/s10877-025-01269-z"},{"key":"2025072216100912000_2025.07.14.664804v1.28","doi-asserted-by":"crossref","unstructured":"Vigueras, G. , Munoz-Gil, L. , Reinisch, V. & Pinto, J.T . A PopPBPK-RL approach for precision dosing of benazepril in renal impaired patients. J Pharmacokinet Pharmacodyn 52, 6 (2024).","DOI":"10.1007\/s10928-024-09953-4"}],"container-title":[],"original-title":[],"link":[{"URL":"https:\/\/syndication.highwire.org\/content\/doi\/10.1101\/2025.07.14.664804","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,6,24]],"date-time":"2026-06-24T05:04:22Z","timestamp":1782277462000},"score":1,"resource":{"primary":{"URL":"http:\/\/biorxiv.org\/lookup\/doi\/10.1101\/2025.07.14.664804"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,7,18]]},"references-count":28,"URL":"https:\/\/doi.org\/10.1101\/2025.07.14.664804","relation":{"is-preprint-of":[{"id-type":"doi","id":"10.1002\/cpt.70118","asserted-by":"subject"}]},"subject":[],"published":{"date-parts":[[2025,7,18]]},"subtype":"preprint"}}