{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,8,17]],"date-time":"2026-08-17T14:50:11Z","timestamp":1786978211388,"version":"build-2736575974"},"reference-count":32,"publisher":"Wiley","issue":"6","license":[{"start":{"date-parts":[[2025,5,21]],"date-time":"2025-05-21T00:00:00Z","timestamp":1747785600000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/creativecommons.org\/licenses\/by-nc-nd\/4.0\/"}],"funder":[{"DOI":"10.13039\/501100009192","name":"Alberta Innovates","doi-asserted-by":"publisher","id":[{"id":"10.13039\/501100009192","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/501100000038","name":"Natural Sciences and Engineering Research Council of Canada","doi-asserted-by":"publisher","id":[{"id":"10.13039\/501100000038","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/100008363","name":"Kr\u00e6ftens Bek\u00e6mpelse","doi-asserted-by":"publisher","award":["R269\u2010A15548"],"award-info":[{"award-number":["R269\u2010A15548"]}],"id":[{"id":"10.13039\/100008363","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":["aapm.onlinelibrary.wiley.com"],"crossmark-restriction":true},"short-container-title":["Medical Physics"],"published-print":{"date-parts":[[2025,6]]},"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:sec>\n                    <jats:title>Background<\/jats:title>\n                    <jats:p>Bayesian networks are seeing increased usage in healthcare, particularly for modeling complex treatment decisions under uncertainty. Bayesian networks offer significant advantages over classical machine learning and deep learning techniques due to their interpretability, with the network visualized through a directed acyclic graph outlining conditional relationships. Prior clinical knowledge can also be incorporated into these networks to enhance their clarity and facilitate integration into clinical workflows. However, out\u2010of\u2010box optimization techniques may produce networks that are not logically coherent or reflective of clinical understanding and may focus solely on optimizing information\u2010based metrics without consideration for performance metrics crucial for developing predictive models. In late morbidity modeling, where the risk factors surrounding an outcome may be complex, intercorrelated, and not yet fully identified, it is important to have a customizable optimization approach to automatically produce logical, interpretable Bayesian networks that outline these complex outcomes.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Purpose<\/jats:title>\n                    <jats:p>Develop a simulated annealing\u2010based framework for developing Bayesian network structures for late morbidity prediction in cervical cancer patients, addressing limitations of traditional optimization techniques and prioritizing interpretability.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Methods<\/jats:title>\n                    <jats:p>\n                      This study utilizes the multi\u2010center EMBRACE I cervical cancer dataset (\n                      <jats:italic>n<\/jats:italic>\n                      \u00a0=\u00a01153) to develop Bayesian network structures for late moderate\u2010to\u2010severe (grade\u00a0\u22652) cystitis (CTCAEv.3) prediction. The dataset was split into training\/validation data (80%) and holdout test data (20%). A process of 10\u00a0\u00d7\u00a05\u2010fold cross\u2010validation was integrated into the optimization framework. A simulated annealing\u2010based optimization method was developed incorporating information\u2014theoretic measures, predictive performance measures, and complexity measures. The different network structures developed by this framework were compared in terms of complexity, interpretability, and predictive performance to optimization methods available out\u2010of\u2010box from the PyAgrum package for Python (Greedy Hill Climbing, Tree\u2010Augmented Na\u00efve Bayes, and Chow\u2010Liu Optimization). Bayesian networks were also compared to conventional machine learning classifiers in terms of feature importance and predictive performance. Differences in model predictions arising from structure differences were assessed with Cochran's\n                      <jats:italic>Q<\/jats:italic>\n                      \u2010test (\n                      <jats:italic>p<\/jats:italic>\n                      \u00a0&lt;\u00a00.05).\n                    <\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Results<\/jats:title>\n                    <jats:p>\n                      The simulated annealing framework demonstrated the ability to produce Bayesian network structures with comparable or superior predictive performance compared to out\u2010of\u2010box models. A statistically significant performance difference was identified between the simulated annealing and out\u2010of\u2010box methods with Cochran's\n                      <jats:italic>Q<\/jats:italic>\n                      \u2010test (\n                      <jats:italic>p<\/jats:italic>\n                      \u00a0=\u00a00.03). The simulated annealing approach equalled or outperformed out\u2010of\u2010box models on a bootstrapped holdout test set, with a balanced accuracy of 64.1%, an F1 macro score of 55.9%, and an ROC\u2010AUC of 0.66. Simulated annealing models also featured fewer arcs and nodes, with this simplification resulting in networks that were easier to interpret without compromising on predictive performance, highlighting the effectiveness of simulated annealing in creating highly interpretable models for clinical use.\n                    <\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Conclusion<\/jats:title>\n                    <jats:p>The proposed simulated annealing\u2010based framework represents a novel method for automatically generating Bayesian network structures for cervical cancer late morbidity modeling. Compared to out\u2010of\u2010box optimization techniques, the simulated annealing Bayesian networks provide comparable or superior predictive performance while constructing a more simple, interpretable network useful for clinical implementation.<\/jats:p>\n                  <\/jats:sec>","DOI":"10.1002\/mp.17881","type":"journal-article","created":{"date-parts":[[2025,5,22]],"date-time":"2025-05-22T00:34:33Z","timestamp":1747874073000},"page":"5051-5063","update-policy":"https:\/\/doi.org\/10.1002\/crossmark_policy","source":"Crossref","is-referenced-by-count":2,"title":["A simulated annealing\u2010based Bayesian network structure optimization framework for late morbidity prediction with a large prospective dataset"],"prefix":"10.1002","volume":"52","author":[{"given":"Kailyn","family":"Stenhouse","sequence":"first","affiliation":[{"name":"Department of Physics and Astronomy University of Calgary Calgary Alberta Canada"},{"name":"Department of Medical Physics Arthur J.E. Child Comprehensive Cancer Centre Calgary Alberta Canada"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Philip","family":"McGeachy","sequence":"additional","affiliation":[{"name":"Department of Physics and Astronomy University of Calgary Calgary Alberta Canada"},{"name":"Department of Medical Physics Arthur J.E. Child Comprehensive Cancer Centre Calgary Alberta Canada"},{"name":"Department of Oncology University of Calgary Calgary Alberta Canada"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Sofia","family":"Spampinato","sequence":"additional","affiliation":[{"name":"Department of Radiotherapy Erasmus MC Cancer Institute University Medical Center Rotterdam Rotterdam The Netherlands"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Kari","family":"Tanderup","sequence":"additional","affiliation":[{"name":"Danish Center for Particle Therapy Aarhus University Hospital Aarhus Denmark"},{"name":"Institute of Clinical Medicine Aarhus University Aarhus Denmark"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Kathrin","family":"Kirchheiner","sequence":"additional","affiliation":[{"name":"Department of Radiation Oncology Medical University Vienna General Hospital Vienna Vienna Austria"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Kevin","family":"Martell","sequence":"additional","affiliation":[{"name":"Department of Oncology University of Calgary Calgary Alberta Canada"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Sarah","family":"Quirk","sequence":"additional","affiliation":[{"name":"Department of Radiation Oncology Brigham and Women's Hospital Dana\u2010Farber Cancer Institute, and Harvard Medical School Boston Massachusetts USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Akila N.","family":"Viswanathan","sequence":"additional","affiliation":[{"name":"Department of Radiation Oncology and Molecular Radiation Sciences Johns Hopkins University Baltimore Maryland USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Michael","family":"Roumeliotis","sequence":"additional","affiliation":[{"name":"Department of Radiation Oncology and Molecular Radiation Sciences Johns Hopkins University Baltimore Maryland USA"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"311","published-online":{"date-parts":[[2025,5,21]]},"reference":[{"key":"e_1_2_8_2_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.artmed.2020.101912"},{"key":"e_1_2_8_3_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.artmed.2003.11.001"},{"key":"e_1_2_8_4_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.neucom.2021.01.138"},{"key":"e_1_2_8_5_1","doi-asserted-by":"publisher","DOI":"10.1002\/mp.15651"},{"key":"e_1_2_8_6_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.radonc.2021.01.019"},{"key":"e_1_2_8_7_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.radonc.2018.05.002"},{"key":"e_1_2_8_8_1","author":"Tanderup K","year":"2016","journal-title":"EMBRACE II Study Protocol"},{"key":"e_1_2_8_9_1","doi-asserted-by":"publisher","DOI":"10.3390\/cancers14092097"},{"key":"e_1_2_8_10_1","doi-asserted-by":"publisher","DOI":"10.1016\/B978-0-12-803732-4.00002-7"},{"key":"e_1_2_8_11_1","doi-asserted-by":"publisher","DOI":"10.1007\/s10462\u2010022\u201010351\u2010w"},{"key":"e_1_2_8_12_1","doi-asserted-by":"publisher","DOI":"10.1109\/TPAMI.2020.2990820"},{"key":"e_1_2_8_13_1","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-319-01766-2_60"},{"key":"e_1_2_8_14_1","doi-asserted-by":"publisher","DOI":"10.1002\/int.20138"},{"key":"e_1_2_8_15_1","doi-asserted-by":"publisher","DOI":"10.1109\/CSB.2004.1332531"},{"key":"e_1_2_8_16_1","doi-asserted-by":"publisher","DOI":"10.9734\/bpi\/cimms\/v4\/16996D"},{"key":"e_1_2_8_17_1","doi-asserted-by":"crossref","unstructured":"WicaksonoYA KurniawanD SantosoHA.Bayesian network optimization for earthquake prediction using simulated annealing. In:2018 International Seminar on Application for Technology of Information and Communication Semarang Indonesia;2018:494\u2010497 doi:10.1109\/ISEMANTIC.2018.8549729","DOI":"10.1109\/ISEMANTIC.2018.8549729"},{"key":"e_1_2_8_18_1","doi-asserted-by":"publisher","DOI":"10.1109\/TKDE.2019.2899096"},{"key":"e_1_2_8_19_1","doi-asserted-by":"crossref","unstructured":"AbdelbarAM HedetniemiSM.A parallel hybrid genetic algorithm simulated annealing approach to finding most probable explanations on Bayesian belief networks. In:Proceedings of International Conference on Neural Networks (ICNN'97) Houston TX USA. Vol.1;1997:450\u2010455. doi:10.1109\/ICNN.1997.611710","DOI":"10.1109\/ICNN.1997.611710"},{"key":"e_1_2_8_20_1","doi-asserted-by":"publisher","DOI":"10.1016\/S1470\u20102045(20)30753\u20101"},{"key":"e_1_2_8_21_1","unstructured":"EMBRACE. Accessed January 18 2024.https:\/\/www.embracestudy.dk\/Public\/Default.aspx?ReturnUrl=%2f&AspxAutoDetectCookieSupport=1"},{"key":"e_1_2_8_22_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.ygyno.2020.07.007"},{"key":"e_1_2_8_23_1","doi-asserted-by":"publisher","DOI":"10.1016\/S1053\u20104296(03)00031\u20106"},{"key":"e_1_2_8_24_1","doi-asserted-by":"crossref","unstructured":"NicholsonA JitnahN.Using mutual information to determine relevance in Bayesian Networks. In:LeeH MotodaH eds.PRICAI\u201998: Topics in Artificial Intelligence. PRICAI 1998. Vol.1531. Lecture Notes in Computer Science.Springer; 1999. doi:10.1007\/BFb0095287","DOI":"10.1007\/BFb0095287"},{"key":"e_1_2_8_25_1","unstructured":"ZaffalonM HutterM.Robust feature selection by mutual information distributions. In:UAI\u201902: Proceedings of the Eighteenth Conference on Uncertainty in Artificial Intelligence;2002:557\u2010584. doi:10.5555\/2073876.2073945"},{"key":"e_1_2_8_26_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.csda.2020.107043"},{"key":"e_1_2_8_27_1","doi-asserted-by":"publisher","DOI":"10.11591\/eei.v10i1.2007"},{"key":"e_1_2_8_28_1","unstructured":"PedregosaF VaroquauxG GramfortA et\u00a0al.Scikit\u2010Learn: Machine Learning in Python;2018. doi:10.48550\/arXiv.1201.0490"},{"key":"e_1_2_8_29_1","unstructured":"DucampG GonzalesC WuilleminPH.aGrUM\/pyAgrum\u00a0: a toolbox to build models and algorithms for probabilistic graphical models in python. In:Proceedings of the 10th International Conference on Probabilistic Graphical Models.PMLR;2020. Accessed January 18 2024.https:\/\/proceedings.mlr.press\/v138\/ducamp20a.html"},{"key":"e_1_2_8_30_1","doi-asserted-by":"publisher","DOI":"10.1016\/S0360\u20103016(96)00365\u20103"},{"key":"e_1_2_8_31_1","unstructured":"RaschkaS.Model Evaluation Model Selection and Algorithm Selection in Machine Learning; 2020. doi:10.48550\/arXiv.1811.12808"},{"key":"e_1_2_8_32_1","doi-asserted-by":"publisher","DOI":"10.3389\/fonc.2023.1129380"},{"key":"e_1_2_8_33_1","doi-asserted-by":"publisher","DOI":"10.1136\/bmj.316.7139.1236"}],"container-title":["Medical Physics"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/aapm.onlinelibrary.wiley.com\/doi\/pdf\/10.1002\/mp.17881","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,11,12]],"date-time":"2025-11-12T08:05:17Z","timestamp":1762934717000},"score":1,"resource":{"primary":{"URL":"https:\/\/aapm.onlinelibrary.wiley.com\/doi\/10.1002\/mp.17881"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,5,21]]},"references-count":32,"journal-issue":{"issue":"6","published-print":{"date-parts":[[2025,6]]}},"alternative-id":["10.1002\/mp.17881"],"URL":"https:\/\/doi.org\/10.1002\/mp.17881","archive":["Portico"],"relation":{},"ISSN":["0094-2405","2473-4209"],"issn-type":[{"value":"0094-2405","type":"print"},{"value":"2473-4209","type":"electronic"}],"subject":[],"published":{"date-parts":[[2025,5,21]]},"assertion":[{"value":"2024-07-02","order":0,"name":"received","label":"Received","group":{"name":"publication_history","label":"Publication History"}},{"value":"2025-02-26","order":2,"name":"accepted","label":"Accepted","group":{"name":"publication_history","label":"Publication History"}},{"value":"2025-05-21","order":3,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}]}}