{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,7]],"date-time":"2026-10-07T21:39:40Z","timestamp":1791409180975,"version":"4.3.3"},"reference-count":39,"publisher":"IOP Publishing","issue":"24","license":[{"start":{"date-parts":[[2025,12,15]],"date-time":"2025-12-15T00:00:00Z","timestamp":1765756800000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"},{"start":{"date-parts":[[2025,12,15]],"date-time":"2025-12-15T00:00:00Z","timestamp":1765756800000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/iopscience.iop.org\/info\/page\/text-and-data-mining"}],"funder":[{"DOI":"10.13039\/100007130","name":"University of Texas System","doi-asserted-by":"crossref","award":["Rising Stars Grant"],"award-info":[{"award-number":["Rising Stars Grant"]}],"id":[{"id":"10.13039\/100007130","id-type":"DOI","asserted-by":"crossref"}]},{"name":"National Cancer Institute","award":["R01CA237269"],"award-info":[{"award-number":["R01CA237269"]}]}],"content-domain":{"domain":["iopscience.iop.org"],"crossmark-restriction":false},"short-container-title":["Phys. Med. Biol."],"published-print":{"date-parts":[[2025,12,21]]},"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:p>\n                    <jats:italic>Objective.<\/jats:italic>\n                    This study aims to uncover the opaque decision-making process of an artificial intelligence (AI) agent for automatic treatment planning.\n                    <jats:italic>Approach.<\/jats:italic>\n                    We examined a previously developed AI agent based on the actor-critic with experience replay (ACER) network, which automatically tunes treatment planning parameters (TPPs) for inverse planning in prostate cancer intensity modulated radiotherapy. We selected multiple checkpoint ACER agents from different stages of training and applied an explainable AI method to analyze the attribution from dose-volume histogram (DVH) inputs to TPP-tuning decisions. We then assessed each agent\u2019s planning efficacy and efficiency, and evaluated their policy space and final TPP tuning space. Combining findings from these approaches, we systematically examined how ACER agents generated high-quality treatment plans in response to different DVH inputs.\n                    <jats:italic>Main results.<\/jats:italic>\n                    Attribution analysis revealed that ACER agents progressively learned to identify dose-violation regions from DVH inputs and promote appropriate TPP-tuning actions to mitigate them. Organ-wise similarities between DVH attributions and dose-violation reductions ranged from 0.25 to 0.5 across tested agents. While all agents achieved comparably high final planning scores, their planning efficiency and stability differed. Agents with stronger attribution-violation similarity required fewer tuning steps ( 12\u201313 vs 22), exhibited a more concentrated TPP-tuning space with lower entropy ( 0.3 vs 0.6), converged on adjusting only a few key TPPs, and showed smaller discrepancies between practical tuning steps and the theoretical steps needed to move from initial values to the final TPP space. Putting together, these findings indicate that high-performing ACER agents can effectively identify dose violations from DVH inputs and employ a global tuning strategy to achieve high-quality treatment planning.\n                    <jats:italic>Significance.<\/jats:italic>\n                    This study demonstrates that the AI agent learns effective TPP-tuning strategies, exhibiting behaviors similar to those of experienced human planners. Improved interpretability of the agent\u2019s decision-making process may enhance clinician trust and inspire new strategies for automatic treatment planning.\n                  <\/jats:p>","DOI":"10.1088\/1361-6560\/ae2561","type":"journal-article","created":{"date-parts":[[2025,11,27]],"date-time":"2025-11-27T22:55:10Z","timestamp":1764284110000},"page":"245016","update-policy":"https:\/\/doi.org\/10.1088\/crossmark-policy","source":"Crossref","is-referenced-by-count":2,"title":["New insights into automatic treatment planning for cancer radiotherapy using explainable artificial intelligence"],"prefix":"10.1088","volume":"70","author":[{"ORCID":"https:\/\/orcid.org\/0009-0009-4737-3254","authenticated-orcid":true,"given":"Md","family":"Mainul Abrar","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-6159-2909","authenticated-orcid":true,"given":"Xun","family":"Jia","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-1828-3251","authenticated-orcid":true,"given":"Yujie","family":"Chi","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"266","published-online":{"date-parts":[[2025,12,15]]},"reference":[{"key":"pmbae2561bib1","doi-asserted-by":"crossref","DOI":"10.1002\/mp.17915","type":"preprint","article-title":"Actor critic with experience replay-based automatic treatment planning for prostate cancer intensity modulated radiotherapy","author":"Abrar","year":"2025"},{"key":"pmbae2561bib2","doi-asserted-by":"publisher","first-page":"1834","DOI":"10.3390\/app12041834","type":"journal-article","article-title":"Torchesegeta: framework for interpretability and explainability of image-based deep learning models","volume":"12","author":"Chatterjee","year":"2022","journal-title":"Appl. Sci."},{"key":"pmbae2561bib3","doi-asserted-by":"publisher","first-page":"e83","DOI":"10.1016\/j.ijrobp.2010.12.007","type":"journal-article","article-title":"Improved planning time and plan quality through multicriteria optimization for intensity-modulated radiotherapy","volume":"82","author":"Craft","year":"2012","journal-title":"Int. J. Radiat. Oncol. Biol. Phys."},{"key":"pmbae2561bib4","doi-asserted-by":"publisher","DOI":"10.1259\/bjr.20230142","type":"journal-article","article-title":"Interpretable artificial intelligence in radiology and radiation oncology","volume":"96","author":"Cui","year":"2023","journal-title":"Br. J. Radiol."},{"key":"pmbae2561bib5","doi-asserted-by":"publisher","first-page":"158","DOI":"10.1109\/TRPMS.2021.3107454","type":"journal-article","article-title":"Artificial intelligence in radiation therapy","volume":"6","author":"Fu","year":"2021","journal-title":"IEEE Trans. Radiat. Plasma Med. Sci."},{"key":"pmbae2561bib6","first-page":"141","type":"journal-article","article-title":"Deep recurrent Q-learning for partially observable MDPs","volume":"45","author":"Hausknecht","year":"2015","journal-title":"AAAI Fall Symposia"},{"key":"pmbae2561bib7","first-page":"pp 217","type":"conference-proceedings","article-title":"Accelerating implementation of artificial intelligence in radiotherapy through explainability","volume":"vol 3554","author":"Heising","year":"2023"},{"key":"pmbae2561bib8","doi-asserted-by":"publisher","DOI":"10.1371\/journal.pmed.1002711","type":"journal-article","article-title":"Deep learning for lung cancer prognostication: a retrospective multi-cohort radiomics study","volume":"15","author":"Hosny","year":"2018","journal-title":"PLoS Med."},{"key":"pmbae2561bib9","article-title":"Self-explainable AI for medical image analysis: a survey and new outlooks","author":"Hou","year":"2024","type":"preprint"},{"key":"pmbae2561bib10","doi-asserted-by":"publisher","DOI":"10.1259\/bjr.20180270","type":"journal-article","article-title":"Automation in intensity modulated radiotherapy treatment planning-a review of recent innovations","volume":"91","author":"Hussein","year":"2018","journal-title":"Br. J. Radiol."},{"key":"pmbae2561bib11","author":"Krishnamurthy","year":"2016","type":"book"},{"key":"pmbae2561bib12","doi-asserted-by":"publisher","first-page":"3853","DOI":"10.21037\/tcr-22-1626","type":"journal-article","article-title":"Utilization of model-agnostic explainable artificial intelligence frameworks in oncology: a narrative review","volume":"11","author":"Ladbury","year":"2022","journal-title":"Transl. Cancer Res."},{"key":"pmbae2561bib13","doi-asserted-by":"publisher","first-page":"8725","DOI":"10.1088\/0031-9155\/58\/24\/8725","type":"journal-article","article-title":"Automatic treatment plan re-optimization for adaptive radiotherapy guided with the initial plan DVHS","volume":"58","author":"Li","year":"2013","journal-title":"Phys. Med. Biol."},{"key":"pmbae2561bib14","doi-asserted-by":"publisher","first-page":"994","DOI":"10.1016\/j.ijrobp.2017.04.023","type":"journal-article","article-title":"Online adaptive radiation therapy","volume":"99","author":"Lim-Reinders","year":"2017","journal-title":"Int. J. Radiat. Oncol. Biol. Phys."},{"key":"pmbae2561bib15","doi-asserted-by":"publisher","first-page":"2877","DOI":"10.1002\/mp.15576","type":"journal-article","article-title":"Automatic inverse treatment planning of gamma knife radiosurgery via deep reinforcement learning","volume":"49","author":"Liu","year":"2022","journal-title":"Med. Phys."},{"key":"pmbae2561bib16","article-title":"Patient-specific deep reinforcement learning for automatic replanning in head-and-neck cancer proton therapy","author":"Madondo","year":"2025","type":"preprint"},{"key":"pmbae2561bib17","doi-asserted-by":"publisher","first-page":"617","DOI":"10.1016\/j.canrad.2021.06.006","type":"journal-article","article-title":"Automation in radiotherapy treatment planning: examples of use in clinical practice and future trends for a complete automated workflow","volume":"25","author":"Meyer","year":"2021","journal-title":"Cancer\/Radioth\u00e9rapie"},{"key":"pmbae2561bib18","article-title":"Guidelines for the choice of the baseline in xai attribution methods","author":"Morasso","year":"2025","type":"preprint"},{"key":"pmbae2561bib19","article-title":"Dynamic deep-reinforcement-learning algorithm in partially observed markov decision processes","author":"Omi","year":"2023","type":"preprint"},{"key":"pmbae2561bib20","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1016\/j.ejmp.2021.12.009","type":"journal-article","article-title":"Deep reinforcement learning for treatment planning in high-dose-rate cervical brachytherapy","volume":"94","author":"Pu","year":"2022","journal-title":"Phys. Medica"},{"key":"pmbae2561bib21","doi-asserted-by":"publisher","first-page":"e1","DOI":"10.1002\/mp.13264","type":"journal-article","article-title":"Deep learning in medical imaging and radiation therapy","volume":"46","author":"Sahiner","year":"2019","journal-title":"Med. Phys."},{"key":"pmbae2561bib22","doi-asserted-by":"publisher","first-page":"84486","DOI":"10.1109\/ACCESS.2022.3197671","type":"journal-article","article-title":"Explainable AI for healthcare 5.0: opportunities and challenges","volume":"10","author":"Saraswat","year":"2022","journal-title":"IEEE Access"},{"key":"pmbae2561bib23","doi-asserted-by":"crossref","first-page":"92","DOI":"10.25080\/Majora-92bf1922-011","type":"journal-article","article-title":"Statsmodels: econometric and statistical modeling with python","volume":"7","author":"Seabold","year":"2010","journal-title":"SciPy"},{"key":"pmbae2561bib24","first-page":"pp 618","type":"conference-proceedings","article-title":"Grad-CAM: visual explanations from deep networks via gradient-based localization","author":"Selvaraju","year":"2017"},{"key":"pmbae2561bib25","doi-asserted-by":"publisher","DOI":"10.1088\/2632-2153\/ab869f","type":"journal-article","article-title":"Synergizing medical imaging and radiotherapy with deep learning","volume":"1","author":"Shan","year":"2020","journal-title":"Mach. Learn.: Sci. Technol."},{"key":"pmbae2561bib26","doi-asserted-by":"publisher","DOI":"10.1088\/1361-6560\/ac09a2","type":"journal-article","article-title":"A hierarchical deep reinforcement learning framework for intelligent automatic treatment planning of prostate cancer intensity modulated radiation therapy","volume":"66","author":"Shen","year":"2021","journal-title":"Phys. Med. Biol."},{"key":"pmbae2561bib27","doi-asserted-by":"publisher","DOI":"10.1088\/1361-6560\/ab18bf","type":"journal-article","article-title":"Intelligent inverse treatment planning via deep reinforcement learning, a proof-of-principle study in high dose-rate brachytherapy for cervical cancer","volume":"64","author":"Shen","year":"2019","journal-title":"Phys. Med. Biol."},{"key":"pmbae2561bib28","doi-asserted-by":"publisher","first-page":"2329","DOI":"10.1002\/mp.14114","type":"journal-article","article-title":"Operating a treatment planning system using a deep-reinforcement learning-based virtual treatment planner for prostate cancer intensity-modulated radiation therapy treatment planning","volume":"47","author":"Shen","year":"2020a","journal-title":"Med. Phys."},{"key":"pmbae2561bib29","doi-asserted-by":"publisher","first-page":"05TR01","DOI":"10.1088\/1361-6560\/ab6f51","type":"journal-article","article-title":"An introduction to deep learning in medical physics: advantages, potential and challenges","volume":"65","author":"Shen","year":"2020b","journal-title":"Phys. Med. Biol."},{"key":"pmbae2561bib30","doi-asserted-by":"publisher","DOI":"10.1073\/pnas.2214840120","type":"journal-article","article-title":"Superhuman artificial intelligence can improve human decision-making by increasing novelty","volume":"120","author":"Shin","year":"2023"},{"key":"pmbae2561bib31","doi-asserted-by":"publisher","first-page":"354","DOI":"10.1038\/nature24270","type":"journal-article","article-title":"Mastering the game of go without human knowledge","volume":"550","author":"Silver","year":"2017","journal-title":"Nature"},{"key":"pmbae2561bib32","doi-asserted-by":"publisher","first-page":"1140","DOI":"10.1126\/science.aar6404","type":"journal-article","article-title":"A general reinforcement learning algorithm that masters chess, shogi and Go through self-play","volume":"362","author":"Silver","year":"2018","journal-title":"Science"},{"key":"pmbae2561bib33","doi-asserted-by":"publisher","DOI":"10.1088\/2057-1976\/ac6d82","type":"journal-article","article-title":"The development of a deep reinforcement learning network for dose-volume-constrained treatment planning in prostate cancer intensity modulated radiotherapy","volume":"8","author":"Sprouts","year":"2022","journal-title":"Biomed. Phys. Eng. Express"},{"key":"pmbae2561bib34","first-page":"pp 3319","type":"conference-proceedings","article-title":"Axiomatic attribution for deep networks","author":"Sundararajan","year":"2017"},{"key":"pmbae2561bib35","doi-asserted-by":"publisher","first-page":"9620","DOI":"10.21037\/qims-24-723","type":"journal-article","article-title":"A literature review of artificial intelligence (AI) for medical image segmentation: from AI and explainable AI to trustworthy ai","volume":"14","author":"Teng","year":"2024","journal-title":"Quant. Imaging Med. Surg."},{"key":"pmbae2561bib36","doi-asserted-by":"publisher","DOI":"10.1177\/1533033819873922","type":"journal-article","article-title":"Artificial intelligence in radiotherapy treatment planning: present and future","volume":"18","author":"Wang","year":"2019","journal-title":"Technol. Cancer Res. Treat."},{"key":"pmbae2561bib37","doi-asserted-by":"publisher","DOI":"10.3389\/fonc.2023.1124458","type":"journal-article","article-title":"An integrated solution of deep reinforcement learning for automatic IMRT treatment planning in non-small-cell lung cancer","volume":"13","author":"Wang","year":"2023","journal-title":"Front. Oncol."},{"key":"pmbae2561bib38","doi-asserted-by":"publisher","DOI":"10.1088\/1361-6560\/ad965d","type":"journal-article","article-title":"Automated treatment planning with deep reinforcement learning for head-and-neck (HN) cancer intensity modulated radiation therapy (IMRT)","volume":"70","author":"Yang","year":"2024","journal-title":"Phys. Med. Biol."},{"key":"pmbae2561bib39","doi-asserted-by":"publisher","DOI":"10.1118\/1.4866886","type":"journal-article","article-title":"A multicriteria framework with voxel-dependent parameters for radiotherapy treatment plan optimization","volume":"41","author":"Zarepisheh","year":"2014","journal-title":"Med. Phys."}],"container-title":["Physics in Medicine &amp; Biology"],"original-title":[],"link":[{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/1361-6560\/ae2561","content-type":"text\/html","content-version":"am","intended-application":"text-mining"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/1361-6560\/ae2561\/pdf","content-type":"application\/pdf","content-version":"am","intended-application":"text-mining"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/1361-6560\/ae2561","content-type":"text\/html","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/1361-6560\/ae2561\/pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/1361-6560\/ae2561\/pdf","content-type":"application\/pdf","content-version":"am","intended-application":"syndication"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/1361-6560\/ae2561\/pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"syndication"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/1361-6560\/ae2561\/pdf","content-type":"application\/pdf","content-version":"am","intended-application":"similarity-checking"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/1361-6560\/ae2561\/pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,12,15]],"date-time":"2025-12-15T10:30:01Z","timestamp":1765794601000},"score":1,"resource":{"primary":{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/1361-6560\/ae2561"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,12,15]]},"references-count":39,"journal-issue":{"issue":"24","published-online":{"date-parts":[[2025,12,15]]},"published-print":{"date-parts":[[2025,12,21]]}},"URL":"https:\/\/doi.org\/10.1088\/1361-6560\/ae2561","relation":{},"ISSN":["0031-9155","1361-6560"],"issn-type":[{"value":"0031-9155","type":"print"},{"value":"1361-6560","type":"electronic"}],"subject":[],"published":{"date-parts":[[2025,12,15]]},"assertion":[{"value":"New insights into automatic treatment planning for cancer radiotherapy using explainable artificial intelligence","name":"article_title","label":"Article Title"},{"value":"Physics in Medicine & Biology","name":"journal_title","label":"Journal Title"},{"value":"paper","name":"article_type","label":"Article Type"},{"value":"\u00a9 2025 The Author(s). Published on behalf of Institute of Physics and Engineering in Medicine by IOP Publishing Ltd","name":"copyright_information","label":"Copyright Information"},{"value":"2025-08-18","name":"date_received","label":"Date Received","group":{"name":"publication_dates","label":"Publication dates"}},{"value":"2025-11-27","name":"date_accepted","label":"Date Accepted","group":{"name":"publication_dates","label":"Publication dates"}},{"value":"2025-12-15","name":"date_epub","label":"Online publication date","group":{"name":"publication_dates","label":"Publication dates"}}]}}