{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,9,1]],"date-time":"2026-09-01T14:19:18Z","timestamp":1788272358682,"version":"build-2803163510"},"reference-count":26,"publisher":"Wiley","issue":"8","license":[{"start":{"date-parts":[[2026,8,13]],"date-time":"2026-08-13T00:00:00Z","timestamp":1786579200000},"content-version":"vor","delay-in-days":12,"URL":"http:\/\/onlinelibrary.wiley.com\/termsAndConditions#vor"},{"start":{"date-parts":[[2026,8,1]],"date-time":"2026-08-01T00:00:00Z","timestamp":1785542400000},"content-version":"tdm","delay-in-days":0,"URL":"http:\/\/doi.wiley.com\/10.1002\/tdm_license_1.1"}],"content-domain":{"domain":["aapm.onlinelibrary.wiley.com"],"crossmark-restriction":true},"short-container-title":["Medical Physics"],"published-print":{"date-parts":[[2026,8]]},"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:sec>\n                    <jats:title>Background<\/jats:title>\n                    <jats:p>Ocular proton therapy (OPT) is the main conservative treatment option for ocular melanoma but requires precise gaze stability monitoring during irradiation. In current clinical workflows at the Centro Nazionale di Adroterapia Oncologica (CNAO) in Pavia, this task is performed manually by trained operators, which can introduce reaction delays and human error.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Purpose<\/jats:title>\n                    <jats:p>\n                      To propose and assess the performance of\n                      <jats:italic>OPT\u2010former<\/jats:italic>\n                      , a transformer\u2010based method for real\u2010time segmentation of the iris and pupil, designed for integration into automated safety interlock systems.\n                    <\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Methods<\/jats:title>\n                    <jats:p>\n                      The model, inspired by the MedFormer architecture, performs direct three\u2010class segmentation (background, iris, pupil) on 256\u00a0\u00d7\u00a0256 images acquired by CNAO's Eye Tracking System. A dataset of 1683 frames from 173 patients was split into independent training (1349 frames), validation (110 frames), and testing (224 frames) sets. Data augmentation was applied to increase variability, resulting in 3902 training images. Performance was evaluated using Intersection over Union (IoU), Dice coefficient, and Szymkiewicz\u2013Simpson coefficient (SSC), assessing the\n                      <jats:italic>OPT\u2010former<\/jats:italic>\n                      performance and comparing them with a previously proposed\n                      <jats:italic>3\u2010step U\u2010Net<\/jats:italic>\n                      and manual expert segmentations. Patient\u2010specific fine\u2010tuning was implemented using pre\u2010treatment dry\u2010run session videos.\n                    <\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Results<\/jats:title>\n                    <jats:p>\n                      The\n                      <jats:italic>OPT\u2010former<\/jats:italic>\n                      achieved segmentation accuracy comparable to the\n                      <jats:italic>3\u2010step U\u2010Net<\/jats:italic>\n                      , but with substantially lower inference times (8.62\u00a0Hz on GPU vs. 3.26\u00a0Hz), enabling near frame\u2010by\u2010frame real\u2010time analysis. Patient\u2010specific fine\u2010tuning significantly improved all metrics for both structures (\n                      <jats:italic>p<\/jats:italic>\n                      \u00a0&lt;\u00a00.05), with the largest gains for the iris. After fine\u2010tuning, the model outperformed manual annotations from two experts under simulated clinical conditions.\n                    <\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Conclusions<\/jats:title>\n                    <jats:p>\n                      The proposed\n                      <jats:italic>OPT\u2010former<\/jats:italic>\n                      achieves high segmentation accuracy, with reduced pipeline complexity and inference time suitable for clinical application. These results endorse its role as a robust support tool in clinical practice, seamlessly integrating into established workflows without impacting the consolidated routine, and enabling automated gaze monitoring and real\u2010time safety interlock activation in OPT.\n                    <\/jats:p>\n                  <\/jats:sec>","DOI":"10.1002\/mp.70526","type":"journal-article","created":{"date-parts":[[2026,8,14]],"date-time":"2026-08-14T03:04:03Z","timestamp":1786676643000},"update-policy":"https:\/\/doi.org\/10.1002\/crossmark_policy","source":"Crossref","is-referenced-by-count":0,"title":["OPT\u2010former: A transformer\u2010based approach for eye motion tracking in ocular proton therapy"],"prefix":"10.1002","volume":"53","author":[{"given":"Celeste","family":"Pinori","sequence":"first","affiliation":[{"name":"Department of Electronics, Information and Bioengineering Politecnico di Milano University  Milan Italy"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Gabriele","family":"Ariata","sequence":"additional","affiliation":[{"name":"Department of Electronics, Information and Bioengineering Politecnico di Milano University  Milan Italy"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Maksym","family":"Hladchuk","sequence":"additional","affiliation":[{"name":"Department of Electronics, Information and Bioengineering Politecnico di Milano University  Milan Italy"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Andrea","family":"Pella","sequence":"additional","affiliation":[{"name":"Bioengineering Unit, Clinical Department National Center for Oncological Hadrontherapy (CNAO)  Pavia Italy"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Maria Rosaria","family":"Fiore","sequence":"additional","affiliation":[{"name":"Radiotherapy Unit, Clinical Department National Center for Oncological Hadrontherapy (CNAO)  Pavia Italy"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Giuseppe","family":"Magro","sequence":"additional","affiliation":[{"name":"Medical Physics Unit, Clinical Department National Center for Oncological Hadrontherapy (CNAO)  Pavia Italy"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Chiara","family":"Paganelli","sequence":"additional","affiliation":[{"name":"Department of Electronics, Information and Bioengineering Politecnico di Milano University  Milan Italy"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Agnieszka","family":"Chalaszczyk","sequence":"additional","affiliation":[{"name":"Radiotherapy Unit, Clinical Department National Center for Oncological Hadrontherapy (CNAO)  Pavia Italy"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Marco","family":"Rotondi","sequence":"additional","affiliation":[{"name":"Radiotherapy Unit, Clinical Department National Center for Oncological Hadrontherapy (CNAO)  Pavia Italy"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Mario","family":"Ciocca","sequence":"additional","affiliation":[{"name":"Medical Physics Unit, Clinical Department National Center for Oncological Hadrontherapy (CNAO)  Pavia Italy"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Guido","family":"Baron","sequence":"additional","affiliation":[{"name":"Department of Electronics, Information and Bioengineering Politecnico di Milano University  Milan Italy"},{"name":"Bioengineering Unit, Clinical Department National Center for Oncological Hadrontherapy (CNAO)  Pavia Italy"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"311","published-online":{"date-parts":[[2026,8,13]]},"reference":[{"key":"e_1_2_8_2_1","doi-asserted-by":"publisher","DOI":"10.3389\/fonc.2023.1133909"},{"key":"e_1_2_8_3_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.addr.2016.11.006"},{"key":"e_1_2_8_4_1","doi-asserted-by":"publisher","DOI":"10.21037\/cco.2016.07.06"},{"key":"e_1_2_8_5_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.ijrobp.2006.01.020"},{"key":"e_1_2_8_6_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.ijrobp.2024.06.017"},{"key":"e_1_2_8_7_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.ajo.2016.02.027"},{"key":"e_1_2_8_8_1","doi-asserted-by":"publisher","DOI":"10.1007\/s00417\u2010019\u201004286\u20102"},{"key":"e_1_2_8_9_1","doi-asserted-by":"publisher","DOI":"10.3390\/s21134400"},{"key":"e_1_2_8_10_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.ejmp.2019.02.020"},{"key":"e_1_2_8_11_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.canrad.2022.03.004"},{"key":"e_1_2_8_12_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.ejmp.2021.01.070"},{"key":"e_1_2_8_13_1","doi-asserted-by":"publisher","DOI":"10.1002\/mp.13389"},{"key":"e_1_2_8_14_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.ejmp.2023.102551"},{"key":"e_1_2_8_15_1","doi-asserted-by":"publisher","DOI":"10.1016\/S0360\u20103016(03)00280\u20103"},{"key":"e_1_2_8_16_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.ijrobp.2016.01.040"},{"key":"e_1_2_8_17_1","first-page":"5998","article-title":"Attention is all you need","volume":"30","author":"Vaswani A","year":"2017","journal-title":"Adv Neural Inf Process Syst"},{"key":"e_1_2_8_18_1","unstructured":"DosovitskiyA BeyerL KolesnikovA et al.An Image Is Worth 16\u00d716 Words: Transformers for Image Recognition at Scale.In: 9th International Conference on Learning Representations (ICLR 2021). OpenReview;2021. Paper ID: 11929. arXiv:2010.11929."},{"key":"e_1_2_8_19_1","doi-asserted-by":"crossref","unstructured":"CaoH WangY ChenJ et al.Swin\u2010Unet: Unet\u2010like Pure Transformer for Medical Image Segmentation.In: Proceedings of the European Conference on Computer Vision Workshops (ECCVW).2022.","DOI":"10.1007\/978-3-031-25066-8_9"},{"key":"e_1_2_8_20_1","unstructured":"GaoY ZhouM LiuD YanZ ZhangS MetaxasDN.A Data\u2010Scalable Transformer for Medical Image Segmentation: Architecture Model Efficiency and Benchmark.arXiv preprint arXiv:2203.00131.2022. doi:10.48550\/arXiv.2203.0013"},{"key":"e_1_2_8_21_1","doi-asserted-by":"crossref","unstructured":"SzegedyC LiuW JiaY et\u00a0al.Going Deeper With Convolutions. In:Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR).2015. doi:10.1109\/CVPR.2015.7298594","DOI":"10.1109\/CVPR.2015.7298594"},{"key":"e_1_2_8_22_1","doi-asserted-by":"publisher","DOI":"10.3390\/info11020125"},{"key":"e_1_2_8_23_1","doi-asserted-by":"publisher","DOI":"10.1186\/s12880\u2010015\u20100068\u2010x"},{"key":"e_1_2_8_24_1","doi-asserted-by":"publisher","DOI":"10.1186\/s12911\u2010023\u201002362\u20106"},{"key":"e_1_2_8_25_1","unstructured":"Tensorflow.tf.keras.losses\u2014TensorFlow API Documentation. Preprint posted online2025.https:\/\/www.tensorflow.org\/api_docs\/python\/tf\/keras\/losses"},{"key":"e_1_2_8_26_1","doi-asserted-by":"publisher","DOI":"10.1109\/ACCESS.2019.2962617"},{"key":"e_1_2_8_27_1","doi-asserted-by":"crossref","unstructured":"WobbrockJO FindlaterL GergleD HigginsJJ The aligned rank transform for nonparametric factorial analyses using only anova procedures. In:Proceedings of the SIGCHI Conference on Human Factors in Computing Systems.2011:143\u2010146. doi:10.1145\/1978942.1978963","DOI":"10.1145\/1978942.1978963"}],"container-title":["Medical Physics"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/aapm.onlinelibrary.wiley.com\/doi\/pdf\/10.1002\/mp.70526","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/aapm.onlinelibrary.wiley.com\/doi\/full-xml\/10.1002\/mp.70526","content-type":"application\/xml","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/aapm.onlinelibrary.wiley.com\/doi\/pdf\/10.1002\/mp.70526","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,9,1]],"date-time":"2026-09-01T13:43:53Z","timestamp":1788270233000},"score":1,"resource":{"primary":{"URL":"https:\/\/aapm.onlinelibrary.wiley.com\/doi\/10.1002\/mp.70526"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,8]]},"references-count":26,"journal-issue":{"issue":"8","published-print":{"date-parts":[[2026,8]]}},"alternative-id":["10.1002\/mp.70526"],"URL":"https:\/\/doi.org\/10.1002\/mp.70526","archive":["Portico"],"relation":{},"ISSN":["0094-2405","2473-4209"],"issn-type":[{"value":"0094-2405","type":"print"},{"value":"2473-4209","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,8]]},"assertion":[{"value":"2025-09-05","order":0,"name":"received","label":"Received","group":{"name":"publication_history","label":"Publication History"}},{"value":"2026-05-20","order":2,"name":"accepted","label":"Accepted","group":{"name":"publication_history","label":"Publication History"}},{"value":"2026-08-13","order":3,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}],"article-number":"e70526"}}