{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,9,28]],"date-time":"2026-09-28T16:08:31Z","timestamp":1790611711976,"version":"4.1.0"},"reference-count":39,"publisher":"Wiley","issue":"7","license":[{"start":{"date-parts":[[2025,4,25]],"date-time":"2025-04-25T00:00:00Z","timestamp":1745539200000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":["aapm.onlinelibrary.wiley.com"],"crossmark-restriction":true},"short-container-title":["Medical Physics"],"published-print":{"date-parts":[[2025,7]]},"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:sec>\n                    <jats:title>Background<\/jats:title>\n                    <jats:p>\n                      Deep learning methods are promising in automating segmentation of organs at risk (OARs) in radiotherapy. However, the lack of a geometric indicator for dosimetry accuracy remains to be a problem. This issue is particularly pronounced in specific radiotherapy treatments where only the proximity of structures to the radiotherapy target affects the dose planning. In cervical cancer high dose\u2010rate (HDR) brachytherapy, treatment planning is motivated by limiting dose to the hottest 2 cubic centimeters (D2cm\n                      <jats:sup>3<\/jats:sup>\n                      ) of the OARs. Similarly, Ethos online adaptive radiotherapy system prioritizes only the closest target structures for adaptive plan generation.\n                    <\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Purpose<\/jats:title>\n                    <jats:p>\n                      We propose a novel geometrically focused deep learning training method and evaluation metric, using cervical brachytherapy as a case study. A distance\u2010penalized (DP) loss function was developed to focus attention on the near\u2010to\u2010target OAR regions. We also introduced and evaluated a novel geometric metric, weighted dice similarity coefficient (wDSC), correlated with OARs D2cm\n                      <jats:sup>3<\/jats:sup>\n                      .\n                    <\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Methods<\/jats:title>\n                    <jats:p>\n                      A model was trained using a 3D U\u2010Net architecture and 170 T2\u2010weighted magnetic resonance (MR) images (56 patients) with clinical contours. The dataset was split into subsets at the patient level: 45 patients (150 scans) as the training set for five\u2010fold cross\u2010validation and 11 patients (20 scans) as the testing set. Another dataset from our institution, consisting of 35 MR scans from 22 cervical cancer patients, was used as an independent internal testing set. A distance map, emphasizing errors near high\u2010risk clinical target volume (CTV\n                      <jats:sub>HR<\/jats:sub>\n                      ), was used to penalize two commonly used loss functions, cross\u2010entropy (CE) loss and DiceCE loss. The wDSC emphasizes the accuracy of OAR regions proximal to CTV\n                      <jats:sub>HR<\/jats:sub>\n                      by incorporating a weighted factor in the original vDSC. The Pearson correlation coefficient (\n                      <jats:italic>r<\/jats:italic>\n                      ) was used to quantify the strength of the relationship between D2cm\n                      <jats:sup>3<\/jats:sup>\n                      accuracy and six evaluation metrics (wDSC and five standard metrics). A physician rated and revised the auto\u2010contours for the clinical acceptability tests.\n                    <\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Results<\/jats:title>\n                    <jats:p>\n                      The wDSC moderately correlated (\n                      <jats:italic>r<\/jats:italic>\n                      \u00a0=\u00a0\u20100.55) with D2cm\n                      <jats:sup>3<\/jats:sup>\n                      accuracy, outperforming standard geometric metrics. Models using DP loss functions consistently yielded higher wDSCs compared to their respective non\u2010DP counterparts. DP loss models also improved D2cm\n                      <jats:sup>3<\/jats:sup>\n                      accuracy, indicating an enhanced accuracy in dosimetry. The clinical acceptability tests revealed that more than 94% of bladder and rectum contours and approximately half of the sigmoid and small bowel contours were clinically accepted.\n                    <\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Conclusion<\/jats:title>\n                    <jats:p>\n                      We developed and evaluated a new geometric metric, wDSC, as a better indicator of D2cm\n                      <jats:sup>3<\/jats:sup>\n                      accuracy, which has the potential to become a surrogate for dosimetric accuracy in cervical brachytherapy. The model with DP loss showed non\u2010statistically significant improvements in geometric and dosimetric performance. This work also holds the potential to be used for precise OARs delineation in adaptive radiotherapy.\n                    <\/jats:p>\n                  <\/jats:sec>","DOI":"10.1002\/mp.17840","type":"journal-article","created":{"date-parts":[[2025,4,25]],"date-time":"2025-04-25T22:15:12Z","timestamp":1745619312000},"update-policy":"https:\/\/doi.org\/10.1002\/crossmark_policy","source":"Crossref","is-referenced-by-count":4,"title":["Geometrically focused training and evaluation of organs\u2010at\u2010risk segmentation via deep learning"],"prefix":"10.1002","volume":"52","author":[{"given":"Ruiyan","family":"Ni","sequence":"first","affiliation":[{"name":"Department of Medical Biophysics University of Toronto Toronto Canada"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Elizabeth","family":"Chuk","sequence":"additional","affiliation":[{"name":"Princess Margaret Cancer Center University Health Network Toronto Canada"},{"name":"Department of Radiation Oncology University of Toronto Toronto Canada"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Kathy","family":"Han","sequence":"additional","affiliation":[{"name":"Princess Margaret Cancer Center University Health Network Toronto Canada"},{"name":"Department of Radiation Oncology University of Toronto Toronto Canada"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Jennifer","family":"Croke","sequence":"additional","affiliation":[{"name":"Princess Margaret Cancer Center University Health Network Toronto Canada"},{"name":"Department of Radiation Oncology University of Toronto Toronto Canada"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Anthony","family":"Fyles","sequence":"additional","affiliation":[{"name":"Princess Margaret Cancer Center University Health Network Toronto Canada"},{"name":"Department of Radiation Oncology University of Toronto Toronto Canada"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Jelena","family":"Lukovic","sequence":"additional","affiliation":[{"name":"Princess Margaret Cancer Center University Health Network Toronto Canada"},{"name":"Department of Radiation Oncology University of Toronto Toronto Canada"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Michael","family":"Milosevic","sequence":"additional","affiliation":[{"name":"Princess Margaret Cancer Center University Health Network Toronto Canada"},{"name":"Department of Radiation Oncology University of Toronto Toronto Canada"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Benjamin","family":"Haibe\u2010Kains","sequence":"additional","affiliation":[{"name":"Department of Medical Biophysics University of Toronto Toronto Canada"},{"name":"Princess Margaret Cancer Center University Health Network Toronto Canada"},{"name":"Vector Institute for Artificial Intelligence Toronto Canada"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Alexandra","family":"Rink","sequence":"additional","affiliation":[{"name":"Department of Medical Biophysics University of Toronto Toronto Canada"},{"name":"Princess Margaret Cancer Center University Health Network Toronto Canada"},{"name":"Department of Radiation Oncology University of Toronto Toronto Canada"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"311","published-online":{"date-parts":[[2025,4,25]]},"reference":[{"key":"e_1_2_8_2_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.clon.2024.03.005"},{"key":"e_1_2_8_3_1","doi-asserted-by":"publisher","DOI":"10.1120\/jacmp.v17i4.6051"},{"key":"e_1_2_8_4_1","doi-asserted-by":"publisher","DOI":"10.1186\/1748\u2010717x\u20107\u2010160"},{"key":"e_1_2_8_5_1","doi-asserted-by":"publisher","DOI":"10.1002\/mp.15490"},{"key":"e_1_2_8_6_1","doi-asserted-by":"publisher","DOI":"10.5114\/jcb.2021.106118"},{"key":"e_1_2_8_7_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.ijrobp.2020.04.011"},{"key":"e_1_2_8_8_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.radonc.2010.06.009"},{"key":"e_1_2_8_9_1","doi-asserted-by":"publisher","DOI":"10.1109\/34.232073"},{"key":"e_1_2_8_10_1","unstructured":"ReinkeA TizabiMD BaumgartnerM et\u00a0al.Understanding metric\u2010related pitfalls in image analysis validation. arXiv preprint arXiv:230201790. Published online 2023."},{"key":"e_1_2_8_11_1","doi-asserted-by":"publisher","DOI":"10.1002\/mp.13200"},{"key":"e_1_2_8_12_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.phro.2019.12.001"},{"key":"e_1_2_8_13_1","doi-asserted-by":"publisher","DOI":"10.2196\/26151"},{"key":"e_1_2_8_14_1","doi-asserted-by":"publisher","DOI":"10.1007\/s10278\u2010021\u201000460\u20103"},{"issue":"2","key":"e_1_2_8_15_1","article-title":"Making on\u2010line adaptive radiotherapy possible using artificial intelligence and machine learning for efficient daily re\u2010planning","volume":"8","author":"Archambault Y","year":"2020","journal-title":"Med Phys Int J"},{"key":"e_1_2_8_16_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.radonc.2005.11.014"},{"key":"e_1_2_8_17_1","unstructured":"CalivaF IriondoC MartinezAM MajumdarS PedoiaV.Distance map loss penalty term for semantic segmentation. arXiv preprint arXiv:190803679. Published online 2019."},{"key":"e_1_2_8_18_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.media.2021.102035"},{"key":"e_1_2_8_19_1","article-title":"ICRU report 89: prescribing, recording, and reporting brachytherapy for cancer of the cervix","author":"Swamidas J","year":"2017","journal-title":"J Med Phys"},{"key":"e_1_2_8_20_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.radonc.2024.110332"},{"key":"e_1_2_8_21_1","doi-asserted-by":"publisher","DOI":"10.1016\/S0734-189X(86)80047-0"},{"key":"e_1_2_8_22_1","doi-asserted-by":"publisher","DOI":"10.1038\/s41592\u2010020\u201001008\u2010z"},{"key":"e_1_2_8_23_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.compmedimag.2019.04.005"},{"key":"e_1_2_8_24_1","doi-asserted-by":"publisher","DOI":"10.1117\/1.JMI.5.1.015006"},{"key":"e_1_2_8_25_1","doi-asserted-by":"crossref","unstructured":"MarsillaJ Won KimJ TkachuckD et\u00a0al.SCARF: Auto\u2010Segmentation Clinical Acceptability & Reproducibility Framework for benchmarking essential radiation therapy targets in head and neck cancer.bioRxiv. Published online January 25 2022. doi:10.1101\/2022.01.15.22269276","DOI":"10.1101\/2022.01.15.22269276"},{"key":"e_1_2_8_26_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.ijrobp.2014.03.005"},{"key":"e_1_2_8_27_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.brachy.2013.08.010"},{"key":"e_1_2_8_28_1","doi-asserted-by":"publisher","DOI":"10.1002\/mp.15506"},{"key":"e_1_2_8_29_1","doi-asserted-by":"publisher","DOI":"10.1002\/acm2.13024"},{"key":"e_1_2_8_30_1","doi-asserted-by":"publisher","DOI":"10.1002\/mp.15268"},{"key":"e_1_2_8_31_1","doi-asserted-by":"publisher","DOI":"10.1002\/acm2.13951"},{"key":"e_1_2_8_32_1","first-page":"755","article-title":"OAR\u2010Weighted Dice Score: a spatially aware, radiosensitivity aware metric for target structure contour quality assessment","volume":"2024","author":"McCullum L","year":"2024","journal-title":"Use Comput Radiat Ther"},{"key":"e_1_2_8_33_1","doi-asserted-by":"publisher","DOI":"10.1002\/acm2.13479"},{"key":"e_1_2_8_34_1","doi-asserted-by":"publisher","DOI":"10.1002\/acm2.14057"},{"key":"e_1_2_8_35_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.brachy.2023.04.003"},{"key":"e_1_2_8_36_1","unstructured":"IsenseeF WaldT UlrichC et\u00a0al.NnU\u2010Net revisited: A call for rigorous validation in 3D medical image segmentation.arXiv [csCV]. Published online April 15 2024.http:\/\/arxiv.org\/abs\/2404.09556"},{"key":"e_1_2_8_37_1","doi-asserted-by":"crossref","unstructured":"HuangL MironA HoneK LiY.Segmenting medical images: from UNet to res\u2010UNet and nnUNet. In:2024 IEEE 37th International Symposium on Computer\u2010Based Medical Systems (CBMS).IEEE;2024. doi:10.1109\/cbms61543.2024.00086","DOI":"10.1109\/CBMS61543.2024.00086"},{"key":"e_1_2_8_38_1","unstructured":"D'AntonoliTA BergerLK IndrakantiAK et\u00a0al.TotalSegmentator MRI: Sequence\u2010Independent Segmentation of 59 Anatomical Structures in MR images.arXiv [eessIV]. Published online May 29 2024.http:\/\/arxiv.org\/abs\/2405.19492"},{"key":"e_1_2_8_39_1","doi-asserted-by":"publisher","DOI":"10.1088\/1361\u20106560\/ad7ad1"},{"key":"e_1_2_8_40_1","doi-asserted-by":"publisher","DOI":"10.1186\/s13014\u2010022\u201002121\u20103"}],"container-title":["Medical Physics"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/aapm.onlinelibrary.wiley.com\/doi\/pdf\/10.1002\/mp.17840","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,12,8]],"date-time":"2025-12-08T10:37:08Z","timestamp":1765190228000},"score":1,"resource":{"primary":{"URL":"https:\/\/aapm.onlinelibrary.wiley.com\/doi\/10.1002\/mp.17840"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,4,25]]},"references-count":39,"journal-issue":{"issue":"7","published-print":{"date-parts":[[2025,7]]}},"alternative-id":["10.1002\/mp.17840"],"URL":"https:\/\/doi.org\/10.1002\/mp.17840","archive":["Portico"],"relation":{},"ISSN":["0094-2405","2473-4209"],"issn-type":[{"value":"0094-2405","type":"print"},{"value":"2473-4209","type":"electronic"}],"subject":[],"published":{"date-parts":[[2025,4,25]]},"assertion":[{"value":"2024-10-27","order":0,"name":"received","label":"Received","group":{"name":"publication_history","label":"Publication History"}},{"value":"2025-04-08","order":2,"name":"accepted","label":"Accepted","group":{"name":"publication_history","label":"Publication History"}},{"value":"2025-04-25","order":3,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}],"article-number":"e17840"}}