{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,7,17]],"date-time":"2026-07-17T11:24:00Z","timestamp":1784287440265,"version":"3.55.0"},"reference-count":24,"publisher":"Wiley","issue":"2","license":[{"start":{"date-parts":[[2023,8,6]],"date-time":"2023-08-06T00:00:00Z","timestamp":1691280000000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/onlinelibrary.wiley.com\/termsAndConditions#vor"}],"content-domain":{"domain":["aapm.onlinelibrary.wiley.com"],"crossmark-restriction":true},"short-container-title":["Medical Physics"],"published-print":{"date-parts":[[2024,2]]},"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:sec>\n                    <jats:title>Background<\/jats:title>\n                    <jats:p>Prostate\u2010specific membrane antigen (PSMA) PET imaging represents a valuable source of information reflecting disease stage, response rate, and treatment optimization options, particularly with PSMA radioligand therapy. Quantification of radiopharmaceutical uptake in healthy organs from PSMA images has the potential to minimize toxicity by extrapolation of the radiation dose delivery towards personalization of therapy. However, segmentation and quantification of uptake in organs requires labor\u2010intensive organ delineations that are often not feasible in the clinic nor scalable for large clinical trials.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Purpose<\/jats:title>\n                    <jats:p>\n                      In this work we develop and test the PSMA Healthy organ segmentation network (PSMA\u2010Hornet), a fully\u2010automated deep neural net for simultaneous segmentation of 14 healthy organs representing the normal biodistribution of [\n                      <jats:sup>18<\/jats:sup>\n                      F]DCFPyL on PET\/CT images. We also propose a modified U\u2010net architecture, a self\u2010supervised pre\u2010training method for PET\/CT images, a multi\u2010target Dice loss, and multi\u2010target batch balancing to effectively train PSMA\u2010Hornet and similar networks.\n                    <\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Methods<\/jats:title>\n                    <jats:p>\n                      The study used manually\u2010segmented [\n                      <jats:sup>18<\/jats:sup>\n                      F]DCFPyL PET\/CT images from 100 subjects, and 526 similar images without segmentations. The unsegmented images were used for self\u2010supervised model pretraining. For supervised training, Monte\u2010Carlo cross\u2010validation was used to evaluate the network performance, with 85 subjects in each trial reserved for model training, 5 for validation, and 10 for testing. Image segmentation and quantification metrics were evaluated on the test folds with respect to manual segmentations by a nuclear medicine physician, and compared to inter\u2010rater agreement. The model's segmentation performance was also evaluated on a separate set of 19 images with high tumor load.\n                    <\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Results<\/jats:title>\n                    <jats:p>With our best model, the lowest mean Dice coefficient on the test set was 0.826 for the sublingual gland, and the highest was 0.964 for liver. The highest mean error in tracer uptake quantification was 13.9% in the sublingual gland. Self\u2010supervised pretraining improved training convergence, train\u2010to\u2010test generalization, and segmentation quality. In addition, we found that a multi\u2010target network produced significantly higher segmentation accuracy than single\u2010organ networks.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Conclusions<\/jats:title>\n                    <jats:p>The developed network can be used to automatically obtain high\u2010quality organ segmentations for PSMA image analysis tasks. It can be used to reproducibly extract imaging data, and holds promise for clinical applications such as personalized radiation dosimetry and improved radioligand therapy.<\/jats:p>\n                  <\/jats:sec>","DOI":"10.1002\/mp.16658","type":"journal-article","created":{"date-parts":[[2023,8,6]],"date-time":"2023-08-06T12:05:07Z","timestamp":1691323507000},"page":"1203-1216","update-policy":"https:\/\/doi.org\/10.1002\/crossmark_policy","source":"Crossref","is-referenced-by-count":12,"title":["PSMA\u2010Hornet: Fully\u2010automated, multi\u2010target segmentation of healthy organs in PSMA PET\/CT images"],"prefix":"10.1002","volume":"51","author":[{"given":"Ivan S.","family":"Klyuzhin","sequence":"first","affiliation":[{"name":"Department of Integrative Oncology BC Cancer Research Institute Vancouver BC Canada"},{"name":"AI for Health Microsoft Redmond Washington USA"},{"name":"Deparment of Radiology University of British Columbia Vancouver BC Canada"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Guillaume","family":"Chauss\u00e9","sequence":"additional","affiliation":[{"name":"Department of Integrative Oncology BC Cancer Research Institute Vancouver BC Canada"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Ingrid","family":"Bloise","sequence":"additional","affiliation":[{"name":"Department of Integrative Oncology BC Cancer Research Institute Vancouver BC Canada"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Sara","family":"Harsini","sequence":"additional","affiliation":[{"name":"Department of Integrative Oncology BC Cancer Research Institute Vancouver BC Canada"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Juan Lavista","family":"Ferres","sequence":"additional","affiliation":[{"name":"AI for Health Microsoft Redmond Washington USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Carlos","family":"Uribe","sequence":"additional","affiliation":[{"name":"Deparment of Radiology University of British Columbia Vancouver BC Canada"},{"name":"Department of Functional Imaging BC Cancer Vancouver BC Canada"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Arman","family":"Rahmim","sequence":"additional","affiliation":[{"name":"Department of Integrative Oncology BC Cancer Research Institute Vancouver BC Canada"},{"name":"Deparment of Radiology University of British Columbia Vancouver BC Canada"},{"name":"Deparment of Physics and Astronomy University of British Columbia Vancouver BC Canada"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"311","published-online":{"date-parts":[[2023,8,6]]},"reference":[{"key":"e_1_2_8_2_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.ijrobp.2018.08.066"},{"key":"e_1_2_8_3_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.eururo.2014.06.045"},{"key":"e_1_2_8_4_1","doi-asserted-by":"publisher","DOI":"10.14740\/wjon1191"},{"key":"e_1_2_8_5_1","doi-asserted-by":"publisher","DOI":"10.3322\/caac.21442"},{"key":"e_1_2_8_6_1","doi-asserted-by":"publisher","DOI":"10.1002\/jcb.10661"},{"key":"e_1_2_8_7_1","doi-asserted-by":"publisher","DOI":"10.1007\/s00259-014-2949-6"},{"key":"e_1_2_8_8_1","doi-asserted-by":"publisher","DOI":"10.18632\/oncotarget.12240"},{"key":"e_1_2_8_9_1","doi-asserted-by":"publisher","DOI":"10.1007\/s00259-011-1937-3"},{"key":"e_1_2_8_10_1","doi-asserted-by":"publisher","DOI":"10.2967\/jnumed.118.219352"},{"key":"e_1_2_8_11_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.euo.2018.11.007"},{"key":"e_1_2_8_12_1","doi-asserted-by":"publisher","DOI":"10.1007\/s00259-020-04723-z"},{"key":"e_1_2_8_13_1","doi-asserted-by":"publisher","DOI":"10.1007\/s00259-020-05040-1"},{"key":"e_1_2_8_14_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.cpet.2021.06.002"},{"key":"e_1_2_8_15_1","doi-asserted-by":"publisher","DOI":"10.2967\/jnumed.119.226381"},{"issue":"1","key":"e_1_2_8_16_1","first-page":"S284","article-title":"A 3\u2010minute semi\u2010automated PSMA PET organ segmentation method to use in the clinic and train neural networks","volume":"48","author":"Chauss\u00e9 G","year":"2021","journal-title":"Eur J Nucl Med Mol Imaging"},{"key":"e_1_2_8_17_1","doi-asserted-by":"crossref","unstructured":"SzegedyC WeiL YangqingJ et\u00a0al.Going deeper with convolutions.Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition;1\u20109;2015.","DOI":"10.1109\/CVPR.2015.7298594"},{"key":"e_1_2_8_18_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.sigpro.2021.108036"},{"key":"e_1_2_8_19_1","unstructured":"IoffeS SzegedyC.Batch normalization: accelerating deep network training by reducing internal covariate shift.Proceedings of the 32nd International Conference on Machine Learning (ICML'15)(Lille France);37:448\u2010456;2015."},{"key":"e_1_2_8_20_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.media.2019.101539"},{"key":"e_1_2_8_21_1","doi-asserted-by":"crossref","unstructured":"LiX SunX MengY LiangJ WuF LiJ.Dice Loss for Data\u2010imbalanced NLP Tasks.Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics;465\u2010476;2020.","DOI":"10.18653\/v1\/2020.acl-main.45"},{"key":"e_1_2_8_22_1","doi-asserted-by":"publisher","DOI":"10.1007\/s00259-016-3573-4"},{"key":"e_1_2_8_23_1","doi-asserted-by":"publisher","DOI":"10.1109\/TMI.2014.2325596"},{"key":"e_1_2_8_24_1","doi-asserted-by":"publisher","DOI":"10.1038\/s41592-020-01008-z"},{"key":"e_1_2_8_25_1","doi-asserted-by":"crossref","unstructured":"SinghS KrishnanS Filter response normalization layer: eliminating batch dependence in the training of deep neural networks.Proceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition (CVPR);11237\u201011246;2020.","DOI":"10.1109\/CVPR42600.2020.01125"}],"container-title":["Medical Physics"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/aapm.onlinelibrary.wiley.com\/doi\/pdf\/10.1002\/mp.16658","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2024,2,8]],"date-time":"2024-02-08T15:15:05Z","timestamp":1707405305000},"score":1,"resource":{"primary":{"URL":"https:\/\/aapm.onlinelibrary.wiley.com\/doi\/10.1002\/mp.16658"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2023,8,6]]},"references-count":24,"journal-issue":{"issue":"2","published-print":{"date-parts":[[2024,2]]}},"alternative-id":["10.1002\/mp.16658"],"URL":"https:\/\/doi.org\/10.1002\/mp.16658","archive":["Portico"],"relation":{"has-preprint":[{"id-type":"doi","id":"10.1101\/2022.02.02.22270344","asserted-by":"object"}]},"ISSN":["0094-2405","2473-4209"],"issn-type":[{"value":"0094-2405","type":"print"},{"value":"2473-4209","type":"electronic"}],"subject":[],"published":{"date-parts":[[2023,8,6]]},"assertion":[{"value":"2022-09-24","order":0,"name":"received","label":"Received","group":{"name":"publication_history","label":"Publication History"}},{"value":"2023-05-26","order":1,"name":"accepted","label":"Accepted","group":{"name":"publication_history","label":"Publication History"}},{"value":"2023-08-06","order":2,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}]}}