{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,8,28]],"date-time":"2026-08-28T08:33:56Z","timestamp":1787906036952,"version":"build-2784847793"},"reference-count":27,"publisher":"Wiley","issue":"11","license":[{"start":{"date-parts":[[2024,8,14]],"date-time":"2024-08-14T00:00:00Z","timestamp":1723593600000},"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":[[2024,11]]},"abstract":"<jats:title>Abstract<\/jats:title><jats:sec><jats:title>Background<\/jats:title><jats:p>Recent advancements in anomaly detection have paved the way for novel radiological reading assistance tools that support the identification of findings, aimed at saving time. The clinical adoption of such applications requires a low rate of false positives while maintaining high sensitivity.<\/jats:p><\/jats:sec><jats:sec><jats:title>Purpose<\/jats:title><jats:p>In light of recent interest and development in multi pathology identification, we present a novel method, based on a recent contrastive self\u2010supervised approach, for multiple chest\u2010related abnormality identification including low lung density area (\u201cLLDA\u201d), consolidation (\u201cCONS\u201d), nodules (\u201cNOD\u201d) and interstitial pattern (\u201cIP\u201d). Our approach alerts radiologists about abnormal regions within a computed tomography (CT) scan by providing 3D\u00a0localization.<\/jats:p><\/jats:sec><jats:sec><jats:title>Methods<\/jats:title><jats:p>We introduce a new method for the classification and localization of multiple chest pathologies in 3D Chest CT scans. Our goal is to distinguish four common chest\u2010related abnormalities: \u201cLLDA\u201d, \u201cCONS\u201d, \u201cNOD\u201d, \u201cIP\u201d and \u201cNORMAL\u201d. This method is based on a 3D patch\u2010based classifier with a Resnet backbone encoder pretrained leveraging recent contrastive self supervised approach and a fine\u2010tuned classification head. We leverage the SimCLR contrastive framework for pretraining on an unannotated dataset of randomly selected patches and we then fine\u2010tune it on a labeled dataset. During inference, this classifier generates probability maps for each abnormality across the CT volume, which are aggregated to produce a multi\u2010label patient\u2010level prediction. We compare different training strategies, including random initialization, ImageNet weight initialization, frozen SimCLR pretrained weights and fine\u2010tuned SimCLR pretrained weights. Each training strategy is evaluated on a validation set for hyperparameter selection and tested on a test set. Additionally, we explore the fine\u2010tuned SimCLR pretrained classifier for 3D pathology localization and conduct qualitative\u00a0evaluation.<\/jats:p><\/jats:sec><jats:sec><jats:title>Results<\/jats:title><jats:p>Validated on 111 chest scans for hyperparameter selection and subsequently tested on 251 chest scans with multi\u2010abnormalities, our method achieves an AUROC of 0.931 (95% confidence interval [CI]: [0.9034, 0.9557], \u2010value &lt; 0.001) and 0.963 (95% CI: [0.952, 0.976], \u2010value &lt; 0.001) in the multi\u2010label and binary (i.e., normal versus abnormal) settings, respectively. Notably, our method surpasses the area under the receiver operating characteristic (AUROC) threshold of 0.9 for two abnormalities: IP (0.974) and LLDA (0.952), while achieving values of 0.853 and 0.791 for NOD and CONS, respectively. Furthermore, our results highlight the superiority of incorporating contrastive pretraining within the patch classifier, outperforming Imagenet pretraining weights and non\u2010pretrained counterparts with uninitialized weights (F1 score = 0.943, 0.792, and 0.677 respectively). Qualitatively, the method achieved a satisfactory 88.8% completeness rate in localization and maintained an 88.3% accuracy rate against false\u00a0positives.<\/jats:p><\/jats:sec><jats:sec><jats:title>Conclusions<\/jats:title><jats:p>The proposed method integrates self\u2010supervised learning algorithms for pretraining, utilizes a patch\u2010based approach for 3D pathology localization and develops an aggregation method for multi\u2010label prediction at patient\u2010level. It shows promise in efficiently detecting and localizing multiple anomalies within a single\u00a0scan.<\/jats:p><\/jats:sec>","DOI":"10.1002\/mp.17302","type":"journal-article","created":{"date-parts":[[2024,8,14]],"date-time":"2024-08-14T13:21:25Z","timestamp":1723641685000},"page":"8272-8282","update-policy":"https:\/\/doi.org\/10.1002\/crossmark_policy","source":"Crossref","is-referenced-by-count":4,"title":["Weakly\u2010supervised learning\u2010based pathology detection and localization in 3D chest CT scans"],"prefix":"10.1002","volume":"51","author":[{"given":"Aissam","family":"Djahnine","sequence":"first","affiliation":[{"name":"CREATIS UMR5220, INSERM U1044 Claude Bernard University Lyon 1 INSA Lyon France"},{"name":"Philips Health Technology innovation Paris France"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Emilien","family":"Jupin\u2010Delevaux","sequence":"additional","affiliation":[{"name":"Department of Radiology Hospices Civils de Lyon Lyon France"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Olivier","family":"Nempont","sequence":"additional","affiliation":[{"name":"Philips Health Technology innovation Paris France"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Salim Aymeric","family":"Si\u2010Mohamed","sequence":"additional","affiliation":[{"name":"CREATIS UMR5220, INSERM U1044 Claude Bernard University Lyon 1 INSA Lyon France"},{"name":"Department of Radiology Hospices Civils de Lyon Lyon France"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Fabien","family":"Craighero","sequence":"additional","affiliation":[{"name":"Department of Radiology Hospices Civils de Lyon Lyon France"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Vincent","family":"Cottin","sequence":"additional","affiliation":[{"name":"National Reference Center for Rare Pulmonary Diseases Louis Pradel Hospital Lyon France"},{"name":"Claude Bernard University Lyon 1 Lyon France"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Philippe","family":"Douek","sequence":"additional","affiliation":[{"name":"CREATIS UMR5220, INSERM U1044 Claude Bernard University Lyon 1 INSA Lyon France"},{"name":"Department of Radiology Hospices Civils de Lyon Lyon France"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Alexandre","family":"Popoff","sequence":"additional","affiliation":[{"name":"Philips Health Technology innovation Paris France"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Loic","family":"Boussel","sequence":"additional","affiliation":[{"name":"CREATIS UMR5220, INSERM U1044 Claude Bernard University Lyon 1 INSA Lyon France"},{"name":"Department of Radiology Hospices Civils de Lyon Lyon France"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"311","published-online":{"date-parts":[[2024,8,14]]},"reference":[{"key":"e_1_2_9_2_1","doi-asserted-by":"publisher","DOI":"10.1038\/s41591-019-0447-x"},{"key":"e_1_2_9_3_1","doi-asserted-by":"publisher","DOI":"10.1515\/med-2020-0028"},{"key":"e_1_2_9_4_1","doi-asserted-by":"publisher","DOI":"10.3389\/fnins.2022.1106937"},{"key":"e_1_2_9_5_1","doi-asserted-by":"publisher","DOI":"10.1007\/s11548-022-02576-5"},{"key":"e_1_2_9_6_1","doi-asserted-by":"publisher","DOI":"10.1007\/s11042-022-13005-9"},{"key":"e_1_2_9_7_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.cmpb.2021.106592"},{"key":"e_1_2_9_8_1","doi-asserted-by":"publisher","DOI":"10.1038\/s41598-019-56989-5"},{"key":"e_1_2_9_9_1","doi-asserted-by":"publisher","DOI":"10.1080\/21681163.2015.1124249"},{"key":"e_1_2_9_10_1","doi-asserted-by":"publisher","DOI":"10.1109\/TMI.2016.2535865"},{"key":"e_1_2_9_11_1","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-319-47157-0_18"},{"key":"e_1_2_9_12_1","doi-asserted-by":"publisher","DOI":"10.1117\/12.2512746"},{"key":"e_1_2_9_13_1","doi-asserted-by":"publisher","DOI":"10.1109\/JBHI.2016.2636929"},{"key":"e_1_2_9_14_1","doi-asserted-by":"publisher","DOI":"10.1117\/1.JMI.8.S1.014502"},{"key":"e_1_2_9_15_1","doi-asserted-by":"publisher","DOI":"10.1007\/s11042-022-12214-6"},{"key":"e_1_2_9_16_1","doi-asserted-by":"publisher","DOI":"10.1007\/s10140-020-01886-y"},{"key":"e_1_2_9_17_1","doi-asserted-by":"publisher","DOI":"10.1038\/s41598-021-99015-3"},{"key":"e_1_2_9_18_1","doi-asserted-by":"publisher","DOI":"10.1038\/s41598-021-93832-2"},{"key":"e_1_2_9_19_1","doi-asserted-by":"publisher","DOI":"10.1038\/s41598-023-27549-9"},{"key":"e_1_2_9_20_1","doi-asserted-by":"publisher","DOI":"10.1007\/s10278-022-00599-7"},{"key":"e_1_2_9_21_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.media.2020.101857"},{"key":"e_1_2_9_22_1","doi-asserted-by":"publisher","DOI":"10.1148\/ryai.210026"},{"key":"e_1_2_9_23_1","doi-asserted-by":"publisher","DOI":"10.1117\/12.2550857"},{"key":"e_1_2_9_24_1","unstructured":"ChenT KornblithS NorouziM HintonG.A simple framework for contrastive learning of visual representations. In:Proceedings of the 37th International Conference on Machine Learning. PMLR;2020:149. doi:10.5555\/3524938.3525087"},{"key":"e_1_2_9_25_1","unstructured":"GeninG CraigheroF.Chapter 2. In: Scanner thoracique guide de lecture rapide: images clefs d\u00e9finitions et \u00e9tiologies. Sauramps Medical. Librairie Dalloz;2020. Available at:https:\/\/www.librairiedalloz.fr\/livre\/9791030302714\u2010scanner\u2010thoracique\u2010guide\u2010de\u2010lecture\u2010rapide\u2010images\u2010clefs\u2010definitions\u2010et\u2010etiologies\u2010gilles\u2010genin\u2010fabien\u2010craighero\/"},{"key":"e_1_2_9_26_1","doi-asserted-by":"crossref","unstructured":"DjahnineA PopoffA Jupin\u2010DelevauxE CotinV NempontO BousselL.Tailored 3D CT contrastive pretraining to improve pulmonary pathology classification. In:2022 16th IEEE International Conference on Signal Processing (ICSP).IEEE;2022:229\u2010234. doi:10.1109\/ICSP56322.2022.9965218","DOI":"10.1109\/ICSP56322.2022.9965218"},{"key":"e_1_2_9_27_1","doi-asserted-by":"publisher","DOI":"10.1007\/s00330\u2010021\u201008482\u20109"},{"key":"e_1_2_9_28_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.redii.2022.100018"}],"container-title":["Medical Physics"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/aapm.onlinelibrary.wiley.com\/doi\/pdf\/10.1002\/mp.17302","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2024,12,2]],"date-time":"2024-12-02T20:01:05Z","timestamp":1733169665000},"score":1,"resource":{"primary":{"URL":"https:\/\/aapm.onlinelibrary.wiley.com\/doi\/10.1002\/mp.17302"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2024,8,14]]},"references-count":27,"journal-issue":{"issue":"11","published-print":{"date-parts":[[2024,11]]}},"alternative-id":["10.1002\/mp.17302"],"URL":"https:\/\/doi.org\/10.1002\/mp.17302","archive":["Portico"],"relation":{},"ISSN":["0094-2405","2473-4209"],"issn-type":[{"value":"0094-2405","type":"print"},{"value":"2473-4209","type":"electronic"}],"subject":[],"published":{"date-parts":[[2024,8,14]]},"assertion":[{"value":"2023-12-11","order":0,"name":"received","label":"Received","group":{"name":"publication_history","label":"Publication History"}},{"value":"2024-07-01","order":2,"name":"accepted","label":"Accepted","group":{"name":"publication_history","label":"Publication History"}},{"value":"2024-08-14","order":3,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}]}}