{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,17]],"date-time":"2026-10-17T00:40:46Z","timestamp":1792197646183,"version":"4.4.0"},"reference-count":33,"publisher":"Springer Science and Business Media LLC","license":[{"start":{"date-parts":[[2026,10,16]],"date-time":"2026-10-16T00:00:00Z","timestamp":1792108800000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"},{"start":{"date-parts":[[2026,10,16]],"date-time":"2026-10-16T00:00:00Z","timestamp":1792108800000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"}],"content-domain":{"domain":["link.springer.com"],"crossmark-restriction":false},"short-container-title":["J Digit Imaging. Inform. med."],"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:p>\n                    To test whether anatomical priors from a general-purpose organ segmentation model (TotalSegmentator, TS), applied at inference or during training, improve an externally validated YOLOv11 detector for six acute abdominal emergencies. The frozen detector was applied to 827 external Stanford Merlin CT studies with language-model report labels; a nested 280-study radiologist-adjudicated subset was the reference standard. TS-derived slice selection (Arm\u00a0A), organ-state integration (Arm\u00a0B), and their combination (Arm\u00a0C) were compared with the detector alone under frozen, recalibrated, and fine-tuned conditions. The detector was also retrained from scratch (identical 10-epoch budget) with image channels only, with the patient\u2019s TS organ maps as added input channels, or with another patient\u2019s maps. AUROC was primary; comparisons used a paired study-level bootstrap. The detector generalized externally (macro AUROC 0.816 cohort; 0.879 adjudicated). Slice selection was non-inferior but not beneficial: nearly all confident detections already lay within the anchored organ\u2019s axial range; organ-state fusion improved AUROC only for cholecystitis (hypothesis-generating). In single runs, training with organ maps raised internal test mAP@0.5 (0.517 versus 0.482), and the model depended on the maps, but external macro AUROC did not improve (0.822 versus 0.829;\n                    <jats:inline-formula>\n                      <jats:alternatives>\n                        <jats:tex-math>$$\\Delta $$<\/jats:tex-math>\n                        <mml:math xmlns:mml=\"http:\/\/www.w3.org\/1998\/Math\/MathML\">\n                          <mml:mi>\u0394<\/mml:mi>\n                        <\/mml:math>\n                      <\/jats:alternatives>\n                    <\/jats:inline-formula>\n                    \u22120.007, 95% CI \u22120.019 to +0.007) and kidney-stone AUROC fell (0.678 versus 0.768; exploratory), unlike with another patient\u2019s maps. Per-site recalibration was the most effective operating-point adjustment (macro F1 +0.021); annotation-free fine-tuning reduced discrimination. Explicit anatomy, applied after detection or during training, did not improve this detector\u2019s overall external discrimination. Per-site recalibration gave the largest operating-point gain.\n                  <\/jats:p>","DOI":"10.1007\/s10278-026-02368-2","type":"journal-article","created":{"date-parts":[[2026,10,16]],"date-time":"2026-10-16T23:56:18Z","timestamp":1792194978000},"update-policy":"https:\/\/doi.org\/10.1007\/springer_crossmark_policy","source":"Crossref","is-referenced-by-count":0,"title":["Evaluating Explicit Anatomical Priors Beyond a Mature CT Detector for Six Acute Abdominal Emergencies"],"prefix":"10.1007","author":[{"ORCID":"https:\/\/orcid.org\/0009-0002-1619-595X","authenticated-orcid":false,"given":"Hasan Mete","family":"Erdo\u011fan","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-8189-2885","authenticated-orcid":false,"given":"Ural","family":"Ko\u00e7","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"297","published-online":{"date-parts":[[2026,10,16]]},"reference":[{"key":"2368_CR1","doi-asserted-by":"publisher","unstructured":"Wasserthal J, Breit HC, Meyer MT, Pradella M, Hinck D, Sauter AW, et\u00a0al. TotalSegmentator: robust segmentation of 104 anatomic structures in CT images. Radiol Artif Intell. 2023;5(5):e230024. https:\/\/doi.org\/10.1148\/ryai.230024","DOI":"10.1148\/ryai.230024"},{"key":"2368_CR2","doi-asserted-by":"publisher","unstructured":"D\u2019Antonoli TA, Berger LK, Indrakanti AK, Vishwanathan N, Weiss J, Jung M, et\u00a0al. TotalSegmentator MRI: robust sequence-independent segmentation of multiple anatomic structures in MRI. Radiology. 2025;314(2):e241613. https:\/\/doi.org\/10.1148\/radiol.241613","DOI":"10.1148\/radiol.241613"},{"key":"2368_CR3","doi-asserted-by":"publisher","unstructured":"Sebro R, Mongan J. TotalSegmentator: a gift to the biomedical imaging community. Radiol Artif Intell. 2023;5(5):e230235. https:\/\/doi.org\/10.1148\/ryai.230235","DOI":"10.1148\/ryai.230235"},{"key":"2368_CR4","doi-asserted-by":"publisher","unstructured":"Balaguer-Montero M, Marcos Morales A, Ligero M, et\u00a0al. A CT-based deep learning-driven tool for automatic liver tumor detection and delineation in patients with cancer. Cell Rep Med. 2025;6(4):102032. https:\/\/doi.org\/10.1016\/j.xcrm.2025.102032","DOI":"10.1016\/j.xcrm.2025.102032"},{"key":"2368_CR5","doi-asserted-by":"publisher","unstructured":"Bartnik K, Krzyzi\u0144ski M, Bartczak T, et\u00a0al. A novel radiomics approach for predicting TACE outcomes in hepatocellular carcinoma patients using deep learning for multi-organ segmentation. Sci Rep. 2024;14:14779. https:\/\/doi.org\/10.1038\/s41598-024-65630-z","DOI":"10.1038\/s41598-024-65630-z"},{"key":"2368_CR6","doi-asserted-by":"publisher","unstructured":"Hata A, Muraguchi Y, Nakatsugawa M, et\u00a0al. Automated chest CT three-dimensional quantification of body composition: adipose tissue and paravertebral muscle. Sci Rep. 2024;14:32117. https:\/\/doi.org\/10.1038\/s41598-024-83897-0","DOI":"10.1038\/s41598-024-83897-0"},{"key":"2368_CR7","doi-asserted-by":"publisher","unstructured":"Urooj B, Ko Y, Na S, et\u00a0al. Implementation of a fully automated AI-integrated system for body composition assessment on CT for opportunistic sarcopenia screening: multicenter prospective study. JMIR Form Res. 2025;9:e69940. https:\/\/doi.org\/10.2196\/69940","DOI":"10.2196\/69940"},{"key":"2368_CR8","doi-asserted-by":"publisher","unstructured":"Seng M, Wasserthal J, Bach M, et\u00a0al. AI-based opportunistic CT risk assessment using TotalSegmentator in osteoporotic vertebral fractures. J Imaging Inform Med. 2026. https:\/\/doi.org\/10.1007\/s10278-026-02049-0","DOI":"10.1007\/s10278-026-02049-0"},{"key":"2368_CR9","doi-asserted-by":"publisher","unstructured":"Bendazzoli S, Tzortzakakis A, Abrahamsson A, et\u00a0al. Anatomy-aware lymphoma lesion detection in whole-body PET\/CT. Front Oncol. 2026;16:1695211. https:\/\/doi.org\/10.3389\/fonc.2026.1695211","DOI":"10.3389\/fonc.2026.1695211"},{"key":"2368_CR10","doi-asserted-by":"publisher","unstructured":"Jaus A, Seibold C, Rei\u00df S, Heine L, Schily A, Kim M, Bahnsen FH, Herrmann K, Stiefelhagen R, Kleesiek J. Anatomy-guided pathology segmentation. In: Medical Image Computing and Computer Assisted Intervention (MICCAI) 2024. LNCS 15008. https:\/\/doi.org\/10.1007\/978-3-031-72111-3_1","DOI":"10.1007\/978-3-031-72111-3_1"},{"key":"2368_CR11","doi-asserted-by":"crossref","unstructured":"Prasad AV, Mathai TS, Mukherjee P, Liu J, Summers RM. Leveraging anatomical priors for automated pancreas segmentation on abdominal CT. arXiv:2504.06921. 2025.","DOI":"10.1117\/12.3048976"},{"key":"2368_CR12","doi-asserted-by":"publisher","unstructured":"Redmon J, Divvala S, Girshick R, Farhadi A. You only look once: unified, real-time object detection. In: Proc IEEE Conf Comput Vis Pattern Recognit (CVPR). 2016. p.\u00a0779\u201388. https:\/\/doi.org\/10.1109\/CVPR.2016.91","DOI":"10.1109\/CVPR.2016.91"},{"key":"2368_CR13","unstructured":"Khanam R, Hussain M. YOLOv11: an overview of the key architectural enhancements. arXiv:2410.17725. 2024."},{"key":"2368_CR14","doi-asserted-by":"publisher","unstructured":"Abd El Bagi ME, Almutairi BM, Alsolamy SJ, et\u00a0al. Imaging of non-traumatic acute abdominal pain in adults presenting to the emergency department. Saudi Med J. 2016;37(3):324\u20137. https:\/\/doi.org\/10.15537\/smj.2016.3.13068","DOI":"10.15537\/smj.2016.3.13068"},{"key":"2368_CR15","doi-asserted-by":"publisher","unstructured":"van Leeuwen MP, van Oudheusden TR, Oei SP, Bosma GPT, Haak KV, Saygili G, Ong LLS. Automatic identification of anatomical locations for bone abnormalities in CT imaging: a multiplanar YOLOv5 detection approach. Annu Int Conf IEEE Eng Med Biol Soc (EMBC). 2025;2025:1\u20137. https:\/\/doi.org\/10.1109\/EMBC58623.2025.11254481","DOI":"10.1109\/EMBC58623.2025.11254481"},{"key":"2368_CR16","doi-asserted-by":"publisher","unstructured":"Zhou B, Khosla A, Lapedriza A, Oliva A, Torralba A. Learning deep features for discriminative localization. In: Proc IEEE Conf Comput Vis Pattern Recognit (CVPR). 2016. p.\u00a02921\u20139. https:\/\/doi.org\/10.1109\/CVPR.2016.319","DOI":"10.1109\/CVPR.2016.319"},{"key":"2368_CR17","unstructured":"Islam MA, Jia S, Bruce NDB. How much position information do convolutional neural networks encode? In: Int Conf Learn Represent (ICLR). 2020. arXiv:2001.08248."},{"key":"2368_CR18","doi-asserted-by":"publisher","unstructured":"Finlayson SG, Subbaswamy A, Singh K, Bowers J, Kupke A, Zittrain J, et\u00a0al. The clinician and dataset shift in artificial intelligence. N Engl J Med. 2021;385(3):283\u20136. https:\/\/doi.org\/10.1056\/NEJMc2104626","DOI":"10.1056\/NEJMc2104626"},{"key":"2368_CR19","doi-asserted-by":"publisher","unstructured":"Yu AC, Mohajer B, Eng J. External validation of deep learning algorithms for radiologic diagnosis: a systematic review. Radiol Artif Intell. 2022;4(3):e210064. https:\/\/doi.org\/10.1148\/ryai.210064","DOI":"10.1148\/ryai.210064"},{"key":"2368_CR20","unstructured":"Dahal L, Bhandari Y, Rubin G, Lo JY. ORACLE-CT: anatomy-aware support pooling for CT classification. arXiv:2606.05460. 2026."},{"key":"2368_CR21","doi-asserted-by":"publisher","unstructured":"Erdo\u011fan HM, Ko\u00e7 U. Anatomically localized detection of six acute abdominal emergencies on CT using multi-window deep learning: development and validation. J Imaging Inform Med. 2026. https:\/\/doi.org\/10.1007\/s10278-026-02084-x","DOI":"10.1007\/s10278-026-02084-x"},{"key":"2368_CR22","doi-asserted-by":"publisher","unstructured":"Tejani AS, Klontzas ME, Gatti AA, Mongan JT, Moy L, Park SH, Kahn CE Jr. Checklist for Artificial Intelligence in Medical\u00a0Imaging (CLAIM): 2024 update. Radiol Artif Intell. 2024;6(4):e240300. https:\/\/doi.org\/10.1148\/ryai.240300","DOI":"10.1148\/ryai.240300"},{"key":"2368_CR23","doi-asserted-by":"publisher","unstructured":"Ko\u00e7 U, Sezer EA, \u00d6zkaya YA, Yarbay Y, Be\u015fler MS, Tayda\u015f O, et\u00a0al. Elevating healthcare through artificial intelligence: analyzing the abdominal emergencies data set (TR_ABDOMEN_RAD_EMERGENCY) at TEKNOFEST-2022. Eur Radiol. 2024;34(6):3588\u201397. https:\/\/doi.org\/10.1007\/s00330-023-10391-y","DOI":"10.1007\/s00330-023-10391-y"},{"key":"2368_CR24","doi-asserted-by":"publisher","unstructured":"Blankemeier L, Kumar A, Cohen JP, et\u00a0al. Merlin: a computed tomography vision\u2013language foundation model and dataset. Nature. 2026;652:1318\u20131328. https:\/\/doi.org\/10.1038\/s41586-026-10181-8","DOI":"10.1038\/s41586-026-10181-8"},{"key":"2368_CR25","unstructured":"Guo C, Pleiss G, Sun Y, Weinberger KQ. On calibration of modern neural networks. In: Proc 34th Int Conf Mach Learn (ICML), PMLR 70. 2017. p.\u00a01321\u201330."},{"key":"2368_CR26","doi-asserted-by":"publisher","unstructured":"Arazo E, Ortego D, Albert P, O\u2019Connor NE, McGuinness K. Pseudo-labeling and confirmation bias in deep semi-supervised learning. In: Int Joint Conf Neural Networks (IJCNN). 2020. https:\/\/doi.org\/10.1109\/IJCNN48605.2020.9207304","DOI":"10.1109\/IJCNN48605.2020.9207304"},{"key":"2368_CR27","doi-asserted-by":"publisher","unstructured":"O\u2019Connor OJ, Maher MM. Imaging of cholecystitis. AJR Am J Roentgenol. 2011;196(4):W367\u201374. https:\/\/doi.org\/10.2214\/AJR.10.4340","DOI":"10.2214\/AJR.10.4340"},{"key":"2368_CR28","doi-asserted-by":"publisher","unstructured":"Van Calster B, McLernon DJ, van Smeden M, Wynants L, Steyerberg EW. Calibration: the Achilles heel of predictive analytics. BMC Med. 2019;17:230. https:\/\/doi.org\/10.1186\/s12916-019-1466-7","DOI":"10.1186\/s12916-019-1466-7"},{"key":"2368_CR29","unstructured":"Lee DH. Pseudo-label: the simple and efficient semi-supervised learning method for deep neural networks. In: ICML Workshop on Challenges in Representation Learning. 2013."},{"key":"2368_CR30","doi-asserted-by":"publisher","unstructured":"Xie Q, Luong MT, Hovy E, Le QV. Self-training with noisy student improves ImageNet classification. In: Proc IEEE\/CVF Conf Comput Vis Pattern Recognit (CVPR). 2020. p.\u00a010684\u201395. https:\/\/doi.org\/10.1109\/CVPR42600.2020.01070","DOI":"10.1109\/CVPR42600.2020.01070"},{"key":"2368_CR31","unstructured":"Mohapatra S, Mohanty P. Tracing like a clinician: anatomy-guided spatial priors for cephalometric landmark detection. arXiv:2605.03358. 2026."},{"key":"2368_CR32","unstructured":"Ge Z, Li A, Al-Kindi S, Xu H, Qiu W. OrganLens: organ-specific representation learning for CT foundation models. arXiv:2607.25164. 2026."},{"key":"2368_CR33","unstructured":"Hudler M, Thaler F, Urschler M. Evaluation of anatomical shape priors in deep learning-based cardiac multi-compartment segmentation. arXiv:2605.15707. 2026."}],"container-title":["Journal of Imaging Informatics in Medicine"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/link.springer.com\/content\/pdf\/10.1007\/s10278-026-02368-2.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/link.springer.com\/article\/10.1007\/s10278-026-02368-2","content-type":"text\/html","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/link.springer.com\/content\/pdf\/10.1007\/s10278-026-02368-2.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,10,16]],"date-time":"2026-10-16T23:56:19Z","timestamp":1792194979000},"score":1,"resource":{"primary":{"URL":"https:\/\/link.springer.com\/article\/10.1007\/s10278-026-02368-2"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,10,16]]},"references-count":33,"alternative-id":["2368"],"URL":"https:\/\/doi.org\/10.1007\/s10278-026-02368-2","relation":{},"ISSN":["2948-2933"],"issn-type":[{"value":"2948-2933","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,10,16]]},"assertion":[{"value":"11 July 2026","order":1,"name":"received","label":"Received","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"16 September 2026","order":2,"name":"revised","label":"Revised","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"24 September 2026","order":3,"name":"accepted","label":"Accepted","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"16 October 2026","order":4,"name":"first_online","label":"First Online","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"This study used publicly available, de-identified open data; ethics committee approval was not required.","order":1,"name":"Ethics","label":"Ethics Approval","group":{"name":"EthicsHeading","label":"Declarations"}},{"value":"Not applicable; all data were de-identified.","order":2,"name":"Ethics","label":"Consent to Participate","group":{"name":"EthicsHeading","label":"Declarations"}},{"value":"Not applicable; all data were de-identified.","order":3,"name":"Ethics","label":"Consent for Publication","group":{"name":"EthicsHeading","label":"Declarations"}},{"value":"The authors declare no competing interests.","order":4,"name":"Ethics","label":"Competing Interests","group":{"name":"EthicsHeading","label":"Declarations"}},{"value":"Large language models were used as a research tool within the study methodology: report-derived labels were extracted with MedGemma-27B, as described in the Methods and Supplementary Methods. Separately, AI-assisted tools were used for language editing and reference verification during manuscript preparation. The authors reviewed all model outputs and take responsibility for the content of the manuscript and all reported analyses.","order":5,"name":"Ethics","label":"Use of AI tools","group":{"name":"EthicsHeading","label":"Declarations"}},{"value":"The 280-study Stanford Merlin radiologist-adjudicated subset evaluated here is the same external-validation cohort as the authors\u2019 foundational detector paper, published in this journal: Erdo\u011fan HM, Ko\u00e7 U.\n                      Anatomically localized detection of six acute abdominal emergencies on CT using multi-window deep learning: development and validation.\n                      J Imaging Inform Med. 2026. doi:10.1007\/s10278-026-02084-x. That paper developed and externally validated the detector; the present study addresses a distinct question, namely whether explicit anatomical priors add value beyond that detector, and reports analyses that are not in the prior work: the anatomy analyses, threshold recalibration and annotation-free fine-tuning on Merlin data, and a retraining experiment that reuses 1240 of the prior work\u2019s 1274 development patients.","order":6,"name":"Ethics","label":"Prior Publication and Patient Overlap","group":{"name":"EthicsHeading","label":"Declarations"}}]}}