{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,11,10]],"date-time":"2026-11-10T17:41:25Z","timestamp":1794332485229,"version":"4.9.0"},"reference-count":23,"publisher":"Springer Science and Business Media LLC","issue":"1","license":[{"start":{"date-parts":[[2025,12,2]],"date-time":"2025-12-02T00:00:00Z","timestamp":1764633600000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"},{"start":{"date-parts":[[2025,12,2]],"date-time":"2025-12-02T00:00:00Z","timestamp":1764633600000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"}],"funder":[{"name":"Open access for this article was provided by Universit\u00e0 degli Studi del Piemonte Orientale."}],"content-domain":{"domain":["link.springer.com"],"crossmark-restriction":false},"short-container-title":["Eur Radiol Exp"],"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:sec>\n                    <jats:title>Background<\/jats:title>\n                    <jats:p>Hormone receptor (HR) status guides breast cancer therapy. Deep learning (DL) applied to contrast-enhanced mammography (CEM) might offer a noninvasive means for HR status prediction, but class imbalance challenges model development and assessment. This preliminary study investigates CEM-based DL for HR status prediction, focusing on class imbalance handling.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Materials and methods<\/jats:title>\n                    <jats:p>\n                      In this retrospective study, CEM tumor crops from 105 patients with invasive breast cancer were used. Patients were randomized into training (\n                      <jats:italic>n<\/jats:italic>\n                      \u2009=\u200968), validation (\n                      <jats:italic>n<\/jats:italic>\n                      \u2009=\u200916), and independent test (\n                      <jats:italic>n<\/jats:italic>\n                      \u2009=\u200921) sets. A \u201cResidual Network 18\u201d (ResNet-18), pretrained on ImageNet, was fine-tuned using weighted cross-entropy loss and an Adam optimizer. Model selection used validation area under the precision-recall curve (AUPRC); output probabilities were calibrated via temperature scaling. Performance was reported with accuracy, area under the receiver operating characteristic curve (AUROC), and imbalance-aware metrics (balanced accuracy, Matthews correlation coefficient (MCC)) with 95% confidence intervals (1,000-iteration bootstrap). Results are presented for standard (0.5) and optimized (validation F1-score for HR-negative class) thresholds.\n                    <\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Results<\/jats:title>\n                    <jats:p>Validation AUPRC (model selection metric) was 0.640 (0.304\u20130.906). On the independent test set (optimized threshold 0.755), the model achieved 91.9% accuracy (86.5\u201397.3%), AUROC 0.808 (0.648\u20130.935), balanced accuracy 0.700 (0.550\u20130.853), and MCC 0.605 (0.296\u20130.818).<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Conclusion<\/jats:title>\n                    <jats:p>A ResNet-18, utilizing patient-level data splitting and imbalance-aware fine-tuning, can capture CEM features for HR status, performing well despite significant class imbalance. Generalizability is limited by dataset characteristics and acquisition specifics, warranting further validation in larger, diverse cohorts to establish clinical applicability.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Relevance statement<\/jats:title>\n                    <jats:p>This work explores whether routinely acquired CEM images contain enough information for DL prediction of HR status. A ResNet-18 was trained with weighted loss and patient-level data splits; performance was quantified with imbalance-aware metrics to provide a realistic assessment in a highly skewed dataset, highlighting both the promise and current constraints of CEM-based molecular imaging.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Key Points<\/jats:title>\n                    <jats:p>\n                      <jats:list list-type=\"bullet\">\n                        <jats:list-item>\n                          <jats:p>A ResNet-18, optimized for class imbalance through weighted training and with calibrated probabilities, predicted HR positivity on CEM with 91.9% accuracy and AUROC 0.81 in an independent test cohort using an F1-tuned threshold.<\/jats:p>\n                        <\/jats:list-item>\n                        <jats:list-item>\n                          <jats:p>Balanced accuracy (0.70) and MCC (0.60) demonstrate maintained discrimination despite an approximate 85% class imbalance (HR-positive cases).<\/jats:p>\n                        <\/jats:list-item>\n                        <jats:list-item>\n                          <jats:p>Patient-level splitting was employed to ensure robust evaluation.<\/jats:p>\n                        <\/jats:list-item>\n                        <jats:list-item>\n                          <jats:p>Limitations related to the dataset\u2019s scope and specific imaging protocols may influence broader generalizability.<\/jats:p>\n                        <\/jats:list-item>\n                      <\/jats:list>\n                    <\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Graphical Abstract<\/jats:title>\n                  <\/jats:sec>","DOI":"10.1186\/s41747-025-00653-3","type":"journal-article","created":{"date-parts":[[2025,12,2]],"date-time":"2025-12-02T08:24:50Z","timestamp":1764663890000},"update-policy":"https:\/\/doi.org\/10.1007\/springer_crossmark_policy","source":"Crossref","is-referenced-by-count":3,"title":["Deep-learning prediction of breast cancer hormone receptor status from CEM: a preliminary study"],"prefix":"10.1186","volume":"9","author":[{"given":"Alessandro","family":"Carriero","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0009-1054-2552","authenticated-orcid":false,"given":"Marco","family":"Albera","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Tomasina","family":"Meloni","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Anna","family":"Rampi","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Renzo Luciano","family":"Boldorini","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Anna","family":"Colarieti","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"297","published-online":{"date-parts":[[2025,12,2]]},"reference":[{"key":"653_CR1","unstructured":"World Health Organization (2025) Breast cancer\u2014key facts. Available via https:\/\/www.who.int\/news-room\/fact-sheets\/detail\/breast-cancer. Accessed 31 May 2025"},{"key":"653_CR2","doi-asserted-by":"publisher","first-page":"1134","DOI":"10.1016\/S0140-6736(16)31891-8","volume":"389","author":"N Harbeck","year":"2017","unstructured":"Harbeck N, Gnant M (2017) Breast cancer. Lancet 389:1134\u20131150. https:\/\/doi.org\/10.1016\/S0140-6736(16)31891-8","journal-title":"Lancet"},{"key":"653_CR3","doi-asserted-by":"publisher","first-page":"x207","DOI":"10.1093\/annonc\/mds326","volume":"23","author":"G Viale","year":"2012","unstructured":"Viale G (2012) The current state of breast cancer classification. Ann Oncol 23:x207\u2013x210. https:\/\/doi.org\/10.1093\/annonc\/mds326","journal-title":"Ann Oncol"},{"key":"653_CR4","doi-asserted-by":"publisher","first-page":"355","DOI":"10.1038\/nature12627","volume":"501","author":"PL Bedard","year":"2013","unstructured":"Bedard PL, Hansen AR, Ratain MJ, Siu LL (2013) Tumour heterogeneity in the clinic. Nature 501:355\u2013364. https:\/\/doi.org\/10.1038\/nature12627","journal-title":"Nature"},{"key":"653_CR5","doi-asserted-by":"publisher","first-page":"8","DOI":"10.1186\/s41747-022-00318-5","volume":"7","author":"DB Taylor","year":"2023","unstructured":"Taylor DB, Burrows S, Saunders CM et al (2023) Contrast-enhanced mammography (CEM) versus MRI for breast cancer staging: detection of additional malignant lesions not seen on conventional imaging. Eur Radiol Exp 7:8. https:\/\/doi.org\/10.1186\/s41747-022-00318-5","journal-title":"Eur Radiol Exp"},{"key":"653_CR6","doi-asserted-by":"publisher","first-page":"32","DOI":"10.1186\/s41747-024-00440-6","volume":"8","author":"A Ancona","year":"2024","unstructured":"Ancona A, Telegrafo M, Fella RR et al (2024) CEM immediately after contrast-enhanced CT: a one-step staging of breast cancer. Eur Radiol Exp 8:32. https:\/\/doi.org\/10.1186\/s41747-024-00440-6","journal-title":"Eur Radiol Exp"},{"key":"653_CR7","doi-asserted-by":"publisher","first-page":"935","DOI":"10.1016\/j.crad.2013.04.009","volume":"68","author":"MBI Lobbes","year":"2013","unstructured":"Lobbes MBI, Smidt ML, Houwers J et al (2013) Contrast enhanced mammography: techniques, current results, and potential indications. Clin Radiol 68:935\u2013944. https:\/\/doi.org\/10.1016\/j.crad.2013.04.009","journal-title":"Clin Radiol"},{"key":"653_CR8","doi-asserted-by":"publisher","first-page":"256","DOI":"10.1007\/s00330-013-3007-7","volume":"24","author":"EM Fallenberg","year":"2014","unstructured":"Fallenberg EM, Dromain C, Diekmann F et al (2014) Contrast-enhanced spectral mammography versus MRI: initial results in the detection of breast cancer and assessment of tumour size. Eur Radiol 24:256\u2013264. https:\/\/doi.org\/10.1007\/s00330-013-3007-7","journal-title":"Eur Radiol"},{"key":"653_CR9","doi-asserted-by":"publisher","first-page":"857","DOI":"10.1007\/s10462-023-10543-y","volume":"56","author":"T Zhang","year":"2023","unstructured":"Zhang T, Tan T, Samperna R et al (2023) Radiomics and artificial intelligence in breast imaging: a survey. Artif Intell Rev 56:857\u2013892. https:\/\/doi.org\/10.1007\/s10462-023-10543-y","journal-title":"Artif Intell Rev"},{"key":"653_CR10","doi-asserted-by":"publisher","first-page":"695","DOI":"10.1093\/bjr\/tqae017","volume":"97","author":"MF Covington","year":"2024","unstructured":"Covington MF, Salmon S, Weaver BD, Fajardo LL (2024) State-of-the-art for contrast-enhanced mammography. Br J Radiol 97:695\u2013704. https:\/\/doi.org\/10.1093\/bjr\/tqae017","journal-title":"Br J Radiol"},{"key":"653_CR11","doi-asserted-by":"publisher","first-page":"2497","DOI":"10.1016\/j.acra.2024.11.035","volume":"32","author":"V Sorin","year":"2025","unstructured":"Sorin V, Sklair-Levy M, Glicksberg BS et al (2025) Deep learning for contrast enhanced mammography\u2014a systematic review. Acad Radiol 32:2497\u20132508. https:\/\/doi.org\/10.1016\/j.acra.2024.11.035","journal-title":"Acad Radiol"},{"key":"653_CR12","doi-asserted-by":"publisher","first-page":"60","DOI":"10.1016\/j.media.2017.07.005","volume":"42","author":"G Litjens","year":"2017","unstructured":"Litjens G, Kooi T, Bejnordi BE et al (2017) A survey on deep learning in medical image analysis. Med Image Anal 42:60\u201388. https:\/\/doi.org\/10.1016\/j.media.2017.07.005","journal-title":"Med Image Anal"},{"key":"653_CR13","doi-asserted-by":"crossref","unstructured":"He K, Zhang X, Ren S, Sun J (2016) Deep residual learning for image recognition. In: Proceedings of the 2016 IEEE conference on computer vision and pattern recognition (CVPR). IEEE, Las Vegas, pp 770\u2013778","DOI":"10.1109\/CVPR.2016.90"},{"key":"653_CR14","doi-asserted-by":"publisher","first-page":"60","DOI":"10.1148\/radiol.2019182716","volume":"292","author":"A Yala","year":"2019","unstructured":"Yala A, Lehman C, Schuster T et al (2019) A deep learning mammography-based model for improved breast cancer risk prediction. Radiology 292:60\u201366. https:\/\/doi.org\/10.1148\/radiol.2019182716","journal-title":"Radiology"},{"key":"653_CR15","doi-asserted-by":"publisher","DOI":"10.1038\/s41598-024-83597-9","volume":"15","author":"S Zeng","year":"2025","unstructured":"Zeng S, Chen H, Jing R et al (2025) An assessment of breast cancer HER2, ER, and PR expressions based on mammography using deep learning with convolutional neural networks. Sci Rep 15:4826. https:\/\/doi.org\/10.1038\/s41598-024-83597-9","journal-title":"Sci Rep"},{"key":"653_CR16","doi-asserted-by":"publisher","DOI":"10.1038\/s41523-023-00517-2","volume":"9","author":"T Zhang","year":"2023","unstructured":"Zhang T, Tan T, Han L et al (2023) Predicting breast cancer types on and beyond molecular level in a multi-modal fashion. NPJ Breast Cancer 9:16. https:\/\/doi.org\/10.1038\/s41523-023-00517-2","journal-title":"NPJ Breast Cancer"},{"key":"653_CR17","doi-asserted-by":"publisher","DOI":"10.3389\/fmed.2023.1140514","volume":"10","author":"T Huang","year":"2023","unstructured":"Huang T, Fan B, Qiu Y et al (2023) Application of DCE-MRI radiomics signature analysis in differentiating molecular subtypes of luminal and non-luminal breast cancer. Front Med 10:1140514. https:\/\/doi.org\/10.3389\/fmed.2023.1140514","journal-title":"Front Med"},{"key":"653_CR18","doi-asserted-by":"publisher","DOI":"10.1016\/j.compbiomed.2022.106147","volume":"150","author":"W Ming","year":"2022","unstructured":"Ming W, Li F, Zhu Y et al (2022) Predicting hormone receptors and PAM50 subtypes of breast cancer from multi-scale lesion images of DCE-MRI with transfer learning technique. Comput Biol Med 150:106147. https:\/\/doi.org\/10.1016\/j.compbiomed.2022.106147","journal-title":"Comput Biol Med"},{"key":"653_CR19","doi-asserted-by":"publisher","first-page":"891","DOI":"10.3390\/jcm8060891","volume":"8","author":"A Fanizzi","year":"2019","unstructured":"Fanizzi A, Losurdo L, Basile TMA et al (2019) Fully automated support system for diagnosis of breast cancer in contrast-enhanced spectral mammography images. J Clin Med 8:891. https:\/\/doi.org\/10.3390\/jcm8060891","journal-title":"J Clin Med"},{"key":"653_CR20","doi-asserted-by":"publisher","first-page":"917","DOI":"10.1007\/s00330-023-10170-9","volume":"34","author":"N Qian","year":"2023","unstructured":"Qian N, Jiang W, Guo Y et al (2023) Breast cancer diagnosis from contrast-enhanced mammography using multi-feature fusion neural network. Eur Radiol 34:917\u2013927. https:\/\/doi.org\/10.1007\/s00330-023-10170-9","journal-title":"Eur Radiol"},{"key":"653_CR21","doi-asserted-by":"publisher","first-page":"914","DOI":"10.1007\/s00330-023-10190-5","volume":"34","author":"T Zhang","year":"2023","unstructured":"Zhang T, Mann RM (2023) Contrast-enhanced mammography: better with AI? Eur Radiol 34:914\u2013916. https:\/\/doi.org\/10.1007\/s00330-023-10190-5","journal-title":"Eur Radiol"},{"key":"653_CR22","doi-asserted-by":"publisher","first-page":"780","DOI":"10.1007\/s11307-019-01423-5","volume":"22","author":"MA Marino","year":"2020","unstructured":"Marino MA, Pinker K, Leithner D et al (2020) Contrast-enhanced mammography and radiomics analysis for noninvasive breast cancer characterization: initial results. Mol Imaging Biol 22:780\u2013787. https:\/\/doi.org\/10.1007\/s11307-019-01423-5","journal-title":"Mol Imaging Biol"},{"key":"653_CR23","doi-asserted-by":"publisher","first-page":"4834","DOI":"10.1007\/s00330-022-08538-4","volume":"32","author":"C Dominique","year":"2022","unstructured":"Dominique C, Callonnec F, Berghian A et al (2022) Deep learning analysis of contrast-enhanced spectral mammography to determine histoprognostic factors of malignant breast tumours. Eur Radiol 32:4834\u20134844. https:\/\/doi.org\/10.1007\/s00330-022-08538-4","journal-title":"Eur Radiol"}],"container-title":["European Radiology Experimental"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/link.springer.com\/content\/pdf\/10.1186\/s41747-025-00653-3.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/link.springer.com\/article\/10.1186\/s41747-025-00653-3\/fulltext.html","content-type":"text\/html","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/link.springer.com\/content\/pdf\/10.1186\/s41747-025-00653-3.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,12,2]],"date-time":"2025-12-02T08:24:52Z","timestamp":1764663892000},"score":1,"resource":{"primary":{"URL":"https:\/\/eurradiolexp.springeropen.com\/articles\/10.1186\/s41747-025-00653-3"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,12,2]]},"references-count":23,"journal-issue":{"issue":"1","published-online":{"date-parts":[[2025,12]]}},"alternative-id":["653"],"URL":"https:\/\/doi.org\/10.1186\/s41747-025-00653-3","relation":{},"ISSN":["2509-9280"],"issn-type":[{"value":"2509-9280","type":"electronic"}],"subject":[],"published":{"date-parts":[[2025,12,2]]},"assertion":[{"value":"24 March 2025","order":1,"name":"received","label":"Received","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"17 October 2025","order":2,"name":"accepted","label":"Accepted","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"2 December 2025","order":3,"name":"first_online","label":"First Online","group":{"name":"ArticleHistory","label":"Article History"}},{"order":1,"name":"Ethics","group":{"name":"EthicsHeading","label":"Declarations"}},{"value":"Data collection was approved by the Ethics Committee (No. 118\/20) of \u201cMaggiore della Carit\u00e0\u201d Hospital in Novara. Written informed consent was obtained from all participants.","order":2,"name":"Ethics","group":{"name":"EthicsHeading","label":"Ethics approval and consent to participate"}},{"value":"Written informed consent was obtained from all participants.","order":3,"name":"Ethics","group":{"name":"EthicsHeading","label":"Consent for publication"}},{"value":"ACo is a member of the Scientific Editorial Board of\n                      European Radiology Experimental\n                      (section editor: Abdomen\/gastrointestinal) and, as such, did not participate in the selection or review processes for this article. The remaining authors declare no conflicts of interest.","order":4,"name":"Ethics","group":{"name":"EthicsHeading","label":"Competing interests"}}],"article-number":"116"}}