{"title":"Multimodal artificial intelligence in breast cancer: towards integrated prediction and personalized care.","authors":"Tianyu Zhang, Xinglong Liang, Ritse M Mann","doi":"10.1007/s00330-025-12085-z","DOIUrl":"https://doi.org/10.1007/s00330-025-12085-z","url":null,"abstract":"","PeriodicalId":12076,"journal":{"name":"European Radiology","volume":" ","pages":""},"PeriodicalIF":4.7,"publicationDate":"2025-10-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145343885","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Yasuhito Mitsuyama, Hirotaka Takita, Shannon L Walston, Ko Watanabe, Shoya Ishimaru, Yukio Miki, Daiju Ueda
{"title":"Deep learning models for radiography body-part classification and chest radiograph projection/orientation classification: a multi-institutional study.","authors":"Yasuhito Mitsuyama, Hirotaka Takita, Shannon L Walston, Ko Watanabe, Shoya Ishimaru, Yukio Miki, Daiju Ueda","doi":"10.1007/s00330-025-12053-7","DOIUrl":"https://doi.org/10.1007/s00330-025-12053-7","url":null,"abstract":"<p><strong>Objectives: </strong>Large-scale radiographic datasets often include errors in labels such as body parts or projection, which can undermine automated image analysis. Therefore, we aimed to develop and externally validate two deep-learning models-one for categorising radiographs by body part, and another for identifying projection and rotation of chest radiographs-using large, diverse datasets.</p><p><strong>Materials and methods: </strong>We retrospectively collected radiographs from multiple institutions and public repositories. For the first model (Xp-Bodypart-Checker), we included seven categories (Head, Neck, Chest, Incomplete Chest, Abdomen, Pelvis, Extremities). For the second model (CXp-Projection-Rotation-Checker), we classified chest radiographs by projection (anterior-posterior, posterior-anterior, lateral) and rotation (upright, inverted, left rotation, right rotation). Both models were trained, tuned, and internally tested on separate data, then externally tested on radiographs from different institutions. Model performance was assessed using overall accuracy (micro, macro, and weighted) as well as one-vs.-all area under the receiver operating characteristic curve (AUC).</p><p><strong>Results: </strong>In the Xp-Bodypart-Checker development phase, we included 429,341 radiographs obtained from Institutions A, B, and MURA. In the CXp-Projection-Rotation-Checker development phase, we included 463,728 chest radiographs from CheXpert, PadChest, and Institution A. The Xp-Bodypart-Checker achieved AUC values of 1.00 (99% CI: 1.00-1.00) for all classes other than Incomplete Chest, which had an AUC value of 0.99 (99% CI: 0.98-1.00). The CXp-Projection-Rotation-Checker demonstrated AUC values of 1.00 (99% CI: 1.00-1.00) across all projection and rotation classifications.</p><p><strong>Conclusion: </strong>These models help automatically verify image labels in large radiographic databases, improving quality control across multiple institutions.</p><p><strong>Key points: </strong>Question This study examines how deep learning can accurately classify radiograph body parts and detect chest projection/orientation in large, multi-institutional datasets, enhancing metadata consistency for clinical and research workflows. Findings Xp-Bodypart-Checker classified radiographs into seven categories with AUC values of over 0.99 for all classes, while CXp-Projection-Rotation-Checker achieved AUC values of 1.00 across all projection and rotation classifications. Clinical relevance Trained on over 860,000 multi-institutional radiographs, our two deep-learning models classify radiograph body-part and chest radiograph projection/rotation, identifying mislabeled data and enhancing data integrity, thereby improving reliability for both clinical use and deep-learning research.</p>","PeriodicalId":12076,"journal":{"name":"European Radiology","volume":" ","pages":""},"PeriodicalIF":4.7,"publicationDate":"2025-10-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145344377","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Myriam Margareta Keymling, Felix Tobias Kurz, Tristan Anselm Kuder, Sebastian Bickelhaupt, Thomas Hielscher, Robert Hog, Theresa Mokry, Tawfik Moher Alsady, Sarah Schott, Christian Kratz, Diane Miriam Renz, Heinz-Peter Schlemmer
{"title":"Optimizing whole-body MRI for early cancer detection in Li-Fraumeni syndrome: a prospective bicentric study.","authors":"Myriam Margareta Keymling, Felix Tobias Kurz, Tristan Anselm Kuder, Sebastian Bickelhaupt, Thomas Hielscher, Robert Hog, Theresa Mokry, Tawfik Moher Alsady, Sarah Schott, Christian Kratz, Diane Miriam Renz, Heinz-Peter Schlemmer","doi":"10.1007/s00330-025-11880-y","DOIUrl":"https://doi.org/10.1007/s00330-025-11880-y","url":null,"abstract":"<p><strong>Objectives: </strong>Annual whole-body MRI (WB-MRI) is recommended for early cancer detection in individuals with Li-Fraumeni syndrome (LFS). However, there is no agreement on a standardized MRI protocol. This study evaluated the diagnostic performance of different MRI sequences to suggest an optimized protocol for LFS surveillance.</p><p><strong>Materials and methods: </strong>In this prospective bicentric study, 113 participants with LFS underwent annual WB-MRI and were included in the analysis. The protocol comprised turbo-spin echo (TSE) T1-weighted and inversion-recovery T2-weighted (TIRM) images of the whole body in coronal orientation, and T2-weighted (HASTE), diffusion-weighted (DWI), and T1-weighted DIXON images (pre- and post-contrast agent administration) from head to thighs in axial orientation. An additional fluid-attenuated inversion recovery (FLAIR) sequence imaged the skull only. Initial clinical interpretation was conducted by staff radiologists. The visibility of reported mass lesions was independently graded in all sequences by three experienced radiologists using a Likert scale. Sequence combinations were compared to inform the design of an optimal MRI protocol.</p><p><strong>Results: </strong>Over 30 months, 189 WB-MRI examinations were performed in 113 participants (mean age 40 years, ±12.7 years [standard deviation], 91 women). 188 mass lesions were detected and confirmed as malignant (n = 38), benign (n = 120) or ambiguous (n = 30). In the multi-reader analysis, all new malignant lesions could have been detected by a combination of cranial FLAIR, whole-body DWI, and whole-body HASTE in the axial direction.</p><p><strong>Conclusion: </strong>A shortened, contrast-agent-free WB-MRI protocol combining cranial FLAIR, WB-HASTE, and WB-DWI promises to be an effective and patient-friendly approach for annual cancer surveillance in LFS.</p><p><strong>Key points: </strong>Question Annual whole-body MRI (WB-MRI) is recommended for early cancer detection for individuals with Li-Fraumeni syndrome (LFS), but a standardized sequence protocol has yet to be established. Findings The combination of cranial FLAIR, whole-body HASTE, and whole-body DWI in the axial plane enabled visualization of all newly developed malignant lesions in our study cohort. Clinical relevance A shortened, standardized WB-MRI protocol enables efficient, sensitive early cancer detection in individuals with LFS, minimizing patient burden by reducing examination time and contrast agent use. This approach may improve surveillance participation while enhancing comparability across centers.</p>","PeriodicalId":12076,"journal":{"name":"European Radiology","volume":" ","pages":""},"PeriodicalIF":4.7,"publicationDate":"2025-10-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145336503","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Estelle C Nijssen, Bibi Martens, Babs M Hendriks, Hester A Gietema, Joachim E Wildberger, Cécile R L P N Jeukens
{"title":"Computed tomography pulmonary angiography in around-the-clock clinical care with individualised scan protocols: a 5-year observational study on incidence and causes of repeat scanning.","authors":"Estelle C Nijssen, Bibi Martens, Babs M Hendriks, Hester A Gietema, Joachim E Wildberger, Cécile R L P N Jeukens","doi":"10.1007/s00330-025-12032-y","DOIUrl":"https://doi.org/10.1007/s00330-025-12032-y","url":null,"abstract":"<p><strong>Objectives: </strong>Elevated repeat-scanning rates are reported for CT pulmonary angiography (CTPA). Individualised protocols optimise contrast- and radiation-doses, but whether this affects repeat scanning is unknown. The current study evaluates repeat-CTPA in a 24/7, state-of-the-art clinical-care setting.</p><p><strong>Materials and methods: </strong>This is a retrospective observational single-centre study of consecutive CTPA acquired over a 5-year period during standard clinical care. The primary outcome is the repeat-scan rate. Repeat- and single-scan groups were compared for initial-scan characteristics (patient-related, CT-scanner, contrast-administration, kV-settings, regular hours/shifts, radiation-dose), and cumulative contrast- and radiation-doses. An expert radiologist panel retrospectively evaluated probable reasons for repeat scanning through visual, subjective assessment of initial-scan images.</p><p><strong>Results: </strong>CTPA repeat rate was 3.1% (139/4467). Repeat- and single-scan groups significantly differed: age (55 ± 18 vs. 63 ± 17 years; p < 0.001), Body Mass Index (27 kg/m<sup>2</sup> (IQR 7) vs. 25 kg/m<sup>2</sup> (IQR 6); p = 0.022), radiation-dose (141 mGy∙cm (IQR 73) vs. 121 mGy∙cm (IQR 70); p < 0.001). Cumulative contrast- and radiation-doses were: 96 mL (IQR 31) vs. 48 mL (IQR 22) (p < 0.001); 0.36 gI/kg (IQR 0.11) vs. 0.18 gI/kg (IQR 0.51) (p < 0.001); 272 mGy∙cm (IQR 69) vs. 121 mGy∙cm (IQR 70) (p < 0.001). Retrospective expert-consensus reasons for repeat scanning were: 31/133 patient-related; 28/133 multifactorial; 12/133 contrast/scan-protocol; 4/133 operator-error; 2/133 unidentified. 56/133 (42%) initial scans were retrospectively deemed diagnostic-quality, and these significantly differed from other repeat-categories in patient characteristics age (51 ± 15 vs. 57 ± 19 years; p = 0.045) and sex (64.3% vs. 50.6% female; p = 0.045), and in contrast volume (48 mL (IQR 17) vs. 46 mL (IQR 24); p = 0.031).</p><p><strong>Conclusion: </strong>Individualised scan protocols yielded diagnostic images around the clock, with repeat scanning well within ranges published in the literature. Retrospective expert evaluation suggests repeat rates as low as 1.2% may be possible. Repeat- and single-scan groups significantly differed in patient characteristics, and repeat-scanning reasons were mostly patient-related. These results suggest further tailoring protocols to (younger, female) patients might be beneficial in helping to further reduce CTPA-repeats.</p><p><strong>Key points: </strong>Question CT pulmonary angiography (CTPA) is subject to relatively high repeat-scanning rates, but it is not known how state-of-the-art CTPA and individualised protocols perform in clinical practice today. Findings During 5 years of clinical practice the repeat rate was 3%; retrospective expert image-evaluation suggests a repeat rate as low as 1.2% may be possible. Clinical relevance Repeat- and single-scan groups significantly ","PeriodicalId":12076,"journal":{"name":"European Radiology","volume":" ","pages":""},"PeriodicalIF":4.7,"publicationDate":"2025-10-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145318257","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Predicting cervical lymph node metastasis in papillary thyroid carcinoma using capsule disruption length measured by 3D-US.","authors":"Ruyu Liu, Yuxin Jiang, Xingjian Lai, Ying Wang, Luying Gao, Ruina Zhao, Xuehua Xi, Bo Zhang","doi":"10.1007/s00330-025-12070-6","DOIUrl":"https://doi.org/10.1007/s00330-025-12070-6","url":null,"abstract":"<p><strong>Objectives: </strong>Papillary thyroid carcinoma (PTC) is a prevalent endocrine malignancy with a propensity for lymph node metastasis (LNM). Extrathyroidal extension (ETE) is a key factor in preoperative LNM prediction. The criteria for ultrasound diagnosis of ETE remain controversial. The aim is to determine if the length of capsule disruption (LCD) on three-dimensional ultrasound (3D-US) can predict cervical LNM in PTC patients.</p><p><strong>Material and methods: </strong>A prospective cohort of 168 patients from Peking Union Medical College Hospital was examined by 3D-US. The LCD was measured using the omniview mode of 3D-US. Statistical analyses included Chi-square tests, T-tests, Mann-Whitney tests, ROC curve analysis, and logistic regression analysis.</p><p><strong>Results: </strong>Of the 126 patients included, 71 had LNM. Younger age, male gender, larger malignant nodules, LCD, echogenic foci, and thyroid capsule invasion were significantly associated with LNM. LCD ≥ 0.42 cm increases LNM risk by 4.097 (p < 0.001). A nomogram was constructed incorporating gender, age, maximum diameter of the largest malignant nodule (MDLM), and LCD to estimate the risk of LNM. The accuracy and AUC of the nomogram were 73.0% and 0.795 (0.718-0.873).</p><p><strong>Conclusions: </strong>LCD on 3D-US is a significant predictor of cervical LNM in PTC patients. This study's nomogram, based on easily measurable parameters, can help in the preoperative assessment of LNM risk, potentially guiding surgical management.</p><p><strong>Key points: </strong>Question Can the LCD measured by 3D-US predict cervical LNM in PTC? Findings LCD ≥ 0.42 cm on 3D-US increase LNM risk by 4.097-fold. The nomogram with LCD, gender, age, and nodule size shows good predictive ability (AUC = 0.795). Clinical relevance LCD is a promising predictor of LNM, an alternative to ultrasound thyroid capsule invasion evaluation. The nomogram enables risk-adapted surgery, reducing unnecessary dissection or missed metastases to improve patient outcomes.</p>","PeriodicalId":12076,"journal":{"name":"European Radiology","volume":" ","pages":""},"PeriodicalIF":4.7,"publicationDate":"2025-10-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145312764","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Letter to the Editor: Early post-treatment imaging enables timely diagnosis of viable hepatocellular carcinomas after selective internal radiation therapy.","authors":"Sung-Hua Chiu, Aya Kamaya, Justin R Tse","doi":"10.1007/s00330-025-11998-z","DOIUrl":"https://doi.org/10.1007/s00330-025-11998-z","url":null,"abstract":"","PeriodicalId":12076,"journal":{"name":"European Radiology","volume":" ","pages":""},"PeriodicalIF":4.7,"publicationDate":"2025-10-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145312770","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Reply to the Letter to the Editor: Respiratory symptoms associated with a new lobe-based bronchial scoring system in an urban Chinese low-dose CT screening population.","authors":"Zhenhui Nie, Monique D Dorrius","doi":"10.1007/s00330-025-12019-9","DOIUrl":"https://doi.org/10.1007/s00330-025-12019-9","url":null,"abstract":"","PeriodicalId":12076,"journal":{"name":"European Radiology","volume":" ","pages":""},"PeriodicalIF":4.7,"publicationDate":"2025-10-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145312717","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}