European Radiology最新文献

{"title":"Contribution of mammographic imaging after initial ultrasound in women with focal breast complaints-systematic review.","authors":"Linda Appelman, Eugenio Gil Quessep, Liesbeth Veenendaal, Carmen Siebers, Marialena Tsarouchi, Wendelien Sanderink, Alexandra Athanasiou, Ritse Mann","doi":"10.1007/s00330-025-12218-4","DOIUrl":"10.1007/s00330-025-12218-4","url":null,"abstract":"<p><strong>Objectives: </strong>Breast cancer is the most common cancer among women, making accurate diagnostic imaging essential for evaluating focal breast complaints. Standard breast imaging includes mammography (MG), digital breast tomosynthesis (DBT), and targeted ultrasound (US). Little information is known about the added value of MG/DBT compared to targeted US alone in women with focal breast complaints. This review provides a systematic overview of the (relative) sensitivity of MG/DBT, next to the use of targeted US in women with focal breast complaints, while emphasizing the variations of the reported guidelines.</p><p><strong>Materials and methods: </strong>A comprehensive PubMed search was conducted to identify articles published between January 2002 and February 2024 that assessed the diagnostic performance of MG/DBT and targeted US in women presenting with focal breast complaints.</p><p><strong>Results: </strong>Of the 985 identified studies, 10 met the inclusion criteria after full-text evaluation. Palpable lumps and focal pain are the more frequent symptoms. Most studies show a high sensitivity (range 84.6-100%, the largest study reported 98.3%) of targeted US alone in relatively young women (30-39 years). Mammograms show added value, particularly when it comes to determining the extent of breast cancer and when confirming a benign abnormality.</p><p><strong>Conclusions: </strong>Targeted US is a reliable initial imaging modality, particularly in younger women (≤ 40). When the US suggests a benign finding in low/average-risk women, the added value of MG is minimal. Routine use should be weighed against the risk of overdiagnosis, unnecessary biopsies, and radiation exposure.</p><p><strong>Key points: </strong>Question The best imaging strategy for women aged 30-40 with focal breast complaints is uncertain, and whether US alone suffices remains unresolved. Findings Initial targeted US shows high accuracy in this diagnostic setting, with MG/DBT adding value mainly in determination of tumor extent and from a screening perspective. Clinical relevance Using US as the primary imaging for women < 40 with focal breast symptoms may reduce MG, radiation, and discomfort, while maintaining cancer detection; however, any discordance between clinical and imaging findings should always prompt further evaluation.</p>","PeriodicalId":12076,"journal":{"name":"European Radiology","volume":" ","pages":"3346-3357"},"PeriodicalIF":4.7,"publicationDate":"2026-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145793510","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":"Diagnostic performance of shear-wave dispersion imaging and shear-wave elastography for differentiating benign and malignant breast masses.","authors":"Eunkyung Woo, Ji Young Lee, Bo Kyoung Seo, Min Sun Bae, Eun Kyung Park, Hangseok Choi, Jiyung Choo, Kyu Ran Cho, Ok Hee Woo, Sung Eun Song, Soo-Yeon Kim","doi":"10.1007/s00330-025-12233-5","DOIUrl":"10.1007/s00330-025-12233-5","url":null,"abstract":"<p><strong>Objectives: </strong>Ultrasound advancements enable the automatic quantification of viscosity by measuring the dispersion slope of shear waves. This study evaluated the diagnostic performance of shear-wave dispersion imaging (SWD) in distinguishing benign from malignant breast masses and its supplementary role compared to shear-wave elastography (SWE).</p><p><strong>Materials and methods: </strong>This retrospective study included 630 pathologically confirmed breast lesions (395 benign, 235 malignant) in 604 women (mean age, 48 ± 11 years) from August 2020 to August 2021. Tumor elasticity (E_tumor) and the tumor-to-fat elasticity ratio (E_ratio) were measured using SWE, while tumor dispersion (D_tumor) and the tumor-to-fat dispersion ratio (D_ratio) were obtained using SWD. Diagnostic performance was evaluated using the AUC, with subgroup analyses based on lesion size, depth, and breast thickness.</p><p><strong>Results: </strong>D_tumor achieved the highest AUC (0.96 [95% CI: 0.95, 0.98]) compared with E_tumor (0.93 [95% CI: 0.91, 0.95]), E_ratio (0.92 [95% CI: 0.90, 0.94]), and D_ratio (0.95 [95% CI: 0.93, 0.96]) (p ≤ 0.003). D_tumor outperformed E_tumor for small lesions (≤ 10 mm and 11-20 mm: AUC 0.84 vs 0.67 and 0.91 vs 0.86, respectively; p ≤ 0.002) and BI-RADS 4B lesions (AUC 0.82 vs 0.70, p < 0.001). D_tumor maintained high performance across varying lesion depth and breast thickness, and it differentiated ductal carcinoma in situ from invasive carcinoma (p = 0.04), which SWE did not achieve.</p><p><strong>Conclusion: </strong>SWD improves the diagnostic performance across varying lesion size, depth, and breast thickness, addressing SWE's limitations. It is particularly effective for small lesions (≤ 20 mm) and BI-RADS 4B masses.</p><p><strong>Key points: </strong>Question To evaluate the diagnostic performance of SWD imaging, an automatic viscosity measurement derived from shear waves, in differentiating benign from malignant breast masses. Findings SWD outperformed SWE, improving diagnostic performance for small (≤ 20 mm) and BI-RADS 4B masses while maintaining consistent performance across lesion depths and breast thicknesses. Clinical relevance SWD, obtained automatically during SWE, enhances the diagnostic performance in distinguishing benign from malignant breast lesions, particularly in small and BI-RADS 4B masses, while addressing the limitations of SWE related to lesion size, depth, and breast thickness.</p>","PeriodicalId":12076,"journal":{"name":"European Radiology","volume":" ","pages":"3378-3391"},"PeriodicalIF":4.7,"publicationDate":"2026-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145818272","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":"ESR Essentials: thyroid imaging-practice recommendations by the European Society of Head and Neck Radiology.","authors":"Edith Vassallo, Anne Péporté, Andrew McQueen, Minerva Becker, Jussi Hirvonen","doi":"10.1007/s00330-025-12101-2","DOIUrl":"10.1007/s00330-025-12101-2","url":null,"abstract":"<p><p>Thyroid nodules are frequently encountered at imaging, yet most are benign and do not require intervention. The clinical challenge lies in distinguishing nodules that warrant further investigation from those that do not, to avoid unnecessary biopsies, anxiety, and overtreatment. Ultrasound (US) is the primary imaging modality for thyroid nodule evaluation, supported by structured risk stratification systems such as ACR TI-RADS and EU-TIRADS, which incorporate specific sonographic features and size thresholds to guide clinical decision-making. Nodules without high-risk features can be safely monitored or ignored, especially in asymptomatic patients. Conversely, suspicious characteristics (e.g. irregular margins, microcalcifications, or marked hypoechogenicity) should prompt further assessment, including fine-needle aspiration (FNA). Diffuse thyroid disorders, including Hashimoto's thyroiditis and Graves' disease, are best assessed using US and thyroid function tests. In thyroid cancer, a multidisciplinary team approach involving radiologists, endocrinologists, pathologists, and nuclear medicine specialists is essential for optimal patient care. To implement these recommendations, radiologists should adopt standardised US reporting systems, apply evidence-based criteria for further workup, and collaborate closely with referring clinicians. This approach ensures accurate diagnosis, reduces unnecessary procedures, and aligns radiological practice with current guidelines to support high-value, patient-centred care. KEY POINTS: Ultrasound (US) is the gold standard imaging modality for evaluating thyroid pathology. Implementation of the EU-TIRADS and ACR-TIRADS constitutes a critical part of the work-up of thyroid nodules and is essential for their effective management. Interdisciplinary discussion with all specialists concerned is the most effective way of ensuring that the best possible management strategy is implemented in thyroid cancers.</p>","PeriodicalId":12076,"journal":{"name":"European Radiology","volume":" ","pages":"3788-3796"},"PeriodicalIF":4.7,"publicationDate":"2026-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13086679/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145548618","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Prostate MRI learning curves: establishing training benchmarks for radiology and urology trainees.","authors":"Pavel Stegarescu, Egon Burian, Amelie Lutz, Nathan Perlis, Ulrich Grosse, Nemanja Avramovic, Stoyan Benev, Constantin Bolz, Pia Götz, Marc Koschler, Joana Kostova, Ana Macek, Abigail Martin Mens, Khashayar Namdar, Ioan Popa, Aileen Satari, Sydney Schmidt, Feri Töckelt, Roman Wiegele, Jan Klein, Thomas Herrmann, Gustav Andreisek, Dominik Deniffel","doi":"10.1007/s00330-025-12177-w","DOIUrl":"10.1007/s00330-025-12177-w","url":null,"abstract":"<p><strong>Objectives: </strong>Evidence-based training benchmarks for prostate multiparametric MRI (mpMRI) interpretation remain undefined amid growing educational demands. We compared learning curves between radiology and urology trainees and quantified the impact of prior radiological experience.</p><p><strong>Materials and methods: </strong>Fourteen trainees (10 radiology, median 2.7 years experience; 4 urology, no imaging experience), all naïve to prostate mpMRI, prospectively interpreted 200 cases using a feedback-based platform. Performance metrics included agreement with expert consensus reference for PI-RADSv2.1 (≥ 3), PI-QUALv2 image quality, extraprostatic extension (EPE) grading, and readout time. Learning curves were modeled using generalized estimating equations; segmented regression identified inflection points; bootstrapping generated 95% CIs.</p><p><strong>Results: </strong>Prior radiological experience showed no significant impact on PI-RADSv2.1 (OR per year 1.06 [95% CI: 0.96, 1.16]) or PI-QUALv2 (1.05 [0.99, 1.23]), with a minor effect on EPE grading (1.11 [1.03, 1.24]). Final PI-RADSv2.1 agreement with reference was similar (urology 80.9%, radiology 77.4%; OR 1.24 [0.55, 3.10]), with sensitivity/specificity 0.84/0.80 and 0.83/0.79, and Cohen's κ values (0.64 and 0.61) matching inter-expert κ = 0.63. Learning plateaued after 69-75 cases. Urology trainees demonstrated higher baseline PI-QUALv2/EPE agreement (OR 2.01 [1.35, 3.02] and 1.90 [1.11, 2.93]), but radiology trainees achieved similar final performance (PI-QUALv2: 88.0% vs 89.9%, OR 0.82 [0.35, 1.72]; EPE: 84.6% vs 90.0%, OR 0.61 [0.31, 1.42]). Readout times decreased markedly in both groups (final difference 53.3 s [-9.4, 95.9]).</p><p><strong>Conclusion: </strong>Feedback-based training enabled similar prostate mpMRI interpretation performance across specialties, with most learning within 75 cases. Prior radiological experience had a limited impact. These empirical benchmarks inform certification standards and early-residency curricula in radiology and urology.</p><p><strong>Key points: </strong>Question Evidence-based training benchmarks for prostate mpMRI interpretation competency in radiology and urology trainees remain undefined, despite growing educational needs and clinical demands. Findings Learning curves of 200 prostate mpMRI cases with feedback showed radiology and urology trainees plateauing after 69-75 cases with similar PI-RADSv2.1, PI-QUALv2, EPE grading performance. Clinical relevance Our findings establish an empirical benchmark (~75 cases) to guide prostate mpMRI certification standards and support the implementation of training curricula early in residency across specialties, regardless of prior radiological experience.</p>","PeriodicalId":12076,"journal":{"name":"European Radiology","volume":" ","pages":"3431-3441"},"PeriodicalIF":4.7,"publicationDate":"2026-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13086796/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145767127","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Replacing true non-contrast imaging with DECT in GI bleeding demonstrates non-inferior diagnostic performance, reading time and confidence.","authors":"Moritz Oberparleiter, Hanns-Christian Breit, Jan Vosshenrich, Alina C Seifert, Paul Hehenkamp, Sonaz Malekzadeh, Adrian Kobe, Daniel T Boll, Christoph J Zech, Markus M Obmann","doi":"10.1007/s00330-025-12191-y","DOIUrl":"10.1007/s00330-025-12191-y","url":null,"abstract":"<p><strong>Objectives: </strong>To determine whether a dual-energy CT (DECT) protocol-including arterial and portal-venous phases, virtual non-contrast (VNC) images, and iodine maps-provides non-inferior diagnostic performance to a conventional triphasic CT protocol for gastrointestinal (GI) bleeding.</p><p><strong>Materials and methods: </strong>In this retrospective single-center diagnostic accuracy study, we included all patients who underwent triphasic abdominal CT for GI bleeding between September 2015 and June 2024. For each case, conventional and DECT datasets were generated. Three fellowship-trained abdominal radiologists and two residents independently assessed all cases for active GI bleeding. A consensus review served as the reference standard. Sensitivity and specificity were compared using the Wald method for paired confidence intervals (non-inferiority margin 3%). Diagnostic confidence and reading time were analyzed using the Wilcoxon signed-rank test. Inter-reader agreement was assessed with Fleiss' kappa.</p><p><strong>Results: </strong>One hundred patients (mean age, 70 ± 14 years; 34 women) were evaluated, including 50 with GI bleeding (21 upper, 29 lower) and 50 controls. With conventional triphasic CT, sensitivity and specificity were 91.6% and 94.4%, respectively. With DECT, they were 94.4% and 96.0%, demonstrating non-inferiority within a 3% margin. Diagnostic confidence increased from 4 (IQR, 4-5) to 5 (IQR, 4-5) (p < 0.01). Mean reading time decreased from 96.3 s to 93.6 s (p < 0.01), also meeting non-inferiority. Inter-reader agreement was almost perfect (κ = 0.82). Total DLP was reduced by 20% when true non-contrast images were omitted.</p><p><strong>Conclusion: </strong>DECT-derived VNC and iodine maps provide non-inferior diagnostic performance to conventional CT for GI bleeding and can replace true non-contrast imaging.</p><p><strong>Key points: </strong>Question Eliminating true non-contrast scans from CT protocols for gastrointestinal bleeding could reduce radiation dose, but its impact on diagnostic performance remains unclear. Findings Dual-energy CT with virtual non-contrast images and iodine maps achieved non-inferior sensitivity, specificity, reading time, and diagnostic confidence compared to conventional triphasic CT. Clinical relevance Dual-energy CT with virtual non-contrast images and iodine maps can reliably replace true non-contrast scans in GI bleeding protocols, maintaining diagnostic performance while reducing radiation exposure-offering a safer and more efficient diagnostic alternative for patients.</p>","PeriodicalId":12076,"journal":{"name":"European Radiology","volume":" ","pages":"3950-3959"},"PeriodicalIF":4.7,"publicationDate":"2026-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13086890/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145767084","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"CT and MRI radiomics in cardiovascular risk prediction: a systematic review and meta-analysis by the EuSoMII Radiomics Auditing Group.","authors":"Armando Ugo Cavallo, Andrea Ponsiglione, Bernardo Pereira, Carlo Di Donna, Emmanouil Koltsakis, Federica Vernuccio, Mario Laudazi, Roberto Cannella, Salvatore Claudio Fanni, Tugba Akinci D'Antonoli, Renato Cuocolo","doi":"10.1007/s00330-025-12236-2","DOIUrl":"10.1007/s00330-025-12236-2","url":null,"abstract":"<p><strong>Objectives: </strong>To conduct a comprehensive systematic review of the studies applying radiomics to CT and MRI for the evaluation of cardiac disease, and to perform a meta-analysis of their diagnostic accuracy, focused on cardiovascular events prediction. A secondary aim was to assess the methodological quality of cardiac imaging radiomics studies using the METRICS score.</p><p><strong>Materials and methods: </strong>Four investigators searched multiple medical literature archives (Scopus, Web of Science, and PubMed). The search was conducted from February 7th, 2021, to March 10th, 2025. Papers were also screened to identify studies for the prediction of cardiovascular events, defined as the occurrence of major cardiovascular events or myocardial ischemia. Methodological quality was assessed by using the METRICS tool. Diagnostic accuracy was estimated with pooled area under the curve (AUC).</p><p><strong>Results: </strong>A total of 202 studies were included in the final analysis. Seventeen papers were identified for the meta-analysis, of which 9 were considered eligible for analysis. 111 papers (55%) had CT as the imaging modality, and 91 (45%) papers had MRI. Overall, the average METRICS total score was 54.52% ± 15.89%. Meta-analysis showed pooled AUC of 0.81 (95% CI: 0.75-0.87), with a high level of heterogeneity (I² = 83.4%, τ² = 0.0068). Egger's test for funnel plot asymmetry was statistically significant (z = -2.39, p = 0.017), suggesting potential publication bias.</p><p><strong>Conclusion: </strong>Radiomics in cardiac imaging holds potential, showing moderate quality and relatively high cumulative performance for the prediction of cardiovascular events.</p><p><strong>Key points: </strong>Question What is the current methodological quality and pooled diagnostic performance of cardiovascular radiomics for predicting clinical events, based on a meta-analysis? Findings The average METRICS quality score was 54.52%. A meta-analysis showed a pooled AUC of 0.81 for event prediction, but with high heterogeneity and publication bias. Clinical relevance Assessing radiomics research methodological quality is crucial to enhance reproducibility and clinical applicability of radiomics pipelines. The evaluation of cumulative evidence for cardiovascular events prediction may guide clinical translation and future study design.</p>","PeriodicalId":12076,"journal":{"name":"European Radiology","volume":" ","pages":"4049-4060"},"PeriodicalIF":4.7,"publicationDate":"2026-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13086800/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145818231","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Moderate and severe traumatic brain injury: lesion frequency distribution maps and prognostic importance of brain contusions on early MRI.","authors":"Tiril Svaasand Eliassen, Lars Eirik Bø, Anne-Mari Holte Flusund, Anne Katharina Köster, Nada Andelic, Toril Skandsen, Anne Vik, Kent Gøran Moen","doi":"10.1007/s00330-025-12159-y","DOIUrl":"10.1007/s00330-025-12159-y","url":null,"abstract":"<p><strong>Objective: </strong>Brain contusion frequency, distribution and prognostic value were investigated on early MRI in moderate-severe traumatic brain injury (TBI).</p><p><strong>Materials and methods: </strong>We prospectively included 301 patients (8-70 years) admitted to Trondheim or Oslo University Hospitals with moderate (n = 123) or severe (n = 178) TBI and brain contusion(s) on MRI within 6 weeks (median 9 days) post-injury. Volumetric segmentation of brain contusions was performed manually on fluid-attenuated inversion recovery (FLAIR) MRI. The segmentations were combined into lesion frequency distribution maps for the whole cohort, moderate and severe TBI separately and three outcome categories. The 12-month outcome was assessed with the Glasgow Outcome Scale-Extended (GOSE). Relationship with the outcome was evaluated visually, with adjusted analyses and with voxel-based lesion-symptom mapping (VLSM).</p><p><strong>Results: </strong>The frontal (75%) and temporal (82%) lobes had the highest frequency of brain contusions. There was no significant difference in total lesion volume between moderate and severe TBI (median: moderate TBI 15.8 cm³, severe TBI 13.6 cm³, p = 0.30). The total brain contusion volume significantly predicted the 12-month outcome in the adjusted models and decreased with increasing outcome category (GOSE score 1-4, median volume: 37.8 cm³; GOSE score 5-6, median volume: 14.5 cm³; GOSE score 7-8, median volume: 7.1 cm³; p = 0.02). There was no significant association between lesion location and the GOSE score in the VLSM analyses.</p><p><strong>Conclusion: </strong>In this MRI cohort study, we found similar distributions and volumes of brain contusions in moderate-severe TBI. The brain contusion volume significantly predicted the 12-month outcome, whereas the lesion location was not associated with the outcome.</p><p><strong>Key points: </strong>Question Accurate prognostic models for TBI are important for clinical decision making, but the value of including MRI measures remains unclear. Findings In moderate and severe TBI, brain contusions had the highest frequency in frontotemporal regions. Contusion volume predicted 12-month functional outcome, whereas there was no association with location. Clinical relevance Our findings indicate that brain contusion volume is more important for functional outcomes than lesion location. The brain contusion volume should be included in future AI models as one predictor of patient outcome after TBI.</p>","PeriodicalId":12076,"journal":{"name":"European Radiology","volume":" ","pages":"3518-3530"},"PeriodicalIF":4.7,"publicationDate":"2026-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13086803/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145793580","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Retractions of publications in radiomics: An underestimated problem?","authors":"Aydin Demircioğlu","doi":"10.1007/s00330-025-12231-7","DOIUrl":"10.1007/s00330-025-12231-7","url":null,"abstract":"<p><p>Radiomics is increasingly explored as a tool for improving diagnosis, prognosis, and treatment planning. However, concerns exist about the reproducibility and methodological rigor of its studies. The integration of high-dimensional radiomic features and machine learning makes the field prone to unintentional errors that may warrant retraction. Despite a rising number of retractions in science overall, no dedicated study has examined retractions specifically within radiomics. Therefore, this study aimed to review retracted radiomics publications and identify the characteristics and reasons for their retraction. We systematically searched six databases (Crossref, Retraction Watch Database, OpenAlex, PubMed, Scopus, Web of Science) and identified 93 retracted radiomics publications, of which 20 were included. These articles were analyzed with respect to publisher, country of origin, dates, citation counts, and reasons for retraction. Retraction rates were then estimated and compared with those in general radiology. Our findings indicate that a disproportionate number of retractions are linked to specific publishers and countries (particularly China and India), with overall low citation counts (median 4.0 citations). Retractions peaked sharply in 2023, followed by a strong decline. Many retraction notes lack a clear explanation for the retraction. Estimated retraction rates in radiomics were lower than in general radiology (6.7 vs 7.4 per 10,000 publications). Notably, no major radiological or oncological journal appears to have retracted a radiomics publication. Given that radiomics demands higher, interdisciplinary expertise, this suggests a gap, implying that flawed research may yet have to be retracted. KEY POINTS: Question Considering the technical complexity of radiomics studies and their susceptibility to unintentional errors, how do their retraction rates compare to those in general radiology? Findings Retractions in radiomics were disproportionately linked to specific publishers and countries; however, no retractions appeared in major journals. Estimated retraction rates were lower than those for general radiology publications. Clinical relevance A potential gap in the number of retracted radiomics studies was identified, implying that flawed research in the field may not yet have been addressed.</p>","PeriodicalId":12076,"journal":{"name":"European Radiology","volume":" ","pages":"3778-3787"},"PeriodicalIF":4.7,"publicationDate":"2026-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13086724/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145793506","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Association between embolic agent choice and complications after transcatheter arterial embolization for colonic diverticular bleeding.","authors":"Takatoshi Kubo, Hayato Yamana, Shotaro Aso, Hiroki Matsui, Kiyohide Fushimi, Hideo Yasunaga, Osamu Abe","doi":"10.1007/s00330-025-12210-y","DOIUrl":"10.1007/s00330-025-12210-y","url":null,"abstract":"<p><strong>Objectives: </strong>Although various embolic agents are used for transcatheter arterial embolization (TAE) of colonic diverticular bleeding (CDB), comparative outcome data for different embolic agents are limited. We aimed to assess the association between embolic agent choice and early rebleeding and intestinal ischemia after TAE for CDB.</p><p><strong>Materials and methods: </strong>We conducted a nationwide retrospective cohort study using the Japanese Diagnosis Procedure Combination database between July 2010 and March 2022. Adults who underwent a first TAE for CDB with coils, gelatin sponge (GS) particles, or n‑butyl‑2‑cyanoacrylate (NBCA) were included. Multivariate logistic regression analyses were performed to evaluate the association of embolic agent choice with early rebleeding requiring intervention and intestinal ischemic complications, while adjusting for covariates and within-hospital clustering.</p><p><strong>Results: </strong>The cohort comprised 5625 patients (mean age 72 years ± 12 [standard deviation], 4020 men). Coils, GS particles, and NBCA were used in 59%, 30%, and 11%, respectively. The overall early incidence of rebleeding and intestinal ischemia was 12% and 1.0%, respectively. With coils as the reference, the adjusted odds ratio for GS particles was 1.38 (95% CI: 1.15-1.66; p = 0.001) for early rebleeding and 2.64 (95% CI: 1.43-4.90; p = 0.002) for intestinal ischemia, and those for NBCA were 0.69 (95% CI: 0.50-0.95; p = 0.03) for early rebleeding and 3.53 (95% CI: 1.72-7.22; p = 0.001) for intestinal ischemia.</p><p><strong>Conclusions: </strong>Compared with coils, GS particles were associated with an increase in both early rebleeding and intestinal ischemia, whereas NBCA was associated with decreased rebleeding and increased ischemia.</p><p><strong>Key points: </strong>Question Transcatheter arterial embolization (TAE) is one of the mainstay treatments for colonic diverticular bleeding, but outcome differences among embolic agents remain unclear. Findings Gelatin sponge particles increased both early rebleeding and intestinal ischemic risks, while coils reduced intestinal ischemic risk and n‑butyl‑2‑cyanoacrylate reduced early rebleeding risk. Clinical relevance Embolic material selection should be individualized for TAE in colonic diverticular bleeding. Coils may be safer in ischemia-prone patients, while n‑butyl‑2‑cyanoacrylate may be suitable for those at high risk of rebleeding.</p>","PeriodicalId":12076,"journal":{"name":"European Radiology","volume":" ","pages":"3960-3968"},"PeriodicalIF":4.7,"publicationDate":"2026-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13086735/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145793043","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Development of PET/CT-clinical nomograms for predicting lymph node metastasis in primary lung cancer.","authors":"Xiaoyu Han, Wentao Li, Julie Lin, Shuoyang Li, Maliazurina B Saad, Yuliya Kitsel, Simon Heeke, Lingzhi Hong, Mohamed Qayati Mohamed, Xiuning Le, Natalie Vokes, Myrna Cb Godoy, Brett W Carter, Girish S Shroff, George Eapen, Lauren A Byers, Ara A Vaporciyan, Don L Gibbons, John Heymach, Carol C Wu, Jianjun Zhang, Jia Wu","doi":"10.1007/s00330-025-12166-z","DOIUrl":"10.1007/s00330-025-12166-z","url":null,"abstract":"<p><strong>Objective: </strong>Accurate staging of primary lung cancer is crucial for optimizing therapeutic strategies but remains challenging in clinical practice. We aimed to develop a nomogram incorporating clinical characteristics with CT and PET findings to predict lymph node metastasis (LNM) in primary lung cancer.</p><p><strong>Materials and methods: </strong>We retrospectively analyzed patients with primary lung cancer and mediastinal and hilar LNs from a tertiary care cancer center. All patients underwent endobronchial ultrasound-guided transbronchial needle aspiration, with diagnostic chest CT and PET-CT. Cytological confirmation of transbronchial needle aspiration samples served as the gold standard for diagnosing LNM. We employed an LN-level modeling approach and constructed five models for independent prediction of LNM: (1) Clinical-CT-PET model, (2) Clinical-CT model, (3) PET model, (4) Clinical-PET model, and (5) CT-PET model. Their performance was further evaluated in the subgroup of LNs < 1 cm.</p><p><strong>Results: </strong>This study included 455 patients (mean age 70 ± 10 years; 55.4% male), predominantly adenocarcinoma (62.0%). Most (68.1%) were stage III-IV. In total, 1391 lymph nodes (1112 training, 279 testing) were analyzed to develop and validate the nomogram. The Clinical-CT-PET model achieved the best diagnostic performance, with AUCs of 0.883 (training cohort) and 0.877 (test cohort), sensitivities of 79.5% and 80.0%, and specificities of 87.1% and 86.9%, respectively. For small LNs, it showed higher AUC (0.797 vs. 0.722, p < 0.001) and sensitivity (71.4% vs. 52%) compared to the PET model.</p><p><strong>Conclusion: </strong>We developed a nomogram that noninvasively estimates the risk of LNM in lung cancer that may inform individualized preoperative assessment and evidence-based decision-making.</p><p><strong>Key points: </strong>Question Can a nomogram integrating clinical, CT, and PET features improve preoperative prediction of lymph node metastasis in primary lung cancer, particularly in small nodes? Findings We developed a Clinical-CT-PET nomogram that achieved the best diagnostic accuracy (AUCs 0.883 and 0.877) among five models, especially for small lymph nodes (< 1 cm). Clinical relevance This noninvasive Clinical-CT-PET nomogram may improve the accuracy of preoperative lymph node staging and guide individualized treatment planning. It may also help avoid unnecessary invasive procedures in lung cancer patients, pending further multi-center validation.</p>","PeriodicalId":12076,"journal":{"name":"European Radiology","volume":" ","pages":"4110-4122"},"PeriodicalIF":4.7,"publicationDate":"2026-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13086717/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145767011","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}