European Journal of Radiology最新文献

{"title":"Complementing interpretable machine learning with synergistic analytical strategies for thyroid cancer recurrence prediction","authors":"Souichi Oka ,&nbsp;Yoshiyasu Takefuji","doi":"10.1016/j.ejrad.2025.112308","DOIUrl":"10.1016/j.ejrad.2025.112308","url":null,"abstract":"<div><div>This correspondence critically examines the methodology of Schindele et al. (2025) on thyroid cancer recurrence prediction. While their interpretable XGBoost model achieved a high predictive accuracy of 95.8% and a 0.947 AUROC, it is crucial to recognize that this predictive power does not justify the reliability of its derived feature importance rankings. As widely acknowledged in the literature, high predictive accuracy does not guarantee unbiased or reliable feature attribution. We underscore that gradient boosting decision tree (GBDT) models, including XGBoost, are prone to inherent biases in feature importance estimation, often due to overfitting. Furthermore, SHapley Additive exPlanations (SHAP), a widely adopted explainable AI (XAI) technique, can inherit and even amplify these biases, given its model-dependent nature. This raises concerns about the interpretive validity of the identified risk factors. To mitigate these methodological limitations, we advocate for integrative analytical frameworks that combine machine learning with robust statistical and non-parametric approaches, such as Highly Variable Feature Selection (HVFS) and Independent Component Analysis (ICA). These multi-faceted strategies are indispensable for obtaining robust and interpretable insights into feature importance, warranting their prioritization in future research efforts.</div></div>","PeriodicalId":12063,"journal":{"name":"European Journal of Radiology","volume":"191 ","pages":"Article 112308"},"PeriodicalIF":3.2,"publicationDate":"2025-07-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144632235","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Impact on MRI technologist of a PI-QUAL v2 educational program on prostate MRI image quality assessment","authors":"Bruno Curia ,&nbsp;Chandra Bortolotto ,&nbsp;Leonardo Brizzi ,&nbsp;Simone Santagostini ,&nbsp;Luca Caffù ,&nbsp;Richard Naspro ,&nbsp;Francesco Alessandrino ,&nbsp;Lorenzo Preda","doi":"10.1016/j.ejrad.2025.112280","DOIUrl":"10.1016/j.ejrad.2025.112280","url":null,"abstract":"<div><h3>Purpose</h3><div>The Prostate Imaging Quality scoring system version 2 (PI-QUAL v2) is a new scoring system used to assess the diagnostic quality of prostate magnetic resonance imaging (pMRI). This study investigated the impact of a focused training module on the ability of MRI technologists (MRI-Tech) and MRI-Tech students to evaluate pMRI image quality.</div></div><div><h3>Methods</h3><div>Thirty-nine subjects including MRI-Tech students and experienced MRI-Tech with different levels of experience in pMRI participated in the study. The image quality of twenty pMRIs was evaluated before and after a dedicated lecture on pMRI image quality assessment using PI-QUAL v2 score. Receiver Operating Characteristic (ROC) curves were calculated for each scorer before and after the lecture, stratified by experience, and compared to the PI-QUAL v2 score assigned by a radiologist specialized in pMRI, using DeLong test.</div></div><div><h3>Results</h3><div>A significant improvement in AUC of pMRI image quality assessment was observed: from the baseline (0.31 ± 0.05) to the post-intervention (0.97 ± 0.01), with an improvement of 0.66 (p &lt; 0.001). The ROC curves stratified by experience, demonstrated an improvement of 0.44 [0.50 ± 0.07–0.94 ± 0.01] for II-year MRI-Tech students (p &lt; 0.001), of 0.48 [0.46 ± 0.09–0.94 ± 0.01] for III-year MRI-Tech students (p &lt; 0.032) and 0.60 [0.33 ± 0.10–0.93 ± 0.01] for board-certified MRI-Techs (p &lt; 0.001).</div></div><div><h3>Conclusions</h3><div>The PI-QUAL v2 training module improved the ability of both MRI-Tech students and board-certified MRI-Techs to assess the quality of pMRI images.</div></div>","PeriodicalId":12063,"journal":{"name":"European Journal of Radiology","volume":"191 ","pages":"Article 112280"},"PeriodicalIF":3.2,"publicationDate":"2025-07-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144632207","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Use of abbreviated breast MRI in benign lesions: how early is enough?","authors":"Deniz Esin Tekcan Sanli","doi":"10.1016/j.ejrad.2025.112314","DOIUrl":"10.1016/j.ejrad.2025.112314","url":null,"abstract":"","PeriodicalId":12063,"journal":{"name":"European Journal of Radiology","volume":"191 ","pages":"Article 112314"},"PeriodicalIF":3.2,"publicationDate":"2025-07-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144632276","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Deep learning reconstruction enhances bone visualization in zero echo time MRI for cervical spondylosis: A prospective study","authors":"Li Peng ,&nbsp;Xuelin Pan ,&nbsp;Huilou Liang ,&nbsp;Tongtong Li ,&nbsp;Siyan Zhou ,&nbsp;Wanjian Gao ,&nbsp;Xi Lu ,&nbsp;Xin Rong","doi":"10.1016/j.ejrad.2025.112310","DOIUrl":"10.1016/j.ejrad.2025.112310","url":null,"abstract":"<div><h3>Objective</h3><div>To evaluate the use of deep learning-based reconstruction (DLR) in zero echo time (ZTE) MRI to improve image quality and reduce scan time for assessing cervical spondylosis.</div></div><div><h3>Methods</h3><div>Forty-three preoperative patients with cervical spondylosis underwent ZTE MRI. ZTE data were acquired with two different numbers of excitations (NEX): NEX3 and NEX8. In addition to the conventional reconstruction, the NEX3 images were also reconstructed using DLR (NEX3-DL). Quantitative comparisons of signal-to-noise ratio (SNR) and contrast-to-noise ratio (CNR) across cortical bone, cancellous bone, spinal canal, and posterior muscles were analyzed via Wilcoxon signed-rank tests. Additionally, non-inferiority testing (NEX3-DL vs. NEX8) and superiority testing (NEX3-DL vs. NEX3) were performed. Two groups of musculoskeletal radiologists independently delineated regions of interest and evaluated the following parameters using validated assessment scales: cortical bone depiction, artifacts, perceived image noise, overall image quality. Inter- and intra-reader agreement were assessed with κ values and intraclass correlation coefficients (ICC).</div></div><div><h3>Results</h3><div>NEX3-DL images demonstrated significantly improved SNR and CNR compared to NEX3, matching the performance of NEX8 images with a 62 % shorter scan time. Subjective quality scores of both NEX3-DL and NEX8 were superior to NEX3 (p &lt; 0.01) with no significant differences between them. Inter- and intra-rater reliability for continuous variables and ordinal assessments showed substantial to excellent agreement (κ or ICC &gt; 0.6).</div></div><div><h3>Conclusion</h3><div>DLR-enhanced ZTE MRI improves bone visualization and can be integrated into routine MRI protocols, facilitating comprehensive evaluation of osseous and soft tissue structures within a single, time-efficient, radiation-free examination, and streamlining clinical workflow.</div></div>","PeriodicalId":12063,"journal":{"name":"European Journal of Radiology","volume":"191 ","pages":"Article 112310"},"PeriodicalIF":3.2,"publicationDate":"2025-07-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144632206","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Recognising errors in AI implementation in radiology: A narrative review","authors":"Nikolaos Stogiannos ,&nbsp;Renato Cuocolo ,&nbsp;Tugba Akinci D’Antonoli ,&nbsp;Daniel Pinto dos Santos ,&nbsp;Hugh Harvey ,&nbsp;Merel Huisman ,&nbsp;Burak Kocak ,&nbsp;Elmar Kotter ,&nbsp;Karim Lekadir ,&nbsp;Susan Cheng Shelmerdine ,&nbsp;Kicky G van Leeuwen ,&nbsp;Peter van Ooijen ,&nbsp;Michail E. Klontzas ,&nbsp;Christina Malamateniou","doi":"10.1016/j.ejrad.2025.112311","DOIUrl":"10.1016/j.ejrad.2025.112311","url":null,"abstract":"<div><div>The implementation of AI can suffer from a wide variety of failures. These failures can impact the performance of AI algorithms, impede the adoption of AI solutions in clinical practice, lead to workflow delays, or create unnecessary costs. This narrative review aims to comprehensively discuss different reasons for AI failures in Radiology through the analysis of published evidence across three main components of AI implementation: (i) the AI models throughout their lifecycle, (ii) the technical infrastructure, including the hardware and software needed to develop and deploy AI models and (iii) the human factors involved. Ultimately, based on the identified errors, this report aims to propose solutions to optimise the use and adoption of AI in radiology.</div></div>","PeriodicalId":12063,"journal":{"name":"European Journal of Radiology","volume":"191 ","pages":"Article 112311"},"PeriodicalIF":3.2,"publicationDate":"2025-07-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144633223","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Response to the Letter to the Editor: Diagnostic performance of abbreviated breast MRI in differentiating intraductal papilloma from ductal secretion","authors":"Gülbanu Güner,&nbsp;Sevde Nur Emir","doi":"10.1016/j.ejrad.2025.112312","DOIUrl":"10.1016/j.ejrad.2025.112312","url":null,"abstract":"","PeriodicalId":12063,"journal":{"name":"European Journal of Radiology","volume":"191 ","pages":"Article 112312"},"PeriodicalIF":3.2,"publicationDate":"2025-07-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144611592","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A novel segmentation-based deep learning model for enhanced scaphoid fracture detection","authors":"A. Bützow ,&nbsp;T.T. Anttila ,&nbsp;V. Haapamäki ,&nbsp;J. Ryhänen","doi":"10.1016/j.ejrad.2025.112309","DOIUrl":"10.1016/j.ejrad.2025.112309","url":null,"abstract":"<div><h3>Purpose</h3><div>To develop a deep learning model to detect apparent and occult scaphoid fractures from plain wrist radiographs and to compare the model’s diagnostic performance with that of a group of experts.</div></div><div><h3>Materials and methods</h3><div>A dataset comprising 408 patients, 410 wrists, and 1011 radiographs was collected. 718 of these radiographs contained a scaphoid fracture, verified by magnetic resonance imaging or computed tomography scans. 58 of these fractures were occult. The images were divided into training, test, and occult fracture test sets. The images were annotated by marking the scaphoid bone and the possible fracture area. The performance of the developed DL model was compared with the ground truth and the assessments of three clinical experts.</div></div><div><h3>Results</h3><div>The DL model achieved a sensitivity of 0.86 (95 % CI: 0.75–0.93) and a specificity of 0.83 (0.64–0.94). The model’s accuracy was 0.85 (0.76–0.92), and the area under the receiver operating characteristics curve was 0.92 (0.86–0.97). The clinical experts’ sensitivity ranged from 0.77 to 0.89, and specificity from 0.83 to 0.97. The DL model detected 24 of 58 (41 %) occult fractures, compared to 10.3 %, 13.7 %, and 6.8 % by the clinical experts.</div></div><div><h3>Conclusion</h3><div>Detecting scaphoid fractures using a segmentation-based DL model is feasible and comparable to previously developed DL models. The model performed similarly to a group of experts in identifying apparent scaphoid fractures and demonstrated higher diagnostic accuracy in detecting occult fractures. The improvement in occult fracture detection could enhance patient care.</div></div>","PeriodicalId":12063,"journal":{"name":"European Journal of Radiology","volume":"191 ","pages":"Article 112309"},"PeriodicalIF":3.2,"publicationDate":"2025-07-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144604412","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Long-term outcomes of carotid artery stenting for carotid near-occlusion","authors":"Cheng-Hsuan Tsai ,&nbsp;Ying-Hsien Chen ,&nbsp;Chun-Kai Chen ,&nbsp;Sheng-Fu Liu ,&nbsp;Tsung-Yu Ko ,&nbsp;Shih-Wei Meng ,&nbsp;Chih-Fan Yeh ,&nbsp;Ching-Chang Huang ,&nbsp;Chi-Sheng Hung ,&nbsp;Mao-Shin Lin ,&nbsp;Hsien-Li Kao","doi":"10.1016/j.ejrad.2025.112297","DOIUrl":"10.1016/j.ejrad.2025.112297","url":null,"abstract":"<div><h3>Objective</h3><div>Carotid artery near-occlusion (CANO) is an underdiagnosed condition, and the benefit of revascularizing CANO unproven. This study investigates the long-term outcomes of carotid artery stenting (CAS) in patients with CANO.</div></div><div><h3>Methods</h3><div>We conducted a retrospective study of patients who underwent CAS for carotid stenosis, including CANO and non-CANO groups. CANO was defined as post-stenotic narrowing of the distal internal carotid artery (ICA). The CANO group was further classified based on the presence or absence of full collapse defined as a distal ICA lumen diameter ≤ 2 mm and/or an ipsilateral-to-contralateral ICA ratio ≤ 0.42. The outcome measures included <em>peri</em>-procedural and long-term events, including stroke, mortality, and major adverse cerebrovascular events (MACE).</div></div><div><h3>Results</h3><div>123 patients with CANO and 173 patients with non-CANO carotid stenosis were retrospectively enrolled. Age and sex were comparable between groups. The CANO group had a higher proportion of patients with symptomatic lesions (36.6%) compared to the non-CANO group (25.4%), with the highest rate observed in CANO patients with full collapse (51.9%). Peri-procedural outcomes were similar between groups. There were no significant differences in long-term outcomes between the CANO and non-CANO groups, nor between CANO patients with or without full collapse. Bilateral significant ICA stenosis was a significant predictor of long-term MACE, whereas the presence of CANO or full collapse was not.</div></div><div><h3>Conclusion</h3><div>CAS is a viable option for patients with CANO, providing comparable long-term outcomes to those with conventional carotid stenosis. The presence of CANO with or without full collapse is not associated with worse outcomes.</div></div>","PeriodicalId":12063,"journal":{"name":"European Journal of Radiology","volume":"191 ","pages":"Article 112297"},"PeriodicalIF":3.2,"publicationDate":"2025-07-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144597446","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Improved accuracy of bone marrow edema detection in patients with hip pain using dual-energy computed tomography: Optimization of blended material density images combining water-calcium and water-hydroxyapatite","authors":"Kosuke Yoshii ,&nbsp;Takanori Masuda ,&nbsp;Shinichi Tanaka ,&nbsp;Yuki Sanda ,&nbsp;Shunsuke Takahara","doi":"10.1016/j.ejrad.2025.112293","DOIUrl":"10.1016/j.ejrad.2025.112293","url":null,"abstract":"<div><h3>Rationale and Objectives</h3><div>This study aimed to optimize newly developed blended material density (MD) images, compare them with conventional water-calcium (Ca) and water-hydroxyapatite (HAP) images, and identify the most suitable MD images for detecting bone marrow edema (BME).</div></div><div><h3>Materials and Methods</h3><div>We retrospectively analyzed the data of 24 patients (47 hips) who underwent dual-energy computed tomography (DECT) and magnetic resonance imaging (MRI) within one week of experiencing hip pain between April 2022 and March 2024. DECT was performed using a single-source fast-kV switching system and two material decomposition algorithms to generate 9 blended MD images by adjusting the mixing ratios of water-Ca and water-HAP (water-9Ca:1HAP to water-1Ca:9HAP). The optimal blended MD image was identified using the area under the curve (AUC) from the receiver operating characteristic (ROC) analysis. Radiologists and orthopedic surgeons visually assessed the concordance between the three types of MD images and MRI findings using a five-point scale.</div></div><div><h3>Results</h3><div>Among the 24 patients, 19 (79.2 %) were diagnosed with BME through MRI. Among the blended MD images, water-2Ca:8HAP demonstrated the highest diagnostic accuracy (AUC: 0.980), significantly outperforming water-Ca (AUC: 0.909, P &lt;0.01) and water-HAP (AUC: 0.898, P = 0.01). At a cut-off value of 969 mg/cm³, water-2Ca:8HAP achieved a sensitivity of 100.0 % and specificity of 91.8 %. Visual assessments further confirmed the highest concordance with MRI findings (both P &lt; 0.01).</div></div><div><h3>Conclusion</h3><div>The blended MD image using water-2Ca:8HAP demonstrated superior diagnostic performance compared to conventional images, offering a promising tool for improved BME detection.</div></div>","PeriodicalId":12063,"journal":{"name":"European Journal of Radiology","volume":"191 ","pages":"Article 112293"},"PeriodicalIF":3.2,"publicationDate":"2025-07-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144579632","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Response to comments regarding our study on BPE in CEM “No, background parenchymal enhancement (BPE) is a problem in contrast-enhanced mammography (CEM)”","authors":"Sonja Bechyna,&nbsp;Ambra Santonocito,&nbsp;Nina Pötsch,&nbsp;Paola Clauser,&nbsp;Thomas H Helbich,&nbsp;Pascal A.T. Baltzer","doi":"10.1016/j.ejrad.2025.112305","DOIUrl":"10.1016/j.ejrad.2025.112305","url":null,"abstract":"","PeriodicalId":12063,"journal":{"name":"European Journal of Radiology","volume":"191 ","pages":"Article 112305"},"PeriodicalIF":3.2,"publicationDate":"2025-07-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144596871","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}