{"title":"Validation of MRI short axis analysis for predicting lymphovascular invasion in endometrial cancer patients","authors":"Amelia Favier , Leo Razakamanantsoa , Julie Mereaux , Samia Lamrabet , Edwige Pottier , Claire Theodore , Cyril Touboul , Bassam Haddad , Yohann Dabi , Isabelle Thomassin-Naggara","doi":"10.1016/j.ejrad.2025.112259","DOIUrl":"10.1016/j.ejrad.2025.112259","url":null,"abstract":"<div><h3>Background</h3><div>In the context of FIGO classification updates in Endometrial Cancer (EC), lymphovascular space invasion (LVSI) is often either missing or wrongly assessed in preoperative histological analysis.</div></div><div><h3>Objective</h3><div>This retrospective study aimed to validate the diagnostic efficacy of systematic short-axis measurement on preoperative MRI for predicting lymphovascular space invasion (LVSI) in patients with EC.</div></div><div><h3>Materials</h3><div>A total of 116 patients who underwent preoperative pelvic MRI between January 2015 and December 2019 were included. Two expert radiologists specializing in female pelvic MRI measured the tumor’s short axis (previously described by Lavaud et al) on all sequences in sagittal axes T2-weighted and post-contrast T1-weighted images fat suppressed. MRI findings were compared with preoperative biopsy results and postoperative histopathology.</div></div><div><h3>Results</h3><div>The analysis revealed the highest discrepancies between preoperative histology combined with MRI images and final pathology in tumor grade (21.6 %), FIGO stage (39.6 %), and myometrial invasion (27.6 %). A 24 mm threshold for the anteroposterior measurement was used as a predictor of LVSI. The model utilizing this cutoff demonstrated good performance (AUC = 0.61, p < 0.001) and correctly reclassified 19.8 % of patients with preoperative FIGO stage I tumors as FIGO stage II or more after surgery.</div></div><div><h3>Conclusion</h3><div>This approach may enhance the preoperative prediction of LVSI and improve the application of the updated FIGO classification in endometrial cancer. The results suggest that MRI-derived short-axis measurement could be a valuable tool for refining the preoperative assessment of LVSI in EC patients.</div></div>","PeriodicalId":12063,"journal":{"name":"European Journal of Radiology","volume":"190 ","pages":"Article 112259"},"PeriodicalIF":3.2,"publicationDate":"2025-06-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144366069","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}
Jesús D. González , Pere Canals , Marc Rodrigo-Gisbert , Jordi Mayol , Alvaro García-Tornel , Marc Ribó
{"title":"Multimodal deep learning for predicting unsuccessful recanalization in refractory large vessel occlusion","authors":"Jesús D. González , Pere Canals , Marc Rodrigo-Gisbert , Jordi Mayol , Alvaro García-Tornel , Marc Ribó","doi":"10.1016/j.ejrad.2025.112254","DOIUrl":"10.1016/j.ejrad.2025.112254","url":null,"abstract":"<div><div><strong><em>Purpose:</em></strong> This study explores a multi-modal deep learning approach that integrates pre-intervention neuroimaging and clinical data to predict endovascular therapy (EVT) outcomes in acute ischemic stroke patients. To this end, consecutive stroke patients undergoing EVT were included in the study, including patients with suspected Intracranial Atherosclerosis-related Large Vessel Occlusion ICAD-LVO and other refractory occlusions. <strong><em>Methods:</em></strong> A retrospective, single-center cohort of patients with anterior circulation LVO who underwent EVT between 2017–2023 was analyzed. Refractory LVO (rLVO) defined class, comprised patients who presented any of the following: final angiographic stenosis > 50 %, unsuccessful recanalization (eTICI 0-2a) or required rescue treatments (angioplasty +/- stenting). Neuroimaging data included non-contrast CT and CTA volumes, automated vascular segmentation, and CT perfusion parameters. Clinical data included demographics, comorbidities and stroke severity. Imaging features were encoded using convolutional neural networks and fused with clinical data using a DAFT module. Data were split 80 % for training (with four-fold cross-validation) and 20 % for testing. Explainability methods were used to analyze the contribution of clinical variables and regions of interest in the images. <strong><em>Results:</em></strong> The final sample comprised 599 patients; 481 for training the model (77, 16.0 % rLVO), and 118 for testing (16, 13.6 % rLVO). The best model predicting rLVO using just imaging achieved an AUC of 0.53 ± 0.02 and F1 of 0.19 ± 0.05 while the proposed multimodal model achieved an AUC of 0.70 ± 0.02 and F1 of 0.39 ± 0.02 in testing. <strong><em>Conclusion:</em></strong> Combining vascular segmentation, clinical variables, and imaging data improved prediction performance over single-source models. This approach offers an early alert to procedural complexity, potentially guiding more tailored, timely intervention strategies in the EVT workflow.</div></div>","PeriodicalId":12063,"journal":{"name":"European Journal of Radiology","volume":"190 ","pages":"Article 112254"},"PeriodicalIF":3.2,"publicationDate":"2025-06-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144329580","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}
Teodoro Martín-Noguerol , Carolina Díaz-Angulo , Antonio Luna , Fermín Segovia , Manuel Gómez-Río , Juan M. Górriz
{"title":"Image-based AI tools in peripheral nerves assessment: Current status and integration strategies − A narrative review","authors":"Teodoro Martín-Noguerol , Carolina Díaz-Angulo , Antonio Luna , Fermín Segovia , Manuel Gómez-Río , Juan M. Górriz","doi":"10.1016/j.ejrad.2025.112255","DOIUrl":"10.1016/j.ejrad.2025.112255","url":null,"abstract":"<div><div>Peripheral Nerves (PNs) are traditionally evaluated using US or MRI, allowing radiologists to identify and classify them as normal or pathological based on imaging findings, symptoms, and electrophysiological tests. However, the anatomical complexity of PNs, coupled with their proximity to surrounding structures like vessels and muscles, presents significant challenges. Advanced imaging techniques, including MR-neurography and Diffusion-Weighted Imaging (DWI) neurography, have shown promise but are hindered by steep learning curves, operator dependency, and limited accessibility. Discrepancies between imaging findings and patient symptoms further complicate the evaluation of PNs, particularly in cases where imaging appears normal despite clinical indications of pathology. Additionally, demographic and clinical factors such as age, sex, comorbidities, and physical activity influence PN health but remain unquantifiable with current imaging methods.</div><div>Artificial Intelligence (AI) solutions have emerged as a transformative tool in PN evaluation. AI-based algorithms offer the potential to transition from qualitative to quantitative assessments, enabling precise segmentation, characterization, and threshold determination to distinguish healthy from pathological nerves. These advances could improve diagnostic accuracy and treatment monitoring.</div><div>This review highlights the latest advances in AI applications for PN imaging, discussing their potential to overcome the current limitations and opportunities to improve their integration into routine radiological practice.</div></div>","PeriodicalId":12063,"journal":{"name":"European Journal of Radiology","volume":"190 ","pages":"Article 112255"},"PeriodicalIF":3.2,"publicationDate":"2025-06-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144321967","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}
Lihua Chen , Xiaosong Lan , Yao Huang , Junli Tao , Xuemei Huang , Yangfan Su , Daihong Liu , Xiangming Fang , Jiuquan Zhang
{"title":"CT-based radiomics models for predicting spread through air space in lung cancer: A systematic review and meta-analysis","authors":"Lihua Chen , Xiaosong Lan , Yao Huang , Junli Tao , Xuemei Huang , Yangfan Su , Daihong Liu , Xiangming Fang , Jiuquan Zhang","doi":"10.1016/j.ejrad.2025.112249","DOIUrl":"10.1016/j.ejrad.2025.112249","url":null,"abstract":"<div><h3>Rationale and objectives</h3><div>Numerous studies have developed and validated models to predict spread through air space (STAS) in lung cancer using preoperative computed tomography (CT), yielding inconsistent results. We aimed to estimate the diagnostic accuracy of CT-based radiomics for predicting spread through air space (STAS) for preoperative prediction of lung cancer.</div></div><div><h3>Materials and methods</h3><div>Original studies published prior to January 2024 were searched in various databases. Only studies that used CT-based radiomics to preoperatively predict STAS in lung cancer patients were included. Two researchers independently extracted data and assessed the methodological quality of the included studies. We estimated summary sensitivity (SEN), specificity (SPE), and the areas under the receiver operating characteristic curve (AUC) of CT-based radiomics for predicting STAS. A head-to-head comparison was performed to evaluate the efficacy of clinical and radiomics models.</div></div><div><h3>Results</h3><div>A total of 17 studies with 6254 participants were included, and the methodological quality was found to be moderate. The <em>meta</em>-analysis comprised 26 datasets and achieved a pooled SEN of 0.82 (95 % CI: 0.78, 0.86), SPE of 0.76 (95 % CI: 0.72, 0.80), and AUC of 0.86 (95 % CI: 0.83, 0.89). In 11 pairwise comparison datasets, the radiomics model outperformed the clinical model with a higher AUC of 0.86 (95 % CI: 0.83, 0.89) compared to 0.80 (95 % CI: 0.76, 0.85), p < 0.001.</div></div><div><h3>Conclusions</h3><div>Due to its moderate diagnostic accuracy, widespread use, and low cost, CT-based radiomics can be used to predict STAS in lung cancer preoperatively. However, further research is required in a large, multicentre, and prospective design.</div></div>","PeriodicalId":12063,"journal":{"name":"European Journal of Radiology","volume":"190 ","pages":"Article 112249"},"PeriodicalIF":3.2,"publicationDate":"2025-06-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144321970","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}
Xiaolong Yang , Jiayang Wang , Ping Wang , Yingjie Li , Zhubin Wen , Jiming Shang , Kaige Chen , Chao Tang , Shuang Liang , Wei Meng
{"title":"Deep learning model using CT images for longitudinal prediction of benign and malignant ground-glass nodules","authors":"Xiaolong Yang , Jiayang Wang , Ping Wang , Yingjie Li , Zhubin Wen , Jiming Shang , Kaige Chen , Chao Tang , Shuang Liang , Wei Meng","doi":"10.1016/j.ejrad.2025.112252","DOIUrl":"10.1016/j.ejrad.2025.112252","url":null,"abstract":"<div><h3>Objectives</h3><div>To develop and validate a CT image-based multiple time-series deep learning model for the longitudinal prediction of benign and malignant pulmonary ground-glass nodules (GGNs).</div></div><div><h3>Methods</h3><div>A total of 486 GGNs from an equal number of patients were included in this research, which took place at two medical centers. Each nodule underwent surgical removal and was confirmed pathologically. The patients were randomly assigned to a training set, validation set, and test set, following a distribution ratio of 7:2:1. We established a transformer-based deep learning framework that leverages multi-temporal CT images for the longitudinal prediction of GGNs, focusing on distinguishing between benign and malignant types. Additionally, we utilized 13 different machine learning algorithms to formulate clinical models, delta-radiomics models, and combined models that merge deep learning with CT semantic features. The predictive capabilities of the models were assessed using the receiver operating characteristic (ROC) curve and the area under the curve (AUC).</div></div><div><h3>Results</h3><div>The multiple time-series deep learning model based on CT images surpassed both the clinical model and the delta-radiomics model, showcasing strong predictive capabilities for GGNs across the training, validation, and test sets, with AUCs of 0.911 (95% CI, 0.879–0.939), 0.809 (95% CI,0.715–0.908), and 0.817 (95% CI,0.680–0.937), respectively. Furthermore, the models that integrated deep learning with CT semantic features achieved the highest performance, resulting in AUCs of 0.960 (95% CI, 0.912–0.977), 0.878 (95% CI,0.801–0.942), and 0.890(95% CI, 0.790–0.968).</div></div><div><h3>Conclusion</h3><div>The multiple time-series deep learning model utilizing CT images was effective in predicting benign and malignant GGNs.</div></div>","PeriodicalId":12063,"journal":{"name":"European Journal of Radiology","volume":"190 ","pages":"Article 112252"},"PeriodicalIF":3.2,"publicationDate":"2025-06-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144335791","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":"The role of neuroimaging in brain death diagnosis: a review of radiological protocols and the need for standardization","authors":"Giulia Iacobellis , Alessia Leggio , Cecilia Salzillo , Miriam Solenne , Andrea Marzullo","doi":"10.1016/j.ejrad.2025.112247","DOIUrl":"10.1016/j.ejrad.2025.112247","url":null,"abstract":"<div><div>Brain death, defined as the irreversible cessation of all brain functions, including the brainstem, is a critical concept in modern medicine, particularly in the context of organ transplantation. The diagnosis of brain death relies primarily on a thorough clinical neurological examination, which assesses the absence of brainstem reflexes, coma, and apnoea. This paper underscores the critical role of neuroimaging-based ancillary tests in enhancing the accuracy of brain death determination and calls for harmonization of protocols to address existing disparities and improve clinical practice worldwide. The use of ancillary tests, particularly neuroimaging techniques plays a crucial role in confirming the diagnosis, especially in cases where clinical examination is inconclusive or confounded by factors such as drug intoxication or hypothermia. These tests provide objective evidence of the absence of cerebral blood, thereby supporting the clinical determination of brain death. Despite global consensus on the importance of the clinical neurological examination in diagnosing brain death, significant variations exist between countries regarding the use of ancillary tests, the number of required clinical examinations, observation periods, and the number of physicians involved in the determination process. The discrepancies among international guidelines, highlight the need for standardized to ensure consistency and reliability in brain death diagnosis, particularly in the context of organ transplantation, where timely and accurate diagnosis is essential. This literature review examines brain death in the context of organ donation, highlighting the differences in diagnostic protocols worldwide. It emphasises the importance of clinical and neuroimaging tests for accurate diagnosis and the need to standardise international guidelines to improve clinical practice and ensure timely and reliable decisions, particularly in organ transplantation.</div></div>","PeriodicalId":12063,"journal":{"name":"European Journal of Radiology","volume":"190 ","pages":"Article 112247"},"PeriodicalIF":3.2,"publicationDate":"2025-06-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144364643","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":"Dynamic contrast-enhanced CT-derived extracellular volume fraction for predicting postoperative oncologic outcomes in pancreatic ductal adenocarcinoma","authors":"Hideyuki Fukui , Yasunari Fukuda , Hiromitsu Onishi , Takashi Ota , Atsushi Nakamoto , Toru Honda , Ryo Aihara , Yukihiro Enchi , Daisaku Yamada , Shogo Kobayashi , Hidetoshi Eguchi , Mitsuaki Tatsumi , Noriyuki Tomiyama","doi":"10.1016/j.ejrad.2025.112246","DOIUrl":"10.1016/j.ejrad.2025.112246","url":null,"abstract":"<div><h3>Objective</h3><div>To evaluate the prognostic value of the extracellular volume fraction (fECV) derived from contrast-enhanced computed tomography (CE-CT) for recurrence-free survival (RFS) and overall survival (OS) rates after pancreatic ductal adenocarcinoma (PDAC) surgery.</div></div><div><h3>Methods</h3><div>This retrospective study evaluated 71 patients diagnosed with PDAC postsurgery who underwent CE-CT with precontrast and equilibrium phases before neoadjuvant chemotherapy (35 males, 36 females; mean age, 70.3 years; 95 % CI, 68.1–72.6; SD, 9.8; range, 45–89 years), were enrolled. Noncancerous pancreatic parenchyma and pancreatic tumors were automatically segmented from nonenhanced and equilibrium-phase images, excluding focal lesions, major-vessel, and ducts. Uni- and multivariate analyses (Cox proportional hazards model) were performed to evaluate fECV [=(100 − hematocrit) × (ΔPancreas/ΔAorta] in the nonaffected pancreas and tumor, with age, sex, chemotherapeutic scheme, tumor marker/location/size, stage, histological type, RFS, and OS as factors. Time-dependent receiver-operating characteristic curves showed the optimal fECV cutoff values for predicting RFS and OS.</div></div><div><h3>Results</h3><div>Adjuvant chemotherapy regimen, histological type, and fECV of noncancerous pancreatic parenchyma were independent prognostic factors of OS (<em>p</em> < 0.001, 0.049, and 0.018, respectively), and TNM stage (IB) was an independent predictor of RFS (<em>p</em> = 0.025). RFS and OS were worse in patients with noncancerous pancreatic tissue with higher fECV than in those with lower fECV (optimal cutoffs: 40.32 % for RFS, <em>p</em> = 0.036; 43.65 % for OS, <em>p</em> < 0.001).</div></div><div><h3>Conclusion</h3><div>The fECV of noncancerous pancreatic parenchyma from CE-CT was a significant predictor of survival outcomes in PDAC.</div></div>","PeriodicalId":12063,"journal":{"name":"European Journal of Radiology","volume":"190 ","pages":"Article 112246"},"PeriodicalIF":3.2,"publicationDate":"2025-06-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144321971","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}
Yassir Edrees Almalki , Mohammad Abd Alkhalik Basha , Mohamad Gamal Nada , Mona Mohammed Refaat , Majed Saeed Alshahrani , Sharifa Khalid Alduraibi , Ziyad A. Almushayti , Ahmed M. Abdelkhalik Basha , El-Shaimaa Mohamed Mohamed
{"title":"Sonographic signs of pediatric ovarian torsion: Multicenter study with surgical correlation","authors":"Yassir Edrees Almalki , Mohammad Abd Alkhalik Basha , Mohamad Gamal Nada , Mona Mohammed Refaat , Majed Saeed Alshahrani , Sharifa Khalid Alduraibi , Ziyad A. Almushayti , Ahmed M. Abdelkhalik Basha , El-Shaimaa Mohamed Mohamed","doi":"10.1016/j.ejrad.2025.112258","DOIUrl":"10.1016/j.ejrad.2025.112258","url":null,"abstract":"<div><h3>Objective</h3><div>Pediatric ovarian torsion is a rare but serious condition that requires prompt diagnosis to prevent ovarian damage. This study evaluated the predictive value and reliability of specific sonographic signs in diagnosing pediatric primary ovarian torsion and correlated these findings with surgical outcomes.</div></div><div><h3>Materials and methods</h3><div>This retrospective multicenter study analyzed ultrasound and surgical records from 103 girls aged 7–14 diagnosed with suspected ovarian torsion across three institutions from November 2021 to June 2024. Three senior radiologists independently reviewed ultrasound findings, and surgical records served as the reference standard. Diagnostic performance was calculated for each ultrasound sign, both individually and in combination. Univariate and multivariate logistic regression analyses were performed to assess the predictive value of these signs. The inter-reviewer agreement (IRA) among three radiologists was also evaluated using Fleiss kappa statistics.</div></div><div><h3>Results</h3><div>Ovarian torsion was confirmed in 87 out of 103 cases (84.5 %). The most predictive sonographic signs were ovarian enlargement (OR = 10.09, 95 %CI: 2.01–50.57, p < 0.005), abnormal ovarian position (OR = 8.65, 95 %CI: 1.89–39.52, p < 0.005), and abnormal blood flow (OR = 5.61, 95 %CI: 1.12–28.17, p = 0.036). Ovarian enlargement had the highest sensitivity (92 %), while follicular ring sign showed perfect specificity (100 %). The combination of abnormal ovarian position and enlargement yielded a high positive predictive value (98.6 %). The IRA was excellent for ovarian enlargement (κ = 0.90) and abnormal position (κ = 0.85).</div></div><div><h3>Conclusion</h3><div>Ovarian enlargement, abnormal ovarian position, and abnormal blood flow are reliable predictors of pediatric ovarian torsion on ultrasound. Combining these signs improves diagnostic accuracy and can guide timely surgical intervention.</div></div>","PeriodicalId":12063,"journal":{"name":"European Journal of Radiology","volume":"190 ","pages":"Article 112258"},"PeriodicalIF":3.2,"publicationDate":"2025-06-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144335792","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}
Johannes H. Reilly , Isabelle van Sloten , Ahmed A. Khalil , Uchralt Temuulen , Elias Kellner , Huma Fatima Ali , Jochen B. Fiebach , Ivana Galinovic
{"title":"Changes in microvessel density in normal-appearing white matter in relation to cerebral small vessel disease: A cohort study","authors":"Johannes H. Reilly , Isabelle van Sloten , Ahmed A. Khalil , Uchralt Temuulen , Elias Kellner , Huma Fatima Ali , Jochen B. Fiebach , Ivana Galinovic","doi":"10.1016/j.ejrad.2025.112250","DOIUrl":"10.1016/j.ejrad.2025.112250","url":null,"abstract":"<div><h3>Background and purpose</h3><div>White matter hyperintensities (WMH) of presumed vascular origin describe structural alterations of cerebral white matter, thought to result from cerebral small vessel disease. However, the in vivo effects of WMH on normal-appearing white matter microvasculature remain elusive. Therefore, we conducted an exploratory investigation of microvascular density in normal-appearing and pathological white matter in patients with WMH.</div></div><div><h3>Methods</h3><div>Using magnetic resonance imaging-based vessel size imaging we investigated an index of microvessel density in vivo in two clinical cohorts with ischaemic events (cohort_A, <em>N</em> = 88, mean age = 77.18) and intracranial neoplasms (cohort_B, <em>N</em> = 58, mean age = 65.45). For analysis, regions of interest were created for the whole normal-appearing white matter, the normal-appearing centrum semiovale, the WMH, the WMH penumbra, and the normal-appearing striatum. The severity of WMH-burden was quantified using the Wahlund score.</div></div><div><h3>Results</h3><div>In both cohorts, the index of microvessel density in the striatum was significantly higher than in normal-appearing white matter. There was no significant difference between WMH and WMH penumbra. However, both WMH and WMH penumbra had a significantly lower index of microvessel density than normal-appearing white matter in the subgroup of patients with high Wahlund scores. Lastly, the index of microvessel density in the normal-appearing centrum semiovale was higher in patients with high compared to low Wahlund scores in both cohorts. This comparison was not significant after adjusting for age.</div></div><div><h3>Conclusions</h3><div>Our results suggest a complex relationship between cerebral small vessel disease-related WMH and microvascular changes in the normal-appearing white matter, potentially indicative of WMH-related angiogenesis.</div></div>","PeriodicalId":12063,"journal":{"name":"European Journal of Radiology","volume":"190 ","pages":"Article 112250"},"PeriodicalIF":3.2,"publicationDate":"2025-06-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144313913","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}
Haijia Mao , Tianhao Feng , Sangying Lv , Dingbo Shu , Fandong Zhu , Jianfeng Yang , Zhenhua Zhao
{"title":"Robustness of radiomic features in photon-counting CT: Impact of radiation dose and virtual monoenergetic reconstructions compared to dual-energy CT","authors":"Haijia Mao , Tianhao Feng , Sangying Lv , Dingbo Shu , Fandong Zhu , Jianfeng Yang , Zhenhua Zhao","doi":"10.1016/j.ejrad.2025.112257","DOIUrl":"10.1016/j.ejrad.2025.112257","url":null,"abstract":"<div><h3>Objectives</h3><div>To evaluate the impact of Virtual Monoenergetic Image (VMI) reconstructions and radiation dose on radiomic feature reproducibility in photon-counting CT (PCCT) and compare its performance with dual-energy CT (DE-CT).</div></div><div><h3>Methods</h3><div>An anthropomorphic abdominal phantom (Kyoto Kagaku CTU-41) simulating liver, kidney, and vertebral tissues was scanned on a PCCT system at four dose levels (1, 3, 6, 12 mGy) with VMI reconstructions (40–100 keV). DE-CT acquisitions (80/140 kVp) at matched doses served as the comparator. Radiomic features were extracted from standardized ROIs using PyRadiomics. Reproducibility was quantified via intraclass correlation coefficient (ICC), concordance correlation coefficient (CCC), and coefficient of variation (CV < 10 %).</div></div><div><h3>Results</h3><div>Demonstrated exceptional reproducibility across all doses (ICC/CCC > 0.96) for liver, with peak stability at 6 mGy (ICC = 0.992, CCC = 0.998) and 44.75 % of features achieving CV ≤ 10 %. Kidneys exhibited inverse dose-reproducibility relationships, with optimal stability at 1 mGy (ICC = 0.783, CCC = 0.858). Vertebrae achieved superior reproducibility at 12 mGy (ICC = 0.631, CCC = 0.859), while 1 mGy and 3 mGy showed lower agreement (ICC < 0.60) due to partial volume effects.<!--> <!-->Liver radiomics showed superior reproducibility at 70–80 keV (ICC/CCC ≈1.00) and low variability (CV>20 %: 35.24–38.10 %). For Kidneys, high consistency was achieved at 70–80 keV (ICC>0.993, CCC>0.997) and persistent variability (CV>20 %: 52.38–68.57 %).</div></div><div><h3>Conclusion</h3><div>PCCT enables robust radiomics for homogeneous tissues (liver) across all doses, while heterogeneous regions (kidney, vertebrae) require energy- and dose-optimized protocols. The inverse dose-reproducibility relationship in renal radiomics highlights PCCT’s unique spectral advantages for low-dose imaging. These findings advocate for clinical adoption of PCCT with protocol standardization to unlock reliable, dose-efficient radiomic biomarkers.</div></div>","PeriodicalId":12063,"journal":{"name":"European Journal of Radiology","volume":"190 ","pages":"Article 112257"},"PeriodicalIF":3.2,"publicationDate":"2025-06-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144364646","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}