Pub Date : 2025-09-20DOI: 10.1016/j.ejro.2025.100687
Ting Li , Nadeer M. Gharaibeh , Gang Wu
Purpose
To explore the feasibility of the You Only Look Once (YOLO) algorithm in the measurement of Carton index.
Methods
1156 knee X-ray images were collected from two centers (960 and 196). Five key points at patella and tibia on knee X-ray were labeled using the software of Labelme. YOLO11 pose models (including YOLO11n, YOLO11m and YOLO11x) were refined by labeled images from center A, and was then used to detect keypoints on images from center B. A line was the line between anterior edge of the tibial plateau and the lower pole of patellar articular surface, and B line was patellar articular surface. Carton index (A/B ratio) of 196 cases was obtained by senior radiologist, junior radiologist and YOLO respectively. The Bland Altman plot, Pearson Correlation test, Mean Absolute Error (MAE) and Intra-class correlation coefficient (ICC) were used to evaluate the agreement in measurement.
Results
Carton index of 196 images were automatically obtained with YOLO11n-pose, YOLO11m-pose and YOLO11x-pose. The ICC between senior and junior radiologists was 0.89. Pearson correlation coefficients were 0.23, 0.43 and 0.73 respectively for YOLO11n, YOLO11m and YOLO11x. ICC were 0.23, 0.42 and 0.72 respectively for YOLO11n, YOLO11m and YOLO11x. MAE were 0.20, 0.17 and 0.10 respectively for YOLO11n, YOLO11m and YOLO11x.
Conclusions
YOLO11x-pose model shows promise in the automatic measurement of Carton index on the knee X-ray image.
目的探讨“You Only Look Once （YOLO）”算法在纸箱指标测量中的可行性。方法从两个中心（960和196）采集膝关节x线片1156张。采用Labelme软件标记膝关节x线髌骨和胫骨的5个关键点。通过A中心的标记图像对YOLO11位姿模型（包括YOLO11n、YOLO11m和YOLO11x）进行精细处理，然后检测B中心图像上的关键点。A线为胫骨平台前缘与髌骨关节面下极之间的直线，B线为髌骨关节面。196例患者分别由高级放射科医师、初级放射科医师和YOLO获得卡尔顿指数（A/B比）。采用Bland Altman图、Pearson相关检验、平均绝对误差（MAE）和类内相关系数（ICC）评价测量一致性。结果采用YOLO11n-pose、YOLO11m-pose和YOLO11x-pose自动获取196幅图像的卡尔顿指数。高级和初级放射科医师之间的ICC为0.89。YOLO11n、YOLO11m和YOLO11x的Pearson相关系数分别为0.23、0.43和0.73。YOLO11n、YOLO11m和YOLO11x的ICC分别为0.23、0.42和0.72。YOLO11n、YOLO11m和YOLO11x的MAE分别为0.20、0.17和0.10。结论syolo11x -pose模型在膝关节x线图像的卡尔顿指数自动测量中具有较好的应用前景。
{"title":"Automatic measurement of Caton index on knee X-ray images using a key point detection model","authors":"Ting Li , Nadeer M. Gharaibeh , Gang Wu","doi":"10.1016/j.ejro.2025.100687","DOIUrl":"10.1016/j.ejro.2025.100687","url":null,"abstract":"<div><h3>Purpose</h3><div>To explore the feasibility of the You Only Look Once (YOLO) algorithm in the measurement of Carton index.</div></div><div><h3>Methods</h3><div>1156 knee X-ray images were collected from two centers (960 and 196). Five key points at patella and tibia on knee X-ray were labeled using the software of Labelme. YOLO11 pose models (including YOLO11n, YOLO11m and YOLO11x) were refined by labeled images from center A, and was then used to detect keypoints on images from center B. A line was the line between anterior edge of the tibial plateau and the lower pole of patellar articular surface, and B line was patellar articular surface. Carton index (A/B ratio) of 196 cases was obtained by senior radiologist, junior radiologist and YOLO respectively. The Bland Altman plot, Pearson Correlation test, Mean Absolute Error (MAE) and Intra-class correlation coefficient (ICC) were used to evaluate the agreement in measurement.</div></div><div><h3>Results</h3><div>Carton index of 196 images were automatically obtained with YOLO11n-pose, YOLO11m-pose and YOLO11x-pose. The ICC between senior and junior radiologists was 0.89. Pearson correlation coefficients were 0.23, 0.43 and 0.73 respectively for YOLO11n, YOLO11m and YOLO11x. ICC were 0.23, 0.42 and 0.72 respectively for YOLO11n, YOLO11m and YOLO11x. MAE were 0.20, 0.17 and 0.10 respectively for YOLO11n, YOLO11m and YOLO11x.</div></div><div><h3>Conclusions</h3><div>YOLO11x-pose model shows promise in the automatic measurement of Carton index on the knee X-ray image.</div></div>","PeriodicalId":38076,"journal":{"name":"European Journal of Radiology Open","volume":"15 ","pages":"Article 100687"},"PeriodicalIF":2.9,"publicationDate":"2025-09-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145095125","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
To evaluate the diagnostic value of cerebral perfusion and its predictive ability of hemorrhagic transformation (HT) in acute ischemic stroke (AIS) after intravenous thrombolysis (IVT) using CT perfusion (CTP).
Methods
Retrospective cohort of 55 AIS patients who underwent CTP before IVT was included. Clinical information, such as the National Institutes of Health Stroke Scale (NIHSS) score and history of atrial fibrillation (AF), were collected. CTP parameters, including cerebral blood flow (CBF), cerebral blood volume (CBV), mean transit time (MTT), permeability surface area product (PS), time-to-maximum (Tmax), time to peak (TTP), were measured in pathological hemispheres. Relative values (rCBV, rCBF, rMTT, rPS, rTmax, rTTP) were calculated as pathological-to-asymptomatic hemisphere ROI ratios. Comparisons between HT and non-HT groups were conducted using Student’s t-Test and Mann-Whitney U test. ROC curve and Logistic regression analysis were used to evaluate model predictive values. Delong's test compared AUC differences among parameters. Dynamic nomogram model was constructed with R-shiny and evaluated.
Results
NIHSS score at admission, NIHSS score before IVT, NIHSS score after IVT, NIHSS score at discharge, AF, PS and rPS were significantly higher than those in the non-HT group (p < 0.005). ROC curve and logistic regression analyses revealed that the combined model including NIHSS score before IVT, AF, and rPS displayed the highest AUC of 0.899 (95 % CI:0.814,0.984; p < 0.001).
Conclusion
Dynamic nomogram model combined NIHSS score before IVT, AF and rPS may act as a real-time visualization tool in the prediction of HT risk after IVT in patients with AIS.
{"title":"A quantitative CT perfusion-derived online dynamic nomogram for predicting hemorrhagic transformation after intravenous thrombolysis in acute ischemic stroke","authors":"Yanping Zheng , Peirong Jiang , Xiuzhu Xu , Liwei Xue , Jialin Chen , Yunjing Xue","doi":"10.1016/j.ejro.2025.100685","DOIUrl":"10.1016/j.ejro.2025.100685","url":null,"abstract":"<div><h3>Purpose</h3><div>To evaluate the diagnostic value of cerebral perfusion and its predictive ability of hemorrhagic transformation (HT) in acute ischemic stroke (AIS) after intravenous thrombolysis (IVT) using CT perfusion (CTP).</div></div><div><h3>Methods</h3><div>Retrospective cohort of 55 AIS patients who underwent CTP before IVT was included. Clinical information, such as the National Institutes of Health Stroke Scale (NIHSS) score and history of atrial fibrillation (AF), were collected. CTP parameters, including cerebral blood flow (CBF), cerebral blood volume (CBV), mean transit time (MTT), permeability surface area product (PS), time-to-maximum (Tmax), time to peak (TTP), were measured in pathological hemispheres. Relative values (rCBV, rCBF, rMTT, rPS, rTmax, rTTP) were calculated as pathological-to-asymptomatic hemisphere ROI ratios. Comparisons between HT and non-HT groups were conducted using Student’s t-Test and Mann-Whitney U test. ROC curve and Logistic regression analysis were used to evaluate model predictive values. Delong's test compared AUC differences among parameters. Dynamic nomogram model was constructed with R-shiny and evaluated.</div></div><div><h3>Results</h3><div>NIHSS score at admission, NIHSS score before IVT, NIHSS score after IVT, NIHSS score at discharge, AF, PS and rPS were significantly higher than those in the non-HT group (<em>p</em> < 0.005). ROC curve and logistic regression analyses revealed that the combined model including NIHSS score before IVT, AF, and rPS displayed the highest AUC of 0.899 (95 % CI:0.814,0.984; <em>p</em> < 0.001).</div></div><div><h3>Conclusion</h3><div>Dynamic nomogram model combined NIHSS score before IVT, AF and rPS may act as a real-time visualization tool in the prediction of HT risk after IVT in patients with AIS.</div></div>","PeriodicalId":38076,"journal":{"name":"European Journal of Radiology Open","volume":"15 ","pages":"Article 100685"},"PeriodicalIF":2.9,"publicationDate":"2025-09-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145095124","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-09-08DOI: 10.1016/j.ejro.2025.100684
Wen-jie Fan , Yu-ru Ma , Quan-meng Liu , Ning Zhang , Yi-yan Liu , Zi-qiang Wen , Bao-lan Lu , Jian-peng Yuan , Shen-ping Yu , Yan Chen
Objectives
To investigate the diagnostic value of dynamic contrast-enhanced MRI (DCE-MRI) quantitative parameters in acute radiation-induced rectal injury (RRI) among patients with rectal cancer.
Methods
This retrospective study included patients confirmed to rectal cancer who underwent rectal MRI (including a DCE-MRI sequence) and endoscopy after neoadjuvant chemoradiotherapy from November 2014 to July 2022. The enrolled patients were divided into an acute RRI group and a non-acute RRI group based on Vienna rectoscopy score. Two radiologists independently measured DCE-MRI quantitative parameters (including the forward volume transfer constant [Ktrans], rate constant [kep], and fractional extravascular extracellular space volume [ve]) and thickness of rectal wall. Receiver operating characteristic curve analysis was performed to analyze statistically significant parameters.
Results
Forty-nine patients (median age, 58 years; interquartile range, 14 years; 34 men) were enrolled, 28 of whom were in the acute RRI group. Ktrans in patients with acute RRI was significantly lower compared to those without acute RRI (0.049 min−1 vs 0.107 min−1; P < 0.001). The area under the receiver operating characteristic curve of Ktrans was 0.80. With a Ktrans cutoff value of 0.079 min−1, the sensitivity and specificity were 93 % and 67 %, respectively.
Conclusion
Ktrans demonstrated moderate performance in diagnosing acute RRI, providing a non-invasive and objective basis for managing and treating rectal cancer patients with acute RRI.
{"title":"Value of dynamic contrast-enhanced MRI in the diagnosis of acute radiation-induced rectal injury in patients with rectal cancer: A comparison with endoscopy","authors":"Wen-jie Fan , Yu-ru Ma , Quan-meng Liu , Ning Zhang , Yi-yan Liu , Zi-qiang Wen , Bao-lan Lu , Jian-peng Yuan , Shen-ping Yu , Yan Chen","doi":"10.1016/j.ejro.2025.100684","DOIUrl":"10.1016/j.ejro.2025.100684","url":null,"abstract":"<div><h3>Objectives</h3><div>To investigate the diagnostic value of dynamic contrast-enhanced MRI (DCE-MRI) quantitative parameters in acute radiation-induced rectal injury (RRI) among patients with rectal cancer.</div></div><div><h3>Methods</h3><div>This retrospective study included patients confirmed to rectal cancer who underwent rectal MRI (including a DCE-MRI sequence) and endoscopy after neoadjuvant chemoradiotherapy from November 2014 to July 2022. The enrolled patients were divided into an acute RRI group and a non-acute RRI group based on Vienna rectoscopy score. Two radiologists independently measured DCE-MRI quantitative parameters (including the forward volume transfer constant [<em>K</em><sup>trans</sup>], rate constant [<em>k</em><sub>ep</sub>], and fractional extravascular extracellular space volume [<em>v</em><sub>e</sub>]) and thickness of rectal wall. Receiver operating characteristic curve analysis was performed to analyze statistically significant parameters.</div></div><div><h3>Results</h3><div>Forty-nine patients (median age, 58 years; interquartile range, 14 years; 34 men) were enrolled, 28 of whom were in the acute RRI group. <em>K</em><sup>trans</sup> in patients with acute RRI was significantly lower compared to those without acute RRI (0.049 min<sup>−1</sup> vs 0.107 min<sup>−1</sup>; <em>P</em> < 0.001). The area under the receiver operating characteristic curve of <em>K</em><sup>trans</sup> was 0.80. With a <em>K</em><sup>trans</sup> cutoff value of 0.079 min<sup>−1</sup>, the sensitivity and specificity were 93 % and 67 %, respectively.</div></div><div><h3>Conclusion</h3><div><em>K</em><sup>trans</sup> demonstrated moderate performance in diagnosing acute RRI, providing a non-invasive and objective basis for managing and treating rectal cancer patients with acute RRI.</div></div>","PeriodicalId":38076,"journal":{"name":"European Journal of Radiology Open","volume":"15 ","pages":"Article 100684"},"PeriodicalIF":2.9,"publicationDate":"2025-09-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145018466","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-09-05DOI: 10.1016/j.ejro.2025.100682
Fengfeng Yang , Zhengyang Li , Haoran Cai, Jing Zhu, Huijia Liu, Yang Zhao
Objectives
This study aimed to determine the efficacy of fat attenuation index (FAI) as a non-invasive diagnostic tool in the precise identification of culprit lesions in individuals diagnosed with acute coronary syndrome (ACS).
Methods
A retrospective analysis of 230 patients with non-ST-segment elevation ACS. PCAT attenuation (FAIstandard) was measured in the proximal 40-mm segment of each major coronary artery. Furthermore, the average PCAT attenuation of the identified lesions was designated as FAIlesion. The average PCAT attenuation across the complete length of coronary artery, referred to as FAIaverage, was computed. Plaque characteristics (volume, composition) were analyzed via coronary computed tomography angiography. Multivariable logistic regression identified predictors of culprit lesions, and diagnostic performance was assessed using area under the curve (AUC) and decision curve analysis.
Results
Culprit lesions exhibited significantly elevated levels of PCAT attenuation across the parameters of FAIstandard, FAIaverage, and FAIlesion. FAIlesion demonstrated superior diagnostic accuracy versus FAIstandard and FAIaverage, and also emerged as the strongest independent predictor (Odds ratio = 2.598, P < 0.001). In training and test sets, a composite model integrating FAIlesion with additional indices demonstrated enhanced diagnostic efficacy for the detection of culprit lesions in patients with ACS (AUC = 0.960, 0.803). Low-attenuation plaque volume (<30 HU) was independently associated with culprit lesions (OR = 3.12, P = 0.002).
Conclusion
FAIlesion, a superior non-invasive biomarker for high-risk ACS lesions compared to traditional FAI, enables earlier precise risk stratification through clinical integration.
{"title":"Non-invasive diagnostic value of pericoronary fat attenuation index for identifying culprit lesions in acute coronary syndrome","authors":"Fengfeng Yang , Zhengyang Li , Haoran Cai, Jing Zhu, Huijia Liu, Yang Zhao","doi":"10.1016/j.ejro.2025.100682","DOIUrl":"10.1016/j.ejro.2025.100682","url":null,"abstract":"<div><h3>Objectives</h3><div>This study aimed to determine the efficacy of fat attenuation index (FAI) as a non-invasive diagnostic tool in the precise identification of culprit lesions in individuals diagnosed with acute coronary syndrome (ACS).</div></div><div><h3>Methods</h3><div>A retrospective analysis of 230 patients with non-ST-segment elevation ACS. PCAT attenuation (FAI<sub>standard</sub>) was measured in the proximal 40-mm segment of each major coronary artery. Furthermore, the average PCAT attenuation of the identified lesions was designated as FAI<sub>lesion</sub>. The average PCAT attenuation across the complete length of coronary artery, referred to as FAI<sub>average</sub>, was computed. Plaque characteristics (volume, composition) were analyzed via coronary computed tomography angiography. Multivariable logistic regression identified predictors of culprit lesions, and diagnostic performance was assessed using area under the curve (AUC) and decision curve analysis.</div></div><div><h3>Results</h3><div>Culprit lesions exhibited significantly elevated levels of PCAT attenuation across the parameters of FAI<sub>standard</sub>, FAI<sub>average</sub>, and FAI<sub>lesion</sub>. FAI<sub>lesion</sub> demonstrated superior diagnostic accuracy versus FAI<sub>standard</sub> and FAI<sub>average</sub>, and also emerged as the strongest independent predictor (Odds ratio = 2.598, P < 0.001). In training and test sets, a composite model integrating FAI<sub>lesion</sub> with additional indices demonstrated enhanced diagnostic efficacy for the detection of culprit lesions in patients with ACS (AUC = 0.960, 0.803). Low-attenuation plaque volume (<30 HU) was independently associated with culprit lesions (OR = 3.12, P = 0.002).</div></div><div><h3>Conclusion</h3><div>FAI<sub>lesion</sub>, a superior non-invasive biomarker for high-risk ACS lesions compared to traditional FAI, enables earlier precise risk stratification through clinical integration.</div></div>","PeriodicalId":38076,"journal":{"name":"European Journal of Radiology Open","volume":"15 ","pages":"Article 100682"},"PeriodicalIF":2.9,"publicationDate":"2025-09-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144996333","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-09-04DOI: 10.1016/j.ejro.2025.100683
Yu Li , Fang Yang , Xuebin Liu , Jiping Luo , Siyu Dan , Xiuli He , Guihao Hu , Ling He , Xiachuan Qin , Tao Wu , Wensheng Yue
Background
Early detection of prostate cancer (PCa) remains challenging, as prostate-specific antigen (PSA) testing and digital rectal examination (DRE) offer limited specificity. Transrectal ultrasound (TRUS) is routinely used for biopsy guidance, but its diagnostic potential for PCa screening is underexplored. We aimed to evaluate TRUS-derived morphological features and develop a nomogram that integrates clinical and TRUS characteristics to improve PCa risk stratification.
Methods
Consecutive patients with suspected PCa were enrolled from two tertiary centers (training cohort: n = 154, October 2021–January 2023; validation cohort: n = 51, December 2021–June 2022). Demographic data, laboratory-derived PSA indices (including PSA density), and TRUS parameters (independently assessed by two blinded sonographers) were collected and analyzed. A predictive nomogram was constructed using multivariate logistic regression and externally validated.
Results
In the training cohort (mean age 70.9 ± 8.0 years; 72 PCa, 82 benign), independent predictors of PCa included elevated PSA density (OR=3.86, 95 % CI: 1.30–11.40, P = 0.015), abnormal DRE (OR=3.06, 95 % CI: 1.09–8.60, P = 0.034), TRUS-defined ill-defined zone boundaries (OR=9.61, 95 % CI: 3.37–39.02, P = 0.002), and hyper-enhancement (OR=7.07, 95 % CI: 2.69–21.89, P < 0.001). The nomogram achieved strong discrimination (training C-index=0.933, 95 % CI: 0.881–0.986; validation C-index=0.907, 95 % CI: 0.792–0.970) with 84.7 % sensitivity, 87.8 % specificity, and 86.4 % accuracy. Pathological concordance was high (kappa=0.726).
Conclusion
TRUS-derived features (ill-defined zones, hyper-enhancement) significantly enhance PCa detection when combined with clinical parameters. Our nomogram provides a practical, visual tool to guide biopsy decisions and demonstrates robust performance across cohorts.
{"title":"Combined predictive model for prostate cancer screening: Development and validation study","authors":"Yu Li , Fang Yang , Xuebin Liu , Jiping Luo , Siyu Dan , Xiuli He , Guihao Hu , Ling He , Xiachuan Qin , Tao Wu , Wensheng Yue","doi":"10.1016/j.ejro.2025.100683","DOIUrl":"10.1016/j.ejro.2025.100683","url":null,"abstract":"<div><h3>Background</h3><div>Early detection of prostate cancer (PCa) remains challenging, as prostate-specific antigen (PSA) testing and digital rectal examination (DRE) offer limited specificity. Transrectal ultrasound (TRUS) is routinely used for biopsy guidance, but its diagnostic potential for PCa screening is underexplored. We aimed to evaluate TRUS-derived morphological features and develop a nomogram that integrates clinical and TRUS characteristics to improve PCa risk stratification.</div></div><div><h3>Methods</h3><div>Consecutive patients with suspected PCa were enrolled from two tertiary centers (training cohort: n = 154, October 2021–January 2023; validation cohort: n = 51, December 2021–June 2022). Demographic data, laboratory-derived PSA indices (including PSA density), and TRUS parameters (independently assessed by two blinded sonographers) were collected and analyzed. A predictive nomogram was constructed using multivariate logistic regression and externally validated.</div></div><div><h3>Results</h3><div>In the training cohort (mean age 70.9 ± 8.0 years; 72 PCa, 82 benign), independent predictors of PCa included elevated PSA density (OR=3.86, 95 % CI: 1.30–11.40, <em>P</em> = 0.015), abnormal DRE (OR=3.06, 95 % CI: 1.09–8.60, <em>P</em> = 0.034), TRUS-defined ill-defined zone boundaries (OR=9.61, 95 % CI: 3.37–39.02, <em>P</em> = 0.002), and hyper-enhancement (OR=7.07, 95 % CI: 2.69–21.89, <em>P</em> < 0.001). The nomogram achieved strong discrimination (training C-index=0.933, 95 % CI: 0.881–0.986; validation C-index=0.907, 95 % CI: 0.792–0.970) with 84.7 % sensitivity, 87.8 % specificity, and 86.4 % accuracy. Pathological concordance was high (kappa=0.726).</div></div><div><h3>Conclusion</h3><div>TRUS-derived features (ill-defined zones, hyper-enhancement) significantly enhance PCa detection when combined with clinical parameters. Our nomogram provides a practical, visual tool to guide biopsy decisions and demonstrates robust performance across cohorts.</div></div>","PeriodicalId":38076,"journal":{"name":"European Journal of Radiology Open","volume":"15 ","pages":"Article 100683"},"PeriodicalIF":2.9,"publicationDate":"2025-09-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144996474","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-09-01DOI: 10.1016/j.ejro.2025.100681
Maki Amano , Jun Ozeki , Yumi Koyama , Xiaoyan Tang , Fumi Nozaki , Mayumi Tani , Yasuo Amano
Purpose
To evaluate the utility of a magnetic resonance imaging (MRI) projection mapping system (PMS) for determining the resection lines during breast-conserving surgery (BCS) in patients with breast cancer presenting with nonmass enhancement (NME) and identify the clinical or MRI variables associated with close or positive margins.
Materials and methods
Forty-one patients with breast cancer exhibiting NME were enrolled. In the operating room, a maximum intensity projection image generated from supine MRI was projected onto the breast using a PMS, which employed a structured light method to measure the surface of the breast. Cancer contours delineated on the MRI-PMS, with an additional safety margin, served as the resection lines for cylindrical BCS. Margins were pathologically categorized as negative (> 2 mm), close (≤ 2 mm), or positive. The association between margin status and clinical or MRI variables was analyzed.
Results
Surgical margins were negative in 24 patients (58.5 %), close in 15 (36.6 %), and positive in 2 (4.9 %). There were significant differences in the maximum diameter of nonmass components (NMCs) shown by pathology, that of NME on MRI, and the discrepancy between the two diameters between patients with negative margin and those with close or positive margin (< 0.05 for all). Receiver operating characteristics revealed that threshold of 40 mm for NMEs provided high specificity of 91.7 %.
Conclusion
The MRI-PMS led to a low rate of positive margins during BCS in patients with breast cancer with NMEs. Large NMCs and NMEs are associated with positive or close margin.
{"title":"Clinical and MRI variables associated with close or positive margins during breast-conserving surgery using MRI projection mapping in breast carcinoma with nonmass enhancement","authors":"Maki Amano , Jun Ozeki , Yumi Koyama , Xiaoyan Tang , Fumi Nozaki , Mayumi Tani , Yasuo Amano","doi":"10.1016/j.ejro.2025.100681","DOIUrl":"10.1016/j.ejro.2025.100681","url":null,"abstract":"<div><h3>Purpose</h3><div>To evaluate the utility of a magnetic resonance imaging (MRI) projection mapping system (PMS) for determining the resection lines during breast-conserving surgery (BCS) in patients with breast cancer presenting with nonmass enhancement (NME) and identify the clinical or MRI variables associated with close or positive margins.</div></div><div><h3>Materials and methods</h3><div>Forty-one patients with breast cancer exhibiting NME were enrolled. In the operating room, a maximum intensity projection image generated from supine MRI was projected onto the breast using a PMS, which employed a structured light method to measure the surface of the breast. Cancer contours delineated on the MRI-PMS, with an additional safety margin, served as the resection lines for cylindrical BCS. Margins were pathologically categorized as negative (> 2 mm), close (≤ 2 mm), or positive. The association between margin status and clinical or MRI variables was analyzed.</div></div><div><h3>Results</h3><div>Surgical margins were negative in 24 patients (58.5 %), close in 15 (36.6 %), and positive in 2 (4.9 %). There were significant differences in the maximum diameter of nonmass components (NMCs) shown by pathology, that of NME on MRI, and the discrepancy between the two diameters between patients with negative margin and those with close or positive margin (< 0.05 for all). Receiver operating characteristics revealed that threshold of 40 mm for NMEs provided high specificity of 91.7 %.</div></div><div><h3>Conclusion</h3><div>The MRI-PMS led to a low rate of positive margins during BCS in patients with breast cancer with NMEs. Large NMCs and NMEs are associated with positive or close margin.</div></div>","PeriodicalId":38076,"journal":{"name":"European Journal of Radiology Open","volume":"15 ","pages":"Article 100681"},"PeriodicalIF":2.9,"publicationDate":"2025-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144922064","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-08-29DOI: 10.1016/j.ejro.2025.100678
Thibault Agripnidis , Angela Ayobi , Sarah Quenet , Yasmina Chaibi , Christophe Avare , Alexis Jacquier , Nadine Girard , Jean-François Hak , Anthony Reyre , Gilles Brun , Ahmed-Ali El Ahmadi
Objective
Several artificial intelligence (AI) tools have been developed to assist in the stroke imaging workflow, which remains a major disease of the 21st century. This study evaluated the combined performance of an FDA-cleared and CE-marked AI-based device with three modules designed to detect intracerebral hemorrhage (ICH), identify large vessel occlusion (LVO), and calculate Alberta Stroke Program Early CT Scores (ASPECTS).
Materials & methods
Non-contrast CT (NCCT) and/or computed tomography angiography (CTA) for suspicion of stroke acquired at La Timone and Nord University hospitals (Marseille, France) between March 2019 and March 2020 were retrospectively collected. The AI tool, CINA-HEAD (Avicenna.AI), processed the data to flag ICH, LVO, and calculate ASPECTS. The results were compared to ground truth evaluations by four expert neuroradiologists to compute diagnostic performances.
Results
A total of 373 NCCT and 331 CTA from 405 patients (mean age 64.9 ± 18.9 SD, 52.6 % female) were included. The AI tool achieved an accuracy of 94.6 % [95 % CI: 91.8 %-96.7 %] for ICH detection on NCCT and of 86.4 % [95 % CI: 82.2 %-89.9 %] for LVO identification on CTA. The region-based ASPECTS analysis yielded an accuracy of 88.6 % [95 % CI: 87.8 %-89.3 %] and the dichotomized ASPECTS classification (ASPECTS ≥ 6) achieved 80.4 % accuracy.
Conclusion
This study demonstrates the reliable, stepwise performance of an AI-based stroke imaging tool across the diagnostic cascade of ICH and LVO detection and ASPECTS scoring. Such robust multi-stage evaluation supports its potential for streamlining acute stroke triage and decision-making.
{"title":"Performance of an artificial intelligence tool for multi-step acute stroke imaging: A multicenter diagnostic study","authors":"Thibault Agripnidis , Angela Ayobi , Sarah Quenet , Yasmina Chaibi , Christophe Avare , Alexis Jacquier , Nadine Girard , Jean-François Hak , Anthony Reyre , Gilles Brun , Ahmed-Ali El Ahmadi","doi":"10.1016/j.ejro.2025.100678","DOIUrl":"10.1016/j.ejro.2025.100678","url":null,"abstract":"<div><h3>Objective</h3><div>Several artificial intelligence (AI) tools have been developed to assist in the stroke imaging workflow, which remains a major disease of the 21st century. This study evaluated the combined performance of an FDA-cleared and CE-marked AI-based device with three modules designed to detect intracerebral hemorrhage (ICH), identify large vessel occlusion (LVO), and calculate Alberta Stroke Program Early CT Scores (ASPECTS).</div></div><div><h3>Materials & methods</h3><div>Non-contrast CT (NCCT) and/or computed tomography angiography (CTA) for suspicion of stroke acquired at La Timone and Nord University hospitals (Marseille, France) between March 2019 and March 2020 were retrospectively collected. The AI tool, CINA-HEAD (Avicenna.AI), processed the data to flag ICH, LVO, and calculate ASPECTS. The results were compared to ground truth evaluations by four expert neuroradiologists to compute diagnostic performances.</div></div><div><h3>Results</h3><div>A total of 373 NCCT and 331 CTA from 405 patients (mean age 64.9 ± 18.9 SD, 52.6 % female) were included. The AI tool achieved an accuracy of 94.6 % [95 % CI: 91.8 %-96.7 %] for ICH detection on NCCT and of 86.4 % [95 % CI: 82.2 %-89.9 %] for LVO identification on CTA. The region-based ASPECTS analysis yielded an accuracy of 88.6 % [95 % CI: 87.8 %-89.3 %] and the dichotomized ASPECTS classification (ASPECTS ≥ 6) achieved 80.4 % accuracy.</div></div><div><h3>Conclusion</h3><div>This study demonstrates the reliable, stepwise performance of an AI-based stroke imaging tool across the diagnostic cascade of ICH and LVO detection and ASPECTS scoring. Such robust multi-stage evaluation supports its potential for streamlining acute stroke triage and decision-making.</div></div>","PeriodicalId":38076,"journal":{"name":"European Journal of Radiology Open","volume":"15 ","pages":"Article 100678"},"PeriodicalIF":2.9,"publicationDate":"2025-08-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144917052","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-08-23DOI: 10.1016/j.ejro.2025.100680
Rui Chen , Hu Zhang , Xingwen Huang , Haitao Han , Jinbo Jian
Objective
To develop and validate a machine learning model based on CT radiomics to improve the ability to differentiate pathological subtypes of pulmonary ground-glass nodules (GGN).
Methods
A retrospective analysis was conducted on clinical data and radiological images from 392 patients with lung adenocarcinoma at Binzhou Medical University Hospital between January 1, 2020 to May 31, 2023. All patients underwent preoperative thin-section chest CT scans and surgical resection. A total of 400 GGNs were included. Regions of interest (ROI) were delineated on the slice showing the largest diameter of the lesions. Based on pathological confirmation, the nodules were divided into two groups: Group 1 (adenocarcinoma in situ, AIS or minimally invasive adenocarcinoma, MIA, 209 nodules) and Group 2 (invasive adenocarcinoma, IAC, 191nodules). The dataset was randomly split into a training set (280 nodules, 70 %) and a validation set (120 nodules, 30 %) at a 7:3 ratio. In the training set, feature dimensionality reduction was performed using minimum redundancy maximum relevance (mRMR) as well as least absolute shrinkage and selection operator (LASSO) to screen out discriminative radiomics features. Then seven machine learning models—logistic regression (LR), support vector machine (SVM), random forest (RF), extra trees, XGBoost, GradientBoosting, and AdaBoost—were constructed. Model performance and prediction efficacy were evaluated based on indicators such as area under the curve (AUC), accuracy, specificity, and sensitivity using receiver operating characteristic (ROC) curves.
Results
Eight radiomics features were ultimately identified. Among the seven models, the GradientBoosting model exhibited the best performance, achieving an AUC of 0.929 (95 % CI: 0.9004–0.9584), accuracy of 0.85, sensitivity of 0.851, and specificity of 0.849 in the training set.
Conclusion
The GradientBoosting model based on CT radiomics features demonstrates superior performance in predicting pathological subtypes of ground glass nodular lung adenocarcinoma, providing a reliable auxiliary tool for clinical diagnosis.
{"title":"CT Radiomics-based machine learning approach for the invasiveness of pulmonary ground-glass nodules prediction","authors":"Rui Chen , Hu Zhang , Xingwen Huang , Haitao Han , Jinbo Jian","doi":"10.1016/j.ejro.2025.100680","DOIUrl":"10.1016/j.ejro.2025.100680","url":null,"abstract":"<div><h3>Objective</h3><div>To develop and validate a machine learning model based on CT radiomics to improve the ability to differentiate pathological subtypes of pulmonary ground-glass nodules (GGN).</div></div><div><h3>Methods</h3><div>A retrospective analysis was conducted on clinical data and radiological images from 392 patients with lung adenocarcinoma at Binzhou Medical University Hospital between January 1, 2020 to May 31, 2023. All patients underwent preoperative thin-section chest CT scans and surgical resection. A total of 400 GGNs were included. Regions of interest (ROI) were delineated on the slice showing the largest diameter of the lesions. Based on pathological confirmation, the nodules were divided into two groups: Group 1 (adenocarcinoma in situ, AIS or minimally invasive adenocarcinoma, MIA, 209 nodules) and Group 2 (invasive adenocarcinoma, IAC, 191nodules). The dataset was randomly split into a training set (280 nodules, 70 %) and a validation set (120 nodules, 30 %) at a 7:3 ratio. In the training set, feature dimensionality reduction was performed using minimum redundancy maximum relevance (mRMR) as well as least absolute shrinkage and selection operator (LASSO) to screen out discriminative radiomics features. Then seven machine learning models—logistic regression (LR), support vector machine (SVM), random forest (RF), extra trees, XGBoost, GradientBoosting, and AdaBoost—were constructed. Model performance and prediction efficacy were evaluated based on indicators such as area under the curve (AUC), accuracy, specificity, and sensitivity using receiver operating characteristic (ROC) curves.</div></div><div><h3>Results</h3><div>Eight radiomics features were ultimately identified. Among the seven models, the GradientBoosting model exhibited the best performance, achieving an AUC of 0.929 (95 % CI: 0.9004–0.9584), accuracy of 0.85, sensitivity of 0.851, and specificity of 0.849 in the training set.</div></div><div><h3>Conclusion</h3><div>The GradientBoosting model based on CT radiomics features demonstrates superior performance in predicting pathological subtypes of ground glass nodular lung adenocarcinoma, providing a reliable auxiliary tool for clinical diagnosis.</div></div>","PeriodicalId":38076,"journal":{"name":"European Journal of Radiology Open","volume":"15 ","pages":"Article 100680"},"PeriodicalIF":2.9,"publicationDate":"2025-08-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144889510","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-08-20DOI: 10.1016/j.ejro.2025.100679
Yuchao Xiong , Wei Guo , Xuwen Zeng , Fan Xu , Li Wu , Jiahui Ou
Background
This study aimed to compare the diagnostic performance of radiomic features derived from dual-layer spectral detector computed tomography (DLSCT) and a deep learning (DL) model applied to conventional CT images in the differentiation of osteoblastic bone metastases (OBM) from bone islands (BI).
Methods
This retrospective study included patients with osteogenic lesions who underwent DLSCT examinations between March 2023 and September 2023. We extracted first-order radiomic features (e.g., mean, maximum, entropy) from both conventional and spectral images. A previously validated DL model was applied to the conventional CT images. We evaluated diagnostic performance using ROC curve analysis, comparing AUC, sensitivity, and specificity.
Results
The study included 216 lesions from 94 patients (66 ± 12 years; 48 males, 46 females): 125 BI and 91 OBM lesions. Significant differences were observed between OBM and BI groups for the mean, maximum, entropy, and uniformity of first-order radiomic features (all P < 0.05). DLSCT (entropy from VMI40keV) and the DL model had comparable AUCs (0.93 vs. 0.96; P = 0.274). However, DLSCT showed superior sensitivity (92 % vs. 62 %; P = 0.002) but comparable specificity (88 % vs. 96 %; P = 0.07) for diagnosing OBM compared to the DL model.
Conclusion
Radiomic features from DLSCT differentiate between BI and OBM with diagnostic performance comparable to that of a DL model. Furthermore, VMI40keV image-derived entropy demonstrated superior sensitivity in diagnosing OBM compared to the DL model.
{"title":"Diagnostic performance of dual-layer spectral CT Radiomics and deep learning for differentiating osteoblastic bone metastases from bone islands","authors":"Yuchao Xiong , Wei Guo , Xuwen Zeng , Fan Xu , Li Wu , Jiahui Ou","doi":"10.1016/j.ejro.2025.100679","DOIUrl":"10.1016/j.ejro.2025.100679","url":null,"abstract":"<div><h3>Background</h3><div>This study aimed to compare the diagnostic performance of radiomic features derived from dual-layer spectral detector computed tomography (DLSCT) and a deep learning (DL) model applied to conventional CT images in the differentiation of osteoblastic bone metastases (OBM) from bone islands (BI).</div></div><div><h3>Methods</h3><div>This retrospective study included patients with osteogenic lesions who underwent DLSCT examinations between March 2023 and September 2023. We extracted first-order radiomic features (e.g., mean, maximum, entropy) from both conventional and spectral images. A previously validated DL model was applied to the conventional CT images. We evaluated diagnostic performance using ROC curve analysis, comparing AUC, sensitivity, and specificity.</div></div><div><h3>Results</h3><div>The study included 216 lesions from 94 patients (66 ± 12 years; 48 males, 46 females): 125 BI and 91 OBM lesions. Significant differences were observed between OBM and BI groups for the mean, maximum, entropy, and uniformity of first-order radiomic features (all P < 0.05). DLSCT (entropy from VMI40keV) and the DL model had comparable AUCs (0.93 vs. 0.96; P = 0.274). However, DLSCT showed superior sensitivity (92 % vs. 62 %; P = 0.002) but comparable specificity (88 % vs. 96 %; P = 0.07) for diagnosing OBM compared to the DL model.</div></div><div><h3>Conclusion</h3><div>Radiomic features from DLSCT differentiate between BI and OBM with diagnostic performance comparable to that of a DL model. Furthermore, VMI40keV image-derived entropy demonstrated superior sensitivity in diagnosing OBM compared to the DL model.</div></div>","PeriodicalId":38076,"journal":{"name":"European Journal of Radiology Open","volume":"15 ","pages":"Article 100679"},"PeriodicalIF":2.9,"publicationDate":"2025-08-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144879230","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
To investigate area under diaphragm (AUD) obtained by dynamic digital radiography (DDR) for the differentiation between normal subjects and chronic obstructive pulmonary disease (COPD) patients.
Methods
This retrospective study included healthy volunteers and COPD patients recruited from 2009 to 2014 at Fukujuji Hospital, who received DDR and pulmonary functional test. AUD was defined as an area under a hemidiaphragm and above the line connecting the ipsilateral costophrenic angle to the top of the hemidiaphragm on DDR image. AUD in full inspiration minus AUD in full expiration (ΔAUD) was also calculated. The diaphragmatic surface was demarcated manually on DDR image to calculate AUD. Three-group comparison of AUD and ΔAUD among normal, mild COPD, and severe COPD subjects was tested with one-way analysis of variance, followed by multiple comparison with Tukey-Kramer method. The diagnostic accuracy of COPD by ΔAUD was assessed using receiver-operating-characteristics (ROC) curve.
Results
Sixty-eight participants (36 men, 29 COPD patients) were enrolled. AUD in full inspiration was larger in healthy volunteers than in COPD patients (right, p < 0.001; left, p = 0.02). ΔAUD were different in the three-group comparison (right, normal, 208.7 ± 184.6 mm2, mild COPD, −18.1 ± 117.5 mm2, severe COPD −97.5 ± 150.0 mm2, p < 0.001; left, normal, 254.9 ± 131.5 mm2, mild COPD, −12.5 ± 136.5 mm2, severe COPD, −100.7 ± 134.1 mm2, p < 0.001). ROC curve showed high diagnostic performance of COPD by unilateral ΔAUD (right, area-under curve 0.942; left, area-under-curve 0.965).
Conclusion
The value of ΔAUD was smaller according to the severity of COPD. ΔAUD can be helpful in distinguishing healthy subjects from COPD patients.
{"title":"Diaphragmatic curvature analysis using dynamic digital radiography","authors":"Takuya Hino , Akinori Tsunomori , Noriaki Wada , Akinori Hata , Taiki Fukuda , Yusei Nakamura , Yoshitake Yamada , Tomoyuki Hida , Mizuki Nishino , Masako Ueyama , Atsuko Kurosaki , Takeshi Kubo , Shoji Kudoh , Kousei Ishigami , Hiroto Hatabu","doi":"10.1016/j.ejro.2025.100676","DOIUrl":"10.1016/j.ejro.2025.100676","url":null,"abstract":"<div><h3>Purpose</h3><div>To investigate area under diaphragm (AUD) obtained by dynamic digital radiography (DDR) for the differentiation between normal subjects and chronic obstructive pulmonary disease (COPD) patients.</div></div><div><h3>Methods</h3><div>This retrospective study included healthy volunteers and COPD patients recruited from 2009 to 2014 at Fukujuji Hospital, who received DDR and pulmonary functional test. AUD was defined as an area under a hemidiaphragm and above the line connecting the ipsilateral costophrenic angle to the top of the hemidiaphragm on DDR image. AUD in full inspiration minus AUD in full expiration (ΔAUD) was also calculated. The diaphragmatic surface was demarcated manually on DDR image to calculate AUD. Three-group comparison of AUD and ΔAUD among normal, mild COPD, and severe COPD subjects was tested with one-way analysis of variance, followed by multiple comparison with Tukey-Kramer method. The diagnostic accuracy of COPD by ΔAUD was assessed using receiver-operating-characteristics (ROC) curve.</div></div><div><h3>Results</h3><div>Sixty-eight participants (36 men, 29 COPD patients) were enrolled. AUD in full inspiration was larger in healthy volunteers than in COPD patients (right, p < 0.001; left, p = 0.02). ΔAUD were different in the three-group comparison (right, normal, 208.7 ± 184.6 mm<sup>2</sup>, mild COPD, −18.1 ± 117.5 mm<sup>2</sup>, severe COPD −97.5 ± 150.0 mm<sup>2</sup>, p < 0.001; left, normal, 254.9 ± 131.5 mm<sup>2</sup>, mild COPD, −12.5 ± 136.5 mm<sup>2</sup>, severe COPD, −100.7 ± 134.1 mm<sup>2</sup>, p < 0.001). ROC curve showed high diagnostic performance of COPD by unilateral ΔAUD (right, area-under curve 0.942; left, area-under-curve 0.965).</div></div><div><h3>Conclusion</h3><div>The value of ΔAUD was smaller according to the severity of COPD. ΔAUD can be helpful in distinguishing healthy subjects from COPD patients.</div></div>","PeriodicalId":38076,"journal":{"name":"European Journal of Radiology Open","volume":"15 ","pages":"Article 100676"},"PeriodicalIF":2.9,"publicationDate":"2025-08-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144773008","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
扫码关注我们
求助内容：
应助结果提醒方式：
京公网安备 11010802042870号