Pub Date : 2025-11-26DOI: 10.1016/j.crad.2025.107192
R. Castro-Zunti , Y.M. Han , K.Y. Kim , A. Vardhan , D.E. Lee , E.S. Ha , Y. Choi , H.S. Chae , G.Y. Jin , S-b. Ko
Aim
Although standardized 3D volume rendering techniques (VRT) and embolization guidance visualize and identify tumor-feeding arteries, current vessel tracking software lacks automatic angle recommendations. This forces an operator, e.g. an interventional radiologist, to leave an ongoing procedure to manually manipulate the system and find the best angle for each feeding vessel—requiring time-consuming re-scrubbing. We propose a computer vision algorithm that suggests a rotation/angle in the VRT where a tumor-feeding artery's view is maximized. We focus on hepatocellular carcinoma.
Methods
Our algorithm accepts a series of post-embolization guidance frames extracted from the 3D VRT; the VRT is rotated in 5° intervals from, e.g., ±15°, fixing one axis (e.g. CRAN/CAUD) and rotating the other (e.g. LAO/RAO). Our algorithm segments the embolization guidance line and recommends 4 views/angles by maximizing the features of line length (contour area) and convex hull area. We developed/iterated our algorithm using 19 patient cases and feedback from various experts.
Results
Over a 50-patient internal validation set, according to an interventional radiologist with 33 years of experience, a view/angle sufficient for the embolization task was always present among the top-4 views/angles suggested by our algorithm (100% retrieval relevance).
Conclusion
Sufficient view/angle selection for hepatic artery embolization can be achieved using traditional computer vision. Our technique is much faster and more explainable than deep learning approaches, and could greatly improve radiologists' procedural efficiency. We recommend conducting a larger study with more patients and further technical iteration.
{"title":"An accurate, straightforward computer vision algorithm for optimal tumor-feeding visualization in cone-beam computed tomography hepatic arteriography: A preliminary study","authors":"R. Castro-Zunti , Y.M. Han , K.Y. Kim , A. Vardhan , D.E. Lee , E.S. Ha , Y. Choi , H.S. Chae , G.Y. Jin , S-b. Ko","doi":"10.1016/j.crad.2025.107192","DOIUrl":"10.1016/j.crad.2025.107192","url":null,"abstract":"<div><h3>Aim</h3><div>Although standardized 3D volume rendering techniques (VRT) and embolization guidance visualize and identify tumor-feeding arteries, current vessel tracking software lacks automatic angle recommendations. This forces an operator, e.g. an interventional radiologist, to leave an ongoing procedure to manually manipulate the system and find the best angle for each feeding vessel—requiring time-consuming re-scrubbing. We propose a computer vision algorithm that suggests a rotation/angle in the VRT where a tumor-feeding artery's view is maximized. We focus on hepatocellular carcinoma.</div></div><div><h3>Methods</h3><div>Our algorithm accepts a series of post-embolization guidance frames extracted from the 3D VRT; the VRT is rotated in 5° intervals from, e.g., ±15°, fixing one axis (e.g. CRAN/CAUD) and rotating the other (e.g. LAO/RAO). Our algorithm segments the embolization guidance line and recommends 4 views/angles by maximizing the features of line length (contour area) and convex hull area. We developed/iterated our algorithm using 19 patient cases and feedback from various experts.</div></div><div><h3>Results</h3><div>Over a 50-patient internal validation set, according to an interventional radiologist with 33 years of experience, a view/angle sufficient for the embolization task was always present among the top-4 views/angles suggested by our algorithm (100% retrieval relevance).</div></div><div><h3>Conclusion</h3><div>Sufficient view/angle selection for hepatic artery embolization can be achieved using traditional computer vision. Our technique is much faster and more explainable than deep learning approaches, and could greatly improve radiologists' procedural efficiency. We recommend conducting a larger study with more patients and further technical iteration.</div></div>","PeriodicalId":10695,"journal":{"name":"Clinical radiology","volume":"93 ","pages":"Article 107192"},"PeriodicalIF":1.9,"publicationDate":"2025-11-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145975513","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}
Pub Date : 2025-11-24DOI: 10.1016/j.crad.2025.107183
Z. Zhu , L. Hou , Y. Zhao , L. Li , X. Zhao
Aim
To assess MRI and clinical features for the differentiation of hepatocellular adenoma (HCA) and well-differentiated hepatocellular carcinoma (WDHCC).
Materials and Methods
Contrast-enhanced MRI images and clinical data of 144 pathologically confirmed HCA or WDHCC enrolled retrospectively from multiple centers between January 2015 and January 2024. Two readers reviewed images to identify imaging features and measure signal intensity on multiple phases images. The predictive model was established using binary Logistic regression, and the predictive ability was evaluated using the area under the curve (AUC), accuracy, sensitivity, and specificity by R software.
Results
Out of 144 eligible patients (35 HCAs, 109 WDHCCs), 23 in 37 indexes showed significant differences. Moreover, 10 parameters remained significant after the univariate regression analysis. To construct a highly accurate predictive model, the significant parameters were further subjected to a multivariate regression model. Six valuable factors (long axis, T1WI, T2WI/FS, capsule enhancement, septa, and cirrhosis) were selected to establish the diagnostic model. Then, a nomogram to discriminate HCA from WDHCC was built on the basis of a multivariate logistic regression model. The AUC of the MRI signal model, the clinical factors model, and the combined model in training sets and validation sets are 0.955, 0.929, 0.962, and 0.898, 0.835, 0.846, respectively. DCA and clinical impact curve was applied to assess the clinical utility of the diagnostic nomogram. Based on the DCA, the MRI signal showed superior clinical utility compared to the other models.
Conclusion
MRI signal-based model provides high diagnostic performance as demonstrated in the differentiation of HCA and WDHCC, supported by a nomogram model.
目的探讨肝细胞腺瘤（HCA）与高分化肝细胞癌（WDHCC）鉴别的MRI及临床特征。材料与方法回顾性研究2015年1月至2024年1月来自多个中心的144例病理证实的HCA或WDHCC的MRI增强图像和临床资料。两位读者回顾了图像以识别成像特征并测量多相图像上的信号强度。采用二元Logistic回归建立预测模型，并通过R软件采用曲线下面积（area under The curve， AUC）、准确性、敏感性和特异性评价预测能力。结果144例符合条件的患者（hca 35例，wdhcc 109例），37项指标中有23项存在显著性差异。单因素回归分析后，10个参数仍然显著。为了构建高精度的预测模型，进一步对显著参数进行多元回归模型。选择6个有价值的因素（长轴、T1WI、T2WI/FS、胶囊增强、间隔、肝硬化）建立诊断模型。然后，在多元逻辑回归模型的基础上，建立了判别HCA和WDHCC的nomogram。MRI信号模型、临床因素模型和联合模型在训练集和验证集上的AUC分别为0.955、0.929、0.962和0.898、0.835、0.846。应用DCA和临床影响曲线评估诊断图的临床应用价值。与其他模型相比，基于DCA的MRI信号具有更好的临床应用价值。结论基于mri信号的模型在HCA和WDHCC的鉴别诊断中具有较高的诊断价值，并得到了nomogram模型的支持。
{"title":"Distinguishing between hepatocellular adenoma and well-differentiated hepatocellular carcinoma using MRI and clinical feature-based nomogram model","authors":"Z. Zhu , L. Hou , Y. Zhao , L. Li , X. Zhao","doi":"10.1016/j.crad.2025.107183","DOIUrl":"10.1016/j.crad.2025.107183","url":null,"abstract":"<div><h3>Aim</h3><div>To assess MRI and clinical features for the differentiation of hepatocellular adenoma (HCA) and well-differentiated hepatocellular carcinoma (WDHCC).</div></div><div><h3>Materials and Methods</h3><div>Contrast-enhanced MRI images and clinical data of 144 pathologically confirmed HCA or WDHCC enrolled retrospectively from multiple centers between January 2015 and January 2024. Two readers reviewed images to identify imaging features and measure signal intensity on multiple phases images. The predictive model was established using binary Logistic regression, and the predictive ability was evaluated using the area under the curve (AUC), accuracy, sensitivity, and specificity by R software.</div></div><div><h3>Results</h3><div>Out of 144 eligible patients (35 HCAs, 109 WDHCCs), 23 in 37 indexes showed significant differences. Moreover, 10 parameters remained significant after the univariate regression analysis. To construct a highly accurate predictive model, the significant parameters were further subjected to a multivariate regression model. Six valuable factors (long axis, T1WI, T2WI/FS, capsule enhancement, septa, and cirrhosis) were selected to establish the diagnostic model. Then, a nomogram to discriminate HCA from WDHCC was built on the basis of a multivariate logistic regression model. The AUC of the MRI signal model, the clinical factors model, and the combined model in training sets and validation sets are 0.955, 0.929, 0.962, and 0.898, 0.835, 0.846, respectively. DCA and clinical impact curve was applied to assess the clinical utility of the diagnostic nomogram. Based on the DCA, the MRI signal showed superior clinical utility compared to the other models.</div></div><div><h3>Conclusion</h3><div>MRI signal-based model provides high diagnostic performance as demonstrated in the differentiation of HCA and WDHCC, supported by a nomogram model.</div></div>","PeriodicalId":10695,"journal":{"name":"Clinical radiology","volume":"94 ","pages":"Article 107183"},"PeriodicalIF":1.9,"publicationDate":"2025-11-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146026043","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}
Pub Date : 2025-11-21DOI: 10.1016/j.crad.2025.107187
M. Nakajo , D. Hirahara , M. Hirahara , Y. Eizuru , A. Tani , F. Kanzaki , K. Takumi , K. Kamimura , T. Yoshiura
Functional and metabolic information provided by positron emission tomography (PET) imaging, such as patient diagnosis, tumour staging, and treatment evaluation, plays an important role in the clinical management of patients with cancer. Nonetheless, its clinical efficacy may be inhibited by differences in image quality and limitations in quantitative robustness. Artificial intelligence (AI) has transformed oncological PET imaging by improving image quality and facilitating a more consistent extraction of quantitative metrics. Recent research emphasises the value of AI in improving diagnostic accuracy and prognostic modelling. However, to ensure that AI-based PET analysis is successfully implemented in clinical practice, challenges such as imaging data standardisation, the development of reliable explainability methods, and the establishment of regulatory frameworks must be addressed. To optimise individualised care, future progress will likely be based on multimodal integration, federated learning, and probabilistic deep learning. Overall, this review highlights both the current progress and the remaining challenges of AI in oncological PET, aiming to provide a balanced perspective for future clinical translation.
{"title":"Artificial intelligence in oncological positron emission tomography: advancing image analysis and interpretation","authors":"M. Nakajo , D. Hirahara , M. Hirahara , Y. Eizuru , A. Tani , F. Kanzaki , K. Takumi , K. Kamimura , T. Yoshiura","doi":"10.1016/j.crad.2025.107187","DOIUrl":"10.1016/j.crad.2025.107187","url":null,"abstract":"<div><div>Functional and metabolic information provided by positron emission tomography (PET) imaging, such as patient diagnosis, tumour staging, and treatment evaluation, plays an important role in the clinical management of patients with cancer. Nonetheless, its clinical efficacy may be inhibited by differences in image quality and limitations in quantitative robustness. Artificial intelligence (AI) has transformed oncological PET imaging by improving image quality and facilitating a more consistent extraction of quantitative metrics. Recent research emphasises the value of AI in improving diagnostic accuracy and prognostic modelling. However, to ensure that AI-based PET analysis is successfully implemented in clinical practice, challenges such as imaging data standardisation, the development of reliable explainability methods, and the establishment of regulatory frameworks must be addressed. To optimise individualised care, future progress will likely be based on multimodal integration, federated learning, and probabilistic deep learning. Overall, this review highlights both the current progress and the remaining challenges of AI in oncological PET, aiming to provide a balanced perspective for future clinical translation.</div></div>","PeriodicalId":10695,"journal":{"name":"Clinical radiology","volume":"92 ","pages":"Article 107187"},"PeriodicalIF":1.9,"publicationDate":"2025-11-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145767384","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}
Pub Date : 2025-11-21DOI: 10.1016/j.crad.2025.107184
C.M. Obhuli , S. Pendem , S. Abhijith , R. Kadavigere , Priyanka , P.S. Priya , C. Chacko
AIM
Computed tomography (CT) plays a central role in thoracic imaging, but maintaining diagnostic image quality at reduced doses remains a challenge. Filtered back projection (FBP) produces high noise, and iterative reconstruction (IR) reduces noise but alters image texture at low dose. Deep learning image reconstruction (DLIR) suppresses noise while preserving detail, yet its diagnostic performance in chest CT remains unclear. This review aimed to evaluate the clinical diagnostic value of DLIR in chest CT imaging.
MATERIALS AND METHODS
A systematic review was conducted according to Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines (International Prospective Register of Systematic Reviews [PROSPERO] registered). The following databases were searched for studies comparing DLIR with IR/FBP in chest CT: PubMed, Embase, Scopus, Web of Science, IEEE, and Cochrane Library. Eligible studies included human participants and reported diagnostic or image-quality outcomes. Quality assessment was performed using the QUADAS-2 tool. Given outcome heterogeneity, results were synthesised qualitatively using effect direction plots and sign tests.
RESULTS
From 1,967 records, 13 studies met the inclusion criteria. DLIR demonstrated superior diagnostic performance compared with IR/FBP and showed higher sensitivity for nodule detection (up to 96.9%), improved area under the curve (AUC) for lung texture analysis (0.97–1.0 vs 0.91–0.97 with hybrid IR), and stronger interobserver agreement for interstitial lung disease (ILD) pattern classification (κ up to 0.992). DLIR achieved substantial dose reductions (up to 97%) and faster reconstruction times while maintaining diagnostic consistency.
CONCLUSION
DLIR demonstrates noninferior to superior diagnostic performance compared with FBP/IR, supporting its role in routine chest CT. Large-scale studies remain essential to establish its impact on patient outcomes and guide clinical adoption.
目的：计算机断层扫描（CT）在胸部成像中起着核心作用，但在低剂量下保持诊断图像质量仍然是一个挑战。滤波后投影（FBP）产生高噪声，迭代重建（IR）在低剂量下降低了噪声，但改变了图像纹理。深度学习图像重建（DLIR）在保留细节的同时抑制了噪声，但其在胸部CT中的诊断性能尚不清楚。本文旨在探讨DLIR在胸部CT成像中的临床诊断价值。材料和方法：根据系统评价和荟萃分析首选报告项目（PRISMA）指南（国际前瞻性系统评价注册[PROSPERO]注册）进行系统评价。我们检索了以下数据库以比较DLIR与IR/FBP在胸部CT中的研究：PubMed， Embase, Scopus, Web of Science， IEEE和Cochrane Library。符合条件的研究包括人类参与者和报告的诊断或图像质量结果。使用QUADAS-2工具进行质量评估。考虑到结果的异质性，使用效应方向图和符号检验对结果进行定性综合。结果：1967项记录中，13项研究符合纳入标准。与IR/FBP相比，DLIR表现出更好的诊断性能，对结节检测的灵敏度更高（高达96.9%），改善肺质地分析的曲线下面积（AUC） （0.97-1.0 vs 0.91-0.97），对间质性肺病（ILD）模式分类的观察者间一致性更强（κ高达0.992）。DLIR在保持诊断一致性的同时实现了剂量的大幅减少（高达97%）和更快的重建时间。结论：与FBP/IR相比，DLIR在常规胸部CT中的诊断价值不亚于FBP/IR。大规模研究对于确定其对患者预后的影响和指导临床应用仍然至关重要。
{"title":"Clinical value of deep learning image reconstruction in chest computed tomography (CT) imaging: a systematic review","authors":"C.M. Obhuli , S. Pendem , S. Abhijith , R. Kadavigere , Priyanka , P.S. Priya , C. Chacko","doi":"10.1016/j.crad.2025.107184","DOIUrl":"10.1016/j.crad.2025.107184","url":null,"abstract":"<div><h3>AIM</h3><div>Computed tomography (CT) plays a central role in thoracic imaging, but maintaining diagnostic image quality at reduced doses remains a challenge. Filtered back projection (FBP) produces high noise, and iterative reconstruction (IR) reduces noise but alters image texture at low dose. Deep learning image reconstruction (DLIR) suppresses noise while preserving detail, yet its diagnostic performance in chest CT remains unclear. This review aimed to evaluate the clinical diagnostic value of DLIR in chest CT imaging.</div></div><div><h3>MATERIALS AND METHODS</h3><div>A systematic review was conducted according to Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines (International Prospective Register of Systematic Reviews [PROSPERO] registered). The following databases were searched for studies comparing DLIR with IR/FBP in chest CT: PubMed, Embase, Scopus, Web of Science, IEEE, and Cochrane Library. Eligible studies included human participants and reported diagnostic or image-quality outcomes. Quality assessment was performed using the QUADAS-2 tool. Given outcome heterogeneity, results were synthesised qualitatively using effect direction plots and sign tests.</div></div><div><h3>RESULTS</h3><div>From 1,967 records, 13 studies met the inclusion criteria. DLIR demonstrated superior diagnostic performance compared with IR/FBP and showed higher sensitivity for nodule detection (up to 96.9%), improved area under the curve (AUC) for lung texture analysis (0.97–1.0 vs 0.91–0.97 with hybrid IR), and stronger interobserver agreement for interstitial lung disease (ILD) pattern classification (κ up to 0.992). DLIR achieved substantial dose reductions (up to 97%) and faster reconstruction times while maintaining diagnostic consistency.</div></div><div><h3>CONCLUSION</h3><div>DLIR demonstrates noninferior to superior diagnostic performance compared with FBP/IR, supporting its role in routine chest CT. Large-scale studies remain essential to establish its impact on patient outcomes and guide clinical adoption.</div></div>","PeriodicalId":10695,"journal":{"name":"Clinical radiology","volume":"92 ","pages":"Article 107184"},"PeriodicalIF":1.9,"publicationDate":"2025-11-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145780515","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}
Pub Date : 2025-11-21DOI: 10.1016/j.crad.2025.107188
Y. Noda , T. Ishihara , N. Kawai , T. Kaga , T. Miyoshi , F. Hyodo , H. Kato , A.R. Kambadakone , M. Matsuo
AIM
The aim of this study is to investigate the relationship between participants' body weights and radiation doses in contrast-enhanced chest-abdomen-pelvis computed tomography (CT) scans, using both older and newer CT scanners that can provide adequate output even at low tube voltages.
MATERIALS AND METHODS
Participants who underwent contrast-enhanced chest-abdomen-pelvis CT from September to December 2021 were prospectively randomised into four groups based on kilovolt peak (kVp) and maximum tube current on two scanners–group A (120 kVp, 835 mA), group B (80 kVp, 700 mA), group C (120 kVp, 900 mA), and group D (80 kVp, 1,300 mA). The relationships between the participants' body weights and CT dose-index volume (CTDIvol) were compared among the four groups using non-linear regression analysis. The background noise was compared between groups A vs B and C vs D.
RESULTS
A total of 118, 104, 100, and 106 participants were included in groups A, B, C, and D, respectively. The CTDIvol was lower in group B than in group A above the body weight of 57 kg (P < .001–.004). Similarly, the CTDIvol was lower in group D than in group C above a body weight of 68 kg (P < .001–.04). The background noise was higher in group B than in group A at abdominal and pelvic regions for participants weighing 57–67 kg (P < .001 for each); however, no difference was found between groups C and D (P = .21–.82).
CONCLUSION
High-tube current output CT scanners necessitate an increase in the participants' body weight threshold for low-kVp scans from 57 kg to 68 kg to achieve a reduction in CTDIvol.
{"title":"Effect of high-tube current x-ray tube on computed tomography (CT) dose-index volume in low kilovolt peak (kVp) contrast-enhanced chest-abdomen-pelvis computed tomography (CT)","authors":"Y. Noda , T. Ishihara , N. Kawai , T. Kaga , T. Miyoshi , F. Hyodo , H. Kato , A.R. Kambadakone , M. Matsuo","doi":"10.1016/j.crad.2025.107188","DOIUrl":"10.1016/j.crad.2025.107188","url":null,"abstract":"<div><h3>AIM</h3><div>The aim of this study is to investigate the relationship between participants' body weights and radiation doses in contrast-enhanced chest-abdomen-pelvis computed tomography (CT) scans, using both older and newer CT scanners that can provide adequate output even at low tube voltages.</div></div><div><h3>MATERIALS AND METHODS</h3><div>Participants who underwent contrast-enhanced chest-abdomen-pelvis CT from September to December 2021 were prospectively randomised into four groups based on kilovolt peak (kVp) and maximum tube current on two scanners–group A (120 kVp, 835 mA), group B (80 kVp, 700 mA), group C (120 kVp, 900 mA), and group D (80 kVp, 1,300 mA). The relationships between the participants' body weights and CT dose-index volume (CTDI<sub>vol</sub>) were compared among the four groups using non-linear regression analysis. The background noise was compared between groups A vs B and C vs D.</div></div><div><h3>RESULTS</h3><div>A total of 118, 104, 100, and 106 participants were included in groups A, B, C, and D, respectively. The CTDI<sub>vol</sub> was lower in group B than in group A above the body weight of 57 kg (<em>P</em> < .001–.004). Similarly, the CTDI<sub>vol</sub> was lower in group D than in group C above a body weight of 68 kg (<em>P</em> < .001–.04). The background noise was higher in group B than in group A at abdominal and pelvic regions for participants weighing 57–67 kg (<em>P</em> < .001 for each); however, no difference was found between groups C and D (<em>P</em> = .21–.82).</div></div><div><h3>CONCLUSION</h3><div>High-tube current output CT scanners necessitate an increase in the participants' body weight threshold for low-kVp scans from 57 kg to 68 kg to achieve a reduction in CTDI<sub>vol</sub>.</div></div>","PeriodicalId":10695,"journal":{"name":"Clinical radiology","volume":"92 ","pages":"Article 107188"},"PeriodicalIF":1.9,"publicationDate":"2025-11-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145780486","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}
Pub Date : 2025-11-21DOI: 10.1016/j.crad.2025.107189
R.L. Cochran , N.D. Mercaldo , N. Nakrour , S. Ghosh , E. Milshteyn , M. Pohl , A. Guidon , D.M. Dahl , A.S. Feldman , M.G. Harisinghani
AIM
Compare the diagnostic performance of whole-gland prostate-specific antigen density (wgPSAD) to zonal volume-adjusted prostate-specific antigen density (PSAD) for predicting prostate cancer in patients who underwent a prostate magnetic resonance imaging (MRI) followed by prostate biopsy.
Materials and Methods
A retrospective study of consecutive patients who underwent prostate MRI followed by systematic biopsy with or without targeted biopsy between January 2019 and December 2020 was performed. Whole-gland (wgPSAD), transition-zone prostate-specific antigen density (tzPSAD), and peripheral-zone prostate-specific antigen density (pzPSAD) were calculated using prostate-specific antigen (PSA) levels drawn before imaging, and volume estimates were derived from MRI using artificial intelligence (AI) software assistance. Diagnostic performance was assessed using logistic regression and estimating internally validated receiver operating characteristic area under the characteristic (AUC) curves.
RESULTS
A total of 551 patients with a median age of 66 years (interquartile range [IQR]: 61–72) were included. The univariable analysis demonstrated superior AUC for wgPSAD (AUC: 0.71 and 0.71) and tzPSAD (AUC: 0.72 and 0.72) compared to pzPSAD (AUC: 0.51 and 0.56) for any cancer and clinically significant prostate cancer (csPCa). The multivariable analysis including age and 5α-reductase inhibitor therapy demonstrated a superior AUC of tzPSAD for predicting csPCa (AUC: 0.77 vs 0.75; P=0.02) compared to both wgPSAD and pzPSAD (AUC: 0.77 vs 0.67; P<0.001). Variable importance analysis suggested prescribed 5α-reductase inhibitor therapy may be protective against csPCa.
CONCLUSION
wgPSAD and tzPSAD are superior to pzPSAD for the detection of csPCa. When accounting for key covariates, tzPSAD may be superior to wgPSAD.
比较全腺体前列腺特异性抗原密度（wgPSAD）与分区体积调整前列腺特异性抗原密度（PSAD）在前列腺磁共振成像（MRI）后前列腺活检患者中预测前列腺癌的诊断性能。材料和方法对2019年1月至2020年12月期间连续接受前列腺MRI检查并进行系统活检（或不进行靶向活检）的患者进行回顾性研究。使用成像前绘制的前列腺特异性抗原（PSA）水平计算全腺体（wgPSAD）、过渡区前列腺特异性抗原密度（tzPSAD）和外周区前列腺特异性抗原密度（pzPSAD），并使用人工智能（AI）软件辅助从MRI中得出体积估计。诊断性能评估采用逻辑回归和估计内部验证的受试者工作特征面积下的特征（AUC）曲线。结果共纳入551例患者，中位年龄66岁（四分位数间距[IQR]: 61-72）。单变量分析显示，与pzPSAD （AUC: 0.51和0.56）相比，wgPSAD （AUC: 0.71和0.71）和tzPSAD （AUC: 0.72和0.72）在任何癌症和临床显著前列腺癌（csPCa）中的AUC均优于pzPSAD （AUC: 0.51和0.56）。包括年龄和5α-还原酶抑制剂治疗在内的多变量分析表明，与wgPSAD和pzPSAD相比，tzPSAD预测csPCa的AUC （AUC: 0.77 vs 0.75； P=0.02）优于pzPSAD （AUC: 0.77 vs 0.67; P<0.001）。变量重要性分析表明，规定的5α-还原酶抑制剂治疗可能对csPCa有保护作用。结论gpsad和tzPSAD检测csPCa优于pzPSAD。在考虑关键协变量时，tzPSAD可能优于wgPSAD。
{"title":"Comparing conventional, peripheral, and transition zone prostate-specific antigen densities for the detection of clinically significant prostate cancer","authors":"R.L. Cochran , N.D. Mercaldo , N. Nakrour , S. Ghosh , E. Milshteyn , M. Pohl , A. Guidon , D.M. Dahl , A.S. Feldman , M.G. Harisinghani","doi":"10.1016/j.crad.2025.107189","DOIUrl":"10.1016/j.crad.2025.107189","url":null,"abstract":"<div><h3>AIM</h3><div>Compare the diagnostic performance of whole-gland prostate-specific antigen density (wgPSAD) to zonal volume-adjusted prostate-specific antigen density (PSAD) for predicting prostate cancer in patients who underwent a prostate magnetic resonance imaging (MRI) followed by prostate biopsy.</div></div><div><h3>Materials and Methods</h3><div>A retrospective study of consecutive patients who underwent prostate MRI followed by systematic biopsy with or without targeted biopsy between January 2019 and December 2020 was performed. Whole-gland (wgPSAD), transition-zone prostate-specific antigen density (tzPSAD), and peripheral-zone prostate-specific antigen density (pzPSAD) were calculated using prostate-specific antigen (PSA) levels drawn before imaging, and volume estimates were derived from MRI using artificial intelligence (AI) software assistance. Diagnostic performance was assessed using logistic regression and estimating internally validated receiver operating characteristic area under the characteristic (AUC) curves.</div></div><div><h3>RESULTS</h3><div>A total of 551 patients with a median age of 66 years (interquartile range [IQR]: 61–72) were included. The univariable analysis demonstrated superior AUC for wgPSAD (AUC: 0.71 and 0.71) and tzPSAD (AUC: 0.72 and 0.72) compared to pzPSAD (AUC: 0.51 and 0.56) for any cancer and clinically significant prostate cancer (csPCa). The multivariable analysis including age and 5α-reductase inhibitor therapy demonstrated a superior AUC of tzPSAD for predicting csPCa (AUC: 0.77 vs 0.75; <em>P</em>=0.02) compared to both wgPSAD and pzPSAD (AUC: 0.77 vs 0.67; <em>P</em><0.001). Variable importance analysis suggested prescribed 5α-reductase inhibitor therapy may be protective against csPCa.</div></div><div><h3>CONCLUSION</h3><div>wgPSAD and tzPSAD are superior to pzPSAD for the detection of csPCa. When accounting for key covariates, tzPSAD may be superior to wgPSAD.</div></div>","PeriodicalId":10695,"journal":{"name":"Clinical radiology","volume":"92 ","pages":"Article 107189"},"PeriodicalIF":1.9,"publicationDate":"2025-11-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145786548","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}
Pub Date : 2025-11-21DOI: 10.1016/j.crad.2025.107186
K.A. Sinci , B. Aksoy , O.D. Aydin , A.I. Akdogan , C. Kazimoglu , O. Tosun
AIM
To evaluate the diagnostic performance of the fat-fluid level (FFL) on standing knee radiographs for detecting intra-articular fractures in acute trauma.
MATERIALS AND METHODS
This retrospective, single-centre study included 76 patients with acute knee trauma who underwent standing radiography and knee computed tomography (CT) within 12 hours. Patients were divided into FFL (+) (n=32) and randomly selected FFL (–) control (n=44) groups. Radiographs were assessed for FFL visibility by projection (anteroposterior [AP], lateral, or both) and for fracture presence. CT served as the reference standard. Diagnostic performance metrics were calculated and interobserver agreement was assessed using Cohen’s kappa.
RESULTS
An FFL was observed in 32 patients (42 %), visible on lateral radiographs in all and on AP views in 56 %. In seven patients, FFL-positive patients (22 %), no fracture line was radiographically visible yet CT-confirmed fractures in all FFL (+) cases (positive predictive value [PPV] = 100 %). Overall, CT-confirmed fractures in 42 patients (55 %); of these, 10 were FFL (–). Using CT as the reference, FFL sensitivity, specificity, PPV, and negative predictive value were 76 %, 100 %, 100 %, and 77 %, respectively. Tibial fractures were most common (69 %), followed by patellar fractures (26 %). AP views were more sensitive for fracture detection, while lateral views better demonstrated the FFL. Interobserver agreement was almost perfect (κ = 0.84-0.90).
CONCLUSION
The FFL on standing lateral radiographs is a highly specific and reproducible indirect indicator of intra-articular fracture. Incorporating standing lateral radiographs into acute knee trauma protocols may improve fracture detection, particularly where CT access is limited.
{"title":"Fat–fluid level on standing lateral knee radiographs as a reliable indicator of occult intra-articular knee fractures in acute trauma evaluation","authors":"K.A. Sinci , B. Aksoy , O.D. Aydin , A.I. Akdogan , C. Kazimoglu , O. Tosun","doi":"10.1016/j.crad.2025.107186","DOIUrl":"10.1016/j.crad.2025.107186","url":null,"abstract":"<div><h3>AIM</h3><div>To evaluate the diagnostic performance of the fat-fluid level (FFL) on standing knee radiographs for detecting intra-articular fractures in acute trauma.</div></div><div><h3>MATERIALS AND METHODS</h3><div>This retrospective, single-centre study included 76 patients with acute knee trauma who underwent standing radiography and knee computed tomography (CT) within 12 hours. Patients were divided into FFL (+) (n=32) and randomly selected FFL (–) control (n=44) groups. Radiographs were assessed for FFL visibility by projection (anteroposterior [AP], lateral, or both) and for fracture presence. CT served as the reference standard. Diagnostic performance metrics were calculated and interobserver agreement was assessed using Cohen’s kappa.</div></div><div><h3>RESULTS</h3><div>An FFL was observed in 32 patients (42 %), visible on lateral radiographs in all and on AP views in 56 %. In seven patients, FFL-positive patients (22 %), no fracture line was radiographically visible yet CT-confirmed fractures in all FFL (+) cases (positive predictive value [PPV] = 100 %). Overall, CT-confirmed fractures in 42 patients (55 %); of these, 10 were FFL (–). Using CT as the reference, FFL sensitivity, specificity, PPV, and negative predictive value were 76 %, 100 %, 100 %, and 77 %, respectively. Tibial fractures were most common (69 %), followed by patellar fractures (26 %). AP views were more sensitive for fracture detection, while lateral views better demonstrated the FFL. Interobserver agreement was almost perfect (κ = 0.84-0.90).</div></div><div><h3>CONCLUSION</h3><div>The FFL on standing lateral radiographs is a highly specific and reproducible indirect indicator of intra-articular fracture. Incorporating standing lateral radiographs into acute knee trauma protocols may improve fracture detection, particularly where CT access is limited.</div></div>","PeriodicalId":10695,"journal":{"name":"Clinical radiology","volume":"92 ","pages":"Article 107186"},"PeriodicalIF":1.9,"publicationDate":"2025-11-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145767405","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}
Pub Date : 2025-11-20DOI: 10.1016/j.crad.2025.107185
D Velazquez-Pimentel, J Pancholi, P Jenkins, N Cinti, M Stephanou, D Kotecha, A White, S Ashraf, O Llewellyn, G Vigneswaran, H Shiwani, J Zhong, Des Alcorn, D J Breen, P Haslam, G Hickson, O Jaffer, P Kennedy, P Littler, P Peddu, N Railton, T M Wah
Aim: This study aims to survey the provision of Interventional Oncology (IO) services in the UK and compare the results to survey data collected in 2016.
Materials and methods: A cross-sectional multicentre study of the provision of IO services was conducted across all interventional radiology (IR) departments in the UK. Data were collected using an electronic survey tool and executed via the UNITE Collaborative. IO procedures were defined using the Royal College of Radiologists classification categories. For each IR department information regarding demographic details, current IO procedures, equipment, and relevant infrastructure was collected. Thereafter, responses were compared to survey data collected in 2016.
Results: A total of 169 hospital boards were invited to participate, 132 (78%) of which responded stating they had an IR department, while 29 (17%) responded stating they had no IR department and 8 (5%) provided no response. Of the hospital boards with IR departments, 49% (n=65/132) provided both disease-modifying and supportive/symptomatic procedures and 51% (n=67/132) offered only supportive/symptomatic procedures. Compared to 2016, there was a modest increase in the provision of disease-modifying procedures with the largest growth seen in transarterial chemoembolisation (+9%), selective internal radiation therapy (+7%), and renal ablation (+8%).
Conclusion: Over the last 8 years, the provision of IO services across the UK has only marginally grown in both supportive and disease-modifying domains. This study highlights the urgent need to identify and address barriers preventing access to IO procedures to ensure the UK population can benefit from modern, evidence-based IO care.
{"title":"IO1-UK: a cross-sectional study to re-evaluate the provision of interventional oncology services across the United Kingdom.","authors":"D Velazquez-Pimentel, J Pancholi, P Jenkins, N Cinti, M Stephanou, D Kotecha, A White, S Ashraf, O Llewellyn, G Vigneswaran, H Shiwani, J Zhong, Des Alcorn, D J Breen, P Haslam, G Hickson, O Jaffer, P Kennedy, P Littler, P Peddu, N Railton, T M Wah","doi":"10.1016/j.crad.2025.107185","DOIUrl":"https://doi.org/10.1016/j.crad.2025.107185","url":null,"abstract":"<p><strong>Aim: </strong>This study aims to survey the provision of Interventional Oncology (IO) services in the UK and compare the results to survey data collected in 2016.</p><p><strong>Materials and methods: </strong>A cross-sectional multicentre study of the provision of IO services was conducted across all interventional radiology (IR) departments in the UK. Data were collected using an electronic survey tool and executed via the UNITE Collaborative. IO procedures were defined using the Royal College of Radiologists classification categories. For each IR department information regarding demographic details, current IO procedures, equipment, and relevant infrastructure was collected. Thereafter, responses were compared to survey data collected in 2016.</p><p><strong>Results: </strong>A total of 169 hospital boards were invited to participate, 132 (78%) of which responded stating they had an IR department, while 29 (17%) responded stating they had no IR department and 8 (5%) provided no response. Of the hospital boards with IR departments, 49% (n=65/132) provided both disease-modifying and supportive/symptomatic procedures and 51% (n=67/132) offered only supportive/symptomatic procedures. Compared to 2016, there was a modest increase in the provision of disease-modifying procedures with the largest growth seen in transarterial chemoembolisation (+9%), selective internal radiation therapy (+7%), and renal ablation (+8%).</p><p><strong>Conclusion: </strong>Over the last 8 years, the provision of IO services across the UK has only marginally grown in both supportive and disease-modifying domains. This study highlights the urgent need to identify and address barriers preventing access to IO procedures to ensure the UK population can benefit from modern, evidence-based IO care.</p>","PeriodicalId":10695,"journal":{"name":"Clinical radiology","volume":" ","pages":"107185"},"PeriodicalIF":1.9,"publicationDate":"2025-11-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146008670","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}
Pub Date : 2025-11-20DOI: 10.1016/j.crad.2025.107181
D.E. Tekcan Sanli , A.N. Sanli
{"title":"Where does spectral computed tomography (CT) fit in breast imaging? Insights and considerations for clinical practice","authors":"D.E. Tekcan Sanli , A.N. Sanli","doi":"10.1016/j.crad.2025.107181","DOIUrl":"10.1016/j.crad.2025.107181","url":null,"abstract":"","PeriodicalId":10695,"journal":{"name":"Clinical radiology","volume":"92 ","pages":"Article 107181"},"PeriodicalIF":1.9,"publicationDate":"2025-11-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145773865","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}
Pub Date : 2025-11-20DOI: 10.1016/j.crad.2025.107180
D. Kajal , A. Nanapragasam , A. Leckie , S. Sagheb , F. Nasri , G. Bouchard-Fortier , J. Solnik , M. Atri
AIM
The aim of this study was validating Ovarian-Adnexal Reporting and Data System (O-RADS) 2022 risk estimates in surgically treated ovarian/adnexal masses comparing accuracy of O-RADS with modified ultrasound simple rules (mUSR) differentiating malignant from benign lesions. The mUSR was a simplified version of the International Ovarian Tumor Analysis (IOTA) using a binary classification of adnexal masses into benign/suspicious for malignancy.
MATERIALS AND METHODS
multisite retrospective study was conducted including patients with pathology-proven adnexal masses between January 2008 and December 2018. All ultrasound (US) video clips reviewed by an experienced radiologist with randomly selected subset were reviewed by two additional radiologists. Areas under receiver operator characteristic curves (AUCs) were compared without and with CA-125.
RESULTS
791 ovarian masses in 765 patients (26 bilateral) (mean age: 44 ± 15 years) (628 benign, 49 borderline, and 114 malignancies) demonstrated malignancy rates of 0.3%, 3.0%, 24.9%, and 82.4% for O-RADS 2, 3, 4, and 5, respectively. O-RADS and mUSR had a sensitivity of 0.96 (confidence interval [CI]: 0.92–0.99) and 0.96 (CI: 0.91–0.98), negative predictive values (NPVs) of 0.99 (CI: 0.97–1.00) and 0.99 (CI: 0.98–1.00) (P>0.05), specificities 0.75 [CI: 0.71–0.78] and 0.88 [CI: 0.85–0.91], and positive predictive values (PPVs) 0.50 (CI: 0.44–0.55) and 0.68 (CI: 0.61–0.74) (P<0.01), respectively. The AUC was 0.855 for O-RADS and 0.920 for mUSR (P=0.005). Interobserver agreement was excellent across all readers for mUSR benign versus mUSR malignant and O-RADS 2/3 versus O-RADS 4/5 (kappa > 0.86). CA 125 improved performance of mUSR (P=0.002) and O-RADS (P=0.005) only in perimenopausal/postmenopausal patients.
CONCLUSION
O-RADS and mUSR both with high sensitivity and NPV for detection of ovarian malignancy but mUSR with significantly higher specificity and PPV than O-RADS. This finding endorses the American College of Radiology (ACR) recommendation for expert sonologist consultation for O-RADS 3 and 4.
{"title":"Ultrasound Ovarian-Adnexal Reporting and Data System (O-RADS) and modified ultrasound simple rules comparison in evaluation of surgically proven adnexal masses","authors":"D. Kajal , A. Nanapragasam , A. Leckie , S. Sagheb , F. Nasri , G. Bouchard-Fortier , J. Solnik , M. Atri","doi":"10.1016/j.crad.2025.107180","DOIUrl":"10.1016/j.crad.2025.107180","url":null,"abstract":"<div><h3>AIM</h3><div>The aim of this study was validating Ovarian-Adnexal Reporting and Data System (O-RADS) 2022 risk estimates in surgically treated ovarian/adnexal masses comparing accuracy of O-RADS with modified ultrasound simple rules (mUSR) differentiating malignant from benign lesions. The mUSR was a simplified version of the International Ovarian Tumor Analysis (IOTA) using a binary classification of adnexal masses into benign/suspicious for malignancy.</div></div><div><h3>MATERIALS AND METHODS</h3><div>multisite retrospective study was conducted including patients with pathology-proven adnexal masses between January 2008 and December 2018. All ultrasound (US) video clips reviewed by an experienced radiologist with randomly selected subset were reviewed by two additional radiologists. Areas under receiver operator characteristic curves (AUCs) were compared without and with CA-125.</div></div><div><h3>RESULTS</h3><div>791 ovarian masses in 765 patients (26 bilateral) (mean age: 44 ± 15 years) (628 benign, 49 borderline, and 114 malignancies) demonstrated malignancy rates of 0.3%, 3.0%, 24.9%, and 82.4% for O-RADS 2, 3, 4, and 5, respectively. O-RADS and mUSR had a sensitivity of 0.96 (confidence interval [CI]: 0.92–0.99) and 0.96 (CI: 0.91–0.98), negative predictive values (NPVs) of 0.99 (CI: 0.97–1.00) and 0.99 (CI: 0.98–1.00) (<em>P></em>0.05), specificities 0.75 [CI: 0.71–0.78] and 0.88 [CI: 0.85–0.91], and positive predictive values (PPVs) 0.50 (CI: 0.44–0.55) and 0.68 (CI: 0.61–0.74) (<em>P<</em>0.01), respectively. The AUC was 0.855 for O-RADS and 0.920 for mUSR (<em>P</em>=0.005). Interobserver agreement was excellent across all readers for mUSR benign versus mUSR malignant and O-RADS 2/3 versus O-RADS 4/5 (kappa > 0.86). CA 125 improved performance of mUSR (<em>P</em>=0.002) and O-RADS (<em>P</em>=0.005) only in perimenopausal/postmenopausal patients.</div></div><div><h3>CONCLUSION</h3><div>O-RADS and mUSR both with high sensitivity and NPV for detection of ovarian malignancy but mUSR with significantly higher specificity and PPV than O-RADS. This finding endorses the American College of Radiology (ACR) recommendation for expert sonologist consultation for O-RADS 3 and 4.</div></div>","PeriodicalId":10695,"journal":{"name":"Clinical radiology","volume":"92 ","pages":"Article 107180"},"PeriodicalIF":1.9,"publicationDate":"2025-11-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145836363","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}
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