Clinical radiology最新文献_第6页

Synthetic diffusion-weighted imaging in prostate cancer diagnosis: a comparison study with different B-value combinations 不同b值组合的合成弥散加权成像在前列腺癌诊断中的比较研究

IF 2.1 3区医学 Q2 RADIOLOGY, NUCLEAR MEDICINE & MEDICAL IMAGING

Clinical radiology

Pub Date : 2025-02-01 DOI: 10.1016/j.crad.2024.106770

L. He , Z. Zhang , J. Zhang , J. Xia , Y. Wang , J. Zhu

AIM

To evaluate the impact of different b-value combinations on synthetic diffusion-weighted imaging (sDWI) and determine the sDWI with an optimal b-value combination for prostatic cancer (PCa) diagnosis.

MATERIAL AND METHODS

A retrospective analysis of 68 patients with abnormal prostate-specific antigen (PSA) was conducted. The sDWI images with b value of 1500 s/mm² were separately reconstructed by the following five b-value combinations: b=0, 200s/mm² (sDWI_0-200); b=600, 800s/mm² (sDWI_600-800); b=0, 600s/mm² (sDWI_0-600); b=200, 800s/mm² sDWI_200-800); b=0, 800s/mm² (sDWI_0-800). Quantitative analysis was performed on the acquired DWI (aDWI) images with b=1500s/mm² (aDWI₁₅₀₀) and all sDWI images. These six image groups were scored in five aspects for image quality and further reviewed by two radiologists via six protocols: Protocol Ⅰ, T₂WI+sDWI_0-200; Protocol Ⅱ, T₂WI+sDWI_600-800; Protocol Ⅲ, T₂WI+sDWI_0-600; Protocol Ⅳ, T₂WI+sDWI_200-800; Protocol Ⅴ, T₂WI+sDWI_0-800; Protocol Ⅵ, T₂WI+aDWI₁₅₀₀. The corresponding diagnostic efficacies for PCa were evaluated using receiver operating characteristic (ROC) curves.

RESULTS

Contrast ratio values of all sDWI images were higher than those of aDWI₁₅₀₀ images. Contrast-to-noise ratio values of sDWI_0-200 and sDWI_600-800 images were lower than those of the rest sDWI images. All subjective quality scores of sDWI_0-600, sDWI_200-800, and sDWI_0-800 were significantly higher than other groups except for background signal suppression. The area under the curve (AUC) of Protocol Ⅲ, Ⅳ, Ⅴ, and Ⅵ was significantly larger than those of other protocols.

CONCLUSION

Different b-value combinations impact the image quality and diagnostic accuracy of sDWI for PCa detection. The combination of b≤200s/mm² and b≥600s/mm² revealed to be optimal.

目的：评价不同b值组合对合成弥散加权成像（sDWI）的影响，确定诊断前列腺癌（PCa）的最佳b值组合sDWI。材料与方法：对68例前列腺特异性抗原（PSA）异常患者进行回顾性分析。对b值为1500 s/mm2的sDWI图像分别采用以下5种b值组合进行重建：b= 0,200 s/mm2 (sDWI0-200)；b= 600,800s /mm2 (sDWI600-800)；b= 0,600 s/mm2 (sDWI0-600)；b= 200,800s /mm2 (sDWI200-800)；b= 0,800s /mm2 （sDWI0-800）。对采集的b=1500s/mm2的DWI （aDWI）图像（aDWI1500）和所有sDWI图像进行定量分析。这六个图像组从五个方面对图像质量进行评分，并由两名放射科医生通过六个协议进行进一步审查：协议Ⅰ，T2WI+sDWI0-200；协议Ⅱ，T2WI+sDWI600-800；协议Ⅲ，T2WI+sDWI0-600；协议Ⅳ，T2WI+sDWI200-800；协议Ⅴ，T2WI+sDWI0-800；协议Ⅵ，T2WI+aDWI1500。采用受试者工作特征（ROC）曲线评价前列腺癌的诊断效果。结果：所有sDWI图像的对比度值均高于aDWI1500图像。sDWI0-200和sDWI600-800图像的噪比值低于其他sDWI图像。除背景信号抑制外，sDWI0-600、sDWI200-800、sDWI0-800主观质量得分均显著高于其他各组。方案Ⅲ、Ⅳ、Ⅴ和Ⅵ的曲线下面积（area under The curve， AUC）明显大于其他方案。结论：不同的b值组合会影响sDWI对PCa检测的图像质量和诊断准确性。b≤200s/mm2和b≥600s/mm2的组合效果最佳。

{"title":"Synthetic diffusion-weighted imaging in prostate cancer diagnosis: a comparison study with different B-value combinations","authors":"L. He , Z. Zhang , J. Zhang , J. Xia , Y. Wang , J. Zhu","doi":"10.1016/j.crad.2024.106770","DOIUrl":"10.1016/j.crad.2024.106770","url":null,"abstract":"<div><h3>AIM</h3><div>To evaluate the impact of different b-value combinations on synthetic diffusion-weighted imaging (sDWI) and determine the sDWI with an optimal b-value combination for prostatic cancer (PCa) diagnosis.</div></div><div><h3>MATERIAL AND METHODS</h3><div>A retrospective analysis of 68 patients with abnormal prostate-specific antigen (PSA) was conducted. The sDWI images with b value of 1500 s/mm<sup>2</sup> were separately reconstructed by the following five b-value combinations: b=0, 200s/mm<sup>2</sup> (sDWI<sub>0-200</sub>); b=600, 800s/mm<sup>2</sup> (sDWI<sub>600-800</sub>); b=0, 600s/mm<sup>2</sup> (sDWI<sub>0-600</sub>); b=200, 800s/mm<sup>2</sup> sDWI<sub>200-800</sub>); b=0, 800s/mm<sup>2</sup> (sDWI<sub>0-800</sub>). Quantitative analysis was performed on the acquired DWI (aDWI) images with b=1500s/mm<sup>2</sup> (aDWI<sub>1500</sub>) and all sDWI images. These six image groups were scored in five aspects for image quality and further reviewed by two radiologists via six protocols: Protocol Ⅰ, T<sub>2</sub>WI+sDWI<sub>0-200</sub>; Protocol Ⅱ, T<sub>2</sub>WI+sDWI<sub>600-800</sub>; Protocol Ⅲ, T<sub>2</sub>WI+sDWI<sub>0-600</sub>; Protocol Ⅳ, T<sub>2</sub>WI+sDWI<sub>200-800</sub>; Protocol Ⅴ, T<sub>2</sub>WI+sDWI<sub>0-800</sub>; Protocol Ⅵ, T<sub>2</sub>WI+aDWI<sub>1500</sub>. The corresponding diagnostic efficacies for PCa were evaluated using receiver operating characteristic (ROC) curves.</div></div><div><h3>RESULTS</h3><div>Contrast ratio values of all sDWI images were higher than those of aDWI<sub>1500</sub> images. Contrast-to-noise ratio values of sDWI<sub>0-200</sub> and sDWI<sub>600-800</sub> images were lower than those of the rest sDWI images. All subjective quality scores of sDWI<sub>0-600</sub>, sDWI<sub>200-800,</sub> and sDWI<sub>0-800</sub> were significantly higher than other groups except for background signal suppression. The area under the curve (AUC) of Protocol Ⅲ, Ⅳ, Ⅴ, and Ⅵ was significantly larger than those of other protocols.</div></div><div><h3>CONCLUSION</h3><div>Different b-value combinations impact the image quality and diagnostic accuracy of sDWI for PCa detection. The combination of b≤200s/mm<sup>2</sup> and b≥600s/mm<sup>2</sup> revealed to be optimal.</div></div>","PeriodicalId":10695,"journal":{"name":"Clinical radiology","volume":"81 ","pages":"Article 106770"},"PeriodicalIF":2.1,"publicationDate":"2025-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142906497","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Improvement of FDG PET/CT and MRI concordance in temporal lobe epilepsy pre-surgical assessment using statistical parametric mapping Z-scores 利用统计参数映射 Z 评分提高颞叶癫痫手术前评估中 FDG PET/CT 和 MRI 的一致性

IF 2.1 3区医学 Q2 RADIOLOGY, NUCLEAR MEDICINE & MEDICAL IMAGING

Clinical radiology

Pub Date : 2025-01-31 DOI: 10.1016/j.crad.2025.106838

M. Kershaw , X. Li , H. Amada , Y. Lu , J. Sawlani , S. Bose , V. Sawlani , S. Hughes

Aim

This retrospective study evaluates the diagnostic performance of statistical parametric mapping (SPM) analysis of interictal F18-fluoro-deoxy-D-glucose positron emission tomography computed tomography (FDG PET/CT) in temporal lobe epilepsy (TLE) patients, aiming to enhance image reporting consistency and correlation between magnetic resonance imaging (MRI) and FDG PET/CT findings and boost confidence in the surgical decision-making.

Materials and Methods

Forty-nine TLE patients undergoing MRI and FDG PET/CT imaging at a tertiary epilepsy service were included. Images were visually interpreted by an experienced radiologist and nuclear medicine physician. SPM-based quantitative analysis for FDG PET/CT including Z score asymmetric index (ZAI) was performed. Statistical analyses include receiver operating characteristic curve and Cohen’s k statistics.

Results

Significant differences in the standardised uptake value (SUV) ratio and ZAI were observed among left TLE, nonepilepsy, and right TLE (p < 0.01). The areas under the curves for left/nonleft and right/nonright groups were 0.838 and 0.780, respectively. The cutoff value to separate left TLE from nonepilepsy and right TLE was 0.305 with 89.7% sensitivity, 80.0% specificity, 94.6% positive predictive value (PPV), 66.7% negative predictive value (NPV), and a 0.697 Youden index for diagnosis. It was 0.190 to separate right TLE from the other 2 with 87.5% sensitivity, 75.6% specificity, 41.2% PPV, 96.9% NPV, and a 0.631 Youden index for diagnosis. The intermethod agreement between MRI and SUV ratio was moderate (k = 0.48; 95% CI, 0.32-0.65) and that between FDG PET/CT qualitative assessment and ZAI was moderate (k = 0.43; 95% CI, 0.10-0.76).

Conclusion

FDG PET/CT-based SUV ratios and ZAI show promising diagnostic value in TLE patients, facilitating the integration of FDG PET/CT practice into presurgical assessment for medically refractory epilepsy.

{"title":"Improvement of FDG PET/CT and MRI concordance in temporal lobe epilepsy pre-surgical assessment using statistical parametric mapping Z-scores","authors":"M. Kershaw , X. Li , H. Amada , Y. Lu , J. Sawlani , S. Bose , V. Sawlani , S. Hughes","doi":"10.1016/j.crad.2025.106838","DOIUrl":"10.1016/j.crad.2025.106838","url":null,"abstract":"<div><h3>Aim</h3><div>This retrospective study evaluates the diagnostic performance of statistical parametric mapping (SPM) analysis of interictal F18-fluoro-deoxy-D-glucose positron emission tomography computed tomography (FDG PET/CT) in temporal lobe epilepsy (TLE) patients, aiming to enhance image reporting consistency and correlation between magnetic resonance imaging (MRI) and FDG PET/CT findings and boost confidence in the surgical decision-making.</div></div><div><h3>Materials and Methods</h3><div>Forty-nine TLE patients undergoing MRI and FDG PET/CT imaging at a tertiary epilepsy service were included. Images were visually interpreted by an experienced radiologist and nuclear medicine physician. SPM-based quantitative analysis for FDG PET/CT including Z score asymmetric index (ZAI) was performed. Statistical analyses include receiver operating characteristic curve and Cohen’s k statistics.</div></div><div><h3>Results</h3><div>Significant differences in the standardised uptake value (SUV) ratio and ZAI were observed among left TLE, nonepilepsy, and right TLE (p < 0.01). The areas under the curves for left/nonleft and right/nonright groups were 0.838 and 0.780, respectively. The cutoff value to separate left TLE from nonepilepsy and right TLE was 0.305 with 89.7% sensitivity, 80.0% specificity, 94.6% positive predictive value (PPV), 66.7% negative predictive value (NPV), and a 0.697 Youden index for diagnosis. It was 0.190 to separate right TLE from the other 2 with 87.5% sensitivity, 75.6% specificity, 41.2% PPV, 96.9% NPV, and a 0.631 Youden index for diagnosis. The intermethod agreement between MRI and SUV ratio was moderate (k = 0.48; 95% CI, 0.32-0.65) and that between FDG PET/CT qualitative assessment and ZAI was moderate (k = 0.43; 95% CI, 0.10-0.76).</div></div><div><h3>Conclusion</h3><div>FDG PET/CT-based SUV ratios and ZAI show promising diagnostic value in TLE patients, facilitating the integration of FDG PET/CT practice into presurgical assessment for medically refractory epilepsy.</div></div>","PeriodicalId":10695,"journal":{"name":"Clinical radiology","volume":"83 ","pages":"Article 106838"},"PeriodicalIF":2.1,"publicationDate":"2025-01-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143487294","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}

引用次数: 0

Automated vertebral compression fracture detection and quantification on opportunistic CT scans: a performance evaluation

IF 2.1 3区医学 Q2 RADIOLOGY, NUCLEAR MEDICINE & MEDICAL IMAGING

Clinical radiology

Pub Date : 2025-01-30 DOI: 10.1016/j.crad.2025.106831

D. Guenoun , M.S. Quemeneur , A. Ayobi , C. Castineira , S. Quenet , J. Kiewsky , M. Mahfoud , C. Avare , Y. Chaibi , P. Champsaur

AIM

Since the majority of vertebral compression fractures (VCFs) are asymptomatic, they often go undetected on opportunistic CT scans. To reduce rates of undiagnosed osteoporosis, we developed a deep learning (DL)-based algorithm using 2D/3D U-Nets convolutional neural networks to opportunistically screen for VCF on CT scans. This study aimed to evaluate the performance of the algorithm using external real-world data.

Materials and Methods

CT scans acquired for various indications other than a suspicion of VCF from January 2019 to August 2020 were retrospectively and consecutively collected. The algorithm was designed to label each vertebra, detect VCF, measure vertebral height loss (VHL) and calculate mean Hounsfield Units (mean HU) for vertebral bone attenuation. For the ground truth, two board-certified radiologists defined if VCF was present and performed the measurements. The algorithm analyzed the scans and the results were compared to the experts' assessments.

Results

A total of 100 patients (mean age: 76.6 years ± 10.1[SD], 72% women) were evaluated. The overall labeling agreement was 94.9% (95%CI: 93.7%–95.9%). Regarding VHL, the 95% limits of agreement (LoA) between the algorithm and the radiologists was [-9.3, 8.6]; 94.1% of the differences lay within the radiologists' LoA and the intraclass correlation coefficient was 0.854 (95%CI: 0.822–0.881). For the mean HU, Pearson's correlation was 0.89 (95%CI: 0.84–0.92; p-value <0.0001). Finally, the algorithm's VCF screening sensitivity and specificity were 92.3% (95%CI: 81.5%–97.9%) and 91.7% (95%CI: 80.0%–97.7%), respectively.

Conclusions

This automated tool for screening and quantification of opportunistic VCF demonstrated high reliability and performance that may facilitate radiologists' task and improve opportunistic osteoporosis assessments.

目的由于大多数椎体压缩性骨折（VCFs）都没有症状，因此往往无法在CT扫描中及时发现。为了降低骨质疏松症的漏诊率，我们开发了一种基于深度学习（DL）的算法，利用二维/三维 U-Nets 卷积神经网络在 CT 扫描中伺机筛查 VCF。本研究旨在使用外部真实世界数据评估该算法的性能。材料与方法回顾性连续收集了 2019 年 1 月至 2020 年 8 月期间因各种适应症（除怀疑 VCF 外）获得的 CT 扫描。该算法旨在标记每个椎体、检测 VCF、测量椎体高度损失（VHL）并计算椎体骨衰减的平均 Hounsfield 单位（平均 HU）。对于地面实况，由两名获得认证的放射科医生确定是否存在 VCF 并进行测量。结果共评估了 100 名患者（平均年龄：76.6 岁 ± 10.1 [SD]，72% 为女性）。总体标记一致率为 94.9%（95%CI：93.7%-95.9%）。在 VHL 方面，算法与放射科医生之间 95% 的一致度（LoA）为[-9.3, 8.6]；94.1% 的差异在放射科医生的 LoA 范围内，类内相关系数为 0.854（95%CI：0.822-0.881）。对于平均 HU 值，皮尔逊相关系数为 0.89（95%CI：0.84-0.92；P 值为 0.0001）。最后，该算法的 VCF 筛查灵敏度和特异性分别为 92.3% (95%CI: 81.5%-97.9%) 和 91.7% (95%CI: 80.0%-97.7%) 。

{"title":"Automated vertebral compression fracture detection and quantification on opportunistic CT scans: a performance evaluation","authors":"D. Guenoun , M.S. Quemeneur , A. Ayobi , C. Castineira , S. Quenet , J. Kiewsky , M. Mahfoud , C. Avare , Y. Chaibi , P. Champsaur","doi":"10.1016/j.crad.2025.106831","DOIUrl":"10.1016/j.crad.2025.106831","url":null,"abstract":"<div><h3>AIM</h3><div>Since the majority of vertebral compression fractures (VCFs) are asymptomatic, they often go undetected on opportunistic CT scans. To reduce rates of undiagnosed osteoporosis, we developed a deep learning (DL)-based algorithm using 2D/3D U-Nets convolutional neural networks to opportunistically screen for VCF on CT scans. This study aimed to evaluate the performance of the algorithm using external real-world data.</div></div><div><h3>Materials and Methods</h3><div>CT scans acquired for various indications other than a suspicion of VCF from January 2019 to August 2020 were retrospectively and consecutively collected. The algorithm was designed to label each vertebra, detect VCF, measure vertebral height loss (VHL) and calculate mean Hounsfield Units (mean HU) for vertebral bone attenuation. For the ground truth, two board-certified radiologists defined if VCF was present and performed the measurements. The algorithm analyzed the scans and the results were compared to the experts' assessments.</div></div><div><h3>Results</h3><div>A total of 100 patients (mean age: 76.6 years ± 10.1[SD], 72% women) were evaluated. The overall labeling agreement was 94.9% (95%CI: 93.7%–95.9%). Regarding VHL, the 95% limits of agreement (LoA) between the algorithm and the radiologists was [-9.3, 8.6]; 94.1% of the differences lay within the radiologists' LoA and the intraclass correlation coefficient was 0.854 (95%CI: 0.822–0.881). For the mean HU, Pearson's correlation was 0.89 (95%CI: 0.84–0.92; p-value <0.0001). Finally, the algorithm's VCF screening sensitivity and specificity were 92.3% (95%CI: 81.5%–97.9%) and 91.7% (95%CI: 80.0%–97.7%), respectively.</div></div><div><h3>Conclusions</h3><div>This automated tool for screening and quantification of opportunistic VCF demonstrated high reliability and performance that may facilitate radiologists' task and improve opportunistic osteoporosis assessments.</div></div>","PeriodicalId":10695,"journal":{"name":"Clinical radiology","volume":"83 ","pages":"Article 106831"},"PeriodicalIF":2.1,"publicationDate":"2025-01-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143480727","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

A machine learning-based 18F-FDG PET/CT multi-modality fusion radiomics model to predict Mediastinal-Hilar lymph node metastasis in NSCLC: a multi-centre study

IF 2.1 3区医学 Q2 RADIOLOGY, NUCLEAR MEDICINE & MEDICAL IMAGING

Clinical radiology

Pub Date : 2025-01-29 DOI: 10.1016/j.crad.2025.106832

W. Zhai , X. Li , T. Zhou , Q. Zhou , X. Lin , X. Jiang , Z. Zhang , Q. Jin , S. Liu , L. Fan

AIM

To develop and validate a machine learning (ML) model based on positron emission tomography/computed tomography (PET/CT) multi-modality fusion radiomics to improve the prediction efficiency of mediastinal-hilar lymph node metastasis (LNM).

MATERIALS AND METHODS

Eighty-eight non-small cell lung cancer (NSCLC) patients with 559 LNs from centre 1 were divided into training and internal validation cohorts (7:3 ratio), and 75 patients with 543 LNs from centre 2 were assigned as external validation cohorts. PET and CT images were fused by wavelet transform. Multi-modality fusion radiomics features from six images of lymph nodes were extracted. The multi-modality fusion radiomics (MFR), multi-modality fusion radiomics + metabolic parameters (MFRM), CT, PET and PET + CT models were developed based on the best one among the 11 ML algorithms. The receiver operating characteristic (ROC) curve and the Delong test were used to assess and compare the performance of the models.

RESULTS

The CatBoost algorithm was chosen, and the MFR, MFRM, CT, PET and PET + CT models were constructed. The MFR and MFRM models showed a high AUC for predicting LNM in centre 1 (AUC = 0.950 and 0.952) and centre 2 (AUC = 0.923 and 0.927), and there were significant differences in centre 2 (P=0.036). The diagnostic efficacy of MFR and MFRM models was significantly higher than CT, PET, PET + CT models and SUVmax≥3.5 (P<0.001). The MFRM prediction was statistically different from the MFR prediction in the hilar/interlobar zone.

CONCLUSION

Both the MFR and MFRM models based on multi-modality fusion radiomics showed great potential for non-invasively predicting mediastinal-hilar LNM in NSCLC.

{"title":"A machine learning-based 18F-FDG PET/CT multi-modality fusion radiomics model to predict Mediastinal-Hilar lymph node metastasis in NSCLC: a multi-centre study","authors":"W. Zhai , X. Li , T. Zhou , Q. Zhou , X. Lin , X. Jiang , Z. Zhang , Q. Jin , S. Liu , L. Fan","doi":"10.1016/j.crad.2025.106832","DOIUrl":"10.1016/j.crad.2025.106832","url":null,"abstract":"<div><h3>AIM</h3><div>To develop and validate a machine learning (ML) model based on positron emission tomography/computed tomography (PET/CT) multi-modality fusion radiomics to improve the prediction efficiency of mediastinal-hilar lymph node metastasis (LNM).</div></div><div><h3>MATERIALS AND METHODS</h3><div>Eighty-eight non-small cell lung cancer (NSCLC) patients with 559 LNs from centre 1 were divided into training and internal validation cohorts (7:3 ratio), and 75 patients with 543 LNs from centre 2 were assigned as external validation cohorts. PET and CT images were fused by wavelet transform. Multi-modality fusion radiomics features from six images of lymph nodes were extracted. The multi-modality fusion radiomics (MFR), multi-modality fusion radiomics + metabolic parameters (MFRM), CT, PET and PET + CT models were developed based on the best one among the 11 ML algorithms. The receiver operating characteristic (ROC) curve and the Delong test were used to assess and compare the performance of the models.</div></div><div><h3>RESULTS</h3><div>The CatBoost algorithm was chosen, and the MFR, MFRM, CT, PET and PET + CT models were constructed. The MFR and MFRM models showed a high AUC for predicting LNM in centre 1 (AUC = 0.950 and 0.952) and centre 2 (AUC = 0.923 and 0.927), and there were significant differences in centre 2 (<em>P</em>=0.036). The diagnostic efficacy of MFR and MFRM models was significantly higher than CT, PET, PET + CT models and SUVmax≥3.5 (<em>P</em><0.001). The MFRM prediction was statistically different from the MFR prediction in the hilar/interlobar zone.</div></div><div><h3>CONCLUSION</h3><div>Both the MFR and MFRM models based on multi-modality fusion radiomics showed great potential for non-invasively predicting mediastinal-hilar LNM in NSCLC.</div></div>","PeriodicalId":10695,"journal":{"name":"Clinical radiology","volume":"83 ","pages":"Article 106832"},"PeriodicalIF":2.1,"publicationDate":"2025-01-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143446122","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}

引用次数: 0

Phase-contrast magnetic resonance imaging-based predictive modelling for surgical outcomes in patients with Chiari malformation type 1 with syringomyelia: a machine learning study

IF 2.1 3区医学 Q2 RADIOLOGY, NUCLEAR MEDICINE & MEDICAL IMAGING

Clinical radiology

Pub Date : 2025-01-29 DOI: 10.1016/j.crad.2025.106829

H. Wang , L. Lu , B. Fan , X. Xiao

Aim

Prospective outcome prediction plays a crucial role in guiding preoperative decision-making in patients with Chiari malformation type I (CM-Ⅰ) with syringomyelia. Here, we aimed to develop a predictive model for postoperative outcomes in patients with CM-Ⅰ with syringomyelia by integrating clinical and radiological parameters.

Materials and Methods

We retrospectively analysed the data of 151 adult patients diagnosed with CM-I with syringomyelia who underwent posterior fossa decompression surgery. Clinical outcomes were assessed using the Chicago Chiari Outcome Scale (CCOS). Predictors were investigated using bivariate and multiple linear regression analyses. Five factors were used to build seven independent machine learning (ML) models: Cat Boost classifier (CatBoost), random forest, light gradient boosting machine, decision tree classifier, logistic regression, K neighbours classifier, and support vector machine. The dataset was randomly divided into training (n = 121, 80%) and test (n = 30, 20%) sets. Model performance was evaluated using precision, recall, F-1 score, and area under the curve (AUC). Shapley additive explanations (SHAP) was used to interpret the feature significance.

Results

The best independent model was the CatBoost model, with an AUC of 0.9583 and an accuracy of 0.9097. The cross-validation results indicated that the accuracy of the CatBoost model was 0.8667. The SHAP plot revealed the important ranking of the features affecting the CCOS score as syrinx diameter, preoperative symptom duration, gait instability, peak diastolic velocity at the foramen magnum, and age.

Conclusion

We successfully developed a model to predict the prognosis of patients with CM-Ⅰ with syringomyelia after posterior fossa decompression.

{"title":"Phase-contrast magnetic resonance imaging-based predictive modelling for surgical outcomes in patients with Chiari malformation type 1 with syringomyelia: a machine learning study","authors":"H. Wang , L. Lu , B. Fan , X. Xiao","doi":"10.1016/j.crad.2025.106829","DOIUrl":"10.1016/j.crad.2025.106829","url":null,"abstract":"<div><h3>Aim</h3><div>Prospective outcome prediction plays a crucial role in guiding preoperative decision-making in patients with Chiari malformation type I (CM-Ⅰ) with syringomyelia. Here, we aimed to develop a predictive model for postoperative outcomes in patients with CM-Ⅰ with syringomyelia by integrating clinical and radiological parameters.</div></div><div><h3>Materials and Methods</h3><div>We retrospectively analysed the data of 151 adult patients diagnosed with CM-I with syringomyelia who underwent posterior fossa decompression surgery. Clinical outcomes were assessed using the Chicago Chiari Outcome Scale (CCOS). Predictors were investigated using bivariate and multiple linear regression analyses. Five factors were used to build seven independent machine learning (ML) models: Cat Boost classifier (CatBoost), random forest, light gradient boosting machine, decision tree classifier, logistic regression, K neighbours classifier, and support vector machine. The dataset was randomly divided into training (n = 121, 80%) and test (n = 30, 20%) sets. Model performance was evaluated using precision, recall, F-1 score, and area under the curve (AUC). Shapley additive explanations (SHAP) was used to interpret the feature significance.</div></div><div><h3>Results</h3><div>The best independent model was the CatBoost model, with an AUC of 0.9583 and an accuracy of 0.9097. The cross-validation results indicated that the accuracy of the CatBoost model was 0.8667. The SHAP plot revealed the important ranking of the features affecting the CCOS score as syrinx diameter, preoperative symptom duration, gait instability, peak diastolic velocity at the foramen magnum, and age.</div></div><div><h3>Conclusion</h3><div>We successfully developed a model to predict the prognosis of patients with CM-Ⅰ with syringomyelia after posterior fossa decompression.</div></div>","PeriodicalId":10695,"journal":{"name":"Clinical radiology","volume":"83 ","pages":"Article 106829"},"PeriodicalIF":2.1,"publicationDate":"2025-01-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143454673","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}

引用次数: 0

A deep learning pipeline for systematic and accurate vertebral fracture reporting in computed tomography 用于在计算机断层扫描中系统、准确地报告脊椎骨折的深度学习管道

IF 2.1 3区医学 Q2 RADIOLOGY, NUCLEAR MEDICINE & MEDICAL IMAGING

Clinical radiology

Pub Date : 2025-01-25 DOI: 10.1016/j.crad.2025.106827

C. Glessgen , J. Cyriac , S. Yang , S. Manneck , H. Wichtmann , A.M. Fischer , H.-C. Breit , D. Harder

AIM

Spine fractures are a frequent and relevant diagnosis, but systematic documentation is time-consuming and sometimes overlooked. A deep learning pipeline for opportunistic fracture detection in computed tomography (CT) spine images of varying field-of-views is introduced.

MATERIALS AND METHODS

This retrospective study builds on 452 CTs of the lumbar/thoracolumbar spine. Patients were included based on the evidence of ≥1 vertebral body fracture and excluded in case of history of spinal surgery or pathologic fractures. The collective was split into training/validation (405) and test (47) sets. An open-source spine dataset was used to train a preliminary segmentation model, which was applied on the training set. The resulting segmentation was post-processed to remove posterior vertebral structures and if needed, manually refined by a radiologist. Using the refined version as new training data, a final segmentation nnU-net was trained. Sagittal slices from each vertebra were labelled individually with regard to fracture evidence. Slices without fracture were used as negative class. Twenty seven thousand nineteen slices (20,396 negative, 6,623 positive) trained a classification algorithm using resnet18. Two senior readers independently assessed fractures in the test set to obtain a consensual ground truth. The segmentation-classification pipeline was applied to the test set and compared with the ground truth.

Results

The segmentation model correctly segmented 330/339 (97%) vertebrae. Considering every segmented vertebra, the classifier detected fractures with 88% sensitivity, 95% specificity, and 93% accuracy.

CONCLUSION

A deep learning pipeline was built and shown to accurately detect fractures on CT images. The final models as well as our code material are available at https://github.com/usb-radiology/VertebraeFx.

{"title":"A deep learning pipeline for systematic and accurate vertebral fracture reporting in computed tomography","authors":"C. Glessgen , J. Cyriac , S. Yang , S. Manneck , H. Wichtmann , A.M. Fischer , H.-C. Breit , D. Harder","doi":"10.1016/j.crad.2025.106827","DOIUrl":"10.1016/j.crad.2025.106827","url":null,"abstract":"<div><h3>AIM</h3><div>Spine fractures are a frequent and relevant diagnosis, but systematic documentation is time-consuming and sometimes overlooked. A deep learning pipeline for opportunistic fracture detection in computed tomography (CT) spine images of varying field-of-views is introduced.</div></div><div><h3>MATERIALS AND METHODS</h3><div>This retrospective study builds on 452 CTs of the lumbar/thoracolumbar spine. Patients were included based on the evidence of ≥1 vertebral body fracture and excluded in case of history of spinal surgery or pathologic fractures. The collective was split into training/validation (405) and test (47) sets. An open-source spine dataset was used to train a preliminary segmentation model, which was applied on the training set. The resulting segmentation was post-processed to remove posterior vertebral structures and if needed, manually refined by a radiologist. Using the refined version as new training data, a final segmentation nnU-net was trained. Sagittal slices from each vertebra were labelled individually with regard to fracture evidence. Slices without fracture were used as negative class. Twenty seven thousand nineteen slices (20,396 negative, 6,623 positive) trained a classification algorithm using resnet18. Two senior readers independently assessed fractures in the test set to obtain a consensual ground truth. The segmentation-classification pipeline was applied to the test set and compared with the ground truth.</div></div><div><h3>Results</h3><div>The segmentation model correctly segmented 330/339 (97%) vertebrae. Considering every segmented vertebra, the classifier detected fractures with 88% sensitivity, 95% specificity, and 93% accuracy.</div></div><div><h3>CONCLUSION</h3><div>A deep learning pipeline was built and shown to accurately detect fractures on CT images. The final models as well as our code material are available at <span><span>https://github.com/usb-radiology/VertebraeFx</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":10695,"journal":{"name":"Clinical radiology","volume":"83 ","pages":"Article 106827"},"PeriodicalIF":2.1,"publicationDate":"2025-01-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143430040","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}

引用次数: 0

Application of Early Dynamic 18F-FDG PET/CT in T1 and T2 Cervical Cancer

IF 2.1 3区医学 Q2 RADIOLOGY, NUCLEAR MEDICINE & MEDICAL IMAGING

Clinical radiology

Pub Date : 2025-01-22 DOI: 10.1016/j.crad.2025.106816

Y. Cui , W. Zhou , B. Zhang, F. Li, Y. Lv

Aim

Early 10-minute dynamic F18-fluorodeoxyglucose positron emission tomography/computed tomography (ED ¹⁸F-FDG PET/CT) can assess blood flow characteristics in cancers such as hepatocellular carcinoma, pancreatic cancer and bladder cancer. However, its application in cervical cancer detection is unexplored. This study evaluated the role of ED ¹⁸F-FDG PET/CT in detecting cervical cancer and the correlation between maximum standardized uptake value (SUVmax) and tumour-to-background ratio (TBR) in patients with stages T1 and T2 cervical cancer.

Materials and Methods

Twenty-six patients with confirmed cervical cancer underwent both ED ¹⁸F-FDG PET/CT and 60-minute static whole-body ¹⁸F-FDG PET/CT (WB ¹⁸F-FDG PET/CT). SUVmax and TBR of cervical cancer lesions were compared between stages T1 and T2.

Results

ED ¹⁸F-FDG PET/CT demonstrated 100% sensitivity (n = 26/26) in detecting cervical cancer. The SUVmax of cervical cancer lesions in ED phases was significantly higher than that of non-cancerous myometrium. A significant positive correlation was observed between the SUVmax of cervical cancer lesions at ED phases (80 seconds and 180–600 seconds) and SUVmax at WB ¹⁸F-FDG PET/CT (P < 0.05). The TBR at ED phases (120 seconds and 240–600 seconds) showed a significant positive correlation with TBR at WB ¹⁸F-FDG PET/CT (P < 0.05). Comparisons between stages T1 and T2 showed significantly higher SUVmax and TBR at various ED phases (300, 360, 480–600 seconds) and WB phase in stage T2 (P < 0.05).

Conclusion

ED ¹⁸F-FDG PET/CT appears to have clinical value in diagnosing cervical cancer. The SUVmax and TBR obtained during the ED ¹⁸F-FDG PET/CT scan may help differentiate between stage T1 and stage T2 cervical cancer.

{"title":"Application of Early Dynamic 18F-FDG PET/CT in T1 and T2 Cervical Cancer","authors":"Y. Cui , W. Zhou , B. Zhang, F. Li, Y. Lv","doi":"10.1016/j.crad.2025.106816","DOIUrl":"10.1016/j.crad.2025.106816","url":null,"abstract":"<div><h3>Aim</h3><div>Early 10-minute dynamic F18-fluorodeoxyglucose positron emission tomography/computed tomography (ED <sup>18</sup>F-FDG PET/CT) can assess blood flow characteristics in cancers such as hepatocellular carcinoma, pancreatic cancer and bladder cancer. However, its application in cervical cancer detection is unexplored. This study evaluated the role of ED <sup>18</sup>F-FDG PET/CT in detecting cervical cancer and the correlation between maximum standardized uptake value (SUVmax) and tumour-to-background ratio (TBR) in patients with stages T1 and T2 cervical cancer.</div></div><div><h3>Materials and Methods</h3><div>Twenty-six patients with confirmed cervical cancer underwent both ED <sup>18</sup>F-FDG PET/CT and 60-minute static whole-body <sup>18</sup>F-FDG PET/CT (WB <sup>18</sup>F-FDG PET/CT). SUVmax and TBR of cervical cancer lesions were compared between stages T1 and T2.</div></div><div><h3>Results</h3><div>ED <sup>18</sup>F-FDG PET/CT demonstrated 100% sensitivity (n = 26/26) in detecting cervical cancer. The SUVmax of cervical cancer lesions in ED phases was significantly higher than that of non-cancerous myometrium. A significant positive correlation was observed between the SUVmax of cervical cancer lesions at ED phases (80 seconds and 180–600 seconds) and SUVmax at WB <sup>18</sup>F-FDG PET/CT (<em>P</em> < 0.05). The TBR at ED phases (120 seconds and 240–600 seconds) showed a significant positive correlation with TBR at WB <sup>18</sup>F-FDG PET/CT (<em>P</em> < 0.05). Comparisons between stages T1 and T2 showed significantly higher SUVmax and TBR at various ED phases (300, 360, 480–600 seconds) and WB phase in stage T2 (<em>P</em> < 0.05).</div></div><div><h3>Conclusion</h3><div>ED <sup>18</sup>F-FDG PET/CT appears to have clinical value in diagnosing cervical cancer. The SUVmax and TBR obtained during the ED <sup>18</sup>F-FDG PET/CT scan may help differentiate between stage T1 and stage T2 cervical cancer.</div></div>","PeriodicalId":10695,"journal":{"name":"Clinical radiology","volume":"83 ","pages":"Article 106816"},"PeriodicalIF":2.1,"publicationDate":"2025-01-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143422336","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}

引用次数: 0

Letter to the Editor: Comments to “The magnetic resonance imaging (MRI) features of intracranial lesions in myelin oligodendrocyte glycoprotein-immunoglobulin G–associated disease (MOGAD)”

IF 2.1 3区医学 Q2 RADIOLOGY, NUCLEAR MEDICINE & MEDICAL IMAGING

Clinical radiology

Pub Date : 2025-01-16 DOI: 10.1016/j.crad.2025.106812

H.C. Manazoğlu, M. Kürtüncü

引用次数: 0

Re: Utility of zero echo time (ZTE) sequence for assessing bony lesions of skull base and calvarium

IF 2.1 3区医学 Q2 RADIOLOGY, NUCLEAR MEDICINE & MEDICAL IMAGING

Clinical radiology

Pub Date : 2025-01-16 DOI: 10.1016/j.crad.2025.106811

M. Jackson

引用次数: 0

Different plaque types and its association with the volume and attenuation of pericoronary adipose tissue as assessed by coronary computed tomography angiography

IF 2.1 3区医学 Q2 RADIOLOGY, NUCLEAR MEDICINE & MEDICAL IMAGING

Clinical radiology

Pub Date : 2025-01-15 DOI: 10.1016/j.crad.2025.106814

Q. Sun , M. Jing , H. Xi , W. Ren , H. Zhu , Y. Wang , Q. Liu , J. Zhou

Aim

To explore the relationship between different plaque types and pericoronary adipose tissue (PCAT) volume and attenuation values in patients with stable coronary artery disease (CAD) based on coronary computed tomographic angiography (CCTA).

Materials and Methods

Three hundred twenty one patients with stable CAD who underwent CCTA from May 2022 to March 2023 were enrolled. Using semi-automatic software, PCAT volumes and CT attenuation values were measured around the plaque and in the segment and proximal coronary artery where the plaque was located. To compare whether there was a statistical difference in PCAT volume and attenuation values among different plaque types in the periplaque, the segment, and proximal coronary artery in which the plaque was positioned.

Results

In total, 552 lesions were included, with 299 calcified plaques (CPs), 174 noncalcified plaques (NCPs), and 79 mixed plaques (MPs). There were excellent agreements between the two radiologists regarding the measured PCAT volumes and attenuation values (all interclass correlation coefficients values > 0.80). The periplaque PCAT volume was larger in CPs and MPs than in NCPs (291.98[213.25,381.03] mm³ vs. 261.00[173.25,377.85] mm³ vs. 206.54[139.72,284.07] mm³, P < 0.05), and the PCAT attenuation values around the plaque and the segment in which the plaque was positioned were higher in NCPs and MPs compared with CPs (-73.00[-79.00,-68.00] HU vs. -76.00[-79.00,-71.00] HU vs. -85.00[-92.00,-80.00] HU, -81.72 ± 0.70 HU vs. –80.73 ± 1.03 HU vs. -84.31 ± 0.49 HU; P < 0.05).

Conclusion

PCAT volume and attenuation values differed significantly among different plaque types, and the differences are particularly significant in measurements around the plaque.

{"title":"Different plaque types and its association with the volume and attenuation of pericoronary adipose tissue as assessed by coronary computed tomography angiography","authors":"Q. Sun , M. Jing , H. Xi , W. Ren , H. Zhu , Y. Wang , Q. Liu , J. Zhou","doi":"10.1016/j.crad.2025.106814","DOIUrl":"10.1016/j.crad.2025.106814","url":null,"abstract":"<div><h3>Aim</h3><div>To explore the relationship between different plaque types and pericoronary adipose tissue (PCAT) volume and attenuation values in patients with stable coronary artery disease (CAD) based on coronary computed tomographic angiography (CCTA).</div></div><div><h3>Materials and Methods</h3><div>Three hundred twenty one patients with stable CAD who underwent CCTA from May 2022 to March 2023 were enrolled. Using semi-automatic software, PCAT volumes and CT attenuation values were measured around the plaque and in the segment and proximal coronary artery where the plaque was located. To compare whether there was a statistical difference in PCAT volume and attenuation values among different plaque types in the periplaque, the segment, and proximal coronary artery in which the plaque was positioned.</div></div><div><h3>Results</h3><div>In total, 552 lesions were included, with 299 calcified plaques (CPs), 174 noncalcified plaques (NCPs), and 79 mixed plaques (MPs). There were excellent agreements between the two radiologists regarding the measured PCAT volumes and attenuation values (all interclass correlation coefficients values > 0.80). The periplaque PCAT volume was larger in CPs and MPs than in NCPs (291.98[213.25,381.03] mm<sup>3</sup> vs. 261.00[173.25,377.85] mm<sup>3</sup> vs. 206.54[139.72,284.07] mm<sup>3</sup>, <em>P</em> < 0.05), and the PCAT attenuation values around the plaque and the segment in which the plaque was positioned were higher in NCPs and MPs compared with CPs (-73.00[-79.00,-68.00] HU vs. -76.00[-79.00,-71.00] HU vs. -85.00[-92.00,-80.00] HU, -81.72 ± 0.70 HU vs. –80.73 ± 1.03 HU vs. -84.31 ± 0.49 HU; <em>P</em> < 0.05).</div></div><div><h3>Conclusion</h3><div>PCAT volume and attenuation values differed significantly among different plaque types, and the differences are particularly significant in measurements around the plaque.</div></div>","PeriodicalId":10695,"journal":{"name":"Clinical radiology","volume":"84 ","pages":"Article 106814"},"PeriodicalIF":2.1,"publicationDate":"2025-01-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143577412","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}

引用次数: 0