Pub Date : 2024-08-07eCollection Date: 2024-01-01DOI: 10.3389/fradi.2024.1327406
Neil D Shah, Mayil Krishnam, Bharat Ambale Venkatesh, Fouzia Khan, Michele Smith, Darwin R Jones, Patrick Koon, Xianglun Mao, Martin A Janich, Anja C S Brau, Michael Salerno, Rajesh Dash, Frandics Chan, Phillip C Yang
Background: Cardiac magnetic resonance is a useful clinical tool to identify late gadolinium enhancement in heart failure patients with implantable electronic devices. Identification of LGE in patients with CIED is limited by artifact, which can be improved with a wide band radiofrequency pulse sequence.
Objective: The authors hypothesize that image quality of LGE images produced using wide-band pulse sequence in patients with devices is comparable to image quality produced using standard LGE sequences in patients without devices.
Methods: Two independent readers reviewed LGE images of 16 patients with CIED and 7 patients without intracardiac devices to assess for image quality, device-related artifact, and presence of LGE using the American Society of Echocardiography/American Heart Association 17 segment model of the heart on a 4-point Likert scale. The mean and standard deviation for image quality and artifact rating were determined. Inter-observer reliability was determined by calculating Cohen's kappa coefficient. Statistical significance was determined by T-test as a p {less than or equal to} 0.05 with a 95% confidence interval.
Results: All patients underwent CMR without any adverse events. Overall IQ of WB LGE images was significantly better in patients with devices compared to standard LGE in patients without devices (p = 0.001) with reduction in overall artifact rating (p = 0.05).
Conclusion: Our study suggests wide-band pulse sequence for LGE can be applied safely to heart failure patients with devices in detection of LV myocardial scar while maintaining image quality, reducing artifact, and following routine imaging protocol after intravenous gadolinium contrast administration.
{"title":"Wideband radiofrequency pulse sequence for evaluation of myocardial scar in patients with cardiac implantable devices.","authors":"Neil D Shah, Mayil Krishnam, Bharat Ambale Venkatesh, Fouzia Khan, Michele Smith, Darwin R Jones, Patrick Koon, Xianglun Mao, Martin A Janich, Anja C S Brau, Michael Salerno, Rajesh Dash, Frandics Chan, Phillip C Yang","doi":"10.3389/fradi.2024.1327406","DOIUrl":"10.3389/fradi.2024.1327406","url":null,"abstract":"<p><strong>Background: </strong>Cardiac magnetic resonance is a useful clinical tool to identify late gadolinium enhancement in heart failure patients with implantable electronic devices. Identification of LGE in patients with CIED is limited by artifact, which can be improved with a wide band radiofrequency pulse sequence.</p><p><strong>Objective: </strong>The authors hypothesize that image quality of LGE images produced using wide-band pulse sequence in patients with devices is comparable to image quality produced using standard LGE sequences in patients without devices.</p><p><strong>Methods: </strong>Two independent readers reviewed LGE images of 16 patients with CIED and 7 patients without intracardiac devices to assess for image quality, device-related artifact, and presence of LGE using the American Society of Echocardiography/American Heart Association 17 segment model of the heart on a 4-point Likert scale. The mean and standard deviation for image quality and artifact rating were determined. Inter-observer reliability was determined by calculating Cohen's kappa coefficient. Statistical significance was determined by <i>T</i>-test as a <i>p</i> {less than or equal to} 0.05 with a 95% confidence interval.</p><p><strong>Results: </strong>All patients underwent CMR without any adverse events. Overall IQ of WB LGE images was significantly better in patients with devices compared to standard LGE in patients without devices (<i>p</i> = 0.001) with reduction in overall artifact rating (<i>p</i> = 0.05).</p><p><strong>Conclusion: </strong>Our study suggests wide-band pulse sequence for LGE can be applied safely to heart failure patients with devices in detection of LV myocardial scar while maintaining image quality, reducing artifact, and following routine imaging protocol after intravenous gadolinium contrast administration.</p>","PeriodicalId":73101,"journal":{"name":"Frontiers in radiology","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-08-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11339872/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142037922","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-07-17DOI: 10.3389/fradi.2024.1403761
Warren A. Campbell, J.F.B. Chick, David S. Shin, M. Makary
Interventional radiology (IR) is a unique specialty that incorporates a diverse set of skills ranging from imaging, procedures, consultation, and patient management. Understanding how IR generates value to the healthcare system is important to review from various perspectives. IR specialists need to understand how to meet demands from various stakeholders to expand their practice improving patient care. Thus, this review discusses the domains of value contributed to medical systems and outlines the parameters of success. IR benefits five distinct parties: patients, practitioners, payers, employers, and innovators. Value to patients and providers is delivered through a wide set of diagnostic and therapeutic interventions. Payers and hospital systems financially benefit from the reduced cost in medical management secondary to fast patient recovery, outpatient procedures, fewer complications, and the prestige of offering diverse expertise for complex patients. Lastly, IR is a field of rapid innovation implementing new procedural technology and techniques. Overall, IR must actively advocate for further growth and influence in the medical field as their value continues to expand in multiple domains. Despite being a nascent specialty, IR has become indispensable to modern medical practice.
介入放射学(IR)是一门独特的专科,融合了成像、手术、会诊和患者管理等多种技能。了解 IR 如何为医疗保健系统创造价值,对于从不同角度审视这一问题非常重要。红外专家需要了解如何满足各利益相关方的需求,以扩大他们的业务范围,改善患者护理。因此,本综述讨论了为医疗系统创造价值的领域,并概述了成功的参数。投资者关系使患者、从业者、支付者、雇主和创新者这五个不同的方面受益。通过一系列广泛的诊断和治疗干预措施,为患者和医疗服务提供者创造价值。由于患者恢复快、门诊手术、并发症少,医疗管理成本降低,以及为复杂病人提供不同的专业技术而获得的声誉,支付方和医院系统也从中获益。最后,IR 是一个快速创新的领域,它采用了新的程序技术和工艺。总之,随着其在多个领域的价值不断扩大,红外技术必须积极倡导在医学领域的进一步发展和影响。尽管 IR 是一个新兴专业,但它已成为现代医疗实践中不可或缺的一部分。
{"title":"Value of interventional radiology and their contributions to modern medical systems","authors":"Warren A. Campbell, J.F.B. Chick, David S. Shin, M. Makary","doi":"10.3389/fradi.2024.1403761","DOIUrl":"https://doi.org/10.3389/fradi.2024.1403761","url":null,"abstract":"Interventional radiology (IR) is a unique specialty that incorporates a diverse set of skills ranging from imaging, procedures, consultation, and patient management. Understanding how IR generates value to the healthcare system is important to review from various perspectives. IR specialists need to understand how to meet demands from various stakeholders to expand their practice improving patient care. Thus, this review discusses the domains of value contributed to medical systems and outlines the parameters of success. IR benefits five distinct parties: patients, practitioners, payers, employers, and innovators. Value to patients and providers is delivered through a wide set of diagnostic and therapeutic interventions. Payers and hospital systems financially benefit from the reduced cost in medical management secondary to fast patient recovery, outpatient procedures, fewer complications, and the prestige of offering diverse expertise for complex patients. Lastly, IR is a field of rapid innovation implementing new procedural technology and techniques. Overall, IR must actively advocate for further growth and influence in the medical field as their value continues to expand in multiple domains. Despite being a nascent specialty, IR has become indispensable to modern medical practice.","PeriodicalId":73101,"journal":{"name":"Frontiers in radiology","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-07-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141829161","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 : 2024-06-28eCollection Date: 2024-01-01DOI: 10.3389/fradi.2024.1416672
Anouk S Verschuur, Chantal M W Tax, Martijn F Boomsma, Helen L Carlson, Gerda van Wezel-Meijler, Regan King, Alexander Leemans, Lara M Leijser
Purpose: The study aimed to (1) assess the feasibility constrained spherical deconvolution (CSD) tractography to reconstruct crossing fiber bundles with unsedated neonatal diffusion MRI (dMRI), and (2) demonstrate the impact of spatial and angular resolution and processing settings on tractography and derived quantitative measures.
Methods: For the purpose of this study, the term-equivalent dMRIs (single-shell b800, and b2000, both 5 b0, and 45 gradient directions) of two moderate-late preterm infants (with and without motion artifacts) from a local cohort [Brain Imaging in Moderate-late Preterm infants (BIMP) study; Calgary, Canada] and one infant from the developing human connectome project with high-quality dMRI (using the b2600 shell, comprising 20 b0 and 128 gradient directions, from the multi-shell dataset) were selected. Diffusion tensor imaging (DTI) and CSD tractography were compared on b800 and b2000 dMRI. Varying image resolution modifications, (pre-)processing and tractography settings were tested to assess their impact on tractography. Each experiment involved visualizing local modeling and tractography for the corpus callosum and corticospinal tracts, and assessment of morphological and diffusion measures.
Results: Contrary to DTI, CSD enabled reconstruction of crossing fibers. Tractography was susceptible to image resolution, (pre-) processing and tractography settings. In addition to visual variations, settings were found to affect streamline count, length, and diffusion measures (fractional anisotropy and mean diffusivity). Diffusion measures exhibited variations of up to 23%.
Conclusion: Reconstruction of crossing fiber bundles using CSD tractography with unsedated neonatal dMRI data is feasible. Tractography settings affected streamline reconstruction, warranting careful documentation of methods for reproducibility and comparison of cohorts.
{"title":"Feasibility study to unveil the potential: considerations of constrained spherical deconvolution tractography with unsedated neonatal diffusion brain MRI data.","authors":"Anouk S Verschuur, Chantal M W Tax, Martijn F Boomsma, Helen L Carlson, Gerda van Wezel-Meijler, Regan King, Alexander Leemans, Lara M Leijser","doi":"10.3389/fradi.2024.1416672","DOIUrl":"10.3389/fradi.2024.1416672","url":null,"abstract":"<p><strong>Purpose: </strong>The study aimed to (1) assess the feasibility constrained spherical deconvolution (CSD) tractography to reconstruct crossing fiber bundles with unsedated neonatal diffusion MRI (dMRI), and (2) demonstrate the impact of spatial and angular resolution and processing settings on tractography and derived quantitative measures.</p><p><strong>Methods: </strong>For the purpose of this study, the term-equivalent dMRIs (single-shell b800, and b2000, both 5 b0, and 45 gradient directions) of two moderate-late preterm infants (with and without motion artifacts) from a local cohort [Brain Imaging in Moderate-late Preterm infants (BIMP) study; Calgary, Canada] and one infant from the developing human connectome project with high-quality dMRI (using the b2600 shell, comprising 20 b0 and 128 gradient directions, from the multi-shell dataset) were selected. Diffusion tensor imaging (DTI) and CSD tractography were compared on b800 and b2000 dMRI. Varying image resolution modifications, (pre-)processing and tractography settings were tested to assess their impact on tractography. Each experiment involved visualizing local modeling and tractography for the corpus callosum and corticospinal tracts, and assessment of morphological and diffusion measures.</p><p><strong>Results: </strong>Contrary to DTI, CSD enabled reconstruction of crossing fibers. Tractography was susceptible to image resolution, (pre-) processing and tractography settings. In addition to visual variations, settings were found to affect streamline count, length, and diffusion measures (fractional anisotropy and mean diffusivity). Diffusion measures exhibited variations of up to 23%.</p><p><strong>Conclusion: </strong>Reconstruction of crossing fiber bundles using CSD tractography with unsedated neonatal dMRI data is feasible. Tractography settings affected streamline reconstruction, warranting careful documentation of methods for reproducibility and comparison of cohorts.</p>","PeriodicalId":73101,"journal":{"name":"Frontiers in radiology","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-06-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11239519/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141617762","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-06-04DOI: 10.3389/fradi.2024.1385424
Patrick Winter, Haben Berhane, Jackson E. Moore, M. Aristova, Teresa Reichl, Julian Wollenberg, Adam Richter, Kelly B. Jarvis, Abhinav Patel, Fan Caprio, Ramez Abdalla, S. Ansari, Michael Markl, Susanne Schnell
Intracranial 4D flow MRI enables quantitative assessment of hemodynamics in patients with intracranial atherosclerotic disease (ICAD). However, quantitative assessments are still challenging due to the time-consuming vessel segmentation, especially in the presence of stenoses, which can often result in user variability. To improve the reproducibility and robustness as well as to accelerate data analysis, we developed an accurate, fully automated segmentation for stenosed intracranial vessels using deep learning.154 dual-VENC 4D flow MRI scans (68 ICAD patients with stenosis, 86 healthy controls) were retrospectively selected. Manual segmentations were used as ground truth for training. For automated segmentation, deep learning was performed using a 3D U-Net. 20 randomly selected cases (10 controls, 10 patients) were separated and solely used for testing. Cross-sectional areas and flow parameters were determined in the Circle of Willis (CoW) and the sinuses. Furthermore, the flow conservation error was calculated. For statistical comparisons, Dice scores (DS), Hausdorff distance (HD), average symmetrical surface distance (ASSD), Bland-Altman analyses, and interclass correlations were computed using the manual segmentations from two independent observers as reference. Finally, three stenosis cases were analyzed in more detail by comparing the 4D flow-based segmentations with segmentations from black blood vessel wall imaging (VWI).Training of the network took approximately 10 h and the average automated segmentation time was 2.2 ± 1.0 s. No significant differences in segmentation performance relative to two independent observers were observed. For the controls, mean DS was 0.85 ± 0.03 for the CoW and 0.86 ± 0.06 for the sinuses. Mean HD was 7.2 ± 1.5 mm (CoW) and 6.6 ± 3.7 mm (sinuses). Mean ASSD was 0.15 ± 0.04 mm (CoW) and 0.22 ± 0.17 mm (sinuses). For the patients, the mean DS was 0.85 ± 0.04 (CoW) and 0.82 ± 0.07 (sinuses), the HD was 8.4 ± 3.1 mm (CoW) and 5.7 ± 1.9 mm (sinuses) and the mean ASSD was 0.22 ± 0.10 mm (CoW) and 0.22 ± 0.11 mm (sinuses). Small bias and limits of agreement were observed in both cohorts for the flow parameters. The assessment of the cross-sectional lumen areas in stenosed vessels revealed very good agreement (ICC: 0.93) with the VWI segmentation but a consistent overestimation (bias ± LOA: 28.1 ± 13.9%).Deep learning was successfully applied for fully automated segmentation of stenosed intracranial vasculatures using 4D flow MRI data. The statistical analysis of segmentation and flow metrics demonstrated very good agreement between the CNN and manual segmentation and good performance in stenosed vessels. To further improve the performance and generalization, more ICAD segmentations as well as other intracranial vascular pathologies will be considered in the future.
{"title":"Automated intracranial vessel segmentation of 4D flow MRI data in patients with atherosclerotic stenosis using a convolutional neural network","authors":"Patrick Winter, Haben Berhane, Jackson E. Moore, M. Aristova, Teresa Reichl, Julian Wollenberg, Adam Richter, Kelly B. Jarvis, Abhinav Patel, Fan Caprio, Ramez Abdalla, S. Ansari, Michael Markl, Susanne Schnell","doi":"10.3389/fradi.2024.1385424","DOIUrl":"https://doi.org/10.3389/fradi.2024.1385424","url":null,"abstract":"Intracranial 4D flow MRI enables quantitative assessment of hemodynamics in patients with intracranial atherosclerotic disease (ICAD). However, quantitative assessments are still challenging due to the time-consuming vessel segmentation, especially in the presence of stenoses, which can often result in user variability. To improve the reproducibility and robustness as well as to accelerate data analysis, we developed an accurate, fully automated segmentation for stenosed intracranial vessels using deep learning.154 dual-VENC 4D flow MRI scans (68 ICAD patients with stenosis, 86 healthy controls) were retrospectively selected. Manual segmentations were used as ground truth for training. For automated segmentation, deep learning was performed using a 3D U-Net. 20 randomly selected cases (10 controls, 10 patients) were separated and solely used for testing. Cross-sectional areas and flow parameters were determined in the Circle of Willis (CoW) and the sinuses. Furthermore, the flow conservation error was calculated. For statistical comparisons, Dice scores (DS), Hausdorff distance (HD), average symmetrical surface distance (ASSD), Bland-Altman analyses, and interclass correlations were computed using the manual segmentations from two independent observers as reference. Finally, three stenosis cases were analyzed in more detail by comparing the 4D flow-based segmentations with segmentations from black blood vessel wall imaging (VWI).Training of the network took approximately 10 h and the average automated segmentation time was 2.2 ± 1.0 s. No significant differences in segmentation performance relative to two independent observers were observed. For the controls, mean DS was 0.85 ± 0.03 for the CoW and 0.86 ± 0.06 for the sinuses. Mean HD was 7.2 ± 1.5 mm (CoW) and 6.6 ± 3.7 mm (sinuses). Mean ASSD was 0.15 ± 0.04 mm (CoW) and 0.22 ± 0.17 mm (sinuses). For the patients, the mean DS was 0.85 ± 0.04 (CoW) and 0.82 ± 0.07 (sinuses), the HD was 8.4 ± 3.1 mm (CoW) and 5.7 ± 1.9 mm (sinuses) and the mean ASSD was 0.22 ± 0.10 mm (CoW) and 0.22 ± 0.11 mm (sinuses). Small bias and limits of agreement were observed in both cohorts for the flow parameters. The assessment of the cross-sectional lumen areas in stenosed vessels revealed very good agreement (ICC: 0.93) with the VWI segmentation but a consistent overestimation (bias ± LOA: 28.1 ± 13.9%).Deep learning was successfully applied for fully automated segmentation of stenosed intracranial vasculatures using 4D flow MRI data. The statistical analysis of segmentation and flow metrics demonstrated very good agreement between the CNN and manual segmentation and good performance in stenosed vessels. To further improve the performance and generalization, more ICAD segmentations as well as other intracranial vascular pathologies will be considered in the future.","PeriodicalId":73101,"journal":{"name":"Frontiers in radiology","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-06-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141266346","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 : 2024-05-22DOI: 10.3389/fradi.2024.1357341
Lidia Luque, Karoline Skogen, Bradley J. MacIntosh, Kyrre E. Emblem, Christopher Larsson, David Bouget, Ragnhild Holden Helland, Ingerid Reinertsen, Ole Solheim, Till Schellhorn, Jonas Vardal, Eduardo E. M. Mireles, Einar O. Vik-Mo, Atle Bjørnerud
Standard treatment of patients with glioblastoma includes surgical resection of the tumor. The extent of resection (EOR) achieved during surgery significantly impacts prognosis and is used to stratify patients in clinical trials. In this study, we developed a U-Net-based deep-learning model to segment contrast-enhancing tumor on post-operative MRI exams taken within 72 h of resection surgery and used these segmentations to classify the EOR as either maximal or submaximal. The model was trained on 122 multiparametric MRI scans from our institution and achieved a mean Dice score of 0.52 ± 0.03 on an external dataset (n = 248), a performance on par with the interrater agreement between expert annotators as reported in literature. We obtained an EOR classification precision/recall of 0.72/0.78 on the internal test dataset (n = 462) and 0.90/0.87 on the external dataset. Furthermore, Kaplan-Meier curves were used to compare the overall survival between patients with maximal and submaximal resection in the internal test dataset, as determined by either clinicians or the model. There was no significant difference between the survival predictions using the model's and clinical EOR classification. We find that the proposed segmentation model is capable of reliably classifying the EOR of glioblastoma tumors on early post-operative MRI scans. Moreover, we show that stratification of patients based on the model's predictions offers at least the same prognostic value as when done by clinicians.
{"title":"Standardized evaluation of the extent of resection in glioblastoma with automated early post-operative segmentation","authors":"Lidia Luque, Karoline Skogen, Bradley J. MacIntosh, Kyrre E. Emblem, Christopher Larsson, David Bouget, Ragnhild Holden Helland, Ingerid Reinertsen, Ole Solheim, Till Schellhorn, Jonas Vardal, Eduardo E. M. Mireles, Einar O. Vik-Mo, Atle Bjørnerud","doi":"10.3389/fradi.2024.1357341","DOIUrl":"https://doi.org/10.3389/fradi.2024.1357341","url":null,"abstract":"Standard treatment of patients with glioblastoma includes surgical resection of the tumor. The extent of resection (EOR) achieved during surgery significantly impacts prognosis and is used to stratify patients in clinical trials. In this study, we developed a U-Net-based deep-learning model to segment contrast-enhancing tumor on post-operative MRI exams taken within 72 h of resection surgery and used these segmentations to classify the EOR as either maximal or submaximal. The model was trained on 122 multiparametric MRI scans from our institution and achieved a mean Dice score of 0.52 ± 0.03 on an external dataset (n = 248), a performance on par with the interrater agreement between expert annotators as reported in literature. We obtained an EOR classification precision/recall of 0.72/0.78 on the internal test dataset (n = 462) and 0.90/0.87 on the external dataset. Furthermore, Kaplan-Meier curves were used to compare the overall survival between patients with maximal and submaximal resection in the internal test dataset, as determined by either clinicians or the model. There was no significant difference between the survival predictions using the model's and clinical EOR classification. We find that the proposed segmentation model is capable of reliably classifying the EOR of glioblastoma tumors on early post-operative MRI scans. Moreover, we show that stratification of patients based on the model's predictions offers at least the same prognostic value as when done by clinicians.","PeriodicalId":73101,"journal":{"name":"Frontiers in radiology","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-05-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141112560","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 : 2024-05-21DOI: 10.3389/fradi.2024.1386906
Rachael Harkness, A. F. Frangi, K. Zucker, Nishant Ravikumar
This study is a retrospective evaluation of the performance of deep learning models that were developed for the detection of COVID-19 from chest x-rays, undertaken with the goal of assessing the suitability of such systems as clinical decision support tools.Models were trained on the National COVID-19 Chest Imaging Database (NCCID), a UK-wide multi-centre dataset from 26 different NHS hospitals and evaluated on independent multi-national clinical datasets. The evaluation considers clinical and technical contributors to model error and potential model bias. Model predictions are examined for spurious feature correlations using techniques for explainable prediction.Models performed adequately on NHS populations, with performance comparable to radiologists, but generalised poorly to international populations. Models performed better in males than females, and performance varied across age groups. Alarmingly, models routinely failed when applied to complex clinical cases with confounding pathologies and when applied to radiologist defined “mild” cases.This comprehensive benchmarking study examines the pitfalls in current practices that have led to impractical model development. Key findings highlight the need for clinician involvement at all stages of model development, from data curation and label definition, to model evaluation, to ensure that all clinical factors and disease features are appropriately considered during model design. This is imperative to ensure automated approaches developed for disease detection are fit-for-purpose in a clinical setting.
{"title":"Multi-centre benchmarking of deep learning models for COVID-19 detection in chest x-rays","authors":"Rachael Harkness, A. F. Frangi, K. Zucker, Nishant Ravikumar","doi":"10.3389/fradi.2024.1386906","DOIUrl":"https://doi.org/10.3389/fradi.2024.1386906","url":null,"abstract":"This study is a retrospective evaluation of the performance of deep learning models that were developed for the detection of COVID-19 from chest x-rays, undertaken with the goal of assessing the suitability of such systems as clinical decision support tools.Models were trained on the National COVID-19 Chest Imaging Database (NCCID), a UK-wide multi-centre dataset from 26 different NHS hospitals and evaluated on independent multi-national clinical datasets. The evaluation considers clinical and technical contributors to model error and potential model bias. Model predictions are examined for spurious feature correlations using techniques for explainable prediction.Models performed adequately on NHS populations, with performance comparable to radiologists, but generalised poorly to international populations. Models performed better in males than females, and performance varied across age groups. Alarmingly, models routinely failed when applied to complex clinical cases with confounding pathologies and when applied to radiologist defined “mild” cases.This comprehensive benchmarking study examines the pitfalls in current practices that have led to impractical model development. Key findings highlight the need for clinician involvement at all stages of model development, from data curation and label definition, to model evaluation, to ensure that all clinical factors and disease features are appropriately considered during model design. This is imperative to ensure automated approaches developed for disease detection are fit-for-purpose in a clinical setting.","PeriodicalId":73101,"journal":{"name":"Frontiers in radiology","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-05-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141114863","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 : 2024-05-01eCollection Date: 2024-01-01DOI: 10.3389/fradi.2024.1327050
M P Belfiore, R Zeccolini, P Roccatagliata, L Gallo, A Fabozzi, S Cappabianca
Aortofemoral bypass surgery is a common procedure for treating aortoiliac occlusive disease, also known as Leriche syndrome, which can cause lower extremity ischemic symptoms. Diagnostic imaging techniques play a crucial role in managing pseudoaneurysms (PSAs), with Duplex ultrasound and Computed Tomography-angiography (CTA) being effective tools for early diagnosis. Pseudoaneurysms (PSAs) present as pulsating masses with various symptoms, and prompt intervention is essential to avoid complications. A case report is presented involving an 82-year-old male who underwent aorto-bifemoral bypass surgery and later developed a pseudoaneurysm (PSA) of the left branch. Surgical treatment involved the removal of the pseudoaneurysm (PSA) and graft replacement. Other cases from the literature are also described, emphasizing the rarity and potential severity of non-anastomotic pseudoaneurysms (PSAs) in reconstructive vascular surgery. Periodic screening of patients who undergo reconstructive vascular surgery is crucial to detect pseudoaneurysms (PSAs) early and prevent complications. Asymptomatic pseudoaneurysms (PSAs) can grow significantly and become life-threatening if not identified in a timely manner. Regular post-operative imaging, such as annual Computed Tomography-angiography (CTA) and/or Duplex ultrasound, is recommended to ensure early diagnosis and appropriate management of complications.
{"title":"Case Report: False aneurysm as a late unusual complication of the aortofemoral bypass graft in a patient with critical leg ischemic symptoms: interesting case.","authors":"M P Belfiore, R Zeccolini, P Roccatagliata, L Gallo, A Fabozzi, S Cappabianca","doi":"10.3389/fradi.2024.1327050","DOIUrl":"10.3389/fradi.2024.1327050","url":null,"abstract":"<p><p>Aortofemoral bypass surgery is a common procedure for treating aortoiliac occlusive disease, also known as Leriche syndrome, which can cause lower extremity ischemic symptoms. Diagnostic imaging techniques play a crucial role in managing pseudoaneurysms (PSAs), with Duplex ultrasound and Computed Tomography-angiography (CTA) being effective tools for early diagnosis. Pseudoaneurysms (PSAs) present as pulsating masses with various symptoms, and prompt intervention is essential to avoid complications. A case report is presented involving an 82-year-old male who underwent aorto-bifemoral bypass surgery and later developed a pseudoaneurysm (PSA) of the left branch. Surgical treatment involved the removal of the pseudoaneurysm (PSA) and graft replacement. Other cases from the literature are also described, emphasizing the rarity and potential severity of non-anastomotic pseudoaneurysms (PSAs) in reconstructive vascular surgery. Periodic screening of patients who undergo reconstructive vascular surgery is crucial to detect pseudoaneurysms (PSAs) early and prevent complications. Asymptomatic pseudoaneurysms (PSAs) can grow significantly and become life-threatening if not identified in a timely manner. Regular post-operative imaging, such as annual Computed Tomography-angiography (CTA) and/or Duplex ultrasound, is recommended to ensure early diagnosis and appropriate management of complications.</p>","PeriodicalId":73101,"journal":{"name":"Frontiers in radiology","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11094235/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140946587","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-04-09DOI: 10.3389/fradi.2024.1335349
L. Abdulaal, A. Maiter, M. Salehi, M. Sharkey, T. Alnasser, Pankaj Garg, S. Rajaram, C. Hill, Christopher Johns, Alex Rothman, K. Dwivedi, D. Kiely, S. Alabed, Andrew J Swift
Background Chronic pulmonary embolism (PE) may result in pulmonary hypertension (CTEPH). Automated CT pulmonary angiography (CTPA) interpretation using artificial intelligence (AI) tools has the potential for improving diagnostic accuracy, reducing delays to diagnosis and yielding novel information of clinical value in CTEPH. This systematic review aimed to identify and appraise existing studies presenting AI tools for CTPA in the context of chronic PE and CTEPH. Methods MEDLINE and EMBASE databases were searched on 11 September 2023. Journal publications presenting AI tools for CTPA in patients with chronic PE or CTEPH were eligible for inclusion. Information about model design, training and testing was extracted. Study quality was assessed using compliance with the Checklist for Artificial Intelligence in Medical Imaging (CLAIM). Results Five studies were eligible for inclusion, all of which presented deep learning AI models to evaluate PE. First study evaluated the lung parenchymal changes in chronic PE and two studies used an AI model to classify PE, with none directly assessing the pulmonary arteries. In addition, a separate study developed a CNN tool to distinguish chronic PE using 2D maximum intensity projection reconstructions. While another study assessed a novel automated approach to quantify hypoperfusion to help in the severity assessment of CTEPH. While descriptions of model design and training were reliable, descriptions of the datasets used in training and testing were more inconsistent. Conclusion In contrast to AI tools for evaluation of acute PE, there has been limited investigation of AI-based approaches to characterising chronic PE and CTEPH on CTPA. Existing studies are limited by inconsistent reporting of the data used to train and test their models. This systematic review highlights an area of potential expansion for the field of AI in medical image interpretation. There is limited knowledge of A systematic review of artificial intelligence tools for chronic pulmonary embolism in CT. This systematic review provides an assessment on research that examined deep learning algorithms in detecting CTEPH on CTPA images, the number of studies assessing the utility of deep learning on CTPA in CTEPH was unclear and should be highlighted.
背景 慢性肺栓塞(PE)可能导致肺动脉高压(CTEPH)。使用人工智能(AI)工具对 CT 肺血管造影(CTPA)进行自动判读有可能提高诊断准确性、减少诊断延误并获得对 CTEPH 有临床价值的新信息。本系统性综述旨在识别和评估在慢性 PE 和 CTEPH 中使用 CTPA 人工智能工具的现有研究。方法 2023 年 9 月 11 日检索了 MEDLINE 和 EMBASE 数据库。符合纳入条件的期刊论文介绍了用于慢性 PE 或 CTEPH 患者 CTPA 的人工智能工具。提取了有关模型设计、训练和测试的信息。根据医学影像人工智能检查表(CLAIM)对研究质量进行评估。结果 有五项研究符合纳入条件,所有这些研究都采用了深度学习人工智能模型来评估肺栓塞。第一项研究评估了慢性 PE 的肺实质变化,两项研究使用人工智能模型对 PE 进行分类,但没有一项研究直接评估肺动脉。此外,另一项研究开发了一种 CNN 工具,利用二维最大强度投影重建来区分慢性 PE。而另一项研究则评估了一种量化低灌注的新型自动方法,以帮助评估 CTEPH 的严重程度。虽然对模型设计和训练的描述是可靠的,但对训练和测试所用数据集的描述却不一致。结论 与评估急性 PE 的人工智能工具不同,基于人工智能的方法对 CTPA 中慢性 PE 和 CTEPH 特征的研究还很有限。现有的研究受到用于训练和测试其模型的数据报告不一致的限制。本系统综述强调了人工智能在医学影像解读领域的潜在扩展领域。对 CT 中慢性肺栓塞人工智能工具的系统综述了解有限。本系统性综述对深度学习算法在 CTPA 图像上检测 CTEPH 的研究进行了评估,但评估深度学习在 CTEPH CTPA 上的实用性的研究数量并不明确,应予以强调。
{"title":"A systematic review of artificial intelligence tools for chronic pulmonary embolism on CT pulmonary angiography","authors":"L. Abdulaal, A. Maiter, M. Salehi, M. Sharkey, T. Alnasser, Pankaj Garg, S. Rajaram, C. Hill, Christopher Johns, Alex Rothman, K. Dwivedi, D. Kiely, S. Alabed, Andrew J Swift","doi":"10.3389/fradi.2024.1335349","DOIUrl":"https://doi.org/10.3389/fradi.2024.1335349","url":null,"abstract":"Background Chronic pulmonary embolism (PE) may result in pulmonary hypertension (CTEPH). Automated CT pulmonary angiography (CTPA) interpretation using artificial intelligence (AI) tools has the potential for improving diagnostic accuracy, reducing delays to diagnosis and yielding novel information of clinical value in CTEPH. This systematic review aimed to identify and appraise existing studies presenting AI tools for CTPA in the context of chronic PE and CTEPH. Methods MEDLINE and EMBASE databases were searched on 11 September 2023. Journal publications presenting AI tools for CTPA in patients with chronic PE or CTEPH were eligible for inclusion. Information about model design, training and testing was extracted. Study quality was assessed using compliance with the Checklist for Artificial Intelligence in Medical Imaging (CLAIM). Results Five studies were eligible for inclusion, all of which presented deep learning AI models to evaluate PE. First study evaluated the lung parenchymal changes in chronic PE and two studies used an AI model to classify PE, with none directly assessing the pulmonary arteries. In addition, a separate study developed a CNN tool to distinguish chronic PE using 2D maximum intensity projection reconstructions. While another study assessed a novel automated approach to quantify hypoperfusion to help in the severity assessment of CTEPH. While descriptions of model design and training were reliable, descriptions of the datasets used in training and testing were more inconsistent. Conclusion In contrast to AI tools for evaluation of acute PE, there has been limited investigation of AI-based approaches to characterising chronic PE and CTEPH on CTPA. Existing studies are limited by inconsistent reporting of the data used to train and test their models. This systematic review highlights an area of potential expansion for the field of AI in medical image interpretation. There is limited knowledge of A systematic review of artificial intelligence tools for chronic pulmonary embolism in CT. This systematic review provides an assessment on research that examined deep learning algorithms in detecting CTEPH on CTPA images, the number of studies assessing the utility of deep learning on CTPA in CTEPH was unclear and should be highlighted.","PeriodicalId":73101,"journal":{"name":"Frontiers in radiology","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-04-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140723848","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 : 2024-04-05DOI: 10.3389/fradi.2024.1283392
Matias Aiskovich, Eduardo Castro, Jenna M. Reinen, S. Fadnavis, Anushree Mehta, Hongyang Li, Amit Dhurandhar, Guillermo Cecchi, Pablo Polosecki
Data collection, curation, and cleaning constitute a crucial phase in Machine Learning (ML) projects. In biomedical ML, it is often desirable to leverage multiple datasets to increase sample size and diversity, but this poses unique challenges, which arise from heterogeneity in study design, data descriptors, file system organization, and metadata. In this study, we present an approach to the integration of multiple brain MRI datasets with a focus on homogenization of their organization and preprocessing for ML. We use our own fusion example (approximately 84,000 images from 54,000 subjects, 12 studies, and 88 individual scanners) to illustrate and discuss the issues faced by study fusion efforts, and we examine key decisions necessary during dataset homogenization, presenting in detail a database structure flexible enough to accommodate multiple observational MRI datasets. We believe our approach can provide a basis for future similarly-minded biomedical ML projects.
数据收集、整理和清理是机器学习(ML)项目的关键阶段。在生物医学 ML 中,通常希望利用多个数据集来增加样本量和多样性,但这带来了独特的挑战,这些挑战来自于研究设计、数据描述符、文件系统组织和元数据的异质性。在本研究中,我们介绍了一种整合多个脑磁共振成像数据集的方法,重点是对其组织和预处理进行同质化,以便进行多重L。我们使用自己的融合实例(来自 54,000 名受试者、12 项研究和 88 台独立扫描仪的约 84,000 张图像)来说明和讨论研究融合工作所面临的问题,并研究了数据集同质化过程中所需的关键决策,详细介绍了可灵活容纳多个观察性 MRI 数据集的数据库结构。我们相信,我们的方法可以为未来类似的生物医学 ML 项目奠定基础。
{"title":"Fusion of biomedical imaging studies for increased sample size and diversity: a case study of brain MRI","authors":"Matias Aiskovich, Eduardo Castro, Jenna M. Reinen, S. Fadnavis, Anushree Mehta, Hongyang Li, Amit Dhurandhar, Guillermo Cecchi, Pablo Polosecki","doi":"10.3389/fradi.2024.1283392","DOIUrl":"https://doi.org/10.3389/fradi.2024.1283392","url":null,"abstract":"Data collection, curation, and cleaning constitute a crucial phase in Machine Learning (ML) projects. In biomedical ML, it is often desirable to leverage multiple datasets to increase sample size and diversity, but this poses unique challenges, which arise from heterogeneity in study design, data descriptors, file system organization, and metadata. In this study, we present an approach to the integration of multiple brain MRI datasets with a focus on homogenization of their organization and preprocessing for ML. We use our own fusion example (approximately 84,000 images from 54,000 subjects, 12 studies, and 88 individual scanners) to illustrate and discuss the issues faced by study fusion efforts, and we examine key decisions necessary during dataset homogenization, presenting in detail a database structure flexible enough to accommodate multiple observational MRI datasets. We believe our approach can provide a basis for future similarly-minded biomedical ML projects.","PeriodicalId":73101,"journal":{"name":"Frontiers in radiology","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-04-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140738325","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 : 2024-03-18DOI: 10.3389/fradi.2024.1345465
Paul Manning, Shanmukha Srinivas, D. Bolar, Matthew K. Rajaratnam, David E. Piccioni, Carrie R. McDonald, J. Hattangadi-Gluth, N. Farid
Conventional contrast-enhanced MRI is currently the primary imaging technique used to evaluate radiation treatment response in meningiomas. However, newer perfusion-weighted MRI techniques, such as 3D pseudocontinuous arterial spin labeling (3D pCASL) MRI, capture physiologic information beyond the structural information provided by conventional MRI and may provide additional complementary treatment response information. The purpose of this study is to assess 3D pCASL for the evaluation of radiation-treated meningiomas.Twenty patients with meningioma treated with surgical resection followed by radiation, or by radiation alone, were included in this retrospective single-institution study. Patients were evaluated with 3D pCASL and conventional contrast-enhanced MRI before and after radiation (median follow up 6.5 months). Maximum pre- and post-radiation ASL normalized cerebral blood flow (ASL-nCBF) was measured within each meningioma and radiation-treated meningioma (or residual resected and radiated meningioma), and the contrast-enhancing area was measured for each meningioma. Wilcoxon signed-rank tests were used to compare pre- and post-radiation ASL-nCBF and pre- and post-radiation area.All treated meningiomas demonstrated decreased ASL-nCBF following radiation (p < 0.001). Meningioma contrast-enhancing area also decreased after radiation (p = 0.008) but only for approximately half of the meningiomas (9), while half (10) remained stable. A larger effect size (Wilcoxon signed-rank effect size) was seen for ASL-nCBF measurements (r = 0.877) compared to contrast-enhanced area measurements (r = 0.597).ASL perfusion may provide complementary treatment response information in radiation-treated meningiomas. This complementary information could aid clinical decision-making and provide an additional endpoint for clinical trials.
{"title":"Arterial spin labeled perfusion MRI for the assessment of radiation-treated meningiomas","authors":"Paul Manning, Shanmukha Srinivas, D. Bolar, Matthew K. Rajaratnam, David E. Piccioni, Carrie R. McDonald, J. Hattangadi-Gluth, N. Farid","doi":"10.3389/fradi.2024.1345465","DOIUrl":"https://doi.org/10.3389/fradi.2024.1345465","url":null,"abstract":"Conventional contrast-enhanced MRI is currently the primary imaging technique used to evaluate radiation treatment response in meningiomas. However, newer perfusion-weighted MRI techniques, such as 3D pseudocontinuous arterial spin labeling (3D pCASL) MRI, capture physiologic information beyond the structural information provided by conventional MRI and may provide additional complementary treatment response information. The purpose of this study is to assess 3D pCASL for the evaluation of radiation-treated meningiomas.Twenty patients with meningioma treated with surgical resection followed by radiation, or by radiation alone, were included in this retrospective single-institution study. Patients were evaluated with 3D pCASL and conventional contrast-enhanced MRI before and after radiation (median follow up 6.5 months). Maximum pre- and post-radiation ASL normalized cerebral blood flow (ASL-nCBF) was measured within each meningioma and radiation-treated meningioma (or residual resected and radiated meningioma), and the contrast-enhancing area was measured for each meningioma. Wilcoxon signed-rank tests were used to compare pre- and post-radiation ASL-nCBF and pre- and post-radiation area.All treated meningiomas demonstrated decreased ASL-nCBF following radiation (p < 0.001). Meningioma contrast-enhancing area also decreased after radiation (p = 0.008) but only for approximately half of the meningiomas (9), while half (10) remained stable. A larger effect size (Wilcoxon signed-rank effect size) was seen for ASL-nCBF measurements (r = 0.877) compared to contrast-enhanced area measurements (r = 0.597).ASL perfusion may provide complementary treatment response information in radiation-treated meningiomas. This complementary information could aid clinical decision-making and provide an additional endpoint for clinical trials.","PeriodicalId":73101,"journal":{"name":"Frontiers in radiology","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-03-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140234058","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}