Article
Radiology, Nuclear Medicine & Medical Imaging
Anushri Parakh, Jinjin Cao, Theodore T. Pierce, Michael A. Blake, Cristy A. Savage, Avinash R. Kambadakone
Summary: The sinogram-based deep learning image reconstructions were both subjectively and objectively preferred over iterative reconstruction due to improved image quality and lower noise, even in large patients. DLIR-H had the best objective scores, indicating potential for clinical use and radiation dose reduction.
EUROPEAN RADIOLOGY
(2021)
Article
Radiology, Nuclear Medicine & Medical Imaging
Hye Joo Park, Seo-Youn Choi, Ji Eun Lee, Sanghyeok Lim, Min Hee Lee, Boem Ha Yi, Jang Gyu Cha, Ji Hye Min, Bora Lee, Yunsub Jung
Summary: This study compared the image quality and radiation dose of a deep learning image reconstruction algorithm (DLIR) with iterative reconstruction (IR) and filtered back projection (FBP) at different tube voltages and tube currents. The results showed that DLIR significantly reduced noise and artifacts and improved overall image quality compared to FBP and hybrid IR. Despite the reduced image sharpness, low-dose CT with DLIR seemed to have a greater potential for dose optimization.
EUROPEAN RADIOLOGY
(2022)
Article
Medicine, General & Internal
Usman Mahmood, David D. B. Bates, Yusuf E. Erdi, Lorenzo Mannelli, Giuseppe Corrias, Christopher Kanan
Summary: This study demonstrates the value of using deep learning to map single energy CT scans to synthetic dual-energy CT material density iodine scans for liver segmentation. The results show that training with synthetic DECT scans can achieve good segmentation accuracy with less data on both the held-out and generalization test sets.
Article
Radiology, Nuclear Medicine & Medical Imaging
Florian Jungmann, Lukas Mueller, Felix Hahn, Maximilian Weustenfeld, Ann-Kathrin Dapper, Aline Maehringer-Kunz, Dirk Graafen, Christoph Dueber, Darius Schafigh, Daniel Pinto dos Santos, Peter Mildenberger, Roman Kloeckner
Summary: This study evaluated the performance of commercial AI solutions in differentiating COVID-19 pneumonia from other lung conditions, revealing variable specificity and low positive predictive value. One solution showed acceptable sensitivity values, suggesting the potential integration of commercial AI solutions as alert tools in clinical routine workflow with further improvement. Randomized trials are needed to fully assess the benefits and risks of AI in image analysis.
EUROPEAN RADIOLOGY
(2022)
Article
Engineering, Biomedical
Alena-Kathrin Golla, Dominik F. Bauer, Ralf Schmidt, Tom Russ, Dominik Norenberg, Khanlian Chung, Christian Toennes, Lothar R. Schad, Frank G. Zoellner
Summary: The study presents an automatic method for extracting and differentiating abdominal blood vessel trees using convolutional neural networks (CNNs), achieving high accuracy and generalizability. Ensemble networks outperform individual CNNs and surpass state-of-the-art methods in segmentation of vascular structures.
IEEE TRANSACTIONS ON BIOMEDICAL ENGINEERING
(2021)
Article
Radiology, Nuclear Medicine & Medical Imaging
Qiao Zhu, Peishuai Che, Meijiao Li, Wei Guo, Kai Ye, Wenyu Yin, Dongheng Chu, Xiaohua Wang, Shufang Li
Summary: This study developed a multi-classification model using case-specific CT information to identify the occurrence of pneumonia, categorize the pneumonia type, and perform lesion segmentation. The AI model achieved high accuracy and precision in pneumonia classification and achieved substantial improvements in segmentation when labeled by senior physicians compared to junior physicians.
QUANTITATIVE IMAGING IN MEDICINE AND SURGERY
(2023)
Article
Veterinary Sciences
Yewon Ji, Hyunwoo Cho, Seungyeob Seon, Kichang Lee, Hakyoung Yoon
Summary: This study developed a deep learning model for automatic kidney detection and volume estimation from CT images of dogs, which showed high accuracy. Furthermore, a reference range for normal kidney volume considering body weight and body condition score was established, aiding in the assessment of kidney in dogs.
FRONTIERS IN VETERINARY SCIENCE
(2022)
Article
Computer Science, Information Systems
Chong Chen, Rui Li, Hong Shen, Liming Xia
Summary: This study conducted an automatic assessment of the disease progression of COVID-19 by combining chest CT scanning and laboratory tests using artificial intelligence methods. Features were extracted from CT imaging and laboratory tests using deep learning methods, and a neural network was used to classify the progression into three categories. The results showed that CT features outperformed laboratory features, and longitudinal assessment using combined features can better predict the progression of COVID-19.
Review
Clinical Neurology
Clement Brossard, Benjamin Lemasson, Arnaud Attye, Jules-Arnaud de Busschere, Jean-Francois Payen, Emmanuel L. Barbier, Jules Greze, Pierre Bouzat
Summary: Computed tomography (CT) has become the gold standard for diagnosing intracerebral lesions after traumatic brain injury (TBI) due to its accessibility and improved image quality. Recent advances in artificial intelligence (AI) have provided opportunities for clinicians to screen more patients with TBI and accurately assess the nature and extent of brain lesions.
FRONTIERS IN NEUROLOGY
(2021)
Article
Radiology, Nuclear Medicine & Medical Imaging
Taek Min Kim, Seung Jae Choi, Ji Yeon Ko, Sungwan Kim, Chang Wook Jeong, Jeong Yeon Cho, Sang Youn Kim, Young-Gon Kim
Summary: This study developed a fully automated deep learning model for adrenal segmentation and achieved good performance in classifying adrenal hyperplasia based on adrenal volume and anthropometric parameters. The proposed segmentation algorithm accurately segmented the adrenal glands on CT scans, which may aid clinicians in identifying possible cases of adrenal hyperplasia.
EUROPEAN RADIOLOGY
(2023)
Article
Radiology, Nuclear Medicine & Medical Imaging
Joel Greffier, Quentin Durand, Julien Frandon, Salim Si-Mohamed, Maeliss Loisy, Fabien de Oliveira, Jean-Paul Beregi, Djamel Dabli
Summary: This study assessed the impact of a new AI reconstruction algorithm (AI-DLR) on image quality in abdominal CT. The results showed that using AI-DLR improved image quality, detectability of lesions, and spatial resolution. The Smooth level was found to be the best compromise between the lowest dose level and adequate image quality.
EUROPEAN RADIOLOGY
(2023)
Article
Medicine, General & Internal
Salvatore Gitto, Renato Cuocolo, Alessio Annovazzi, Vincenzo Anelli, Marzia Acquasanta, Antonino Cincotta, Domenico Albano, Vito Chianca, Virginia Ferraresi, Carmelo Messina, Carmine Zoccali, Elisabetta Armiraglio, Antonina Parafioriti, Rosa Sciuto, Alessandro Luzzati, Roberto Biagini, Massimo Imbriaco, Luca Maria Sconfienza
Summary: This study investigated the performance of CT radiomics-based machine learning for classifying atypical cartilaginous tumors and higher-grade chondrosarcomas of long bones. The classifier showed good accuracy in identifying the lesions in both the training and external test cohorts, indicating its potential value in clinical practice. Preoperative biopsy had lower accuracy compared to the machine learning classifier, suggesting the potential benefit of this approach in improving diagnostic accuracy for these types of tumors.
Article
Radiology, Nuclear Medicine & Medical Imaging
Fanyang Meng, Jonathan Kottlors, Rahil Shahzad, Haifeng Liu, Philipp Fervers, Yinhua Jin, Miriam Rinneburger, Dou Le, Mathilda Weisthoff, Wenyun Liu, Mengzhe Ni, Ye Sun, Liying An, Xiaochen Huai, Dorottya More, Athanasios Giannakis, Isabel Kaltenborn, Andreas Bucher, David Maintz, Lei Zhang, Frank Thiele, Mingyang Li, Michael Perkuhn, Huimao Zhang, Thorsten Persigehl
Summary: This study aims to develop an AI algorithm for differentiating COVID-19 from CAP in CT scans and evaluate its performance. The results show that using AI assistance can improve radiologists' diagnostic accuracy and speed, as well as increase diagnostic confidence.
EUROPEAN RADIOLOGY
(2023)
Article
Radiology, Nuclear Medicine & Medical Imaging
Marly van Assen, Simon S. Martin, Akos Varga-Szemes, Saikiran Rapaka, Serkan Cimen, Puneet Sharma, Pooyan Sahbaee, Carlo N. De Cecco, Rozemarjin Vliegenthart, Tyler J. Leonard, Jeremy R. Burt, U. Joseph Schoepf
Summary: This study demonstrates a strong correlation between deep-learning based calcium quantification on chest CT scans and manual evaluation, as well as traditional Agatston score on Cardiac CT scans. The results show promising performance in risk classifications, suggesting potential for automating the calcium scoring process in clinical practice.
EUROPEAN JOURNAL OF RADIOLOGY
(2021)
Article
Radiology, Nuclear Medicine & Medical Imaging
Jack Junchi Xu, Lars Lonn, Esben Budtz-Jorgensen, Kristoffer L. Hansen, Peter S. Ulriksen
Summary: The study demonstrates that DLIR significantly reduces image noise compared to ASIR-V in DECT image reconstruction, with significant improvements in certain measurements. Additionally, qualitative assessment shows substantial enhancement in image quality with DLIR, especially in thin sliced images.
EUROPEAN RADIOLOGY
(2022)
Article
Oncology
Ariane Lapierre, Lionel Badet, Olivier Rouviere, Gilles Crehange, Julien Berthiller, Philippe Paparel, Olivier Chapet
Summary: The purpose of this study was to determine the maximum tolerated dose for primary renal cell carcinoma treated with stereotactic body radiation therapy (SBRT). A multicentric phase 1 study was conducted using dose escalation, and it was found that there was no dose-limiting toxicity at the prescribed dose levels. The highest dose of 48 Gy in 4 fractions was deemed safe for further studies.
PRACTICAL RADIATION ONCOLOGY
(2023)
Editorial Material
Urology & Nephrology
Johan Stranne, Nicolas Mottet, Olivier Rouviere
Review
Radiology, Nuclear Medicine & Medical Imaging
Olivier Rouviere, Tristan Jaouen, Pierre Baseilhac, Mohammed Lamine Benomar, Raphael Escande, Sebastien Crouzet, Remi Souchon
Summary: The purpose of this study was to review the diagnostic performance of AI-based algorithms for characterizing/detecting prostate cancer on MRI. The algorithms showed promising results and performed well as standalone diagnostic tools. However, the management of discrepancies between algorithm findings and human reading needs further investigation.
DIAGNOSTIC AND INTERVENTIONAL IMAGING
(2023)
Article
Radiology, Nuclear Medicine & Medical Imaging
Benoit Cosset, Monica Sigovan, Sara Boccalini, Fadi Farhat, Philippe Douek, Loic Boussel, Salim Aymeric Si-Mohamed
Summary: This study aimed to investigate the feasibility of identifying and characterizing the three most common types of endoleaks in a thoracic aorta aneurysm model using bicolor K-edge imaging and a biphasic contrast agent injection. The results showed that bicolor K-edge imaging allows for the bicolor characterization of thoracic aorta endoleaks in a single acquisition.
DIAGNOSTIC AND INTERVENTIONAL IMAGING
(2023)
Review
Radiology, Nuclear Medicine & Medical Imaging
Joel Greffier, Nicolas Villani, Didier Defez, Djamel Dabli, Salim Si-Mohamed
Summary: Spectral computed tomography (CT) imaging is a generation of CT systems that utilize the energy-dependent information in CT images. The introduction of dual-energy CT systems expanded this principle. The first generation of spectral CT systems overcame limitations of tissue characterization encountered with conventional CT and provided a new imaging approach. The global expansion of spectral CT is seen as an important leverage for extending CT towards multi-energy CT imaging based on photon counting CT systems.
DIAGNOSTIC AND INTERVENTIONAL IMAGING
(2023)
Review
Radiology, Nuclear Medicine & Medical Imaging
Salim A. A. Si-Mohamed, Sara Boccalini, Marjorie Villien, Yoad Yagil, Klaus Erhard, Loic Boussel, Philippe C. C. Douek
Summary: Spectral photon-counting computed tomography (SPCCT) technology shows many advantages over conventional CT imaging, such as better spatial resolution, greater dose efficiency, and superior tissue contrast. It also utilizes known approaches like virtual monochromatic imaging and material decomposition imaging, as well as a new approach called K-edge imaging. Due to its high potential, SPCCT systems are being researched in clinical settings.
INVESTIGATIVE RADIOLOGY
(2023)
Editorial Material
Cardiac & Cardiovascular Systems
Bertrand Scheppler, Salim A. Si-Mohamed, Simon Leboube, Fadi Farhat, Thomas Bochaton
EUROPEAN HEART JOURNAL-CARDIOVASCULAR IMAGING
(2023)
Article
Medicine, General & Internal
Salim A. Si-Mohamed, Lea Zumbihl, Segolene Turquier, Sara Boccalini, Jean-Francois Mornex, Philippe Douek, Vincent Cottin, Loic Boussel
Summary: This study aimed to evaluate quantitative lung perfusion blood volume (PBV) as a marker of severity in chronic thromboembolic pulmonary hypertension (CTEPH). The study found that quantitative PBV correlated positively with cardiac index (CI), while qualitative PBV showed no correlation. Therefore, quantitative PBV may serve as a non-invasive marker of severity in CTEPH patients.
Article
Respiratory System
Vincent Cottin, Elodie Blanchard, Mallorie Kerjouan, Romain Lazor, Martine Reynaud-Gaubert, Camille Taille, Yurdagul Uzunhanj, Lidwine Wemeau, Claire Andrejak, Dany Baud, Philippe Bonniaud, Pierre-Yves Brillet, Alain Calender, Lara Chalabreysse, Isabelle Court-Fortune, Nicolas Pierre Desbaillets, Gilbert Ferretti, Anne Guillemot, Laurane Hardelin, Marianne Kambouchner, Violette Leclerc, Mathieu Lederlin, Marie-Claire Malinge, Alain Mancel, Sylvain Marchand-Adam, Jean-Michel Maury, Jean-Marc Naccache, Mouhamad Nasser, Hilario Nunes, Gaele Pagnoux, Gregoire Prevot, Christine Rousset-Jablonski, Olivier Rouviere, Salim Si-Mohamed, Renaud Touraine, Julie Traclet, Segolene Turquier, Stephane Vagnarelli, Kais Ahmad
Summary: This article presents the English-language version of the French National Diagnostic and Care Protocol for lymphangioleiomyomatosis, a rare lung disease. The protocol provides practical recommendations for diagnosing, managing, and following up with patients. The recommendations cover various aspects of the disease, including diagnosis criteria, management principles, and use of pharmaceutical specialties and treatments.
RESPIRATORY MEDICINE AND RESEARCH
(2023)
Review
Urology & Nephrology
Giancarlo Marra, Geert J. L. H. van Leenders, Fabio Zattoni, Claudia Kesch, Pawel Rajwa, Philip Cornford, Theodorus van der Kwast, Roderick C. N. van den Bergh, Erik Briers, Thomas Van den Broeck, Gert De Meerleer, Maria De Santis, Daniel Eberli, Andrea Farolfi, Silke Gillessen, Nikolaos Grivas, Jeremy P. Grummet, Ann M. Henry, Michael Lardas, Matt Lieuw, Estefania Linares Espinos, Malcolm D. Mason, Shane O'Hanlon, Inge M. van Oort, Daniela E. Oprea-Lager, Guillaume Ploussard, Olivier Rouviere, Ivo. G. Schoots, Johan Stranne, Derya Tilki, Thomas Wiegel, Peter-Paul M. Willemse, Nicolas Mottet, Giorgio Gandaglia
Summary: This study aimed to compare the oncological outcomes of conventional prostate cancer and unconventional histologic subtypes in localized disease. Unconventional histologic subtypes, such as cribriform, intraductal, and ductal prostate cancer, had worse oncological outcomes compared to conventional prostate cancer. However, mucinous and prostatic intraepithelial neoplasia (PIN)-like subtypes showed similar outcomes to conventional prostate cancer.
Letter
Radiology, Nuclear Medicine & Medical Imaging
Lorraine Martineau, Arthur Branchu, Sara Boccalini, Loic Boussel, Philippe Douek, Salim A. Si-Mohamed
DIAGNOSTIC AND INTERVENTIONAL IMAGING
(2023)
Article
Radiology, Nuclear Medicine & Medical Imaging
Gregoire Cazalas, Clement Klein, Gilles Piana, Eric De Kerviler, Afshin Gangi, Philippe Puech, Cosmina Nedelcu, Remi Grange, Xavier Buy, Marc-Antoine Jegonday, Pierre Bigot, Charles Karim Bensalah, Victor Gaillard, Geraldine Pignot, Philippe Paparel, Lionel Badet, Clement Michiels, Jean Christophe Bernhard, Olivier Rouviere, Nicolas Grenier, Clement Marcelin
Summary: Percutaneous-guided thermal ablation and robotic-assisted partial nephrectomy are effective treatments for T1b renal cancer; however, thermal ablation has a higher local recurrence rate.
EUROPEAN RADIOLOGY
(2023)
Article
Radiology, Nuclear Medicine & Medical Imaging
Joel Greffier, Salim A. Si-Mohamed, Hugo Lacombe, Joey Labour, Djamel Djabli, Sara Boccalini, Mohammad Varasteh, Marjorie Villien, Yoad Yagil, Klaus Erhard, Loic Boussel, Jean-Paul Beregi, Philippe C. Douek
Summary: This study aimed to evaluate the quality of virtual monochromatic images (VMIs) from spectral photon-counting CT (SPCCT) and two energy-integrating detector dual-energy CT (EID-DECT) scanners from the same manufacturer, for the coronary lumen. The results showed that compared to the EID-DECT systems, HR and UHR-SPCCT images exhibited greater detectability of the coronary lumen for 40 to 90 keV VMIs, with higher lumen sharpness and overall quality.
EUROPEAN RADIOLOGY
(2023)
Article
Materials Science, Multidisciplinary
Yasmine Sebti, Salim Si-Mohamed, Rachida Aid, Frederic Geinguenaud, Mohand Chalal, Yoann Lalatonne, Frederic Chaubet, Phalla Ou, Laurence Motte
Summary: In this study, hafnium oxide nanoparticles were designed as a CT contrast agent for spectral photon counting computed tomography. The nanoparticles were surface functionalized with fucoidan, a sulfated polysaccharide that has high affinity for P-selectin, and citrate ions as a control ligand. The in vitro flow adhesion assay showed specific binding of the nanoparticles to activated platelets under arterial flow. SPECTCT experiments demonstrated that the nanoparticles had a higher absorption rate and a good correlation between the measured concentration and the references. Therefore, these functionalized hafnium oxide nanoparticles have the potential to be used for X-ray imaging and molecular-scale diagnosis of atherothrombosis.
MATERIALS ADVANCES
(2023)
Article
Radiology, Nuclear Medicine & Medical Imaging
Joel Greffier, Quentin Durand, Julien Frandon, Salim Si-Mohamed, Maeliss Loisy, Fabien de Oliveira, Jean-Paul Beregi, Djamel Dabli
Summary: This study assessed the impact of a new AI reconstruction algorithm (AI-DLR) on image quality in abdominal CT. The results showed that using AI-DLR improved image quality, detectability of lesions, and spatial resolution. The Smooth level was found to be the best compromise between the lowest dose level and adequate image quality.
EUROPEAN RADIOLOGY
(2023)