Review
Radiology, Nuclear Medicine & Medical Imaging
Vito Chianca, Domenico Albano, Carmelo Messina, Gabriele Vincenzo, Stefania Rizzo, Filippo Del Grande, Luca Maria Sconfienza
Summary: In the last two decades, significant progress has been made in the diagnosis of musculoskeletal tumors with the development of new imaging tools and artificial intelligence software, which have advanced imaging techniques and provided advantages in clinical practice.
Article
Oncology
Maikel Verduin, Sergey Primakov, Inge Compter, Henry C. Woodruff, Sander M. J. van Kuijk, Bram L. T. Ramaekers, Maarten te Dorsthorst, Elles G. M. Revenich, Mark ter Laan, Sjoert A. H. Pegge, Frederick J. A. Meijer, Jan Beckervordersandforth, Ernst Jan Speel, Benno Kusters, Wendy W. J. de Leng, Monique M. Anten, Martijn P. G. Broen, Linda Ackermans, Olaf E. M. G. Schijns, Onno Teernstra, Koos Hovinga, Marc A. Vooijs, Vivianne C. G. Tjan-Heijnen, Danielle B. P. Eekers, Alida A. Postma, Philippe Lambin, Ann Hoeben
Summary: This study demonstrates the potential of combining MRI and clinical features for predicting prognosis in GBM patients, with promising but yet to be validated results in predicting tumor markers. Further optimization and prospective studies are needed.
Review
Radiology, Nuclear Medicine & Medical Imaging
Xiaoge Liu, Zhiqing Duan, Shaobo Fang, Shaowu Wang
Summary: Recent studies have found that MRI shows promising results in evaluating the effectiveness of chemotherapy in bone sarcomas. This article provides a review of current methods for evaluating the efficacy of malignant bone tumors and discusses the application of MRI in this field, highlighting the advantages and limitations of each technique.
JOURNAL OF MAGNETIC RESONANCE IMAGING
(2023)
Review
Biochemical Research Methods
Muhammad Mohsin Qureshi, Nader Allam, Jeongmyo Im, Hyuk-Sang Kwon, Euiheon Chung, I. Alex Vitkin
Summary: Laser speckle imaging (LSI) is a promising method for visualizing and quantifying blood vessels and tissue perfusion. It analyzes the speckle patterns generated by coherent light interacting with biological tissue. LSI techniques can be categorized into qualitative, semi-quantitative, and quantitative based on their degree of quantification. Qualitative LSI generates microvascular maps by capturing speckle contrast variations. Semi-quantitative techniques consider the effect of static scattering on spatiotemporal parameters for more accurate blood flow analysis. Quantitative LSI provides quantitative flow velocity measurements. The future prospects of LSI techniques for optimizing vascular flow quantification are also discussed.
JOURNAL OF BIOPHOTONICS
(2023)
Article
Multidisciplinary Sciences
Shengkun Peng, Lingai Pan, Yang Guo, Bo Gong, Xiaobo Huang, Siyun Liu, Jianxin Huang, Hong Pu, Jie Zeng
Summary: This study evaluates the effectiveness of image-based quantitative CT features in discriminating COVID-19 from non-COVID-19 pneumonia. The results show that radiomics features outperformed traditional CT quantitative features, with higher accuracy and discrimination. Radiomics features can significantly improve the differentiation between COVID-19 and non-COVID-19 pneumonia.
Review
Biochemistry & Molecular Biology
Matteo Ferro, Ottavio de Cobelli, Mihai Dorin Vartolomei, Giuseppe Lucarelli, Felice Crocetto, Biagio Barone, Alessandro Sciarra, Francesco Del Giudice, Matteo Muto, Martina Maggi, Giuseppe Carrieri, Gian Maria Busetto, Ugo Falagario, Daniela Terracciano, Luigi Cormio, Gennaro Musi, Octavian Sabin Tataru
Summary: Radiomics and genomics play crucial roles in prostate cancer research, enhancing clinical value through mathematical analysis and machine learning. Validation of recent findings in large, randomized cohorts can establish the role of radiogenomics in the future.
INTERNATIONAL JOURNAL OF MOLECULAR SCIENCES
(2021)
Article
Radiology, Nuclear Medicine & Medical Imaging
Hanna Muenzfeld, Claus Nowak, Stefanie Riedlberger, Alexander Hartenstein, Bernd Hamm, Paul Jahnke, Tobias Penzkofer
Summary: This study evaluated the impact of clinical imaging techniques on the stability of radiomics features using a 3D printed anthropomorphic CT phantom. Results showed that most radiomics features did not have sufficient intra-scanner repeatability to serve as reliable diagnostic tools.
EUROPEAN JOURNAL OF RADIOLOGY
(2021)
Review
Oncology
Hishan Tharmaseelan, Alexander Hertel, Shereen Rennebaum, Dominik Noerenberg, Verena Haselmann, Stefan O. Schoenberg, Matthias F. Froelich
Summary: Modern personalized therapy approaches are transforming advanced cancer into a chronic disease. Novel omics methodologies in molecular biology enable individual characterization of cancerous lesions. Combined with other non-invasive methods, such as liquid profiling, quantitative imaging biomarkers provide a more individual assessment of tumor biology and potential therapies. Emerging techniques for evaluating oncologic imaging are transitioning from traditional response assessment to comprehensive cancer characterization via imaging, allowing for truly personalized and optimized cancer diagnosis and treatment.
Article
Oncology
Divya Bhardwaj, Archya Dasgupta, Daniel DiCenzo, Stephen Brade, Kashuf Fatima, Karina Quiaoit, Maureen Trudeau, Sonal Gandhi, Andrea Eisen, Frances Wright, Nicole Look-Hong, Belinda Curpen, Lakshmanan Sannachi, Gregory J. Czarnota
Summary: This study explored the use of quantitative ultrasound (QUS) in predicting recurrence for patients with locally advanced breast cancer (LABC) early during neoadjuvant chemotherapy (NAC). The ultrasound imaging was analyzed using advanced computational techniques and artificial intelligence to develop radiomic models. The inclusion of texture derivatives in the analysis improved the performance of the classifier in predicting recurrence. The study highlights the utility of QUS radiomics in predicting clinical outcomes during the treatment of LABC.
Article
Computer Science, Artificial Intelligence
Hui Qu, Ruichuan Shi, Shuqin Li, Fengying Che, Jian Wu, Haoran Li, Weixing Chen, Hao Zhang, Zhi Li, Xiaoyu Cui
Summary: The traditional static radiology feature extraction method is insufficient to fully evaluate a patient's condition. This study proposes a new dynamic radiomics feature extraction method, which achieves higher accuracy and robustness in predicting different clinical questions.
APPLIED INTELLIGENCE
(2022)
Article
Radiology, Nuclear Medicine & Medical Imaging
Jia-wei Li, Yu-cheng Cao, Zhi-jin Zhao, Zhao-ting Shi, Xiao-qian Duan, Cai Chang, Jian-gang Chen
Summary: Through a retrospective review of 252 female TNBC patients, it was found that high-throughput quantitative sonographic features are superior to traditional qualitative features in predicting the biological behavior of TNBC.
EUROPEAN RADIOLOGY
(2022)
Article
Chemistry, Multidisciplinary
Alistair J. Sterling, Stamatia Zavitsanou, Joseph Ford, Fernanda Duarte
Summary: Recent advancements in experimental and computational techniques have revolutionized the field of chemistry, allowing for the development of predictive models to guide catalyst design. The field of organocatalysis faces challenges and promising future directions, with the combination of traditional physical organic chemistry tools and machine learning models providing a powerful approach for deeper understanding and enhanced predictive capabilities.
WILEY INTERDISCIPLINARY REVIEWS-COMPUTATIONAL MOLECULAR SCIENCE
(2021)
Article
Radiology, Nuclear Medicine & Medical Imaging
Loic Duron, Alexandre Heraud, Frederique Charbonneau, Mathieu Zmuda, Julien Savatovsky, Laure Fournier, Augustin Lecler
Summary: An MRI radiomics signature can effectively differentiate between benign and malignant orbital lesions, outperforming expert radiologists. The model had an area under the receiver operating characteristic curve of 0.869 on the test set, with high accuracy, sensitivity, and specificity values. Additional clinical and imaging data did not significantly impact the algorithm's performance.
INVESTIGATIVE RADIOLOGY
(2021)
Article
Radiology, Nuclear Medicine & Medical Imaging
Marco Caballo, Domenico R. Pangallo, Wendelien Sanderink, Andrew M. Hernandez, Su Hyun Lyu, Filippo Molinari, John M. Boone, Ritse M. Mann, Ioannis Sechopoulos
Summary: The study developed and evaluated an algorithm for breast mass classification using multi-marker radiomic approach, achieving high diagnostic accuracy through multiple-step feature selection process and testing on an independent dataset.
Article
Radiology, Nuclear Medicine & Medical Imaging
Di Giuliano Francesca, Minosse Silvia, Picchi Eliseo, Ferrazzoli Valentina, Da Ros Valerio, Muto Massimo, Pistolese Chiara Adriana, Garaci Francesco, Floris Roberto
Summary: The study compared brain MRI using different sequences and found that while BRAVO had higher image quality scores than Silent qualitatively, the diagnostic use of Silent images was similar to BRAVO. Quantitatively, the contrast-to-noise ratio for grey matter and cerebrospinal fluid was comparable in the two sequences, with the signal-to-noise ratio in cerebrospinal fluid being higher in Silent than BRAVO. The acoustic noise of the Silent sequence was significantly lower compared to BRAVO.
Review
Biology
Zohaib Salahuddin, Henry C. Woodruff, Avishek Chatterjee, Philippe Lambin
Summary: AI is increasingly used in clinical applications for diagnosis and treatment decisions, with deep neural networks showing equal or better performance than clinicians. However, their lack of interpretability calls for the development of methods to ensure their trustworthiness. Nine different types of interpretability methods have been identified for understanding deep learning models in medical image analysis, with ongoing research on improving interpretability and evaluation methods for deep neural networks.
COMPUTERS IN BIOLOGY AND MEDICINE
(2022)
Article
Radiology, Nuclear Medicine & Medical Imaging
Chenggong Yan, Lingfeng Wang, Jie Lin, Jun Xu, Tianjing Zhang, Jin Qi, Xiangying Li, Wei Ni, Guangyao Wu, Jianbin Huang, Yikai Xu, Henry C. Woodruff, Philippe Lambin
Summary: This study developed an AI-based fully automated CT image analysis system for detection, diagnosis, and burden quantification of pulmonary TB, achieving human-level diagnostic performance through deep learning technology.
EUROPEAN RADIOLOGY
(2022)
Article
Biochemical Research Methods
Yanchao Zhang, Tom S. Bailey, Aleksandra M. Kubiak, Philippe Lambin, Jan Theys
Summary: Several species from the Clostridium genus have potential as industrial solvent producers and cancer therapeutic delivery vehicles. Researchers have developed libraries of promoters, UTRs, and start codons to improve gene regulation in Clostridium species and E. coli cloning strains. These findings provide valuable insights for future applications in industry and medicine, as well as fundamental research on Clostridium biology.
ACS SYNTHETIC BIOLOGY
(2022)
Review
Oncology
Erich P. Huang, James P. B. O'Connor, Lisa M. McShane, Maryellen L. Giger, Philippe Lambin, Paul E. Kinahan, Eliot L. Siegel, Lalitha K. Shankar
Summary: Integration of computer-extracted tumour characteristics into medical imaging computer-aided diagnosis (CAD) algorithms has been a long-standing practice. However, the translation of radiomics, an extension of CAD involving quantitative characterization of healthy or pathological structures and processes captured by medical imaging, into clinically useful tools has been limited. This may be due to factors such as varying imaging and radiomic feature extraction protocols, potential pitfalls in radiomic data analysis, and a lack of evidence demonstrating the clinical benefit of radiomic-based tools. The authors provide 16 criteria to guide the clinical translation of radiomics, with the aim of accelerating the use of this technology to improve patient outcomes.
NATURE REVIEWS CLINICAL ONCOLOGY
(2023)
Article
Oncology
Loeki Aldenhoven, B. Ramaekers, J. Degens, C. Oberije, J. van Loon, A. C. Dingemans, D. De Ruysscher, M. Joore
Summary: For patients with stage III NSCLC, the strategy of photon radiotherapy for all patients (XRTAll) is the most cost-effective strategy. The strategy of proton therapy for all patients (PTAll) has the best performance in terms of initial cost and quality-adjusted life years (QALYs), but it is also the most expensive. The proton therapy strategy for selected patients (PTIndividualized) is cost-effective at specific thresholds.
RADIOTHERAPY AND ONCOLOGY
(2023)
Article
Oncology
Abdalla Ibrahim, Akshayaa Vaidyanathan, Sergey Primakov, Flore Belmans, Fabio Bottari, Turkey Refaee, Pierre Lovinfosse, Alexandre Jadoul, Celine Derwael, Fabian Hertel, Henry C. C. Woodruff, Helle D. D. Zacho, Sean Walsh, Wim Vos, Mariaelena Occhipinti, Francois-Xavier Hanin, Philippe Lambin, Felix M. M. Mottaghy, Roland Hustinx
Summary: The study aims to develop a deep learning (DL) algorithm for classifying areas of increased uptake on bone scintigraphy scans. The algorithm was trained and validated on a dataset of 2365 scans, and its performance was evaluated on an external testing set of 998 scans. The results showed that the DL algorithm achieved higher specificity and sensitivity compared to nuclear medicine physicians, and it can detect metastatic bone disease (MBD) in a shorter time.
Review
Dermatology
Y. Widaatalla, T. Wolswijk, F. Adan, L. M. Hillen, H. C. Woodruff, I. Halilaj, A. Ibrahim, P. Lambin, K. Mosterd
Summary: Basal cell carcinoma (BCC) is a common type of cancer, and the application of artificial intelligence techniques for detecting and classifying BCC is necessary. This article reviews the current evidence on the use of handcrafted and deep radiomics models for BCC detection and classification in dermoscopy, optical coherence tomography, and reflectance confocal microscopy.
JOURNAL OF THE EUROPEAN ACADEMY OF DERMATOLOGY AND VENEREOLOGY
(2023)
Article
Radiology, Nuclear Medicine & Medical Imaging
Manon P. L. Beuque, Marc B. I. Lobbes, Yvonka van Wijk, Yousif Widaatalla, Sergey Primakov, Michael Majer, Corinne Balleyguier, Henry C. Woodruff, Philippe Lambin
Summary: This study developed a comprehensive machine learning tool that can automatically identify, segment, and classify breast lesions on contrast-enhanced mammography (CEM) images in recall patients. The results showed that the combined output of the handcrafted radiomics and deep learning models achieved good diagnostic performance in lesion identification and classification, outperforming the individual models.
Article
Oncology
Zohaib Salahuddin, Yi Chen, Xian Zhong, Henry C. Woodruff, Nastaran Mohammadian Rad, Shruti Atul Mali, Philippe Lambin
Summary: Automatic delineation and detection of primary tumour (GTVp) and lymph nodes (GTVn) in head and neck cancer using PET and CT can help diagnose and stratify patient risk. This study utilized data from nine centres and developed a segmentation model that estimated uncertainty for false positive reduction. Radiomics features extracted from GTVp and GTVn in PET and CT were found to be prognostic for recurrence-free survival prediction. The framework incorporated uncertainty estimation, fairness, and explainability.
Article
Oncology
Annie Y. Ng, Ben Glocker, Cary Oberije, Georgia Fox, Nisha Sharma, Jonathan J. James, Eva Ambrozay, Jonathan Nash, Edith Karpati, Sarah Kerruish, Peter D. Kecskemethy
Summary: This study evaluated a new strategy of using artificial intelligence (AI) as a supporting reader for mammography-based double reading screening practice. The results showed that AI as a supporting reader was superior or non-inferior to human double reading on all screening metrics, while reducing the number of cases requiring second human reading by up to 87%.
JOURNAL OF BREAST IMAGING
(2023)
Article
Imaging Science & Photographic Technology
Rihab Hami, Sena Apeke, Pascal Redou, Laurent Gaubert, Ludwig J. Dubois, Philippe Lambin, Dimitris Visvikis, Nicolas Boussion
Summary: The effectiveness of radiotherapy depends on various factors, and the tumor response to radiation varies among patients. This study developed a multi-scale model that integrated five major biological concepts in radiotherapy to predict the effects of radiation on tumor growth. The model considered factors such as oxygen level, cell cycle position, cellular sensitivity, and repair, providing a basis for a more personalized clinical tool.
JOURNAL OF IMAGING
(2023)
Article
Oncology
Xian Zhong, Zohaib Salahuddin, Yi Chen, Henry C. Woodruff, Haiyi Long, Jianyun Peng, Xiaoyan Xie, Manxia Lin, Philippe Lambin
Summary: This study developed and validated an interpretable radiomics model based on two-dimensional shear wave elastography (2D-SWE) for predicting symptomatic post-hepatectomy liver failure in patients with hepatocellular carcinoma. The clinical-radiomics model outperformed the clinical model and radiomics model, and the first-order radiomics features were identified as the most important for PHLF prediction.
Article
Biochemistry & Molecular Biology
Annie Y. Ng, Cary J. G. Oberije, Eva Ambrozay, Endre Szabo, Orsolya Serfozo, Edit Karpati, Georgia Fox, Ben Glocker, Elizabeth A. Morris, Gabor Forrai, Peter D. Kecskemethy
Summary: Implementing AI as an additional reader can improve early detection of breast cancer, with higher positive predictive value and minimal to no unnecessary recalls.
Article
Microbiology
Yanchao Zhang, Aleksandra M. Kubiak, Tom S. Bailey, Luuk Claessen, Philip Hittmeyer, Ludwig Dubois, Jan Theys, Philippe Lambin
Summary: Clostridium species have gained attention in industrial and medical applications, with the development of genetic tools enabling the advancement of the CRISPR-Cas systems. This study demonstrated the establishment of a CRISPR-Cas12a system in clostridia with two different cas12a genes, allowing for efficient and rapid genome modification. The results showed that the CRISPR-FnCas12a system offers flexible target selection in clostridia, with a specific folding pattern of the precursor crRNA being important for high mutation generation efficiency.
MICROBIOLOGY SPECTRUM
(2023)