4.7 Article

Advanced imaging tools for childhood tuberculosis: potential applications and research needs

期刊

LANCET INFECTIOUS DISEASES
卷 20, 期 11, 页码 E289-E297

出版社

ELSEVIER SCI LTD
DOI: 10.1016/S1473-3099(20)30177-8

关键词

-

资金

  1. Maternal Adolescent Pediatric Research Branch, National Institute of Allergy and Infectious Diseases, NIH, Bethesda, MD, USA [HHSN27220160001G]

向作者/读者索取更多资源

Tuberculosis is the leading cause of death globally that is due to a single pathogen, and up to a fifth of patients with tuberculosis in high-incidence countries are children younger than 16 years. Unfortunately, the diagnosis of childhood tuberculosis is challenging because the disease is often paucibacillary and it is difficult to obtain suitable specimens, causing poor sensitivity of currently available pathogen-based tests. Chest radiography is important for diagnostic evaluations because it detects abnormalities consistent with childhood tuberculosis, but several limitations exist in the interpretation of such results. Therefore, other imaging methods need to be systematically evaluated in children with tuberculosis, although current data suggest that when available, cross-sectional imaging, such as CT, should be considered in the diagnostic evaluation for tuberculosis in a symptomatic child. Additionally, much of the understanding of childhood tuberculosis stems from clinical specimens that might not accurately represent the lesional biology at infection sites. By providing non-invasive measures of lesional biology, advanced imaging tools could enhance the understanding of basic biology and improve on the poor sensitivity of current pathogen detection systems. Finally, there are key knowledge gaps regarding the use of imaging tools for childhood tuberculosis that we outlined in this Personal View, in conjunction with a proposed roadmap for future research.

作者

我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。

评论

主要评分

4.7
评分不足

次要评分

新颖性
-
重要性
-
科学严谨性
-
评价这篇论文

推荐

Article Oncology

The development of automated visual evaluation for cervical cancer screening: The promise and challenges in adapting deep-learning for clinical testing

Kanan T. Desai, Brian Befano, Zhiyun Xue, Helen Kelly, Nicole G. Campos, Didem Egemen, Julia C. Gage, Ana-Cecilia Rodriguez, Vikrant Sahasrabuddhe, David Levitz, Paul Pearlman, Jose Jeronimo, Sameer Antani, Mark Schiffman, Silvia de Sanjose

Summary: Limited access to effective cervical cancer screening programs in resource-limited settings leads to high cervical cancer burden. Human papillomavirus (HPV) testing is recognized as the preferable primary screening approach, providing long-term reassurance and adaptability to self-sampling. Visual inspection with acetic acid (VIA) is widely used in resource-limited settings, but it is subjective and inaccurate.

INTERNATIONAL JOURNAL OF CANCER (2022)

Article Multidisciplinary Sciences

Deep learning model calibration for improving performance in class-imbalanced medical image classification tasks

Sivaramakrishnan Rajaraman, Prasanth Ganesan, Sameer Antani

Summary: In this study, the effect of model calibration on the performance of medical image classification tasks was systematically analyzed. The results show that calibration can significantly improve performance at the default classification threshold, but the differences are not significant at the PR-guided threshold. This observation holds for different image modalities and degrees of class imbalance.

PLOS ONE (2022)

Article Multidisciplinary Sciences

DeBoNet: A deep bone suppression model ensemble to improve disease detection in chest radiographs

Sivaramakrishnan Rajaraman, Gregg Cohen, Lillian Spear, Les Folio, Sameer Antani

Summary: Automated bone suppression methods can enhance the visibility of soft tissues in chest X-ray images and improve automated disease detection. The DeBoNet ensemble model, constructed using top-performing convolutional neural network models, outperformed individual models in terms of various metrics. Applying the best-performing bone-suppression model to CXR images showed improved performance in detecting pulmonary abnormalities consistent with COVID-19. Automatic bone suppression brings benefits to disease classification.

PLOS ONE (2022)

Article Genetics & Heredity

Detecting Tuberculosis-Consistent Findings in Lateral Chest X-Rays Using an Ensemble of CNNs and Vision Transformers

Sivaramakrishnan Rajaraman, Ghada Zamzmi, Les R. Folio, Sameer Antani

Summary: By constructing an ensemble of CNN and ViT models, this study successfully detected TB-consistent findings in lateral CXRs, achieving significant performance improvement. The interpretation of CNN and ViT models' decisions also highlighted the discriminative image regions contributing to the final output.

FRONTIERS IN GENETICS (2022)

Article Oncology

Unsupervised Deep Learning Registration of Uterine Cervix Sequence Images

Peng Guo, Zhiyun Xue, Sandeep Angara, Sameer K. Antani

Summary: This study proposes a novel deep learning-based image registration method to automatically align a sequence of cervical images. The proposed method achieves significant improvement in cervical boundary detection compared to unregistered images and maintains image integrity.

CANCERS (2022)

Article Infectious Diseases

Adequacy of the 10 mg/kg Daily Dose of Antituberculosis Drug Isoniazid in Infants under 6 Months of Age

Maria Goretti Lopez-Ramos, Joan Vinent, Rob Aarnoutse, Angela Colbers, Eneritz Velasco-Arnaiz, Loreto Martorell, Lola Falcon-Neyra, Olaf Neth, Luis Prieto, Sara Guillen, Fernando Baquero-Artigao, Ana Mendez-Echevarria, David Gomez-Pastrana, Ana Belen Jimenez, Rebeca Lahoz, Jose Tomas Ramos-Amador, Antoni Soriano-Arandes, Begona Santiago, Rosa Farre, Claudia Fortuny, Dolors Soy, Antoni Noguera-Julian

Summary: In 2010, the WHO recommended increasing the daily doses of first-line anti-tuberculosis medicines in children. This study aimed to investigate the pharmacokinetics of a once-daily dose of isoniazid (INH) in infants under 6 months of age. The study found that the target adult levels were not reached in a few cases, but overall, the treatment was well tolerated and no major safety concerns were raised.

ANTIBIOTICS-BASEL (2023)

Editorial Material Computer Science, Information Systems

Guest Editorial Multimodal Learning in Medical Imaging Informatics

K. C. Santosh, Sameer Antani

IEEE JOURNAL OF BIOMEDICAL AND HEALTH INFORMATICS (2023)

Article Medicine, General & Internal

Assessing the Impact of Image Resolution on Deep Learning for TB Lesion Segmentation on Frontal Chest X-rays

Sivaramakrishnan Rajaraman, Feng Yang, Ghada Zamzmi, Zhiyun Xue, Sameer Antani

Summary: Deep learning models have achieved state-of-the-art performance in segmenting anatomical and disease regions of interest (ROIs) in medical images. This study investigates the performance variations of an Inception-V3 UNet model using different image resolutions, lung ROI cropping, and aspect ratio adjustments in segmenting tuberculosis-consistent lesions in chest X-rays (CXRs), and identifies the optimal image resolution to improve segmentation performance.

DIAGNOSTICS (2023)

Article Medicine, General & Internal

Cross Dataset Analysis of Domain Shift in CXR Lung Region Detection

Zhiyun Xue, Feng Yang, Sivaramakrishnan Rajaraman, Ghada Zamzmi, Sameer Antani

Summary: This paper investigates domain shift in medical imaging-based machine learning predictions. It proposes a new feature visualization method to explain the performance of object detection networks. The results provide valuable guidance for the analysis of training data and domain shift analysis in medical imaging machine learning research.

DIAGNOSTICS (2023)

Article Computer Science, Artificial Intelligence

Can deep adult lung segmentation models generalize to the pediatric population?

Sivaramakrishnan Rajaraman, Feng Yang, Ghada Zamzmi, Zhiyun Xue, Sameer Antani

Summary: Lung segmentation in chest X-rays is crucial for accurate diagnosis of cardiopulmonary diseases. However, the shape of the lungs varies significantly across different developmental stages, which can impact the performance of deep learning models trained on adult populations when applied to pediatric lung segmentation. This study aims to analyze the generalizability of deep adult lung segmentation models to the pediatric population and proposes a systematic approach to improve performance. Novel evaluation metrics are introduced to assess segmentation performance and cross-domain generalization. The results show significant improvement in cross-domain generalization through the proposed approach.

EXPERT SYSTEMS WITH APPLICATIONS (2023)

Article Oncology

Artificial intelligence-based image analysis in clinical testing: lessons from cervical cancer screening

Didem Egemen, Rebecca B. Perkins, Li C. Cheung, Brian Befano, Ana Cecilia Rodriguez, Kanan Desai, Andreanne Lemay, Syed Rakin Ahmed, Sameer Antani, Jose Jeronimo, Nicolas Wentzensen, Jayashree Kalpathy-Cramer, Silvia De Sanjose, Mark Schiffman

Summary: This article emphasizes the importance of novel screening and diagnostic tests based on artificial intelligence image recognition algorithms. It provides a conceptual step-by-step approach to bridge the gap between the creation of AI algorithms and clinical efficacy. The article also highlights the need for rigorous evaluation, risk estimation, and development of guidelines for clinical use.

JNCI-JOURNAL OF THE NATIONAL CANCER INSTITUTE (2023)

Editorial Material Medicine, General & Internal

The Need for Artificial Intelligence Curriculum in Military Medical Education

Jonathan R. Spirnak, Sameer Antani

Summary: There has been an exponential increase in research on the applications of artificial intelligence (AI) in medicine in the past decade. The release of large language models like ChatGPT has sparked discussions on whether machine intelligence can surpass human capability. However, concerns have been raised regarding the social, legal, and moral implications of this powerful technology. The challenge in medicine is to harness AI to improve patient outcomes. Military medicine can greatly benefit from AI, but this requires the understanding and collaboration of the rising generations of military medical professionals.

MILITARY MEDICINE (2023)

Article Computer Science, Information Systems

Assessing Inter-Annotator Agreement for Medical Image Segmentation

Feng Yang, Ghada Zamzmi, Sandeep Angara, Sivaramakrishnan Rajaraman, Andre Aquilina, Zhiyun Xue, Stefan Jaeger, Emmanouil Papagiannakis, Sameer K. K. Antani

Summary: This study aims to assess the inter-annotator agreement among multiple expert annotators when segmenting the same lesion(s)/abnormalities on medical images, and proposes three metrics for assessment.

IEEE ACCESS (2023)

Proceedings Paper Computer Science, Interdisciplinary Applications

Data Characterization for Reliable AI in Medicine

Sivaramakrishnan Rajaraman, Ghada Zamzmi, Feng Yang, Zhiyun Xue, Sameer K. Antani

Summary: This article discusses the impact of underlying data characteristics on AI-based medical computer vision algorithms and presents recent works conducted in a research lab to enhance understanding of how these characteristics influence the design of medical decision-making algorithms and outcome reliability.

RECENT TRENDS IN IMAGE PROCESSING AND PATTERN RECOGNITION, RTIP2R 2022 (2023)

Article Computer Science, Information Systems

Annotations of Lung Abnormalities in the Shenzhen Chest X-ray Dataset for Computer-Aided Screening of Pulmonary Diseases

Feng Yang, Pu Xuan Lu, Min Deng, Yi Xiang J. Wang, Sivaramakrishnan Rajaraman, Zhiyun Xue, Les R. Folio, Sameer K. Antani, Stefan Jaeger

Summary: This study presents a collection of annotations/segmentations of pulmonary radiological manifestations consistent with tuberculosis (TB), with the goal of advancing image segmentation methods and improving the fine-grained segmentation of TB findings in digital chest X-ray images.
暂无数据