Attention gated networks: Learning to leverage salient regions in medical images.

Background
Chest radiograph interpretation is critical for the detection of thoracic diseases, including tuberculosis and lung cancer, which affect millions of people worldwide each year. This time-consuming task typically requires expert radiologists to read the images, leading to fatigue-based diagnostic error and lack of diagnostic expertise in areas of the world where radiologists are not available. Recently, deep learning approaches have been able to achieve expert-level performance in medical image interpretation tasks, powered by large network architectures and fueled by the emergence of large labeled datasets. The purpose of this study is to investigate the performance of a deep learning algorithm on the detection of pathologies in chest radiographs compared with practicing radiologists.

/pdf/deep-learning-for-chest-radiograph-diagnosis-a-retrospective-1tsxdvti5i.pdf

Deep learning for chest radiograph diagnosis: A retrospective comparison of the CheXNeXt algorithm to practicing radiologists

This paper considers the task of thorax disease classification on chest X-ray images Existing methods generally use the global image as input for network learning Such a strategy is limited in two aspects 1) A thorax disease usually happens in (small) localized areas which are disease specific Training CNNs using global image may be affected by the (excessive) irrelevant noisy areas 2) Due to the poor alignment of some CXR images, the existence of irregular borders hinders the network performance In this paper, we address the above problems by proposing a three-branch attention guided convolution neural network (AG-CNN) AG-CNN 1) learns from disease-specific regions to avoid noise and improve alignment, 2) also integrates a global branch to compensate the lost discriminative cues by local branch Specifically, we first learn a global CNN branch using global images Then, guided by the attention heat map generated from the global branch, we inference a mask to crop a discriminative region from the global image The local region is used for training a local CNN branch Lastly, we concatenate the last pooling layers of both the global and local branches for fine-tuning the fusion branch The Comprehensive experiment is conducted on the ChestX-ray14 dataset We first report a strong global baseline producing an average AUC of 0841 with ResNet-50 as backbone After combining the local cues with the global information, AG-CNN improves the average AUC to 0868 While DenseNet-121 is used, the average AUC achieves 0871, which is a new state of the art in the community

Diagnose like a Radiologist: Attention Guided Convolutional Neural Network for Thorax Disease Classification.

Novel coronavirus disease (nCOVID-19) is the most challenging problem for the world. The disease is caused by severe acute respiratory syndrome coronavirus-2 (SARS-COV-2), leading to high morbidity and mortality worldwide. The study reveals that infected patients exhibit distinct radiographic visual characteristics along with fever, dry cough, fatigue, dyspnea, etc. Chest X-Ray (CXR) is one of the important, non-invasive clinical adjuncts that play an essential role in the detection of such visual responses associated with SARS-COV-2 infection. However, the limited availability of expert radiologists to interpret the CXR images and subtle appearance of disease radiographic responses remains the biggest bottlenecks in manual diagnosis. In this study, we present an automatic COVID screening (ACoS) system that uses radiomic texture descriptors extracted from CXR images to identify the normal, suspected, and nCOVID-19 infected patients. The proposed system uses two-phase classification approach (normal vs. abnormal and nCOVID-19 vs. pneumonia) using majority vote based classifier ensemble of five benchmark supervised classification algorithms. The training-testing and validation of the ACoS system are performed using 2088 (696 normal, 696 pneumonia and 696 nCOVID-19) and 258 (86 images of each category) CXR images, respectively. The obtained validation results for phase-I (accuracy (ACC) = 98.062%, area under curve (AUC) = 0.956) and phase-II (ACC = 91.329% and AUC = 0.831) show the promising performance of the proposed system. Further, the Friedman post-hoc multiple comparisons and z-test statistics reveals that the results of ACoS system are statistically significant. Finally, the obtained performance is compared with the existing state-of-the-art methods.

Coronavirus disease (COVID-19) detection in Chest X-Ray images using majority voting based classifier ensemble

Multi-label chest X-ray image classification via category-wise residual attention learning

Learning to detect chest radiographs containing pulmonary lesions using visual attention networks

Pheochromocytoma (PC) and paraganglioma (PGL) are rare neuroendocrine tumors that occur throughout the body from the base of the skull to the pelvis. Sympathetic catecholamine-secreting tumors may be associated with hyperadrenergic symptoms and long-term morbidity if they are untreated. Typically biochemically silent, head and neck PGLs may result in cranial nerve palsies and symptoms due to localized mass effect. Tumors can arise sporadically or as part of an inheritable PC-PGL syndrome. Up to 40% of tumors are recognized to be associated with germline mutations in an increasing array of susceptibility genes, including those that appear to arise sporadically. Most commonly, up to 25% of all PC-PGLs are associated with mutations in one of the succinate dehydrogenase (SDH) enzyme subunit genes. The resulting familial PC-PGL syndrome varies according to the affected enzyme subunit (most commonly SDHB and SDHD mutations) with respect to tumor prevalence, location, age of onset, and risk of malignancy. Patients with SDH enzyme mutations have increased lifetime risk of developing multifocal tumors and malignancy. Early recognition of individuals at high risk, genetic testing, screening of family members, and lifelong surveillance programs are recommended, but not without health, economic, and psychologic implications. Anatomic and functional imaging is key to diagnosis, staging, treatment planning, and lifelong surveillance of these individuals. Radiologists must be aware of the imaging appearance of these varied tumors.©RSNA, 2019.

https://pubs.rsna.org/doi/pdf/10.1148/rg.2019180151

Imaging Features of Succinate Dehydrogenase–deficient Pheochromocytoma-Paraganglioma Syndromes

Indeterminate nonpalpable focal testicular lesions have emerged as a clinical problem with the increasing use of scrotal ultrasound, particularly in the context of infertility. Conventional morphological ultrasound and color Doppler have been unreliable at differentiating benign from malignant lesions. Multiparametric ultrasound (mpUS) comprises real-time elastography and contrast-enhanced ultrasound as adjunctive tools, and is ready for use in most state-of-the-art ultrasound systems. Initial experience with mpUS from selected specialist centers shows promise for lesion characterization, and potential for affecting management and improving outcomes. This article provides a summary of the existing literature on testicular mpUS, and outlines the clinical context from a urological and histopathological perspective.

Multiparametric Ultrasound of Nonpalpable Focal Testicular Lesions.

Renal cysts are a common imaging finding, often incidental. Ultrasound, CT and MRI are the main modalities responsible for renal cyst detection and characterization. These modalities often play a complementary role in modern radiological practice, each of them with strengths and limitations. In view of a recently proposed 'multimodality' update to the historical Bosniak classification, this article provides a general overview of the current imaging approach to renal cysts, and outlines some of the diverse pathologic entities responsible for renal cyst formation.

Samuel Joseph Withey

Papers

Learning to detect chest radiographs containing pulmonary lesions using visual attention networks

Imaging Features of Succinate Dehydrogenase–deficient Pheochromocytoma-Paraganglioma Syndromes

Multiparametric Ultrasound of Nonpalpable Focal Testicular Lesions.

Multimodality Assessment of Cystic Renal Masses.