Sivan Lieberman

Chest pathology detection using deep learning with non-medical training

In this work, we examine the strength of deep learning approaches for pathology detection in chest radiographs. Convolutional neural networks (CNN) deep architecture classification approaches have gained popularity due to their ability to learn mid and high level image representations. We explore the ability of CNN learned from a non-medical dataset to identify different types of pathologies in chest x-rays. We tested our algorithm on a 433 image dataset. The best performance was achieved using CNN and GIST features. We obtained an area under curve (AUC) of 0.87-0.94 for the different pathologies. The results demonstrate the feasibility of detecting pathology in chest x-rays using deep learning approaches based on non-medical learning. This is a first-of-its-kind experiment that shows that Deep learning with ImageNet, a large scale non-medical image database may be a good substitute to domain specific representations, which are yet to be available, for general medical image recognition tasks.

Chest pathology identification using deep feature selection with non-medical training

We demonstrate the feasibility of detecting pathology in chest X-rays using deep learning approaches based on non-medical learning. Convolutional neural networks (CNN) learn higher level image representations. In this work, we explore the features extracted from layers of the CNN along with a set of classical features, including GIST and bag-of-words. We show results of classification using each feature set as well as fusing among the features. Finally, we perform feature selection on the collection of features to show the most informative feature set for the task. Results of 0.78–0.95 AUC for various pathologies are shown on a data-set of more than 600 radiographs. This study shows the strength and robustness of the CNN features. We conclude that deep learning with large-scale non- medical image databases may be a good substitute, or addition to domain-specific representations which are yet to be available for general medical image recognition tasks.

Feasibility of Confocal Laser Microscopy in CT-Guided Needle Biopsy of Pulmonary and Mediastinal Tumors: A Proof-of-Concept Pilot Study

This report describes the use of confocal laser microscopy (CLM) with CT-guided transthoracic needle biopsy (TTNB) for the diagnosis of heterogeneous large mediastinal and lung tumors. The procedure was performed in five patients diagnosed with a mediastinal mass and five patients diagnosed with a lung mass. CLM was used before CT-guided TTNB. Fluorescein administration allowed for the identification of blood vessels in both locations. Malignant cells were identified in mediastinal masses. Complications included one case of pneumothorax. In large tumors, CLM allows vascularized tissue to be differentiated from necrotic and fibrotic areas before biopsy.

Enfermedad pulmonar relacionada con IgG4 - Tres casos no tratados con resultado benigno

IgG4-related disease is a fibroinflammatory disease in which the organs involved share similar pathological findings. Chest disease has been recently clinically and radiologically characterized. Most reports advocate prompt immunosuppressive therapy and describe a fast and good response. We report 3 cases of untreated IgG4-related lung disease that on follow-up have been clinically asymptomatic and radiologically stable or improved. In some cases of IgG4-related lung disease immunosuppressive therapy may not be warranted.

Localized Fisher vector representation for pathology detection in chest radiographs

In this work, we present a novel framework for automatic detection of abnormalities in chest radiographs. The representation model is based on the Fisher Vector encoding method. In the representation process, we encode each chest radiograph using a set of extracted local descriptors. These include localized texture features that address typical local texture abnormalities as well as spatial features. Using a Gaussian Mixture Model, a rich image descriptor is generated for each chest radiograph. An improved representation is obtained by selection of features that correspond to the relevant region of interest for each pathology. Categorization of the X-ray images is conducted using supervised learning and the SVM classifier. The proposed system was tested on a dataset of 636 chest radiographs taken from a real clinical environment. We measured the performance in terms of area (AUC) under the receiver operating characteristic (ROC) curve. Results show an AUC value of 0.878 for abnormal mediastinum detection, and AUC values of 0.827 and 0.817 for detection of right and left lung opacities, respectively. These results improve upon the state-of-the-art as compared with two alternative representation models.

Chest Radiograph Pathology Categorization via Transfer Learning

Abstract The goal of this chapter is to give an overview of the research we have been conducting in automated X-ray pathology detection for the past 10 years, from bag-of-visual-words (BoVW) models to the Convolutional Neural Network (CNN) Deep Learning schemes. Our study was one of the first to suggest the possibility of using non-medical training, using transfer learning from the general imagery to the medical domain. In this chapter we explore deep features that are extracted from intermediate CNN layers in comparison to a set of classical shallow features, including GLCM, PHOG, GABOR, GIST and the more recent, state-of-the-art BoVW model. We investigate the possible benefits of using feature selection techniques on the Deep CNN feature layers. Average AUC results of close to 90% are shown for categorization of 6 different pathologies in a dataset of more than 600 radiographs. This study shows the strength and robustness of the CNN features. We conclude that deep learning with large scale non-medical image databases may be a good substitute, or addition, to domain specific representations which are yet to be available for general medical image recognition tasks. With the BoW schemes, our method won first place in ImageClef competitions. With the DL architectures, we are now able to use the system in real clinical settings.

Mediastinal Imaging Pitfalls.

Potential pitfalls in the interpretation of diseases involving the mediastinum are seen when imaging with computed tomography and [18F]-fluoro-2-deoxy-d-glucose positron emission tomography. These pitfalls can involve any mediastinal structure, including the mediastinal vessels, heart, lymph nodes, thymus, trachea, esophagus, and fat. Misinterpretation of normal variants or benign conditions as pathology can affect staging and alter treatment. After reading this review, the reader should be able to identify common mediastinal imaging pitfalls and apply ancillary measures to confirm the correct diagnosis and thus reach an accurate diagnosis to facilitate correct patient treatment.

Long-term Imaging of the Lungs After Sealant Bronchoscopic Lung Volume Reduction.

Purpose: The aim of the study was to assess the pulmonary temporal changes after bronchoscopic lung volume reduction (BLVR) using sealants for treatment of emphysema. Materials and Methods: We retrospectively assessed all chest computerized tomography (CT) and F-18 fluorodeoxyglucose (FDG) positron emission tomography CT scans of patients treated at our institution with BLVR. Results: Eleven patients were treated with sealants: 4 with biological sealants and 7 with synthetic sealants. The first CT scan after biological sealant treatment showed no abnormalities in 8 lobes and 5 nodules, and 3 consolidations in 7 lobes. All findings resolved within 3 months, except for a nodule that decreased after 2 months and remained stable for 9 years. The first CT scan after utilizing the synthetic sealant showed abnormalities in each treated lobe: 19 nodules/masses (16 cavitary, 3 solid) and 3 consolidations. Follow-up CT scans were available for 16 nodules/masses: 1 resolved, 12 slowly decreased in size, 1 remained unchanged, and 2 grew. Of 3 consolidations 1 resolved and 2 decreased in size. FDG positron emission tomography CT scans performed in 2 patients showed FDG uptake higher than mediastinal background activity in 2 nodules in the same patient. Conclusions: Pulmonary changes after BLVR are variable. After treatment with biological sealants, most findings resolve within 3 months. In contrast, after synthetic sealants, although the majority regress over time, some show waxing and waning in growth that can mimic malignancy. FDG uptake in some of these lesions is suggestive of chronic inflammation. Radiologists should be aware of the spectrum of these pulmonary changes to avoid misdiagnosis of lung cancer.

Pneumothorax detection in chest radiographs using local and global texture signatures

A novel framework for automatic detection of pneumothorax abnormality in chest radiographs is presented. The suggested method is based on a texture analysis approach combined with supervised learning techniques. The proposed framework consists of two main steps: at first, a texture analysis process is performed for detection of local abnormalities. Labeled image patches are extracted in the texture analysis procedure following which local analysis values are incorporated into a novel global image representation. The global representation is used for training and detection of the abnormality at the image level. The presented global representation is designed based on the distinctive shape of the lung, taking into account the characteristics of typical pneumothorax abnormalities. A supervised learning process was performed on both the local and global data, leading to trained detection system. The system was tested on a dataset of 108 upright chest radiographs. Several state of the art texture feature sets were experimented with (Local Binary Patterns, Maximum Response filters). The optimal configuration yielded sensitivity of 81% with specificity of 87%. The results of the evaluation are promising, establishing the current framework as a basis for additional improvements and extensions.