Thoracic Disease Classification Research Articles

Learning to Generalize Towards Unseen Domains via a Content-Aware Style Invariant Model for Disease Detection From Chest X-Rays.

Performance degradation due to distribution discrepancy is a longstanding challenge in intelligent imaging, particularly for chest X-rays (CXRs). Recent studies have demonstrated that CNNs are biased toward styles (e.g., uninformative textures) rather than content (e.g., shape), in stark contrast to the human vision system. Radiologists tend to learn visual cues from CXRs and thus perform well across multiple domains. Motivated by this, we employ the novel on-the-fly style randomization modules at both image (SRM-IL) and feature (SRM-FL) levels to create rich style perturbed features while keeping the content intact for robust cross-domain performance. Previous methods simulate unseen domains by constructing new styles via interpolation or swapping styles from existing data, limiting them to available source domains during training. However, SRM-IL samples the style statistics from the possible value range of a CXR image instead of the training data to achieve more diversified augmentations. Moreover, we utilize pixel-wise learnable parameters in the SRM-FL compared to pre-defined channel-wise mean and standard deviations as style embeddings for capturing more representative style features. Additionally, we leverage consistency regularizations on global semantic features and predictive distributions from with and without style-perturbed versions of the same CXR to tweak the model's sensitivity toward content markers for accurate predictions. Our proposed method, trained on CheXpert and MIMIC-CXR datasets, achieves 77.32±0.35, 88.38±0.19, 82.63±0.13 AUCs(%) on the unseen domain test datasets, i.e., BRAX, VinDr-CXR, and NIH chest X-ray14, respectively, compared to 75.56±0.80, 87.57±0.46, 82.07±0.19 from state-of-the-art models on five-fold cross-validation with statistically significant results in thoracic disease classification.

Robust Stochastic Neural Ensemble Learning With Noisy Labels for Thoracic Disease Classification.

Chest radiography is the most common radiology examination for thoracic disease diagnosis, such as pneumonia. A tremendous number of chest X-rays prompt data-driven deep learning models in constructing computer-aided diagnosis systems for thoracic diseases. However, in realistic radiology practice, a deep learning-based model often suffers from performance degradation when trained on data with noisy labels possibly caused by different types of annotation biases. To this end, we present a novel stochastic neural ensemble learning (SNEL) framework for robust thoracic disease diagnosis using chest X-rays. The core idea of our method is to learn from noisy labels by constructing model ensembles and designing noise-robust loss functions. Specifically, we propose a fast neural ensemble method that collects parameters simultaneously across model instances and along optimization trajectories. Moreover, we propose a loss function that both optimizes a robust measure and characterizes a diversity measure of ensembles. We evaluated our proposed SNEL method on three publicly available hospital-scale chest X-ray datasets. The experimental results indicate that our method outperforms competing methods and demonstrate the effectiveness and robustness of our method in learning from noisy labels. Our code is available at https://github.com/hywang01/SNEL.

Patch-Level Feature Selection for Thoracic Disease Classification by Chest X-ray Images Using Information Bottleneck.

Chest X-ray (CXR) examination serves as a widely employed clinical test in medical diagnostics. Many studied have tried to apply artificial intelligence (AI) programs to analyze CXR images. Despite numerous positive outcomes, assessing the applicability of AI models for comprehensive diagnostic support remains a formidable challenge. We observed that, even when AI models exhibit high accuracy on one dataset, their performance may deteriorate when tested on another. To address this issue, we propose incorporating a variational information bottleneck (VIB) at the patch level to enhance the generalizability of diagnostic support models. The VIB introduces a probabilistic model aimed at approximating the posterior distribution of latent variables given input data, thereby enhancing the model's generalization capabilities on unseen data. Unlike the conventional VIB approaches that flatten features and use a re-parameterization trick to sample a new latent feature, our method applies the trick to 2D feature maps. This design allows only important pixels to respond, and the model will select important patches in an image. Moreover, the proposed patch-level VIB seamlessly integrates with various convolutional neural networks, offering a versatile solution to improve performance. Experimental results illustrate enhanced accuracy in standard experiment settings. In addition, the method shows robust improvement when training and testing on different datasets.

Learning group-wise spatial attention and label dependencies for multi-task thoracic disease classification

This paper considers the multi-label thoracic abnormality classification with chest X-ray images. In clinical settings, Chest X-ray imaging is a general diagnostic tool applied to visualize numerous thoracic pathological changes. While deep learning techniques have been extensively tested in this field, certain challenges persist. The data in existing thoracic abnormality datasets is insufficient, and some diseases are extremely imbalanced. Meanwhile, the dependencies between different labels are often ignored. To tackle these issues head-on, this paper introduces two crucial modules: the group-wise spatial attention (GWSA) module and the label co-occurrence dependency (LCD) module, integrated with DenseNet121 backbone. Specifically, GWSA enhances the spatial features within distinct groups while keeping the between-group feature discrimination. LCD models the correlations between different thoracic abnormalities to refine the predicted probabilities. In conjunction with the DenseNet121 backbone, these two modules reach an average AUC score of 0.818 on Chest X-ray14 dataset, achieving state-of-the-art. Source code is available at https://github.com/YujiaKCL/Chest-Xray14-GWSA-LCD.

Artificially Generated Visual Scanpath Improves Multi-label Thoracic Disease Classification in Chest X-Ray Images

Artificially Generated Visual Scanpath Improves Multi-label Thoracic Disease Classification in Chest X-Ray Images

Effsemble: faster, smaller and more accurate ensemble networks for thoracic disease classification

Effsemble: faster, smaller and more accurate ensemble networks for thoracic disease classification

Multi-Level Residual Feature Fusion Network for Thoracic Disease Classification in Chest X-Ray Images

Motivation: Automated identification of thoracic diseases from chest X-ray images (CXR) is a significant area in computer-aided diagnosis. However, most existing methods have limited ability to extract multi-scale features and accurately capture the spatial location of lesions when dealing with thoracic diseases that exhibit concurrency and large variations in lesion size. Method: Based on the above problems, we propose a multi-level residual feature fusion network (MLRFNet) for classifying thoracic diseases. Our approach can quickly capture receptive field information across different lesion sizes and enhance disease-specific features within the spatial domain on feature maps. The MLRFNet comprises two main components: a feature extractor that learns multi-scale semantic information from chest X-ray images and a multi-level residual feature classifier (MRFC) that refines disease-specific pathological features at spatial locations to reduce interference from irrelevant regions. Additionally, the ECA attention modules connect both components to enable flexible channel-wise focus on critical pathological information. Results: We evaluated the performance of MLRFNet through a series of experiments on two datasets: ChestX-Ray14 and CheXpert. Our results show that MLRFNet achieves an average AUC of 0.853 on the ChestX-Ray14 dataset and 0.904 on the CheXpert dataset. Conclusion: The results of experiments demonstrate that our proposed method works better than the current state-of-the-art baselines. Future work will focus on investigating the interdependencies among labels for thoracic diseases and techniques for model compression.

Effsemble: faster, smaller and more accurate ensemble networks for thoracic disease classification

Effsemble: faster, smaller and more accurate ensemble networks for thoracic disease classification

Domain-ensemble learning with cross-domain mixup for thoracic disease classification in unseen domains

Chest radiography is the most common radiology examination for thoracic disease diagnosis, such as pneumonia. A tremendous number of chest X-rays prompt data-driven deep learning models in constructing computer-aided diagnosis (CAD) systems for thoracic disease diagnosis. However, in clinically realistic practice, a trained deep learning model often suffers from performance degradation when applied to unseen data, mainly due to domain shifts caused by various scanner vendors, imaging protocols, and patient demographics etc. To this end, we present a novel domain-ensemble learning with cross-domain mixup (DELCOM) method for thoracic disease diagnosis in unseen domains using chest X-rays. The core idea of our method is to learn multiple domain-specific models using data from both virtual and source domains and to generalize them to unseen domains. Specifically, we first introduce cross-domain mixup to construct virtual samples not belonging to any source domain by mixing up the pairs of samples from source domains. We then model the simulation process of domain shift through domain-ensemble learning. The results indicate that our method not only outperforms other state-of-the-art domain generalization (DG) methods on unseen datasets but also achieves comparable performance across all source domains.

PCAN: Pixel-wise classification and attention network for thoracic disease classification and weakly supervised localization.

Automatic chest X-ray (CXR) disease classification has drawn increasing public attention as CXR is widely used in thoracic disease diagnosis. Existing classification networks typically employ a global average pooling layer to produce the final feature for the subsequent classifier. This limits the classification performance owing to the characteristics of lesions in CXR images, including small relative sizes, varied absolute sizes, and different occurrence locations. In this study, we propose a pixel-wise classification and attention network (PCAN) to simultaneously perform disease classification and weakly supervised localization, which provides interpretability for disease classification. The PCAN comprises a backbone network for extracting mid-level features, a pixel-wise classification branch (pc-branch) for generating pixel-wise diagnoses, and a pixel-wise attention branch (pa-branch) for producing pixel-wise weights. The pc-branch is capable of explicitly detecting small lesions, and the pa-branch is capable of adaptively focusing on different regions when classifying different thoracic diseases. Then, the pixel-wise diagnoses are multiplied with the pixel-wise weights to obtain the disease localization map, which provides the sizes and locations of lesions in a manner of weakly supervised learning. The final image-wise diagnosis is obtained by summing up the disease localization map at the spatial dimension. Comprehensive experiments conducted on the ChestX-ray14 and CheXpert datasets demonstrate the effectiveness of the proposed PCAN, which has great potential for thoracic disease diagnosis and treatment. The source codes are available at https://github.com/fzfs/PCAN.

Classification of Thoracic Diseases Based on Chest X-ray Images Using Kernel Support Vector Machine

Machine learning is the leading field of artificial intelligence that has achieved expert-level performance. Diagnosis and treatment of various medical diseases have led to advancements in medical imaging. Chest X-ray-based thoracic disease classification or identification is one of the potential applications in medical imaging based on machine learning. The study consists of 112,120 images of 30,804 individual patients with fourteen thoracic disease labels, which encapsulated the support vector machine (SVM). We have considered 04 kernels in SVM, namely, linear (L-SVM), polynomial (P-SVM), radial basis (R-SVM), and hyperbolic tangent (H-SVM) for classification of thoracic diseases based on X-ray images. To reduce the dimensionality and outliers from the SVM, variants are coupled with novel fast principal component analysis (FPCA). It appears that there is a significant p ≤ 0.05 difference between SVM variants where P-SVM and R-SVM next in order outperforms on most of the disease identification models with average validated classification accuracy ranging from 92% to 98%. The average calibrated accuracy ranges from 99.5% and reaches to 100% in most of the cases. The study is worth investigating as it is good for radiologists as they will be able to classify the diseases and it will help in improving and enhancing different medical techniques.

Anatomy-XNet: An Anatomy Aware Convolutional Neural Network for Thoracic Disease Classification in Chest X-Rays

Thoracic disease detection from chest radiographs using deep learning methods has been an active area of research in the last decade. Most previous methods attempt to focus on the diseased organs of the image by identifying spatial regions responsible for significant contributions to the model's prediction. In contrast, expert radiologists first locate the prominent anatomical structures before determining if those regions are anomalous. Therefore, integrating anatomical knowledge within deep learning models could bring substantial improvement in automatic disease classification. Motivated by this, we propose Anatomy-XNet, an anatomy-aware attention-based thoracic disease classification network that prioritizes the spatial features guided by the pre-identified anatomy regions. We adopt a semi-supervised learning method by utilizing available small-scale organ-level annotations to locate the anatomy regions in large-scale datasets where the organ-level annotations are absent. The proposed Anatomy-XNet uses the pre-trained DenseNet-121 as the backbone network with two corresponding structured modules, the Anatomy Aware Attention (A3) and Probabilistic Weighted Average Pooling, in a cohesive framework for anatomical attention learning. We experimentally show that our proposed method sets a new state-of-the-art benchmark by achieving an AUC score of 85.78%, 92.07%, and, 84.04% on three publicly available large-scale CXR datasets-NIH, Stanford CheXpert, and MIMIC-CXR, respectively. This not only proves the efficacy of utilizing the anatomy segmentation knowledge to improve the thoracic disease classification but also demonstrates the generalizability of the proposed framework.

A Weakly supervised approach for thoracic diseases detection

Medical image analysis has become increasingly important as it can effectively assist health care providers with precise clinical diagnosis decisions. Thoracic diseases are one of the leading causes of death in the U.S. Accurate and reliable thoracic disease classification plays a crucial role in early diagnosis and consequently proper treatment plans by providing great assistance to physicians. In the health care domain, clinical diagnoses and decisions are made based on specialist’s opinions which is a tedious and error-prone process. Therefore, computer-aided systems are significantly applied to help health care administrations in decision making as a decision support. Deep learning approaches have shown promising potential in clinical image processing. In this study, an online feature selection method for efficient sampling of high-resolution X-ray images is proposed. A weakly-supervised approach that combines DenseNet-121 layers with a saliency sampling approach to extract the most informative parts or pixels of the input images is proposed. Unlike uniform sampling, the proposed approach preserves the spatial information of input images contributing to an accurate prediction. Performance of the proposed model is evaluated using two public chest X-ray datasets including ChestX-ray14 and CheXpert. The final model that aggregates the predictions over multiple sampling of the input image achieves the averaged AUC of 0.890 and 0.921 for ChestX-ray14 and CheXpert datasets, respectively. Our results for thoracic disease detection demonstrate that the proposed approach can effectively improve the performance of the thoracic disease prediction by identifying and preserving the most important parts of the input images.

DuETNet: Dual Encoder based Transfer Network for thoracic disease classification

• Practical way of achieving memory efficiency under restricted resources is proposed. • Resolution calibration guideline with regard to input image sizes is given. • Entropy based label smoothing considering high class imbalance is proposed. • Channel- and spatial-wise attention modules use query, key, and value matrices. • The attention modules are highly compatible with CNN architectures. In thoracic disease classification , the original chest X-ray images are high resolution images. Nevertheless, in existing convolution neural network (CNN) models, the original images are resized to 224 × 224 before use. Diseases in local areas may not be sufficiently represented because the chest X-ray images have been resized, which compresses information excessively. Therefore, a higher resolution is required to focus on the local representations. Although the large input resolution reduces memory efficiency, previous studies have investigated using CNNs with the large input for classification performance improvement. Moreover, optimization for imbalanced classes is required because chest X-ray images have highly imbalanced pathology labels. Hence, this study proposes the Dual Encoder based Transfer Network (DuETNet) to counter the inefficiency caused by large input resolution and improve classification performance by adjusting the input size based on the RandomResizedCrop method. This image transformation method crops a random area of a given image and resizes it to a given size. Thus, a resolution calibration guideline is a practical way to achieve memory efficiency and performance gains under restricted resources by adjusting the scale factor σ on the training and test images. To treat high class imbalance, we propose entropy based label smoothing method. The method enhances generalization performance for the imbalanced minor classes by penalizing the major classes. The dual encoder comprises channel and spatial encoders, which apply channel- and spatial-wise attention to enhance the relatively significant features from the adjusted images. To evaluate the performance of DuETNet, we used the ChestX-ray14 and MIMIC-CXR-JPG datasets, and DuETNet achieved a new state-of-the-art method.

Consistent response for automated multilabel thoracic disease classification

SummaryWhile recent studies on automated multilabel chest X‐ray (CXR) images classification have shown remarkable progress in leveraging complicated network and attention mechanisms, the automated detection on chest radiographs is still challenging because the pathological patterns are usually highly diverse in their sizes and locations. The CNN model will suffer from the complicated background and high diversity of diseases, which reduce the generalization and performance of the model. To solve these problems, we propose a dual‐distribution consistency (DDC) model, which increases the consistency from two aspects, that is, feature‐level and label‐level. This model integrates two novel loss functions: multilabel response consistency (MRC) loss and distribution consistency (DC) loss. Specifically, we use the original image and its transformed image as inputs to imitate different views of CXR images. The MRC loss encourages the multilabel‐wise attention maps to be consistent between the original CXR image and its transformed counterpart. And the DC loss can force their output probability distributions to be uniform. In this manner, we can make sure that the model can learn discriminative features by using a different view of CXR images. Experiments conducted on the ChestX‐ray14 dataset show the effectiveness of the proposed method.

Modeling long-range dependencies for weakly supervised disease classification and localization on chest X-ray.

Computer-aided diagnosis based on chest X-ray (CXR) is an exponentially growing field of research owing to the development of deep learning, especially convolutional neural networks (CNNs). However, due to the intrinsic locality of convolution operations, CNNs cannot model long-range dependencies. Although vision transformers (ViTs) have recently been proposed to alleviate this limitation, those trained on patches cannot learn any dependencies for inter-patch pixels and thus, are insufficient for medical image detection. To address this problem, in this paper, we propose a CXR detection method which integrates CNN with a ViT for modeling patch-wise and inter-patch dependencies. We experimented on the ChestX-ray14 dataset and followed the official training-test set split. Because the training data only had global annotations, the detection network was weakly supervised. A DenseNet with a feature pyramid structure was designed and integrated with an adaptive ViT to model inter-patch and patch-wise long-range dependencies and obtain fine-grained feature maps. We compared the performance using our method with that of other disease detection methods. For disease classification, our method achieved the best result among all the disease detection methods, with a mean area under the curve (AUC) of 0.829. For lesion localization, our method achieved significantly higher intersection of the union (IoU) scores on the test images with bounding box annotations than did the other detection methods. The visualized results showed that our predictions were more accurate and detailed. Furthermore, evaluation of our method in an external validation dataset demonstrated its generalization ability. Our proposed method achieves the new state of the art for thoracic disease classification and weakly supervised localization. It has potential to assist in clinical decision-making.

A deep learning approach for the classification of TB from NIH CXR dataset

In this research, a novel customized deep learning model is proposed to detect Tuberculosis (TB) from chest X-rays (CXR). The model is utilized for three experimentations: (i) classification of CXR image as healthy or TB infected, (ii) sub-classification of infected images to TB specific manifestations, and (iii) classification of CXR image to thoracic disease manifestations. The National Institute of Health (NIH) CXR is used for experimentation. For the first two experimentations, the subset of the dataset is used containing only 10 TB specific manifestations, whereas, the entire NIH CXR dataset is used for the third experiment. The F1 score for binary classification of TB in experiment 1 is calculated as 0.92 which is higher than the average F1 score of the radiologists. The average accuracy for classifying TB specific manifestations in experiment 2 is recorded as 0.84. Finally, the average accuracy of the thoracic disease classification is recorded as 0.82 in experiment 3. The proposed system outperformed the existing approaches reporting higher AUC for each manifestation. Whereas, to the best of knowledge it is the first such attempt on NIH CXR dataset for TB and TB specific manifestation classification and the proposed system showed promising results.

Weighing features of lung and heart regions for thoracic disease classification

BackgroundChest X-rays are the most commonly available and affordable radiological examination for screening thoracic diseases. According to the domain knowledge of screening chest X-rays, the pathological information usually lay on the lung and heart regions. However, it is costly to acquire region-level annotation in practice, and model training mainly relies on image-level class labels in a weakly supervised manner, which is highly challenging for computer-aided chest X-ray screening. To address this issue, some methods have been proposed recently to identify local regions containing pathological information, which is vital for thoracic disease classification. Inspired by this, we propose a novel deep learning framework to explore discriminative information from lung and heart regions.ResultWe design a feature extractor equipped with a multi-scale attention module to learn global attention maps from global images. To exploit disease-specific cues effectively, we locate lung and heart regions containing pathological information by a well-trained pixel-wise segmentation model to generate binarization masks. By introducing element-wise logical AND operator on the learned global attention maps and the binarization masks, we obtain local attention maps in which pixels are are 1 for lung and heart region and 0 for other regions. By zeroing features of non-lung and heart regions in attention maps, we can effectively exploit their disease-specific cues in lung and heart regions. Compared to existing methods fusing global and local features, we adopt feature weighting to avoid weakening visual cues unique to lung and heart regions. Our method with pixel-wise segmentation can help overcome the deviation of locating local regions. Evaluated by the benchmark split on the publicly available chest X-ray14 dataset, the comprehensive experiments show that our method achieves superior performance compared to the state-of-the-art methods.ConclusionWe propose a novel deep framework for the multi-label classification of thoracic diseases in chest X-ray images. The proposed network aims to effectively exploit pathological regions containing the main cues for chest X-ray screening. Our proposed network has been used in clinic screening to assist the radiologists. Chest X-ray accounts for a significant proportion of radiological examinations. It is valuable to explore more methods for improving performance.

Deep Mining External Imperfect Data for Chest X-Ray Disease Screening.

Deep learning approaches have demonstrated remarkable progress in automatic Chest X-ray analysis. The data-driven feature of deep models requires training data to cover a large distribution. Therefore, it is substantial to integrate knowledge from multiple datasets, especially for medical images. However, learning a disease classification model with extra Chest X-ray (CXR) data is yet challenging. Recent researches have demonstrated that performance bottleneck exists in joint training on different CXR datasets, and few made efforts to address the obstacle. In this paper, we argue that incorporating an external CXR dataset leads to imperfect training data, which raises the challenges. Specifically, the imperfect data is in two folds: domain discrepancy, as the image appearances vary across datasets; and label discrepancy, as different datasets are partially labeled. To this end, we formulate the multi-label thoracic disease classification problem as weighted independent binary tasks according to the categories. For common categories shared across domains, we adopt task-specific adversarial training to alleviate the feature differences. For categories existing in a single dataset, we present uncertainty-aware temporal ensembling of model predictions to mine the information from the missing labels further. In this way, our framework simultaneously models and tackles the domain and label discrepancies, enabling superior knowledge mining ability. We conduct extensive experiments on three datasets with more than 360,000 Chest X-ray images. Our method outperforms other competing models and sets state-of-the-art performance on the official NIH test set with 0.8349 AUC, demonstrating its effectiveness of utilizing the external dataset to improve the internal classification.

Lesion Location Attention Guided Network for Multi-Label Thoracic Disease Classification in Chest X-Rays.

Traditional clinical experiences have shown the benefit of lesion location attention for improving clinical diagnosis tasks. Inspired by this point of interest, in this paper we propose a novel lesion location attention guided network named LLAGnet to focus on the discriminative features from lesion locations for multi-label thoracic disease classification in chest X-rays (CXRs). By revealing the equivalence of the region-level attention (RLA) and channel-level attention (CLA), we find that the RLA is available as priors for object localization while the CLA implicitly provides high weights to the attractive channels, which both enable lesion location attention excitation. To integrate the advantages from both mechanisms, the proposed LLAGnet is structured with two corresponding attention modules, i.e., the RLA and CLA modules. Specifically, the RLA module consists of the global and local branches. And the weakly supervised attention mechanism embedded in the global branch can obtain visual regions of lesion locations by back-propagating gradients. Then the optimal attention region is amplified and applied to the local branch to provide more fine-grained features for the image classification. Finally, the CLA module adaptively enhances the weights of channel-wise features from the lesion locations by modeling interdependencies among channels. Extensive experiments on the ChestX-ray14 dataset clearly substantiate the effectiveness of LLAGnet as compared with the state-of-the-art baselines.