A survey of recent methods for addressing AI fairness and bias in biomedicine.

Objectives: Artificial intelligence (AI) systems have the potential to revolutionize clinical practices, including improving diagnostic accuracy and surgical decision-making, while also reducing costs and manpower. However, it is important to recognize that these systems may perpetuate social inequities or demonstrate biases, such as those based on race or gender. Such biases can occur before, during, or after the development of AI models, making it critical to understand and address potential biases to enable the accurate and reliable application of AI models in clinical settings. To mitigate bias concerns during model development, we surveyed recent publications on different debiasing methods in the fields of biomedical natural language processing (NLP) or computer vision (CV). Then we discussed the methods, such as data perturbation and adversarial learning, that have been applied in the biomedical domain to address bias. Methods: We performed our literature search on PubMed, ACM digital library, and IEEE Xplore of relevant articles published between January 2018 and December 2023 using multiple combinations of keywords. We then filtered the result of 10,041 articles automatically with loose constraints, and manually inspected the abstracts of the remaining 890 articles to identify the 55 articles included in this review. Additional articles in the references are also included in this review. We discuss each method and compare its strengths and weaknesses. Finally, we review other potential methods from the general domain that could be applied to biomedicine to address bias and improve fairness. Results: The bias of AIs in biomedicine can originate from multiple sources such as insufficient data, sampling bias and the use of health-irrelevant features or race-adjusted algorithms. Existing debiasing methods that focus on algorithms can be categorized into distributional or algorithmic. Distributional methods include data augmentation, data perturbation, data reweighting methods, and federated learning. Algorithmic approaches include unsupervised representation learning, adversarial learning, disentangled representation learning, loss-based methods and causality-based methods.

Towards long-tailed, multi-label disease classification from chest X-ray: Overview of the CXR-LT challenge.

Many real-world image recognition problems, such as diagnostic medical imaging exams, are "long-tailed" - there are a few common findings followed by many more relatively rare conditions. In chest radiography, diagnosis is both a long-tailed and multi-label problem, as patients often present with multiple findings simultaneously. While researchers have begun to study the problem of long-tailed learning in medical image recognition, few have studied the interaction of label imbalance and label co-occurrence posed by long-tailed, multi-label disease classification. To engage with the research community on this emerging topic, we conducted an open challenge, CXR-LT, on long-tailed, multi-label thorax disease classification from chest X-rays (CXRs). We publicly release a large-scale benchmark dataset of over 350,000 CXRs, each labeled with at least one of 26 clinical findings following a long-tailed distribution. We synthesize common themes of top-performing solutions, providing practical recommendations for long-tailed, multi-label medical image classification. Finally, we use these insights to propose a path forward involving vision-language foundation models for few- and zero-shot disease classification.

Improving model fairness in image-based computer-aided diagnosis.

Deep learning has become a popular tool for computer-aided diagnosis using medical images, sometimes matching or exceeding the performance of clinicians. However, these models can also reflect and amplify human bias, potentially resulting inaccurate missed diagnoses. Despite this concern, the problem of improving model fairness in medical image classification by deep learning has yet to be fully studied. To address this issue, we propose an algorithm that leverages the marginal pairwise equal opportunity to reduce bias in medical image classification. Our evaluations across four tasks using four independent large-scale cohorts demonstrate that our proposed algorithm not only improves fairness in individual and intersectional subgroups but also maintains overall performance. Specifically, the relative change in pairwise fairness difference between our proposed model and the baseline model was reduced by over 35%, while the relative change in AUC value was typically within 1%. By reducing the bias generated by deep learning models, our proposed approach can potentially alleviate concerns about the fairness and reliability of image-based computer-aided diagnosis.

Carotid-cavernous fistula following mechanical thrombectomy of the tortuous internal carotid artery: A case report.

BACKGROUND: A carotid-cavernous fistula (CCF) is an abnormal connection between the internal carotid artery (ICA) and the cavernous sinus. Although direct CCFs typically result from trauma or as an iatrogenic complication of neuroendovascular procedures, they can occur as surgery-related complications after mechanical thrombectomy (MT). With the widespread use of MT in patients with acute ischemic stroke complicated with large vessel occlusion, it is important to document CCF following MT and how to avoid them. In this study, we present a case of a patient who developed a CCF following MT and describe in detail the characteristics of ICA tortuosity in this case. CASE SUMMARY: A 60-year-old woman experienced weakness in the left upper and lower limbs as well as difficulty speaking for 4 h. The neurological examination revealed left central facial paralysis and left hemiplegia, with a National Institutes of Health Stroke Scale score of 9. Head magnetic resonance imaging revealed an acute cerebral infarction in the right basal ganglia and radial crown. Magnetic resonance angiography demonstrated an occlusion of the right ICA and middle cerebral artery. Digital subtraction angiography demonstrated distal occlusion of the cervical segment of the right ICA. We performed suction combined with stent thrombectomy. Then, postoperative angiography was performed, which showed a right CCF. One month later, CCF embolization was performed, and the patient's clinical symptoms have significantly improved 5 mo after the operation. CONCLUSION: Although a CCF is a rare complication after MT, it should be considered. Understanding the tortuosity of the internal carotid-cavernous sinus may help predict the complexity of MT and avoid this complication.

A scoping review on multimodal deep learning in biomedical images and texts.

Computer-assisted diagnostic and prognostic systems of the future should be capable of simultaneously processing multimodal data. Multimodal deep learning (MDL), which involves the integration of multiple sources of data, such as images and text, has the potential to revolutionize the analysis and interpretation of biomedical data. However, it only caught researchers' attention recently. To this end, there is a critical need to conduct a systematic review on this topic, identify the limitations of current work, and explore future directions. In this scoping review, we aim to provide a comprehensive overview of the current state of the field and identify key concepts, types of studies, and research gaps with a focus on biomedical images and texts joint learning, mainly because these two were the most commonly available data types in MDL research. This study reviewed the current uses of multimodal deep learning on five tasks: (1) Report generation, (2) Visual question answering, (3) Cross-modal retrieval, (4) Computer-aided diagnosis, and (5) Semantic segmentation. Our results highlight the diverse applications and potential of MDL and suggest directions for future research in the field. We hope our review will facilitate the collaboration of natural language processing (NLP) and medical imaging communities and support the next generation of decision-making and computer-assisted diagnostic system development.

A scoping review on multimodal deep learning in biomedical images and texts.

OBJECTIVE: Computer-assisted diagnostic and prognostic systems of the future should be capable of simultaneously processing multimodal data. Multimodal deep learning (MDL), which involves the integration of multiple sources of data, such as images and text, has the potential to revolutionize the analysis and interpretation of biomedical data. However, it only caught researchers' attention recently. To this end, there is a critical need to conduct a systematic review on this topic, identify the limitations of current work, and explore future directions. METHODS: In this scoping review, we aim to provide a comprehensive overview of the current state of the field and identify key concepts, types of studies, and research gaps with a focus on biomedical images and texts joint learning, mainly because these two were the most commonly available data types in MDL research. RESULT: This study reviewed the current uses of multimodal deep learning on five tasks: (1) Report generation, (2) Visual question answering, (3) Cross-modal retrieval, (4) Computer-aided diagnosis, and (5) Semantic segmentation. CONCLUSION: Our results highlight the diverse applications and potential of MDL and suggest directions for future research in the field. We hope our review will facilitate the collaboration of natural language processing (NLP) and medical imaging communities and support the next generation of decision-making and computer-assisted diagnostic system development.

A new classification method for diagnosing COVID-19 pneumonia based on joint CNN features of chest X-ray images and parallel pyramid MLP-mixer module.

During the past three years, the coronavirus disease 2019 (COVID-19) has swept the world. The rapid and accurate recognition of covid-19 pneumonia are ,therefore, of great importance. To handle this problem, we propose a new pipeline of deep learning framework for diagnosing COVID-19 pneumonia via chest X-ray images from normal, COVID-19, and other pneumonia patients. In detail, the self-trained YOLO-v4 network was first used to locate and segment the thoracic region, and the output images were scaled to the same size. Subsequently, the pre-trained convolutional neural network was adopted to extract the features of X-ray images from 13 convolutional layers, which were fused with the original image to form a 14-dimensional image matrix. It was then put into three parallel pyramid multi-layer perceptron (MLP)-Mixer modules for comprehensive feature extraction through spatial fusion and channel fusion based on different scales so as to grasp more extensive feature correlation. Finally, by combining all image features from the 14-channel output, the classification task was achieved using two fully connected layers as well as Softmax classifier for classification. Extensive simulations based on a total of 4099 chest X-ray images were conducted to verify the effectiveness of the proposed method. Experimental results indicated that our proposed method can achieve the best performance in almost all cases, which is good for auxiliary diagnosis of COVID-19 and has great clinical application potential.

Evaluate underdiagnosis and overdiagnosis bias of deep learning model on primary open-angle glaucoma diagnosis in under-served populations.

In the United States, primary open-angle glaucoma (POAG) is the leading cause of blindness, especially among African American and Hispanic individuals. Deep learning has been widely used to detect POAG using fundus images as its performance is comparable to or even surpasses diagnosis by clinicians. However, human bias in clinical diagnosis may be reflected and amplified in the widely-used deep learning models, thus impacting their performance. Biases may cause (1) underdiagnosis, increasing the risks of delayed or inadequate treatment, and (2) overdiagnosis, which may increase individuals' stress, fear, well-being, and unnecessary/costly treatment. In this study, we examined the underdiagnosis and overdiagnosis when applying deep learning in POAG detection based on the Ocular Hypertension Treatment Study (OHTS) from 22 centers across 16 states in the United States. Our results show that the widely-used deep learning model can underdiagnose or overdiagnose under-served populations. The most underdiagnosed group is female younger (< 60 yrs) group, and the most overdiagnosed group is Black older (≥ 60 yrs) group. Biased diagnosis through traditional deep learning methods may delay disease detection, treatment and create burdens among under-served populations, thereby, raising ethical concerns about using deep learning models in ophthalmology clinics.

Enhancing thoracic disease detection using chest X-rays from PubMed Central Open Access.

Large chest X-rays (CXR) datasets have been collected to train deep learning models to detect thorax pathology on CXR. However, most CXR datasets are from single-center studies and the collected pathologies are often imbalanced. The aim of this study was to automatically construct a public, weakly-labeled CXR database from articles in PubMed Central Open Access (PMC-OA) and to assess model performance on CXR pathology classification by using this database as additional training data. Our framework includes text extraction, CXR pathology verification, subfigure separation, and image modality classification. We have extensively validated the utility of the automatically generated image database on thoracic disease detection tasks, including Hernia, Lung Lesion, Pneumonia, and pneumothorax. We pick these diseases due to their historically poor performance in existing datasets: the NIH-CXR dataset (112,120 CXR) and the MIMIC-CXR dataset (243,324 CXR). We find that classifiers fine-tuned with additional PMC-CXR extracted by the proposed framework consistently and significantly achieved better performance than those without (e.g., Hernia: 0.9335 vs 0.9154; Lung Lesion: 0.7394 vs. 0.7207; Pneumonia: 0.7074 vs. 0.6709; Pneumothorax 0.8185 vs. 0.7517, all in AUC with p< 0.0001) for CXR pathology detection. In contrast to previous approaches that manually submit the medical images to the repository, our framework can automatically collect figures and their accompanied figure legends. Compared to previous studies, the proposed framework improved subfigure segmentation and incorporates our advanced self-developed NLP technique for CXR pathology verification. We hope it complements existing resources and improves our ability to make biomedical image data findable, accessible, interoperable, and reusable.

Person re-identification via semi-supervised adaptive graph embedding.

Video surveillance is an indispensable part of the smart city for public safety and security. Person Re-Identification (Re-ID), as one of elementary learning tasks for video surveillance, is to track and identify a given pedestrian in a multi-camera scene. In general, most existing methods has firstly adopted a CNN based detector to obtain the cropped pedestrian image, it then aims to learn a specific distance metric for retrieval. However, unlabeled gallery images are generally overlooked and not utilized in the training. On the other hands, Manifold Embedding (ME) has well been applied to Person Re-ID as it is good to characterize the geometry of database associated with the query data. However, ME has its limitation to be scalable to large-scale data due to the huge computational complexity for graph construction and ranking. To handle this problem, we in this paper propose a novel scalable manifold embedding approach for Person Re-ID task. The new method is to incorporate both graph weight construction and manifold regularized term in the same framework. The graph we developed is discriminative and doubly-stochastic so that the side information has been considered so that it can enhance the clustering performances. The doubly-stochastic property can also guarantee the graph is highly robust and less sensitive to the parameters. Meriting from such a graph, we then incorporate the graph construction, the subspace learning method in the unified loss term. Therefore, the subspace results can be utilized into the graph construction, and the updated graph can in turn incorporate discriminative information for graph embedding. Extensive simulations is conducted based on three benchmark Person Re-ID datasets and the results verify that the proposed method can achieve better ranking performance compared with other state-of-the-art graph-based methods.

Primary Open-Angle Glaucoma Diagnosis from Optic Disc Photographs Using a Siamese Network.

Purpose: Primary open-angle glaucoma (POAG) is one of the leading causes of irreversible blindness in the United States and worldwide. Although deep learning methods have been proposed to diagnose POAG, these methods all used a single image as input. Contrastingly, glaucoma specialists typically compare the follow-up image with the baseline image to diagnose incident glaucoma. To simulate this process, we proposed a Siamese neural network, POAGNet, to detect POAG from optic disc photographs. Design: The POAGNet, an algorithm for glaucoma diagnosis, is developed using optic disc photographs. Participants: The POAGNet was trained and evaluated on 2 data sets: (1) 37 339 optic disc photographs from 1636 Ocular Hypertension Treatment Study (OHTS) participants and (2) 3684 optic disc photographs from the Sequential fundus Images for Glaucoma (SIG) data set. Gold standard labels were obtained using reading center grades. Methods: We proposed a Siamese network model, POAGNet, to simulate the clinical process of identifying POAG from optic disc photographs. The POAGNet consists of 2 side outputs for deep supervision and uses convolution to measure the similarity between 2 networks. Main Outcome Measures: The main outcome measures are the area under the receiver operating characteristic curve, accuracy, sensitivity, and specificity. Results: In POAG diagnosis, extensive experiments show that POAGNet performed better than the best state-of-the-art model on the OHTS test set (area under the curve [AUC] 0.9587 versus 0.8750). It also outperformed the baseline models on the SIG test set (AUC 0.7518 versus 0.6434). To assess the transferability of POAGNet, we also validated the impact of cross-data set variability on our model. The model trained on OHTS achieved an AUC of 0.7490 on SIG, comparable to the previous model trained on the same data set. When using the combination of SIG and OHTS for training, our model achieved superior AUC to the single-data model (AUC 0.8165 versus 0.7518). These demonstrate the relative generalizability of POAGNet. Conclusions: By simulating the clinical grading process, POAGNet demonstrated high accuracy in POAG diagnosis. These results highlight the potential of deep learning to assist and enhance clinical POAG diagnosis. The POAGNet is publicly available on https://github.com/bionlplab/poagnet.

Radiology Text Analysis System (RadText): Architecture and Evaluation.

Analyzing radiology reports is a time-consuming and error-prone task, which raises the need for an efficient automated radiology report analysis system to alleviate the workloads of radiologists and encourage precise diagnosis. In this work, we present RadText, a high-performance open-source Python radiology text analysis system. RadText offers an easy-to-use text analysis pipeline, including de-identification, section segmentation, sentence split and word tokenization, named entity recognition, parsing, and negation detection. Superior to existing widely used toolkits, RadText features a hybrid text processing schema, supports raw text processing and local processing, which enables higher accuracy, better usability and improved data privacy. RadText adopts BioC as the unified interface, and also standardizes the output into a structured representation that is compatible with Observational Medical Outcomes Partnership (OMOP) Common Data Model (CDM), which allows for a more systematic approach to observational research across multiple, disparate data sources. We evaluated RadText on the MIMIC-CXR dataset, with five new disease labels that we annotated for this work. RadText demonstrates highly accurate classification performances, with a 0.91 average precision, 0.94 average recall and 0.92 average F-1 score. We also annotated a test set for the five new disease labels to facilitate future research or applications. We have made our code, documentations, examples and the test set available at https://github.com/bionlplab/radtext.

Supervised Pretraining through Contrastive Categorical Positive Samplings to Improve COVID-19 Mortality Prediction.

Clinical EHR data is naturally heterogeneous, where it contains abundant sub-phenotype. Such diversity creates challenges for outcome prediction using a machine learning model since it leads to high intra-class variance. To address this issue, we propose a supervised pre-training model with a unique embedded k-nearest-neighbor positive sampling strategy. We demonstrate the enhanced performance value of this framework theoretically and show that it yields highly competitive experimental results in predicting patient mortality in real-world COVID-19 EHR data with a total of over 7,000 patients admitted to a large, urban health system. Our method achieves a better AUROC prediction score of 0.872, which outperforms the alternative pre-training models and traditional machine learning methods. Additionally, our method performs much better when the training data size is small (345 training instances).

Automated diagnosing primary open-angle glaucoma from fundus image by simulating human's grading with deep learning.

Primary open-angle glaucoma (POAG) is a leading cause of irreversible blindness worldwide. Although deep learning methods have been proposed to diagnose POAG, it remains challenging to develop a robust and explainable algorithm to automatically facilitate the downstream diagnostic tasks. In this study, we present an automated classification algorithm, GlaucomaNet, to identify POAG using variable fundus photographs from different populations and settings. GlaucomaNet consists of two convolutional neural networks to simulate the human grading process: learning the discriminative features and fusing the features for grading. We evaluated GlaucomaNet on two datasets: Ocular Hypertension Treatment Study (OHTS) participants and the Large-scale Attention-based Glaucoma (LAG) dataset. GlaucomaNet achieved the highest AUC of 0.904 and 0.997 for POAG diagnosis on OHTS and LAG datasets. An ensemble of network architectures further improved diagnostic accuracy. By simulating the human grading process, GlaucomaNet demonstrated high accuracy with increased transparency in POAG diagnosis (comprehensiveness scores of 97% and 36%). These methods also address two well-known challenges in the field: the need for increased image data diversity and relying heavily on perimetry for POAG diagnosis. These results highlight the potential of deep learning to assist and enhance clinical POAG diagnosis. GlaucomaNet is publicly available on https://github.com/bionlplab/GlaucomaNet .

Prior Knowledge Enhances Radiology Report Generation.

Radiology report generation aims to produce computer-aided diagnoses to alleviate the workload of radiologists and has drawn increasing attention recently. However, previous deep learning methods tend to neglect the mutual influences between medical findings, which can be the bottleneck that limits the quality of generated reports. In this work, we propose to mine and represent the associations among medical findings in an informative knowledge graph and incorporate this prior knowledge with radiology report generation to help improve the quality of generated reports. Experiment results demonstrate the superior performance of our proposed method on the IU X-ray dataset with a ROUGE-L of 0.384±0.007 and CIDEr of 0.340±0.011. Compared with previous works, our model achieves an average of 1.6% improvement (2.0% and 1.5% improvements in CIDEr and ROUGE-L, respectively). The experiments suggest that prior knowledge can bring performance gains to accurate radiology report generation. We will make the code publicly available at https://github.com/bionlplab/report_generation_amia2022.

Knowledge Graph Applications in Medical Imaging Analysis: A Scoping Review.

Background: There is an increasing trend to represent domain knowledge in structured graphs, which provide efficient knowledge representations for many downstream tasks. Knowledge graphs are widely used to model prior knowledge in the form of nodes and edges to represent semantically connected knowledge entities, which several works have adopted into different medical imaging applications. Methods: We systematically searched over five databases to find relevant articles that applied knowledge graphs to medical imaging analysis. After screening, evaluating, and reviewing the selected articles, we performed a systematic analysis. Results: We looked at four applications in medical imaging analysis, including disease classification, disease localization and segmentation, report generation, and image retrieval. We also identified limitations of current work, such as the limited amount of available annotated data and weak generalizability to other tasks. We further identified the potential future directions according to the identified limitations, including employing semisupervised frameworks to alleviate the need for annotated data and exploring task-agnostic models to provide better generalizability. Conclusions: We hope that our article will provide the readers with aggregated documentation of the state-of-the-art knowledge graph applications for medical imaging to encourage future research.

Cascaded Triplanar Autoencoder M-Net for Fully Automatic Segmentation of Left Ventricle Myocardial Scar From Three-Dimensional Late Gadolinium-Enhanced MR Images.

While three-dimensional (3D) late gadolinium-enhanced (LGE) magnetic resonance (MR) imaging provides good conspicuity of small myocardial lesions with short acquisition time, it poses a challenge for image analysis as a large number of axial images are required to be segmented. We developed a fully automatic convolutional neural network (CNN) called cascaded triplanar autoencoder M-Net (CTAEM-Net) to segment myocardial scar from 3D LGE MRI. Two sub-networks were cascaded to segment the left ventricle (LV) myocardium and then the scar within the pre-segmented LV myocardium. Each sub-network contains three autoencoder M-Nets (AEM-Nets) segmenting the axial, sagittal and coronal slices of the 3D LGE MR image, with the final segmentation determined by voting. The AEM-Net integrates three features: (1) multi-scale inputs, (2) deep supervision and (3) multi-tasking. The multi-scale inputs allow consideration of the global and local features in segmentation. Deep supervision provides direct supervision to deeper layers and facilitates CNN convergence. Multi-task learning reduces segmentation overfitting by acquiring additional information from autoencoder reconstruction, a task closely related to segmentation. The framework provides an accuracy of 86.43% and 90.18% for LV myocardium and scar segmentation, respectively, which are the highest among existing methods to our knowledge. The time required for CTAEM-Net to segment LV myocardium and the scar was 49.72 ± 9.69s and 120.25 ± 23.18s per MR volume, respectively. The accuracy and efficiency afforded by CTAEM-Net will make possible future large population studies. The generalizability of the framework was also demonstrated by its competitive performance in two publicly available datasets of different imaging modalities.

Multi-scale Multi-structure Siamese Network (MMSNet) for Primary Open-Angle Glaucoma Prediction.

Primary open-angle glaucoma (POAG) is one of the leading causes of irreversible blindness in the United States and worldwide. POAG prediction before onset plays an important role in early treatment. Although deep learning methods have been proposed to predict POAG, these methods mainly focus on current status prediction. In addition, all these methods used a single image as input. On the other hand, glaucoma specialists determine a glaucomatous eye by comparing the follow-up optic nerve image with the baseline along with supplementary clinical data. To simulate this process, we proposed a Multi-scale Multi-structure Siamese Network (MMSNet) to predict future POAG event from fundus photographs. The MMSNet consists of two side-outputs for deep supervision and 2D blocks to utilize two-dimensional features to assist classification. The MMSNet network was trained and evaluated on a large dataset: 37,339 fundus photographs from 1,636 Ocular Hypertension Treatment Study (OHTS) participants. Extensive experiments show that MMSNet outperforms the state-of-the-art on two "POAG prediction before onset" tasks. Our AUC are 0.9312 and 0.9507, which are 0.2204 and 0.1490 higher than the state-of-the-art, respectively. In addition, an ablation study is performed to check the contribution of different components. These results highlight the potential of deep learning to assist and enhance the prediction of future POAG event. The proposed network will be publicly available on https://github.com/bionlplab/MMSNet.

Multimodal Image Classification by Multiview Latent Pattern Extraction, Selection, and Correlation.

The large amount of data available in the modern big data era opens new opportunities to expand our knowledge by integrating information from heterogeneous sources. Multiview learning has recently achieved tremendous success in deriving complementary information from multiple data modalities. This article proposes a framework called multiview latent space projection (MVLSP) to integrate features extracted from multiple sources in a discriminative way to facilitate binary and multiclass classifications. Our approach is associated with three innovations. First, most existing multiview learning algorithms promote pairwise consistency between two views and do not have a natural extension to applications with more than two views. MVLSP finds optimum mappings from a common latent space to match the feature space in each of the views. As the matching is performed on a view-by-view basis, the framework can be readily extended to multiview applications. Second, feature selection in the common latent space can be readily achieved by adding a class view, which matches the latent space representations of training samples with their corresponding labels. Then, high-order view correlations are extracted by considering feature-label correlations. Third, a technique is proposed to optimize the integration of different latent patterns based on their correlations. The experimental results on the prostate image dataset demonstrate the effectiveness of the proposed method.

Artificial intelligence in tumor subregion analysis based on medical imaging: A review.

Medical imaging is widely used in the diagnosis and treatment of cancer, and artificial intelligence (AI) has achieved tremendous success in medical image analysis. This paper reviews AI-based tumor subregion analysis in medical imaging. We summarize the latest AI-based methods for tumor subregion analysis and their applications. Specifically, we categorize the AI-based methods by training strategy: supervised and unsupervised. A detailed review of each category is presented, highlighting important contributions and achievements. Specific challenges and potential applications of AI in tumor subregion analysis are discussed.