On Combining Multiple-Instance Learning and Active Learning for Computer-Aided Detection of Tuberculosis

AI for Detection of Tuberculosis: Implications for Global Health

Tuberculosis, which primarily affects developing countries, remains a significant global health concern. Since the 2010s, the role of chest radiography has expanded in tuberculosis triage and screening beyond its traditional complementary role in the diagnosis of tuberculosis. Computer-aided diagnosis (CAD) systems for tuberculosis detection on chest radiographs have recently made substantial progress in diagnostic performance, thanks to deep learning technologies. The current performance of CAD systems for tuberculosis has approximated that of human experts, presenting a potential solution to the shortage of human readers to interpret chest radiographs in low- or middle-income, high-tuberculosis-burden countries. This article provides a critical appraisal of developmental process reporting in extant CAD software for tuberculosis, based on the Checklist for Artificial Intelligence in Medical Imaging. It also explores several considerations to scale up CAD solutions, encompassing manufacturer-independent CAD validation, economic and political aspects, and ethical concerns, as well as the potential for broadening radiography-based diagnosis to other nontuberculosis diseases. Collectively, CAD for tuberculosis will emerge as a representative deep learning application, catalyzing advances in global health and health equity. Keywords: Computer-aided Diagnosis (CAD), Conventional Radiography, Thorax, Lung, Machine Learning Supplemental material is available for this article. © RSNA, 2024.

Quantitative Chest X-ray Radiomics for Therapy Response Monitoring in Patients with Pulmonary Tuberculosis

Tuberculosis (TB) remains the second leading cause of death globally from a single infectious agent, and there is a critical need to develop improved imaging biomarkers and aid rapid assessments of responses to therapy. We aimed to utilize radiomics, a rapidly developing image analysis tool, to develop a scoring system for this purpose. A chest X-ray radiomics score (RadScore) was developed by implementing a unique segmentation method, followed by feature extraction and parameter map construction. Signature parameter maps that showed a high correlation to lung pathology were consolidated into four frequency bins to obtain the RadScore. A clinical score (TBscore) and a radiological score (RLscore) were also developed based on existing scoring algorithms. The correlation between the change in the three scores, calculated from serial X-rays taken while patients received TB therapy, was evaluated using Spearman's correlation. Poor correlations were observed between the changes in the TBscore and the RLscore (0.09 (p-value = 0.36)) and the TBscore and the RadScore (0.02 (p-value = 0.86)). The changes in the RLscore and the RadScore had a much stronger correlation of 0.22, which is statistically significant (p-value = 0.02). This shows that the developed RadScore has the potential to be a quantitative monitoring tool for responses to therapy.

Introducing a secondary segmentation to construct a radiomics model for pulmonary tuberculosis cavities

PURPOSE

Accurate segmentation (separating diseased portions of the lung from normal appearing lung) is a challenge in radiomic studies of non-neoplastic diseases, such as pulmonary tuberculosis (PTB). In this study, we developed a segmentation method, applicable to chest X-rays (CXR), that can eliminate the need for precise disease delineation, and that is effective for constructing radiomic models for automatic PTB cavity classification.

METHODS

This retrospective study used a dataset of 266 posteroanterior CXR of patients diagnosed with laboratory confirmed PTB. The lungs were segmented using a U-net-based in-house automatic segmentation model. A secondary segmentation was developed using a sliding window, superimposed on the primary lung segmentation. Pyradiomics was used for feature extraction from every window which increased the dimensionality of the data, but this allowed us to accurately capture the spread of the features across the lung. Two separate measures (standard-deviation and variance) were used to consolidate the features. Pearson's correlation analysis (with a 0.8 cut-off value) was then applied for dimensionality reduction followed by the construction of Random Forest radiomic models.

RESULTS

Two almost identical radiomic signatures consisting of 10 texture features each (9 were the same plus 1 other feature) were identified using the two separate consolidation measures. Two well performing random forest models were constructed from these signatures. The standard-deviation model (AUC = 0.9444 (95% CI, 0.8762; 0.9814)) performed marginally better than the variance model (AUC = 0.9288 (95% CI, 0.9046; 0.9843)).

CONCLUSION

The introduction of the secondary sliding window segmentation on CXR could eliminate the need for disease delineation in pulmonary radiomic studies, and it could improve the accuracy of CXR reporting currently regaining prominence as a high-volume screening tool as the developed radiomic models correctly classify cavities from normal CXR.

Deep joint learning of pathological region localization and Alzheimer's disease diagnosis

The identification of Alzheimer's disease (AD) using structural magnetic resonance imaging (sMRI) has been studied based on the subtle morphological changes in the brain. One of the typical approaches is a deep learning-based patch-level feature representation. For this approach, however, the predetermined patches before learning the diagnostic model can limit classification performance. To mitigate this problem, we propose the BrainBagNet with a position-based gate (PG), which applies position information of brain images represented through the 3D coordinates. Our proposed method represents the patch-level class evidence based on both MR scan and position information for image-level prediction. To validate the effectiveness of our proposed framework, we conducted comprehensive experiments comparing it with state-of-the-art methods, utilizing two publicly available datasets: the Alzheimer's Disease Neuroimaging Initiative (ADNI) and the Australian Imaging, Biomarkers and Lifestyle (AIBL) dataset. Furthermore, our experimental results demonstrate that our proposed method outperforms the existing competing methods in terms of classification performance for both AD diagnosis and mild cognitive impairment conversion prediction tasks. In addition, we performed various analyses of the results from diverse perspectives to obtain further insights into the underlying mechanisms and strengths of our proposed framework. Based on the results of our experiments, we demonstrate that our proposed framework has the potential to advance deep-learning-based patch-level feature representation studies for AD diagnosis and MCI conversion prediction. In addition, our method provides valuable insights, such as interpretability, and the ability to capture subtle changes, into the underlying pathological processes of AD and MCI, benefiting both researchers and clinicians.

Tuberculosis Disease Diagnosis Based on an Optimized Machine Learning Model

Computer science plays an important role in modern dynamic health systems. Given the collaborative nature of the diagnostic process, computer technology provides important services to healthcare professionals and organizations, as well as to patients and their families, researchers, and decision-makers. Thus, any innovations that improve the diagnostic process while maintaining quality and safety are crucial to the development of the healthcare field. Many diseases can be tentatively diagnosed during their initial stages. In this study, all developed techniques were applied to tuberculosis (TB). Thus, we propose an optimized machine learning-based model that extracts optimal texture features from TB-related images and selects the hyper-parameters of the classifiers. Increasing the accuracy rate and minimizing the number of characteristics extracted are our goals. In other words, this is a multitask optimization issue. A genetic algorithm (GA) is used to choose the best features, which are then fed into a support vector machine (SVM) classifier. Using the ImageCLEF 2020 data set, we conducted experiments using the proposed approach and achieved significantly higher accuracy and better outcomes in comparison with the state-of-the-art works. The obtained experimental results highlight the efficiency of modified SVM classifier compared with other standard ones.

Weakly Supervised Liver Tumor Segmentation Using Couinaud Segment Annotation

Automatic liver tumor segmentation is of great importance for assisting doctors in liver cancer diagnosis and treatment planning. Recently, deep learning approaches trained with pixel-level annotations have contributed many breakthroughs in image segmentation. However, acquiring such accurate dense annotations is time-consuming and labor-intensive, which limits the performance of deep neural networks for medical image segmentation. We note that Couinaud segment is widely used by radiologists when recording liver cancer-related findings in the reports, since it is well-suited for describing the localization of tumors. In this paper, we propose a novel approach to train convolutional networks for liver tumor segmentation using Couinaud segment annotations. Couinaud segment annotations are image-level labels with values ranging from 1 to 8, indicating a specific region of the liver. Our proposed model, namely CouinaudNet, can estimate pseudo tumor masks from the Couinaud segment annotations as pixel-wise supervision for training a fully supervised tumor segmentation model, and it is composed of two components: 1) an inpainting network with Couinaud segment masks which can effectively remove tumors for pathological images by filling the tumor regions with plausible healthy-looking intensities; 2) a difference spotting network for segmenting the tumors, which is trained with healthy-pathological pairs generated by an effective tumor synthesis strategy. The proposed method is extensively evaluated on two liver tumor segmentation datasets. The experimental results demonstrate that our method can achieve competitive performance compared to the fully supervised counterpart and the state-of-the-art methods while requiring significantly less annotation effort.

Pulmonary tuberculosis diagnosis, differentiation and disease management: A review of radiomics applications

Pulmonary tuberculosis is a worldwide epidemic that can only be fought effectively with early and accurate diagnosis and proper disease management. The means of diagnosis and disease management should be easily accessible, cost effective and be readily available in the high tuberculosis burdened countries where it is most needed. Fortunately, the fast development of computer science in recent years has ensured that medical images can accurately be quantified. Radiomics is one such tool that can be used to quantify medical images. This review article focuses on the literature currently available on the application of radiomics explicitly for the purpose of diagnosis, differentiation from other pulmonary diseases and disease management of pulmonary tuberculosis. Despite using a formal search strategy, only five articles could be found on the application of radiomics to pulmonary tuberculosis. In all five articles reviewed, radiomic feature extraction was successfully used to quantify digital medical images for the purpose of comparing, or differentiating, pulmonary tuberculosis from other pulmonary diseases. This demonstrates that the use of radiomics for the purpose of tuberculosis disease management and diagnosis remains a valuable data mining opportunity not yet realised.

Deep embeddings and logistic regression for rapid active learning in histopathological images

BACKGROUND AND OBJECTIVE

Recognizing different tissue components is one of the most fundamental and essential works in digital pathology. Current methods are often based on convolutional neural networks (CNNs), which need numerous annotated samples for training. Creating large-scale histopathological datasets is labor-intensive, where interactive data annotation is a potential solution.

METHODS

We propose DELR (Deep Embedding-based Logistic Regression) to enable rapid model training and inference for histopathological image analysis. DELR utilizes a pretrained CNN to encode images as compact embeddings with low computational cost. The embeddings are then used to train a Logistic Regression model efficiently. We implemented DELR in an active learning framework, and validated it on three histopathological problems (binary, 4-category, and 8-category classification challenge for lung, breast, and colorectal cancer, respectively). We also investigated the influence of active learning strategy and type of the encoder.

RESULTS

On all the three datasets, DELR can achieve an area under curve (AUC) metric higher than 0.95 with only 100 image patches per class. Although its AUC is slightly lower than a fine-tuned CNN counterpart, DELR can be 536, 316, and 1481 times faster after pre-encoding. Moreover, DELR is proved to be compatible with a variety of active learning strategies and encoders.

CONCLUSIONS

DELR can achieve comparable accuracy to CNN with rapid running speed. These advantages make it a potential solution for real-time interactive data annotation.

A multi-resolution model for histopathology image classification and localization with multiple instance learning

Large numbers of histopathological images have been digitized into high resolution whole slide images, opening opportunities in developing computational image analysis tools to reduce pathologists' workload and potentially improve inter- and intra-observer agreement. Most previous work on whole slide image analysis has focused on classification or segmentation of small pre-selected regions-of-interest, which requires fine-grained annotation and is non-trivial to extend for large-scale whole slide analysis. In this paper, we proposed a multi-resolution multiple instance learning model that leverages saliency maps to detect suspicious regions for fine-grained grade prediction. Instead of relying on expensive region- or pixel-level annotations, our model can be trained end-to-end with only slide-level labels. The model is developed on a large-scale prostate biopsy dataset containing 20,229 slides from 830 patients. The model achieved 92.7% accuracy, 81.8% Cohen's Kappa for benign, low grade (i.e. Grade group 1) and high grade (i.e. Grade group ≥ 2) prediction, an area under the receiver operating characteristic curve (AUROC) of 98.2% and an average precision (AP) of 97.4% for differentiating malignant and benign slides. The model obtained an AUROC of 99.4% and an AP of 99.8% for cancer detection on an external dataset.

Use of Digital Technology to Enhance Tuberculosis Control: Scoping Review

Background

Tuberculosis (TB) is the leading cause of death from a single infectious agent, with around 1.5 million deaths reported in 2018, and is a major contributor to suffering worldwide, with an estimated 10 million new cases every year. In the context of the World Health Organization’s End TB strategy and the quest for digital innovations, there is a need to understand what is happening around the world regarding research into the use of digital technology for better TB care and control.

Objective

The purpose of this scoping review was to summarize the state of research on the use of digital technology to enhance TB care and control. This study provides an overview of publications covering this subject and answers 3 main questions: (1) to what extent has the issue been addressed in the scientific literature between January 2016 and March 2019, (2) which countries have been investing in research in this field, and (3) what digital technologies were used?

Methods

A Web-based search was conducted on PubMed and Web of Science. Studies that describe the use of digital technology with specific reference to keywords such as TB, digital health, eHealth, and mHealth were included. Data from selected studies were synthesized into 4 functions using narrative and graphical methods. Such digital health interventions were categorized based on 2 classifications, one by function and the other by targeted user.

Results

A total of 145 relevant studies were identified out of the 1005 published between January 2016 and March 2019. Overall, 72.4% (105/145) of the research focused on patient care and 20.7% (30/145) on surveillance and monitoring. Other programmatic functions 4.8% (7/145) and electronic learning 2.1% (3/145) were less frequently studied. Most digital health technologies used for patient care included primarily diagnostic 59.4% (63/106) and treatment adherence tools 40.6% (43/106). On the basis of the second type of classification, 107 studies targeted health care providers (107/145, 73.8%), 20 studies targeted clients (20/145, 13.8%), 17 dealt with data services (17/145, 11.7%), and 1 study was on the health system or resource management. The first authors’ affiliations were mainly from 3 countries: the United States (30/145 studies, 20.7%), China (20/145 studies, 13.8%), and India (17/145 studies, 11.7%). The researchers from the United States conducted their research both domestically and abroad, whereas researchers from China and India conducted all studies domestically.

Conclusions

The majority of research conducted between January 2016 and March 2019 on digital interventions for TB focused on diagnostic tools and treatment adherence technologies, such as video-observed therapy and SMS. Only a few studies addressed interventions for data services and health system or resource management.

Uncertainty Assisted Robust Tuberculosis Identification With Bayesian Convolutional Neural Networks

Tuberculosis (TB) is an infectious disease that can lead towards death if left untreated. TB detection involves extraction of complex TB manifestation features such as lung cavity, air space consolidation, endobronchial spread, and pleural effusions from chest x-rays (CXRs). Deep learning based approach named convolutional neural network (CNN) has the ability to learn complex features from CXR images. The main problem is that CNN does not consider uncertainty to classify CXRs using softmax layer. It lacks in presenting the true probability of CXRs by differentiating confusing cases during TB detection. This paper presents the solution for TB identification by using Bayesian-based convolutional neural network (B-CNN). It deals with the uncertain cases that have low discernibility among the TB and non-TB manifested CXRs. The proposed TB identification methodology based on B-CNN is evaluated on two TB benchmark datasets, i.e., Montgomery and Shenzhen. For training and testing of proposed scheme we have utilized Google Colab platform which provides NVidia Tesla K80 with 12 GB of VRAM, single core of 2.3 GHz Xeon Processor, 12 GB RAM and 320 GB of disk. B-CNN achieves 96.42% and 86.46% accuracy on both dataset, respectively as compared to the state-of-the-art machine learning and CNN approaches. Moreover, B-CNN validates its results by filtering the CXRs as confusion cases where the variance of B-CNN predicted outputs is more than a certain threshold. Results prove the supremacy of B-CNN for the identification of TB and non-TB sample CXRs as compared to counterparts in terms of accuracy, variance in the predicted probabilities and model uncertainty.

A systematic review of the diagnostic accuracy of artificial intelligence-based computer programs to analyze chest x-rays for pulmonary tuberculosis

We undertook a systematic review of the diagnostic accuracy of artificial intelligence-based software for identification of radiologic abnormalities (computer-aided detection, or CAD) compatible with pulmonary tuberculosis on chest x-rays (CXRs). We searched four databases for articles published between January 2005-February 2019. We summarized data on CAD type, study design, and diagnostic accuracy. We assessed risk of bias with QUADAS-2. We included 53 of the 4712 articles reviewed: 40 focused on CAD design methods (“Development” studies) and 13 focused on evaluation of CAD (“Clinical” studies). Meta-analyses were not performed due to methodological differences. Development studies were more likely to use CXR databases with greater potential for bias as compared to Clinical studies. Areas under the receiver operating characteristic curve (median AUC [IQR]) were significantly higher: in Development studies AUC: 0.88 [0.82–0.90]) versus Clinical studies (0.75 [0.66–0.87]; p-value 0.004); and with deep-learning (0.91 [0.88–0.99]) versus machine-learning (0.82 [0.75–0.89]; p = 0.001). We conclude that CAD programs are promising, but the majority of work thus far has been on development rather than clinical evaluation. We provide concrete suggestions on what study design elements should be improved.

Bag-Level Aggregation for Multiple-Instance Active Learning in Instance Classification Problems

A growing number of applications, e.g., video surveillance and medical image analysis, require training recognition systems from large amounts of weakly annotated data, while some targeted interactions with a domain expert are allowed to improve the training process. In such cases, active learning (AL) can reduce labeling costs for training a classifier by querying the expert to provide the labels of most informative instances. This paper focuses on AL methods for instance classification problems in multiple instance learning (MIL), where data are arranged into sets, called bags, which are weakly labeled. Most AL methods focus on single-instance learning problems. These methods are not suitable for MIL problems because they cannot account for the bag structure of data. In this paper, new methods for bag-level aggregation of instance informativeness are proposed for multiple instance AL (MIAL). The aggregated informativeness method identifies the most informative instances based on classifier uncertainty and queries bags incorporating the most information. The other proposed method, called cluster-based aggregative sampling, clusters data hierarchically in the instance space. The informativeness of instances is assessed by considering bag labels, inferred instance labels, and the proportion of labels that remain to be discovered in clusters. Both proposed methods significantly outperform reference methods in extensive experiments using benchmark data from several application domains. Results indicate that using an appropriate strategy to address MIAL problems yields a significant reduction in the number of queries needed to achieve the same level of performance as single-instance AL methods.

A Cinematic View of Tissue Microbiology in the Live Infected Host

Tissue microbiology allows for the study of bacterial infection in the most clinically relevant microenvironment, the living host. Advancements in techniques and technology have facilitated the development of novel ways of studying infection. Many of these advancements have come from outside the field of microbiology. In this article, we outline the progression from bacteriology through cellular microbiology to tissue microbiology, highlighting seminal studies along the way. We outline the enormous potential but also some of the challenges of the tissue microbiology approach. We focus on the role of emerging technologies in the continual development of infectious disease research and highlight future possibilities in our ongoing quest to understand host-pathogen interaction.

Not-so-supervised: A survey of semi-supervised, multi-instance, and transfer learning in medical image analysis

Machine learning (ML) algorithms have made a tremendous impact in the field of medical imaging. While medical imaging datasets have been growing in size, a challenge for supervised ML algorithms that is frequently mentioned is the lack of annotated data. As a result, various methods that can learn with less/other types of supervision, have been proposed. We give an overview of semi-supervised, multiple instance, and transfer learning in medical imaging, both in diagnosis or segmentation tasks. We also discuss connections between these learning scenarios, and opportunities for future research. A dataset with the details of the surveyed papers is available via https://figshare.com/articles/Database_of_surveyed_literature_in_Not-so-supervised_a_survey_of_semi-supervised_multi-instance_and_transfer_learning_in_medical_image_analysis_/7479416.