Search

Background: Large vision models (LVM) pretrained by large datasets have demonstrated their enormous capacity to understand visual patterns and capture semantic information from images. We proposed a novel method of knowledge domain adaptation with pretrained LVM for a low-cost artificial intelligence (AI) model to quantify the severity of SARS-CoV-2 pneumonia based on frontal chest X-ray (CXR) images. Methods: Our method used the pretrained LVMs as the primary feature extractor and self-supervised contrastive learning for domain adaptation. An encoder with a 2048-dimensional feature vector output was first trained by self-supervised learning for knowledge domain adaptation. Then a multi-layer perceptron (MLP) was trained for the final severity prediction. A dataset with 2599 CXR images was used for model training and evaluation. Results: The model based on the pretrained vision transformer (ViT) and self-supervised learning achieved the best performance in cross validation, with mean squared error (MSE) of 23.83 (95 % CI 22.67–25.00) and mean absolute error (MAE) of 3.64 (95 % CI 3.54–3.73). Its prediction correlation has the R2 of 0.81 (95 % CI 0.79–0.82) and Spearman ρ of 0.80 (95 % CI 0.77–0.81), which are comparable to the current state-of-the-art (SOTA) methods trained by much larger CXR datasets. Conclusion: The proposed new method has achieved the SOTA performance to quantify the severity of SARS-CoV-2 pneumonia at a significantly lower cost. The method can be extended to other infectious disease detection or quantification to expedite the application of AI in medical research.

IntroductionDeep learning (DL) has significantly advanced medical image classification. However, it often relies on transfer learning (TL) from models pretrained on large, generic non-medical image datasets like ImageNet. Conversely, medical images possess unique visual characteristics that such general models may not adequately capture.MethodsThis study examines the effectiveness of modality-specific pretext learning strengthened by image denoising and deblurring in enhancing the classification of pediatric chest X-ray (CXR) images into those exhibiting no findings, i.e., normal lungs, or with cardiopulmonary disease manifestations. Specifically, we use a VGG-16-Sharp-U-Net architecture and leverage its encoder in conjunction with a classification head to distinguish normal from abnormal pediatric CXR findings. We benchmark this performance against the traditional TL approach, viz., the VGG-16 model pretrained only on ImageNet. Measures used for performance evaluation are balanced accuracy, sensitivity, specificity, F-score, Matthew’s Correlation Coefficient (MCC), Kappa statistic, and Youden’s index.ResultsOur findings reveal that models developed from CXR modality-specific pretext encoders substantially outperform the ImageNet-only pretrained model, viz., Baseline, and achieve significantly higher sensitivity (p

Abstract Cervical cancer is a leading cause of cancer mortality, with approximately 90% of the 250,000 deaths per year occurring in low- and middle-income countries (LMIC). Secondary prevention with cervical screening involves detecting and treating precursor lesions; however, scaling screening efforts in LMIC has been hampered by infrastructure and cost constraints. Recent work has supported the development of an artificial intelligence (AI) pipeline on digital images of the cervix to achieve an accurate and reliable diagnosis of treatable precancerous lesions. In particular, WHO guidelines emphasize visual triage of women testing positive for human papillomavirus (HPV) as the primary screen, and AI could assist in this triage task. In this work, we implemented a comprehensive deep-learning model selection and optimization study on a large, collated, multi-geography, multi-institution, and multi-device dataset of 9462 women (17,013 images). We evaluated relative portability, repeatability, and classification performance. The top performing model, when combined with HPV type, achieved an area under the Receiver Operating Characteristics (ROC) curve (AUC) of 0.89 within our study population of interest, and a limited total extreme misclassification rate of 3.4%, on held-aside test sets. Our model also produced reliable and consistent predictions, achieving a strong quadratic weighted kappa (QWK) of 0.86 and a minimal %2-class disagreement (% 2-Cl. D.) of 0.69%, between image pairs across women. Our work is among the first efforts at designing a robust, repeatable, accurate and clinically translatable deep-learning model for cervical screening.

Abstract Background WHO has recommended HPV testing for cervical screening where it is practical and affordable. If used, it is important to both clarify and implement the clinical management of positive results. We estimated the performance in Lusaka, Zambia of a novel screening/triage approach combining HPV typing with visual assessment assisted by a deep-learning approach called automated visual evaluation (AVE). Methods In this well-established cervical cancer screening program nested inside public sector primary care health facilities, experienced nurses examined women with high-quality digital cameras; the magnified illuminated images permit inspection of the surface morphology of the cervix and expert telemedicine quality assurance. Emphasizing sensitive criteria to avoid missing precancer/cancer, ~ 25% of women screen positive, reflecting partly the high HIV prevalence. Visual screen-positive women are treated in the same visit by trained nurses using either ablation (~ 60%) or LLETZ excision, or referred for LLETZ or more extensive surgery as needed. We added research elements (which did not influence clinical care) including collection of HPV specimens for testing and typing with BD Onclarity™ with a five channel output (HPV16, HPV18/45, HPV31/33/52/58, HPV35/39/51/56/59/66/68, human DNA control), and collection of triplicate cervical images with a Samsung Galaxy J8 smartphone camera™ that were analyzed using AVE, an AI-based algorithm pre-trained on a large NCI cervical image archive. The four HPV groups and three AVE classes were crossed to create a 12-level risk scale, ranking participants in order of predicted risk of precancer. We evaluated the risk scale and assessed how well it predicted the observed diagnosis of precancer/cancer. Results HPV type, AVE classification, and the 12-level risk scale all were strongly associated with degree of histologic outcome. The AVE classification showed good reproducibility between replicates, and added finer predictive accuracy to each HPV type group. Women living with HIV had higher prevalence of precancer/cancer; the HPV-AVE risk categories strongly predicted diagnostic findings in these women as well. Conclusions These results support the theoretical efficacy of HPV-AVE-based risk estimation for cervical screening. If HPV testing can be made affordable, cost-effective and point of care, this risk-based approach could be one management option for HPV-positive women.

In an era of rapid advancements in artificial intelligence (AI) technologies, particularly in medical imaging and natural language processing, strategic efforts to leverage AI’s capabilities in analyzing complex medical data and integrating it into clinical workflows have emerged as a key driver of innovation in healthcare [...]

Model initialization techniques are vital for improving the performance and reliability of deep learning models in medical computer vision applications. While much literature exists on non-medical images, the impacts on medical images, particularly chest X-rays (CXRs) are less understood. Addressing this gap, our study explores three deep model initialization techniques: Cold-start, Warm-start, and Shrink and Perturb start, focusing on adult and pediatric populations. We specifically focus on scenarios with periodically arriving data for training, thereby embracing the real-world scenarios of ongoing data influx and the need for model updates. We evaluate these models for generalizability against external adult and pediatric CXR datasets. We also propose novel ensemble methods: F-score-weighted Sequential Least-Squares Quadratic Programming (F-SLSQP) and Attention-Guided Ensembles with Learnable Fuzzy Softmax to aggregate weight parameters from multiple models to capitalize on their collective knowledge and complementary representations. We perform statistical significance tests with 95% confidence intervals and p-values to analyze model performance. Our evaluations indicate models initialized with ImageNet-pretrained weights demonstrate superior generalizability over randomly initialized counterparts, contradicting some findings for non-medical images. Notably, ImageNet-pretrained models exhibit consistent performance during internal and external testing across different training scenarios. Weight-level ensembles of these models show significantly higher recall (p

Data integration is a well-motivated problem in the clinical data science domain. Availability of patient data, reference clinical cases, and datasets for research have the potential to advance the healthcare industry. However, the unstructured (text, audio, or video data) and heterogeneous nature of the data, the variety of data standards and formats, and patient privacy constraint make data interoperability and integration a challenge. The clinical text is further categorized into different semantic groups and may be stored in different files and formats. Even the same organization may store cases in different data structures, making data integration more challenging. With such inherent complexity, domain experts and domain knowledge are often necessary to perform data integration. However, expert human labor is time and cost prohibitive. To overcome the variability in the structure, format, and content of the different data sources, we map the text into common categories and compute similarity within those. In this paper, we present a method to categorize and merge clinical data by considering the underlying semantics behind the cases and use reference information about the cases to perform data integration. Evaluation shows that we were able to merge 88% of clinical data from five different sources.

Cervical cancer is caused by the persistent infection of certain types of the Human Papillomavirus (HPV) and is a leading cause of female mortality particularly in low and middle-income countries (LMIC). Visual inspection of the cervix with acetic acid (VIA) is a commonly used technique in cervical screening. While this technique is inexpensive, clinical assessment is highly subjective, and relatively poor reproducibility has been reported. A deep learning-based algorithm for automatic visual evaluation (AVE) of aceto-whitened cervical images was shown to be effective in detecting confirmed precancer (i.e. direct precursor to invasive cervical cancer). The images were selected from a large longitudinal study conducted by the National Cancer Institute in the Guanacaste province of Costa Rica. The training of AVE used annotation for cervix boundary, and the data scarcity challenge was dealt with manually optimized data augmentation. In contrast, we present a novel approach for cervical precancer detection using a deep metric learning-based (DML) framework which does not incorporate any effort for cervix boundary marking. The DML is an advanced learning strategy that can deal with data scarcity and bias training due to class imbalance data in a better way. Three different widely-used state-of-the-art DML techniques are evaluated- (a) Contrastive loss minimization, (b) N-pair embedding loss minimization, and, (c) Batch-hard loss minimization. Three popular Deep Convolutional Neural Networks (ResNet-50, MobileNet, NasNet) are configured for training with DML to produce class-separated (i.e. linearly separable) image feature descriptors. Finally, a K-Nearest Neighbor (KNN) classifier is trained with the extracted deep features. Both the feature quality and classification performance are quantitatively evaluated on the same data set as used in AVE. It shows that, unlike AVE, without using any data augmentation, the best model produced from our research improves specificity in disease detection without compromising sensitivity. The present research thus paves the way for new research directions for the related field.

The accurate segmentation of cardiac images into anatomically meaningful regions is critical for the extraction of quantitative cardiac indices. The common pipeline for segmentation comprises regions of interest (ROIs) localization and segmentation stages that are independent of each other and typically performed using separate models. In this paper, we propose an end-to-end network, called Trilateral Attention Network (TaNet), for real-time region localization and segmentation. TaNet has a module for ROIs localization and three segmentation pathways: spatial pathway, handcrafted pathway, and context pathway. The localization module focuses segmentation attention on the desired region while learning the context relationship between different regions in the image. The localized regions are then sent to the three pathways for segmentation. The spatial pathway, which has regular convolutional kernels, is used to extract deep features at different levels of abstraction. The handcrafted pathway, which has hand-designed convolutional kernels, is used to extract a unique set of features complementary to the deep features. Finally, the context (or global) pathway is used to enlarge the receptive field. By jointly training TaNet for localization and segmentation, TaNet achieved superior performance, in terms of accuracy and speed, when evaluated on two echocardiography datasets for cardiac region segmentation.

Abstract Background Light microscopy is often used for malaria diagnosis in the field. However, it is time-consuming and quality of the results depends heavily on the skill of microscopists. Automating malaria light microscopy is a promising solution, but it still remains a challenge and an active area of research. Current tools are often expensive and involve sophisticated hardware components, which makes it hard to deploy them in resource-limited areas. Results We designed an Android mobile application called Malaria Screener, which makes smartphones an affordable yet effective solution for automated malaria light microscopy. The mobile app utilizes high-resolution cameras and computing power of modern smartphones to screen both thin and thick blood smear images for P. falciparum parasites. Malaria Screener combines image acquisition, smear image analysis, and result visualization in its slide screening process, and is equipped with a database to provide easy access to the acquired data. Conclusion Malaria Screener makes the screening process faster, more consistent, and less dependent on human expertise. The app is modular, allowing other research groups to integrate their methods and models for image processing and machine learning, while acquiring and analyzing their data.

Abstract Background Accelerated global control of cervical cancer would require primary prevention with human papillomavirus (HPV) vaccination in addition to novel screening program strategies that are simple, inexpensive, and effective. We present the feasibility and outcome of a community-based HPV self-sampled screening program. Methods In Ile Ife, Nigeria, 9406 women aged 30–49 years collected vaginal self-samples, which were tested for HPV in the local study laboratory using Hybrid Capture-2 (HC2) (Qiagen). HPV-positive women were referred to the colposcopy clinic. Gynecologist colposcopic impression dictated immediate management; biopsies were taken when definite acetowhitening was present to produce a histopathologic reference standard of precancer (and to determine final clinical management). Retrospective linkage to the medical records identified 442 of 9406 women living with HIV (WLWH). Results With self-sampling, it was possible to screen more than 100 women per day per clinic. Following an audio-visual presentation and in-person instructions, overall acceptability of self-sampling was very high (81.2% women preferring self-sampling over clinician collection). HPV positivity was found in 17.3% of women. Intensive follow-up contributed to 85.9% attendance at the colposcopy clinic. Of those referred, 8.2% were initially treated with thermal ablation and 5.6% with large loop excision of transformation zone (LLETZ). Full visibility of the squamocolumnar junction, necessary for optimal visual triage and ablation, declined from 68.5% at age 30 to 35.4% at age 49. CIN2+ and CIN3+ (CIN- Cervical intraepithelial neoplasia), including five cancers, were identified by histology in 5.9 and 3.2% of the HPV-positive women, respectively (0.9 and 0.5% of the total screening population), leading to additional treatment as indicated. The prevalences of HPV infection and CIN2+ were substantially higher (40.5 and 2.5%, respectively) among WLWH. Colposcopic impression led to over- and under-treatment compared to the histopathology reference standard. Conclusion A cervical cancer screening program using self-sampled HPV testing, with colposcopic immediate management of women positive for HPV, proved feasible in Nigeria. Based on the collected specimens and images, we are now evaluating the use of a combination of partial HPV typing and automated visual evaluation (AVE) of cervical images to improve the accuracy of the screening program.

Research on detecting Tuberculosis (TB) findings on chest radiographs (or Chest X-rays: CXR) using convolutional neural networks (CNNs) has demonstrated superior performance due to the emergence of publicly available, large-scale datasets with expert annotations and availability of scalable computational resources. However, these studies use only the frontal CXR projections, i.e., the posterior-anterior (PA), and the anterior-posterior (AP) views for analysis and decision-making. Lateral CXRs which are heretofore not studied help detect clinically suspected pulmonary TB, particularly in children. Further, Vision Transformers (ViTs) with built-in self-attention mechanisms have recently emerged as a viable alternative to the traditional CNNs. Although ViTs demonstrated notable performance in several medical image analysis tasks, potential limitations exist in terms of performance and computational efficiency, between the CNN and ViT models, necessitating a comprehensive analysis to select appropriate models for the problem under study. This study aims to detect TB-consistent findings in lateral CXRs by constructing an ensemble of the CNN and ViT models. Several models are trained on lateral CXR data extracted from two large public collections to transfer modality-specific knowledge and fine-tune them for detecting findings consistent with TB. We observed that the weighted averaging ensemble of the predictions of CNN and ViT models using the optimal weights computed with the Sequential Least-Squares Quadratic Programming method delivered significantly superior performance (MCC: 0.8136, 95% confidence intervals (CI): 0.7394, 0.8878, p < 0.05) compared to the individual models and other ensembles. We also interpreted the decisions of CNN and ViT models using class-selective relevance maps and attention maps, respectively, and combined them to highlight the discriminative image regions contributing to the final output. We observed that (i) the model accuracy is not related to disease region of interest (ROI) localization and (ii) the bitwise-AND of the heatmaps of the top-2-performing models delivered significantly superior ROI localization performance in terms of mean average precision [mAP@(0.1 0.6) = 0.1820, 95% CI: 0.0771,0.2869, p < 0.05], compared to other individual models and ensembles. The code is available at https://github.com/sivaramakrishnan-rajaraman/Ensemble-of-CNN-and-ViT-for-TB-detection-in-lateral-CXR.

In medical image classification tasks, it is common to find that the number of normal samples far exceeds the number of abnormal samples. In such class-imbalanced situations, reliable training of deep neural networks continues to be a major challenge, therefore biasing the predicted class probabilities toward the majority class. Calibration has been proposed to alleviate some of these effects. However, there is insufficient analysis explaining whether and when calibrating a model would be beneficial. In this study, we perform a systematic analysis of the effect of model calibration on its performance on two medical image modalities, namely, chest X-rays and fundus images, using various deep learning classifier backbones. For this, we study the following variations: (i) the degree of imbalances in the dataset used for training; (ii) calibration methods; and (iii) two classification thresholds, namely, default threshold of 0.5, and optimal threshold from precision-recall (PR) curves. Our results indicate that at the default classification threshold of 0.5, the performance achieved through calibration is significantly superior (p < 0.05) to using uncalibrated probabilities. However, at the PR-guided threshold, these gains are not significantly different (p > 0.05). This observation holds for both image modalities and at varying degrees of imbalance. The code is available at https://github.com/sivaramakrishnan-rajaraman/Model_calibration.

Automatic detection of some pulmonary abnormalities using chest X-rays may be impacted adversely due to obscuring by bony structures like the ribs and the clavicles. Automated bone suppression methods would increase soft tissue visibility and enhance automated disease detection. We evaluate this hypothesis using a custom ensemble of convolutional neural network models, which we call DeBoNet, that suppresses bones in frontal CXRs. First, we train and evaluate variants of U-Nets, Feature Pyramid Networks, and other proposed custom models using a private collection of CXR images and their bone-suppressed counterparts. The DeBoNet, constructed using the top-3 performing models, outperformed the individual models in terms of peak signal-to-noise ratio (PSNR) (36.7977±1.6207), multi-scale structural similarity index measure (MS-SSIM) (0.9848±0.0073), and other metrics. Next, the best-performing bone-suppression model is applied to CXR images that are pooled from several sources, showing no abnormality and other findings consistent with COVID-19. The impact of bone suppression is demonstrated by evaluating the gain in performance in detecting pulmonary abnormality consistent with COVID-19 disease. We observe that the model trained on bone-suppressed CXRs (MCC: 0.9645, 95% confidence interval (0.9510, 0.9780)) significantly outperformed (p < 0.05) the model trained on non-bone-suppressed images (MCC: 0.7961, 95% confidence interval (0.7667, 0.8255)) in detecting findings consistent with COVID-19 indicating benefits derived from automatic bone suppression on disease classification. The code is available at https://github.com/sivaramakrishnan-rajaraman/Bone-Suppresion-Ensemble.

Domain shift is one of the key challenges affecting reliability in medical imaging-based machine learning predictions. It is of significant importance to investigate this issue to gain insights into its characteristics toward determining controllable parameters to minimize its impact. In this paper, we report our efforts on studying and analyzing domain shift in lung region detection in chest radiographs. We used five chest X-ray datasets, collected from different sources, which have manual markings of lung boundaries in order to conduct extensive experiments toward this goal. We compared the characteristics of these datasets from three aspects: information obtained from metadata or an image header, image appearance, and features extracted from a pretrained model. We carried out experiments to evaluate and compare model performances within each dataset and across datasets in four scenarios using different combinations of datasets. We proposed a new feature visualization method to provide explanations for the applied object detection network on the obtained quantitative results. We also examined chest X-ray modality-specific initialization, catastrophic forgetting, and model repeatability. We believe the observations and discussions presented in this work could help to shed some light on the importance of the analysis of training data for medical imaging machine learning research, and could provide valuable guidance for domain shift analysis.

Deep learning (DL) models are state-of-the-art in segmenting anatomical and disease regions of interest (ROIs) in medical images. Particularly, a large number of DL-based techniques have been reported using chest X-rays (CXRs). However, these models are reportedly trained on reduced image resolutions for reasons related to the lack of computational resources. Literature is sparse in discussing the optimal image resolution to train these models for segmenting the tuberculosis (TB)-consistent lesions in CXRs. In this study, we investigated the performance variations with an Inception-V3 UNet model using various image resolutions with/without lung ROI cropping and aspect ratio adjustments and identified the optimal image resolution through extensive empirical evaluations to improve TB-consistent lesion segmentation performance. We used the Shenzhen CXR dataset for the study, which includes 326 normal patients and 336 TB patients. We proposed a combinatorial approach consisting of storing model snapshots, optimizing segmentation threshold and test-time augmentation (TTA), and averaging the snapshot predictions, to further improve performance with the optimal resolution. Our experimental results demonstrate that higher image resolutions are not always necessary; however, identifying the optimal image resolution is critical to achieving superior performance.

We propose a new generative model named adaptive cycle-consistent generative adversarial network, or Ad CycleGAN to perform image translation between normal and COVID-19 positive chest X-ray images. An independent pre-trained criterion is added to the conventional Cycle GAN architecture to exert adaptive control on image translation. The performance of Ad CycleGAN is compared with the Cycle GAN without the external criterion. The quality of the synthetic images is evaluated by quantitative metrics including Mean Squared Error (MSE), Root Mean Squared Error (RMSE), Peak Signal-to-Noise Ratio (PSNR), Universal Image Quality Index (UIQI), visual information fidelity (VIF), Frechet Inception Distance (FID), and translation accuracy. The experimental results indicate that the synthetic images generated either by the Cycle GAN or by the Ad CycleGAN have lower MSE and RMSE, and higher scores in PSNR, UIQI, and VIF in homogenous image translation (i.e., Y → Y) compared to the heterogenous image translation process (i.e., X → Y). The synthetic images by Ad CycleGAN through the heterogeneous image translation have significantly higher FID score compared to Cycle GAN (p < 0.01). The image translation accuracy of Ad CycleGAN is higher than that of Cycle GAN when normal images are converted to COVID-19 positive images (p < 0.01). Therefore, we conclude that the Ad CycleGAN with the independent criterion can improve the accuracy of GAN image translation. The new architecture has more control on image synthesis and can help address the common class imbalance issue in machine learning methods and artificial intelligence applications with medical images.

Cardiopulmonary diseases are a significant cause of mortality and morbidity worldwide [...]

Medical image analysis typically includes several tasks such as enhancement, segmentation, and classification. Traditionally, these tasks are implemented using separate deep learning models for separate tasks, which is not efficient because it involves unnecessary training repetitions, demands greater computational resources, and requires a relatively large amount of labeled data. In this paper, we propose a multi-task training approach for medical image analysis, where individual tasks are fine-tuned simultaneously through relevant knowledge transfer using a unified modality-specific feature representation (UMS-Rep). We explore different fine-tuning strategies to demonstrate the impact of the strategy on the performance of target medical image tasks. We experiment with different visual tasks (e.g., image denoising, segmentation, and classification) to highlight the advantages offered with our approach for two imaging modalities, chest X-ray and Doppler echocardiography. Our results demonstrate that the proposed approach reduces the overall demand for computational resources and improves target task generalization and performance. Specifically, the proposed approach improves accuracy (up to ~ 9% ↑) and decreases computational time (up to ~ 86% ↓) as compared to the baseline approach. Further, our results prove that the performance of target tasks in medical images is highly influenced by the utilized fine-tuning strategy.

Chest X-rays (CXRs) are the most commonly performed diagnostic examination to detect cardiopulmonary abnormalities. However, the presence of bony structures such as ribs and clavicles can obscure subtle abnormalities, resulting in diagnostic errors. This study aims to build a deep learning (DL)-based bone suppression model that identifies and removes these occluding bony structures in frontal CXRs to assist in reducing errors in radiological interpretation, including DL workflows, related to detecting manifestations consistent with tuberculosis (TB). Several bone suppression models with various deep architectures are trained and optimized using the proposed combined loss function and their performances are evaluated in a cross-institutional test setting using several metrics such as mean absolute error (MAE), peak signal-to-noise ratio (PSNR), structural similarity index measure (SSIM), and multiscale structural similarity measure (MS–SSIM). The best-performing model (ResNet–BS) (PSNR = 34.0678; MS–SSIM = 0.9828) is used to suppress bones in the publicly available Shenzhen and Montgomery TB CXR collections. A VGG-16 model is pretrained on a large collection of publicly available CXRs. The CXR-pretrained model is then fine-tuned individually on the non-bone-suppressed and bone-suppressed CXRs of Shenzhen and Montgomery TB CXR collections to classify them as showing normal lungs or TB manifestations. The performances of these models are compared using several performance metrics such as accuracy, the area under the curve (AUC), sensitivity, specificity, precision, F-score, and Matthews correlation coefficient (MCC), analyzed for statistical significance, and their predictions are qualitatively interpreted through class-selective relevance maps (CRMs). It is observed that the models trained on bone-suppressed CXRs (Shenzhen: AUC = 0.9535 ± 0.0186; Montgomery: AUC = 0.9635 ± 0.0106) significantly outperformed (p < 0.05) the models trained on the non-bone-suppressed CXRs (Shenzhen: AUC = 0.8991 ± 0.0268; Montgomery: AUC = 0.8567 ± 0.0870).. Models trained on bone-suppressed CXRs improved detection of TB-consistent findings and resulted in compact clustering of the data points in the feature space signifying that bone suppression improved the model sensitivity toward TB classification.

Background: Cervical intraepithelial neoplasia (CIN) is regarded as a potential precancerous state of the uterine cervix. Timely and appropriate early treatment of CIN can help reduce cervical cancer mortality. Accurate estimation of CIN grade correlated with human papillomavirus type, which is the primary cause of the disease, helps determine the patient's risk for developing the disease. Colposcopy is used to select women for biopsy. Expert pathologists examine the biopsied cervical epithelial tissue under a microscope. The examination can take a long time and is prone to error and often results in high inter-and intra-observer variability in outcomes. Methodology: We propose a novel image analysis toolbox that can automate CIN diagnosis using whole slide image (digitized biopsies) of cervical tissue samples. The toolbox is built as a four-step deep learning model that detects the epithelium regions, segments the detected epithelial portions, analyzes local vertical segment regions, and finally classifies each epithelium block with localized attention. We propose an epithelium detection network in this study and make use of our earlier research on epithelium segmentation and CIN classification to complete the design of the end-to-end CIN diagnosis toolbox. Results: The results show that automated epithelium detection and segmentation for CIN classification yields comparable results to manually segmented epithelium CIN classification. Conclusion: This highlights the potential as a tool for automated digitized histology slide image analysis to assist expert pathologists.

Automated Visual Examination (AVE) is a deep learning algorithm that aims to improve the effectiveness of cervical precancer screening, particularly in low- and medium-resource regions. It was trained on data from a large longitudinal study conducted by the National Cancer Institute (NCI) and has been shown to accurately identify cervices with early stages of cervical neoplasia for clinical evaluation and treatment. The algorithm processes images of the uterine cervix taken with a digital camera and alerts the user if the woman is a candidate for further evaluation. This requires that the algorithm be presented with images of the cervix, which is the object of interest, of acceptable quality, i.e., in sharp focus, with good illumination, without shadows or other occlusions, and showing the entire squamo-columnar transformation zone. Our prior work has addressed some of these constraints to help discard images that do not meet these criteria. In this work, we present a novel algorithm that determines that the image contains the cervix to a sufficient extent. Non-cervix or other inadequate images could lead to suboptimal or wrong results. Manual removal of such images is labor intensive and time-consuming, particularly in working with large retrospective collections acquired with inadequate quality control. In this work, we present a novel ensemble deep learning method to identify cervix images and non-cervix images in a smartphone-acquired cervical image dataset. The ensemble method combined the assessment of three deep learning architectures, RetinaNet, Deep SVDD, and a customized CNN (Convolutional Neural Network), each using a different strategy to arrive at its decision, i.e., object detection, one-class classification, and binary classification. We examined the performance of each individual architecture and an ensemble of all three architectures. An average accuracy and F-1 score of 91.6% and 0.890, respectively, were achieved on a separate test dataset consisting of more than 30,000 smartphone-captured images.

Medical images are acquired at different resolutions based on clinical goals or available technology. In general, however, high-resolution images with fine structural details are preferred for visual task analysis. Recognizing this significance, several deep learning networks have been proposed to enhance medical images for reliable automated interpretation. These deep networks are often computationally complex and require a massive number of parameters, which restrict them to highly capable computing platforms with large memory banks. In this paper, we propose an efficient deep learning approach, called Hydra, which simultaneously reduces computational complexity and improves performance. The Hydra consists of a trunk and several computing heads. The trunk is a super-resolution model that learns the mapping from low-resolution to high-resolution images. It has a simple architecture that is trained using multiple scales at once to minimize a proposed learning-loss function. We also propose to append multiple task-specific heads to the trained Hydra trunk for simultaneous learning of multiple visual tasks in medical images. The Hydra is evaluated on publicly available chest X-ray image collections to perform image enhancement, lung segmentation, and abnormality classification. Our experimental results support our claims and demonstrate that the proposed approach can improve the performance of super-resolution and visual task analysis in medical images at a remarkably reduced computational cost.

The novel severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has caused a pandemic resulting in over 2.7 million infected individuals and over 190,000 deaths and growing. Assertions in the literature suggest that respiratory disorders due to COVID-19 commonly present with pneumonia-like symptoms which are radiologically confirmed as opacities. Radiology serves as an adjunct to the reverse transcription-polymerase chain reaction test for confirmation and evaluating disease progression. While computed tomography (CT) imaging is more specific than chest X-rays (CXR), its use is limited due to cross-contamination concerns. CXR imaging is commonly used in high-demand situations, placing a significant burden on radiology services. The use of artificial intelligence (AI) has been suggested to alleviate this burden. However, there is a dearth of sufficient training data for developing image-based AI tools. We propose increasing training data for recognizing COVID-19 pneumonia opacities using weakly labeled data augmentation. This follows from a hypothesis that the COVID-19 manifestation would be similar to that caused by other viral pathogens affecting the lungs. We expand the training data distribution for supervised learning through the use of weakly labeled CXR images, automatically pooled from publicly available pneumonia datasets, to classify them into those with bacterial or viral pneumonia opacities. Next, we use these selected images in a stage-wise, strategic approach to train convolutional neural network-based algorithms and compare against those trained with non-augmented data. Weakly labeled data augmentation expands the learned feature space in an attempt to encompass variability in unseen test distributions, enhance inter-class discrimination, and reduce the generalization error. Empirical evaluations demonstrate that simple weakly labeled data augmentation (Acc: 0.5555 and Acc: 0.6536) is better than baseline non-augmented training (Acc: 0.2885 and Acc: 0.5028) in identifying COVID-19 manifestations as viral pneumonia. Interestingly, adding COVID-19 CXRs to simple weakly labeled augmented training data significantly improves the performance (Acc: 0.7095 and Acc: 0.8889), suggesting that COVID-19, though viral in origin, creates a uniquely different presentation in CXRs compared with other viral pneumonia manifestations.

Background: Cervical cancer is one of the deadliest cancers affecting women globally. Cervical intraepithelial neoplasia (CIN) assessment using histopathological examination of cervical biopsy slides is subject to interobserver variability. Automated processing of digitized histopathology slides has the potential for more accurate classification for CIN grades from normal to increasing grades of pre-malignancy: CIN1, CIN2, and CIN3. Methodology: Cervix disease is generally understood to progress from the bottom (basement membrane) to the top of the epithelium. To model this relationship of disease severity to spatial distribution of abnormalities, we propose a network pipeline, DeepCIN, to analyze high-resolution epithelium images (manually extracted from whole-slide images) hierarchically by focusing on localized vertical regions and fusing this local information for determining Normal/CIN classification. The pipeline contains two classifier networks: (1) a cross-sectional, vertical segment-level sequence generator is trained using weak supervision to generate feature sequences from the vertical segments to preserve the bottom-to-top feature relationships in the epithelium image data and (2) an attention-based fusion network image-level classifier predicting the final CIN grade by merging vertical segment sequences. Results: The model produces the CIN classification results and also determines the vertical segment contributions to CIN grade prediction. Conclusion: Experiments show that DeepCIN achieves pathologist-level CIN classification accuracy.

Background: Automated pathology techniques for detecting cervical cancer at the premalignant stage have advantages for women in areas with limited medical resources. Methods: This article presents EpithNet, a deep learning approach for the critical step of automated epithelium segmentation in digitized cervical histology images. EpithNet employs three regression networks of varying dimensions of image input blocks (patches) surrounding a given pixel, with all blocks at a fixed resolution, using varying network depth. Results: The proposed model was evaluated on 311 digitized histology epithelial images and the results indicate that the technique maximizes region-based information to improve pixel-wise probability estimates. EpithNet-mc model, formed by intermediate concatenation of the convolutional layers of the three models, was observed to achieve 94% Jaccard index (intersection over union) which is 26.4% higher than the benchmark model. Conclusions: EpithNet yields better epithelial segmentation results than state-of-the-art benchmark methods.

Evidence from recent research shows that automatic visual evaluation (AVE) of photographic images of the uterine cervix using deep learning-based algorithms presents a viable solution for improving cervical cancer screening by visual inspection with acetic acid (VIA). However, a significant performance determinant in AVE is the photographic image quality. While this includes image sharpness and focus, an important criterion is the localization of the cervix region. Deep learning networks have been successfully applied for object localization and segmentation in images, providing impetus for studying their use for fine contour segmentation of the cervix. In this paper, we present an evaluation of two state-of-the-art deep learning-based object localization and segmentation methods, viz., Mask R-convolutional neural network (CNN) and MaskX R-CNN, for automatic cervix segmentation using three datasets. We carried out extensive experimental tests and algorithm comparisons on each individual dataset and across datasets, and achieved performance either notably higher than, or comparable to, that reported in the literature. The highest Dice and intersection-over-union (IoU) scores that we obtained using Mask R-CNN were 0.947 and 0.901, respectively.

Deep learning (DL) methods are increasingly being applied for developing reliable computer-aided detection (CADe), diagnosis (CADx), and information retrieval algorithms. However, challenges in interpreting and explaining the learned behavior of the DL models hinders their adoption and use in real-world systems. In this study, we propose a novel method called “Class-selective Relevance Mapping„ (CRM) for localizing and visualizing discriminative regions of interest (ROI) within a medical image. Such visualizations offer improved explanation of the convolutional neural network (CNN)-based DL model predictions. We demonstrate CRM effectiveness in classifying medical imaging modalities toward automatically labeling them for visual information retrieval applications. The CRM is based on linear sum of incremental mean squared errors (MSE) calculated at the output layer of the CNN model. It measures both positive and negative contributions of each spatial element in the feature maps produced from the last convolution layer leading to correct classification of an input image. A series of experiments on a “multi-modality„ CNN model designed for classifying seven different types of image modalities shows that the proposed method is significantly better in detecting and localizing the discriminative ROIs than other state of the art class-activation methods. Further, to visualize its effectiveness we generate “class-specific„ ROI maps by averaging the CRM scores of images in each modality class, and characterize the visual explanation through their different size, shape, and location for our multi-modality CNN model that achieved over 98% performance on a dataset constructed from publicly available images.

Vast amounts of clinical and biomedical research data are produced daily. These data can help enable data driven healthcare through novel biomedical discoveries, improved diagnostics processes, epidemiology, and education. However, finding, and gaining access to these data and relevant metadata that are necessary to achieve these goals remains a challenge. Furthermore, data management and enabling widespread, albeit controlled, use poses a major challenge for data producers. These data sources are often geographically distributed, with diverse characteristics, and are controlled by a host of logistical and legal factors that require appropriate governance and access control guarantees. To overcome these obstacles, a set of guiding principles under the term FAIR has been previously introduced. The primary desirable dataset properties are thus that the data should be Findable, Accessible, Interoperable, and Reusable (FAIR). In this paper, we introduce and describe an abstract framework that models these ideal goals, and could be a step toward supporting data driven research. We also develop a system instantiated on our framework called the Data integration and indexing System (DiiS). The system provides an integration model for making healthcare data available on a global scale. Our research work describes the challenges inhibiting data producers, data stewards, and data brokers in achieving FAIR goals for sharing biomedical data. We attempt to address some of the key challenges through the proposed system. We evaluated our framework using the software architecture testing technique and also looked at how different challenges in data integration are addressed by our system. Our evaluation shows that the DiiS framework is a user friendly data integration system that would greatly contribute to biomedical research.

Pneumonia affects 7% of the global population, resulting in 2 million pediatric deaths every year. Chest X-ray (CXR) analysis is routinely performed to diagnose the disease. Computer-aided diagnostic (CADx) tools aim to supplement decision-making. These tools process the handcrafted and/or convolutional neural network (CNN) extracted image features for visual recognition. However, CNNs are perceived as black boxes since their performance lack explanations. This is a serious bottleneck in applications involving medical screening/diagnosis since poorly interpreted model behavior could adversely affect the clinical decision. In this study, we evaluate, visualize, and explain the performance of customized CNNs to detect pneumonia and further differentiate between bacterial and viral types in pediatric CXRs. We present a novel visualization strategy to localize the region of interest (ROI) that is considered relevant for model predictions across all the inputs that belong to an expected class. We statistically validate the models’ performance toward the underlying tasks. We observe that the customized VGG16 model achieves 96.2% and 93.6% accuracy in detecting the disease and distinguishing between bacterial and viral pneumonia respectively. The model outperforms the state-of-the-art in all performance metrics and demonstrates reduced bias and improved generalization.

Background: Advances in image analysis and computational techniques have facilitated automatic detection of critical features in histopathology images. Detection of nuclei is critical for squamous epithelium cervical intraepithelial neoplasia (CIN) classification into normal, CIN1, CIN2, and CIN3 grades. Methods: In this study, a deep learning (DL)-based nuclei segmentation approach is investigated based on gathering localized information through the generation of superpixels using a simple linear iterative clustering algorithm and training with a convolutional neural network. Results: The proposed approach was evaluated on a dataset of 133 digitized histology images and achieved an overall nuclei detection (object-based) accuracy of 95.97%, with demonstrated improvement over imaging-based and clustering-based benchmark techniques. Conclusions: The proposed DL-based nuclei segmentation Method with superpixel analysis has shown improved segmentation results in comparison to state-of-the-art methods.

Background: In previous research, we introduced an automated, localized, fusion-based approach for classifying uterine cervix squamous epithelium into Normal, CIN1, CIN2, and CIN3 grades of cervical intraepithelial neoplasia (CIN) based on digitized histology image analysis. As part of the CIN assessment process, acellular and atypical cell concentration features were computed from vertical segment partitions of the epithelium region to quantize the relative distribution of nuclei. Methods: Feature data was extracted from 610 individual segments from 61 images for epithelium classification into categories of Normal, CIN1, CIN2, and CIN3. The classification results were compared against CIN labels obtained from two pathologists who visually assessed abnormality in the digitized histology images. In this study, individual vertical segment CIN classification accuracy improvement is reported using the logistic regression classifier for an expanded data set of 118 histology images. Results: We analyzed the effects on classification using the same pathologist labels for training and testing versus using one pathologist labels for training and the other for testing. Based on a leave-one-out approach for classifier training and testing, exact grade CIN accuracies of 81.29% and 88.98% were achieved for individual vertical segment and epithelium whole-image classification, respectively. Conclusions: The Logistic and Random Tree classifiers outperformed the benchmark SVM and LDA classifiers from previous research. The Logistic Regression classifier yielded an improvement of 10.17% in CIN Exact grade classification results based on CIN labels for training-testing for the individual vertical segments and the whole image from the same single expert over the baseline approach using the reduced features. Overall, the CIN classification rates tended to be higher using the training-testing labels for the same expert than for training labels from one expert and testing labels from the other expert. The Exact class fusion- based CIN discrimination results obtained in this study are similar to the Exact class expert agreement rate.