Ha Q. Nguyen

Hieu H. Pham, Ngoc H. Nguyen, Thanh T. Tran et al.

The development of diagnostic models for detecting and diagnosing pediatric diseases in CXR scans is undertaken due to the lack of high-quality physician-annotated datasets. To overcome this challenge, we introduce and release PediCXR, a new pediatric CXR dataset of 9,125 studies retrospectively collected from a major pediatric hospital in Vietnam between 2020 and 2021. Each scan was manually annotated by a pediatric radiologist with more than ten years of experience. The dataset was labeled for the presence of 36 critical findings and 15 diseases. In particular, each abnormal finding was identified via a rectangle bounding box on the image. To the best of our knowledge, this is the first and largest pediatric CXR dataset containing lesion-level annotations and image-level labels for the detection of multiple findings and diseases. For algorithm development, the dataset was divided into a training set of 7,728 and a test set of 1,397. To encourage new advances in pediatric CXR interpretation using data-driven approaches, we provide a detailed description of the PediCXR data sample and make the dataset publicly available on https://physionet.org/content/pedicxr/1.0.0/

Khiem H. Le, Tuan V. Tran, Hieu H. Pham et al.

Building an accurate computer-aided diagnosis system based on data-driven approaches requires a large amount of high-quality labeled data. In medical imaging analysis, multiple expert annotators often produce subjective estimates about "ground truth labels" during the annotation process, depending on their expertise and experience. As a result, the labeled data may contain a variety of human biases with a high rate of disagreement among annotators, which significantly affect the performance of supervised machine learning algorithms. To tackle this challenge, we propose a simple yet effective approach to combine annotations from multiple radiology experts for training a deep learning-based detector that aims to detect abnormalities on medical scans. The proposed method first estimates the ground truth annotations and confidence scores of training examples. The estimated annotations and their scores are then used to train a deep learning detector with a re-weighted loss function to localize abnormal findings. We conduct an extensive experimental evaluation of the proposed approach on both simulated and real-world medical imaging datasets. The experimental results show that our approach significantly outperforms baseline approaches that do not consider the disagreements among annotators, including methods in which all of the noisy annotations are treated equally as ground truth and the ensemble of different models trained on different label sets provided separately by annotators.

Hieu H. Pham, Ha Q. Nguyen, Hieu T. Nguyen et al.

Recent artificial intelligence (AI) algorithms have achieved radiologist-level performance on various medical classification tasks. However, only a few studies addressed the localization of abnormal findings from CXR scans, which is essential in explaining the image-level classification to radiologists. We introduce in this paper an explainable deep learning system called VinDr-CXR that can classify a CXR scan into multiple thoracic diseases and, at the same time, localize most types of critical findings on the image. VinDr-CXR was trained on 51,485 CXR scans with radiologist-provided bounding box annotations. It demonstrated a comparable performance to experienced radiologists in classifying 6 common thoracic diseases on a retrospective validation set of 3,000 CXR scans, with a mean area under the receiver operating characteristic curve (AUROC) of 0.967 (95% confidence interval [CI]: 0.958-0.975). The VinDr-CXR was also externally validated in independent patient cohorts and showed its robustness. For the localization task with 14 types of lesions, our free-response receiver operating characteristic (FROC) analysis showed that the VinDr-CXR achieved a sensitivity of 80.2% at the rate of 1.0 false-positive lesion identified per scan. A prospective study was also conducted to measure the clinical impact of the VinDr-CXR in assisting six experienced radiologists. The results indicated that the proposed system, when used as a diagnosis supporting tool, significantly improved the agreement between radiologists themselves with an increase of 1.5% in mean Fleiss' Kappa. We also observed that, after the radiologists consulted VinDr-CXR's suggestions, the agreement between each of them and the system was remarkably increased by 3.3% in mean Cohen's Kappa.

Thao T. B. Nguyen, Tam M. Vo, Thang V. Nguyen et al.

We propose a data collecting and annotation pipeline that extracts information from Vietnamese radiology reports to provide accurate labels for chest X-ray (CXR) images. This can benefit Vietnamese radiologists and clinicians by annotating data that closely match their endemic diagnosis categories which may vary from country to country. To assess the efficacy of the proposed labeling technique, we built a CXR dataset containing 9,752 studies and evaluated our pipeline using a subset of this dataset. With an F1-score of at least 0.9923, the evaluation demonstrates that our labeling tool performs precisely and consistently across all classes. After building the dataset, we train deep learning models that leverage knowledge transferred from large public CXR datasets. We employ a variety of loss functions to overcome the curse of imbalanced multi-label datasets and conduct experiments with various model architectures to select the one that delivers the best performance. Our best model (CheXpert-pretrained EfficientNet-B2) yields an F1-score of 0.6989 (95% CI 0.6740, 0.7240), AUC of 0.7912, sensitivity of 0.7064 and specificity of 0.8760 for the abnormal diagnosis in general. Finally, we demonstrate that our coarse classification (based on five specific locations of abnormalities) yields comparable results to fine classification (twelve pathologies) on the benchmark CheXpert dataset for general anomaly detection while delivering better performance in terms of the average performance of all classes.

Dat T. Ngo, Thao T. B. Nguyen, Hieu T. Nguyen et al.

The rapid development in representation learning techniques such as deep neural networks and the availability of large-scale, well-annotated medical imaging datasets have to a rapid increase in the use of supervised machine learning in the 3D medical image analysis and diagnosis. In particular, deep convolutional neural networks (D-CNNs) have been key players and were adopted by the medical imaging community to assist clinicians and medical experts in disease diagnosis and treatment. However, training and inferencing deep neural networks such as D-CNN on high-resolution 3D volumes of Computed Tomography (CT) scans for diagnostic tasks pose formidable computational challenges. This challenge raises the need of developing deep learning-based approaches that are robust in learning representations in 2D images, instead 3D scans. In this work, we propose for the first time a new strategy to train \emph{slice-level} classifiers on CT scans based on the descriptors of the adjacent slices along the axis. In particular, each of which is extracted through a convolutional neural network (CNN). This method is applicable to CT datasets with per-slice labels such as the RSNA Intracranial Hemorrhage (ICH) dataset, which aims to predict the presence of ICH and classify it into 5 different sub-types. We obtain a single model in the top 4% best-performing solutions of the RSNA ICH challenge, where model ensembles are allowed. Experiments also show that the proposed method significantly outperforms the baseline model on CQ500. The proposed method is general and can be applied to other 3D medical diagnosis tasks such as MRI imaging. To encourage new advances in the field, we will make our codes and pre-trained model available upon acceptance of the paper.

Hieu H. Pham, Khiem H. Le, Tuan V. Tran et al.

The work discusses the use of machine learning algorithms for anomaly detection in medical image analysis and how the performance of these algorithms depends on the number of annotators and the quality of labels. To address the issue of subjectivity in labeling with a single annotator, we introduce a simple and effective approach that aggregates annotations from multiple annotators with varying levels of expertise. We then aim to improve the efficiency of predictive models in abnormal detection tasks by estimating hidden labels from multiple annotations and using a re-weighted loss function to improve detection performance. Our method is evaluated on a real-world medical imaging dataset and outperforms relevant baselines that do not consider disagreements among annotators.

Nhan T. Nguyen, Dat Q. Tran, Nghia T. Nguyen et al.

We propose a novel method that combines a convolutional neural network (CNN) with a long short-term memory (LSTM) mechanism for accurate prediction of intracranial hemorrhage on computed tomography (CT) scans. The CNN plays the role of a slice-wise feature extractor while the LSTM is responsible for linking the features across slices. The whole architecture is trained end-to-end with input being an RGB-like image formed by stacking 3 different viewing windows of a single slice. We validate the method on the recent RSNA Intracranial Hemorrhage Detection challenge and on the CQ500 dataset. For the RSNA challenge, our best single model achieves a weighted log loss of 0.0522 on the leaderboard, which is comparable to the top 3% performances, almost all of which make use of ensemble learning. Importantly, our method generalizes very well: the model trained on the RSNA dataset significantly outperforms the 2D model, which does not take into account the relationship between slices, on CQ500. Our codes and models is publicly avaiable at https://github.com/VinBDI-MedicalImagingTeam/midl2020-cnnlstm-ich.

Hieu H. Pham, Dung V. Do, Ha Q. Nguyen

X-ray imaging in DICOM format is the most commonly used imaging modality in clinical practice, resulting in vast, non-normalized databases. This leads to an obstacle in deploying AI solutions for analyzing medical images, which often requires identifying the right body part before feeding the image into a specified AI model. This challenge raises the need for an automated and efficient approach to classifying body parts from X-ray scans. Unfortunately, to the best of our knowledge, there is no open tool or framework for this task to date. To fill this lack, we introduce a DICOM Imaging Router that deploys deep CNNs for categorizing unknown DICOM X-ray images into five anatomical groups: abdominal, adult chest, pediatric chest, spine, and others. To this end, a large-scale X-ray dataset consisting of 16,093 images has been collected and manually classified. We then trained a set of state-of-the-art deep CNNs using a training set of 11,263 images. These networks were then evaluated on an independent test set of 2,419 images and showed superior performance in classifying the body parts. Specifically, our best performing model achieved a recall of 0.982 (95% CI, 0.977-0.988), a precision of 0.985 (95% CI, 0.975-0.989) and a F1-score of 0.981 (95% CI, 0.976-0.987), whilst requiring less computation for inference (0.0295 second per image). Our external validity on 1,000 X-ray images shows the robustness of the proposed approach across hospitals. These remarkable performances indicate that deep CNNs can accurately and effectively differentiate human body parts from X-ray scans, thereby providing potential benefits for a wide range of applications in clinical settings. The dataset, codes, and trained deep learning models from this study will be made publicly available on our project website at https://vindr.ai/.

Thanh T. Tran, Hieu H. Pham, Thang V. Nguyen et al.

Chest radiograph (CXR) interpretation in pediatric patients is error-prone and requires a high level of understanding of radiologic expertise. Recently, deep convolutional neural networks (D-CNNs) have shown remarkable performance in interpreting CXR in adults. However, there is a lack of evidence indicating that D-CNNs can recognize accurately multiple lung pathologies from pediatric CXR scans. In particular, the development of diagnostic models for the detection of pediatric chest diseases faces significant challenges such as (i) lack of physician-annotated datasets and (ii) class imbalance problems. In this paper, we retrospectively collect a large dataset of 5,017 pediatric CXR scans, for which each is manually labeled by an experienced radiologist for the presence of 10 common pathologies. A D-CNN model is then trained on 3,550 annotated scans to classify multiple pediatric lung pathologies automatically. To address the high-class imbalance issue, we propose to modify and apply "Distribution-Balanced loss" for training D-CNNs which reshapes the standard Binary-Cross Entropy loss (BCE) to efficiently learn harder samples by down-weighting the loss assigned to the majority classes. On an independent test set of 777 studies, the proposed approach yields an area under the receiver operating characteristic (AUC) of 0.709 (95% CI, 0.690-0.729). The sensitivity, specificity, and F1-score at the cutoff value are 0.722 (0.694-0.750), 0.579 (0.563-0.595), and 0.389 (0.373-0.405), respectively. These results significantly outperform previous state-of-the-art methods on most of the target diseases. Moreover, our ablation studies validate the effectiveness of the proposed loss function compared to other standard losses, e.g., BCE and Focal Loss, for this learning task. Overall, we demonstrate the potential of D-CNNs in interpreting pediatric CXRs.

Hoang C. Nguyen, Tung T. Le, Hieu H. Pham et al.

We introduce a new benchmark dataset, namely VinDr-RibCXR, for automatic segmentation and labeling of individual ribs from chest X-ray (CXR) scans. The VinDr-RibCXR contains 245 CXRs with corresponding ground truth annotations provided by human experts. A set of state-of-the-art segmentation models are trained on 196 images from the VinDr-RibCXR to segment and label 20 individual ribs. Our best performing model obtains a Dice score of 0.834 (95% CI, 0.810--0.853) on an independent test set of 49 images. Our study, therefore, serves as a proof of concept and baseline performance for future research.

Ngoc Huy Nguyen, Ha Quy Nguyen, Nghia Trung Nguyen et al.

Computer-Aided Diagnosis (CAD) systems for chest radiographs using artificial intelligence (AI) have recently shown a great potential as a second opinion for radiologists. The performances of such systems, however, were mostly evaluated on a fixed dataset in a retrospective manner and, thus, far from the real performances in clinical practice. In this work, we demonstrate a mechanism for validating an AI-based system for detecting abnormalities on X-ray scans, VinDr-CXR, at the Phu Tho General Hospital - a provincial hospital in the North of Vietnam. The AI system was directly integrated into the Picture Archiving and Communication System (PACS) of the hospital after being trained on a fixed annotated dataset from other sources. The performance of the system was prospectively measured by matching and comparing the AI results with the radiology reports of 6,285 chest X-ray examinations extracted from the Hospital Information System (HIS) over the last two months of 2020. The normal/abnormal status of a radiology report was determined by a set of rules and served as the ground truth. Our system achieves an F1 score - the harmonic average of the recall and the precision - of 0.653 (95% CI 0.635, 0.671) for detecting any abnormalities on chest X-rays. Despite a significant drop from the in-lab performance, this result establishes a high level of confidence in applying such a system in real-life situations.

Hieu H. Pham, Tung T. Le, Dat T. Ngo et al.

The chest X-rays (CXRs) is one of the views most commonly ordered by radiologists (NHS),which is critical for diagnosis of many different thoracic diseases. Accurately detecting thepresence of multiple diseases from CXRs is still a challenging task. We present a multi-labelclassification framework based on deep convolutional neural networks (CNNs) for diagnos-ing the presence of 14 common thoracic diseases and observations. Specifically, we trained astrong set of CNNs that exploit dependencies among abnormality labels and used the labelsmoothing regularization (LSR) for a better handling of uncertain samples. Our deep net-works were trained on over 200,000 CXRs of the recently released CheXpert dataset (Irvinandal., 2019) and the final model, which was an ensemble of the best performing networks,achieved a mean area under the curve (AUC) of 0.940 in predicting 5 selected pathologiesfrom the validation set. To the best of our knowledge, this is the highest AUC score yetreported to date. More importantly, the proposed method was also evaluated on an inde-pendent test set of the CheXpert competition, containing 500 CXR studies annotated by apanel of 5 experienced radiologists. The reported performance was on average better than2.6 out of 3 other individual radiologists with a mean AUC of 0.930, which had led to thecurrent state-of-the-art performance on the CheXpert test set.

Hieu H. Pham, Tung T. Le, Dat Q. Tran et al.

Chest radiography is one of the most common types of diagnostic radiology exams, which is critical for screening and diagnosis of many different thoracic diseases. Specialized algorithms have been developed to detect several specific pathologies such as lung nodule or lung cancer. However, accurately detecting the presence of multiple diseases from chest X-rays (CXRs) is still a challenging task. This paper presents a supervised multi-label classification framework based on deep convolutional neural networks (CNNs) for predicting the risk of 14 common thoracic diseases. We tackle this problem by training state-of-the-art CNNs that exploit dependencies among abnormality labels. We also propose to use the label smoothing technique for a better handling of uncertain samples, which occupy a significant portion of almost every CXR dataset. Our model is trained on over 200,000 CXRs of the recently released CheXpert dataset and achieves a mean area under the curve (AUC) of 0.940 in predicting 5 selected pathologies from the validation set. This is the highest AUC score yet reported to date. The proposed method is also evaluated on the independent test set of the CheXpert competition, which is composed of 500 CXR studies annotated by a panel of 5 experienced radiologists. The performance is on average better than 2.6 out of 3 other individual radiologists with a mean AUC of 0.930, which ranks first on the CheXpert leaderboard at the time of writing this paper.