Dataset Information

Development and Evaluation of a Leukemia Diagnosis System Using Deep Learning in Real Clinical Scenarios.

ABSTRACT: Leukemia is the most common malignancy affecting children. The morphologic analysis of bone marrow smears is an important initial step for diagnosis. Recent publications demonstrated that artificial intelligence is able to classify blood cells but a long way from clinical use. A total of 1,732 bone marrow images were used for the training of a convolutional neural network (CNN). New techniques of deep learning were integrated and an end-to-end leukemia diagnosis system was developed by using raw images without pre-processing. The system creatively imitated the workflow of a hematologist by detecting and excluding uncountable and crushed cells, then classifying and counting the remain cells to make a diagnosis. The performance of the CNN in classifying WBCs achieved an accuracy of 82.93%, precision of 86.07% and F1 score of 82.02%. And the performance in diagnosing acute lymphoid leukemia achieved an accuracy of 89%, sensitivity of 86% and specificity of 95%. The system also performs well at detecting the bone marrow metastasis of lymphoma and neuroblastoma, achieving an average accuracy of 82.93%. This is the first study which included a wider variety of cell types in leukemia diagnosis, and achieved a relatively high performance in real clinical scenarios.

SUBMITTER: Zhou M

PROVIDER: S-EPMC8264256 | biostudies-literature | 2021

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Development and Evaluation of a Leukemia Diagnosis System Using Deep Learning in Real Clinical Scenarios.

Zhou Min M Wu Kefei K Yu Lisha L Xu Mengdi M Yang Junjun J Shen Qing Q Liu Bo B Shi Lei L Wu Shuang S Dong Bin B Wang Hansong H Yuan Jiajun J Shen Shuhong S Zhao Liebin L

Frontiers in pediatrics 20210624

Leukemia is the most common malignancy affecting children. The morphologic analysis of bone marrow smears is an important initial step for diagnosis. Recent publications demonstrated that artificial intelligence is able to classify blood cells but a long way from clinical use. A total of 1,732 bone marrow images were used for the training of a convolutional neural network (CNN). New techniques of deep learning were integrated and an end-to-end leukemia diagnosis system was developed by using raw ...[more]

PMID: 34249819

Similar Datasets

Project description:BackgroundAccurately diagnosing supraspinatus tears based on magnetic resonance imaging (MRI) is challenging and time-combusting due to the experience level variability of the musculoskeletal radiologists and orthopedic surgeons. We developed a deep learning-based model for automatically diagnosing supraspinatus tears (STs) using shoulder MRI and validated its feasibility in clinical practice.Materials and methodsA total of 701 shoulder MRI data (2804 images) were retrospectively collected for model training and internal test. An additional 69 shoulder MRIs (276 images) were collected from patients who underwent shoulder arthroplasty and constituted the surgery test set for clinical validation. Two advanced convolutional neural networks (CNN) based on Xception were trained and optimized to detect STs. The diagnostic performance of the CNN was evaluated according to its sensitivity, specificity, precision, accuracy, and F1 score. Subgroup analyses were performed to verify its robustness, and we also compared the CNN's performance with that of 4 radiologists and 4 orthopedic surgeons on the surgery and internal test sets.ResultsOptimal diagnostic performance was achieved on the 2D model, from which F1-scores of 0.824 and 0.75, and areas under the ROC curves of 0.921 (95% confidence interval, 0.841-1.000) and 0.882 (0.817-0.947) were observed on the surgery and internal test sets. For the subgroup analysis, the 2D CNN model demonstrated a sensitivity of 0.33-1.000 and 0.625-1.000 for different degrees of tears on the surgery and internal test sets, and there was no significant performance difference between 1.5 and 3.0 T data. Compared with eight clinicians, the 2D CNN model exhibited better diagnostic performance than the junior clinicians and was equivalent to senior clinicians.ConclusionsThe proposed 2D CNN model realized the adequate and efficient automatic diagnoses of STs, which achieved a comparable performance of junior musculoskeletal radiologists and orthopedic surgeons. It might be conducive to assisting poor-experienced radiologists, especially in community scenarios lacking consulting experts.

Project description:IntroductionDespite the existence of numerous clinical techniques for identifying neurological brain disorders in their early stages, Electroencephalogram (EEG) data shows great promise as a means of detecting Alzheimer's disease (AD) at an early stage. The main goal of this research is to create a reliable and accurate clinical decision support system leveraging EEG signal processing to detect AD in its initial phases.MethodsThe research utilized a dataset consisting of 35 neurotypical individuals, 31 patients with mild AD, and 22 patients with moderate AD. Data were collected while participants were at rest. To extract features from the EEG signals, a band-pass filter was applied to the dataset and the Empirical Mode Decomposition (EMD) technique was employed to decompose the filtered signals. The EMD technique was then leveraged to generate feature vectors by combining multiple signal features, thereby enhancing diagnostic performance. Various artificial intelligence approaches were also explored and compared to identify features of the extracted EEG signals distinguishing mild AD, moderate AD, and neurotypical cases. The performance of the classifiers was evaluated using k-fold cross-validation and leave-one-subject-out (LOSO) cross-validation methods.ResultsThe results of this study provided valuable insights into potential avenues for the early diagnosis of AD. The performance of the various offered methodologies has been compared and evaluated by computing the overall diagnosis precision, recall, and accuracy. The proposed methodologies achieved a maximum classification accuracy of 99.9 and 94.8% for k-fold and LOSO cross-validation techniques, respectively.ConclusionThe study aims to assess and compare different proposed methodologies and determine the most effective combination approach for the early detection of AD. Our research findings strongly suggest that the proposed diagnostic support technique is a highly promising supplementary tool for discovering prospective diagnostic biomarkers that can greatly aid in the early clinical diagnosis of AD.

Project description:BackgroundAccurate assessment of the percentage total body surface area (%TBSA) of burn wounds is crucial in the management of burn patients. The resuscitation fluid and nutritional needs of burn patients, their need for intensive unit care, and probability of mortality are all directly related to %TBSA. It is difficult to estimate a burn area of irregular shape by inspection. Many articles have reported discrepancies in estimating %TBSA by different doctors.ObjectiveWe propose a method, based on deep learning, for burn wound detection, segmentation, and calculation of %TBSA on a pixel-to-pixel basis.MethodsA 2-step procedure was used to convert burn wound diagnosis into %TBSA. In the first step, images of burn wounds were collected from medical records and labeled by burn surgeons, and the data set was then input into 2 deep learning architectures, U-Net and Mask R-CNN, each configured with 2 different backbones, to segment the burn wounds. In the second step, we collected and labeled images of hands to create another data set, which was also input into U-Net and Mask R-CNN to segment the hands. The %TBSA of burn wounds was then calculated by comparing the pixels of mask areas on images of the burn wound and hand of the same patient according to the rule of hand, which states that one's hand accounts for 0.8% of TBSA.ResultsA total of 2591 images of burn wounds were collected and labeled to form the burn wound data set. The data set was randomly split into training, validation, and testing sets in a ratio of 8:1:1. Four hundred images of volar hands were collected and labeled to form the hand data set, which was also split into 3 sets using the same method. For the images of burn wounds, Mask R-CNN with ResNet101 had the best segmentation result with a Dice coefficient (DC) of 0.9496, while U-Net with ResNet101 had a DC of 0.8545. For the hand images, U-Net and Mask R-CNN had similar performance with DC values of 0.9920 and 0.9910, respectively. Lastly, we conducted a test diagnosis in a burn patient. Mask R-CNN with ResNet101 had on average less deviation (0.115% TBSA) from the ground truth than burn surgeons.ConclusionsThis is one of the first studies to diagnose all depths of burn wounds and convert the segmentation results into %TBSA using different deep learning models. We aimed to assist medical staff in estimating burn size more accurately, thereby helping to provide precise care to burn victims.

Dataset Information

Development and Evaluation of a Leukemia Diagnosis System Using Deep Learning in Real Clinical Scenarios.

Publications

Development and Evaluation of a Leukemia Diagnosis System Using Deep Learning in Real Clinical Scenarios.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets