An Efficient Blood-Cell Segmentation for the Detection of Hematological Disorders

CoTCoNet: An optimized coupled transformer-convolutional network with an adaptive graph reconstruction for leukemia detection

BACKGROUND

Swift and accurate blood smear analyses are crucial for diagnosing leukemia and other hematological malignancies. However, manual leukocyte count and morphological evaluation remain time-consuming and prone to errors. Additionally, conventional image processing methods struggle to differentiate cells due to visual similarities between malignant and benign cell morphology.

METHOD

In response to above challenges, we propose Coupled Transformer Convolutional Network (CoTCoNet) framework for leukemia classification. CoTCoNet integrates dual-feature extraction to capture long-range global features and fine-grained spatial patterns, facilitating the identification of complex hematological characteristics. Additionally, the framework employs a graph-based module to uncover hidden, biologically relevant features of leukocyte cells, along with a Population-based Meta-Heuristic Algorithm for feature selection and optimization. Furthermore, we introduce a novel combination of leukocyte segmentation and synthesis, which isolates relevant regions while augmenting the training dataset with realistic leukocyte samples. This strategy isolates relevant regions while augmenting the training data with realistic leukocyte samples, enhancing feature extraction, and addressing data scarcity without compromising data integrity.

RESULTS

We evaluated CoTCoNet on a dataset of 16,982 annotated cells, achieving an accuracy of 0.9894 and an F1-Score of 0.9893. We tested CoTCoNet on four diverse, publicly available datasets (including those above) to assess generalizability. Results demonstrate a significant performance improvement over existing state-of-the-art approaches.

CONCLUSIONS

CoTCoNet represents a significant advancement in leukemia classification, offering enhanced accuracy and efficiency compared to traditional methods. By incorporating explainable visualizations that closely align with cell annotations, the framework provides deeper insights into its decision-making process, further solidifying its potential in clinical settings.

Deep self-supervised machine learning algorithms with a novel feature elimination and selection approaches for blood test-based multi-dimensional health risks classification

BACKGROUND

Blood test is extensively performed for screening, diagnoses and surveillance purposes. Although it is possible to automatically evaluate the raw blood test data with the advanced deep self-supervised machine learning approaches, it has not been profoundly investigated and implemented yet.

RESULTS

This paper proposes deep machine learning algorithms with multi-dimensional adaptive feature elimination, self-feature weighting and novel feature selection approaches. To classify the health risks based on the processed data with the deep layers, four machine learning algorithms having various properties from being utterly model free to gradient driven are modified.

CONCLUSIONS

The results show that the proposed deep machine learning algorithms can remove the unnecessary features, assign self-importance weights, selects their most informative ones and classify the health risks automatically from the worst-case low to worst-case high values.

Allogeneic hematopoietic stem cell transplantation for pediatric acute lymphoblastic leukemia

This study was aimed to explore the prognosis of allogenic hematopoietic stem cell transplantation (allo-HSCT) in pediatric patients with acute lymphoblastic leukemia (ALL). This retrospective case series study included children with ALL who underwent allo-HSCT at Beijing Children's Hospital of Capital Medical University, Beijing, China, between January 2009 and December 2019. The outcomes included 5-year overall survival (OS) and event-free survival (EFS). A total of 75 children (52 males) were included. The median age at presentation was 5.30 years, and the median time from diagnosis to transplantation was 1.64 years. There were 15 human leukocyte antigen (HLA)-matched and 60 HLA-semi-matched transplants, 73 complete remissions (CR), and 2 MRD-positive transplants. The median follow-up time was 41 months. Out of 75 patients, 51 children survived, and 24 died/given up at the terminal stage. The 5-year OS and EFS rates were 67.77% and 57.30%, respectively, whereas the 5-year recurrence rate was 35.69%. Acute and chronic graft versus host diseases occurred in 40 and 28 cases, respectively. Children with MLL gene fusion had higher survival rates compared to other subgroups. Haplo-HSCT is not inferior to HLA-matched transplant. The children with MLL rearrangement had an acceptable 5-year OS, while complications and relapse should be monitored.

Deep learning enhances acute lymphoblastic leukemia diagnosis and classification using bone marrow images

Acute lymphoblastic leukemia (ALL) poses a significant health challenge, particularly in pediatric cases, requiring precise and rapid diagnostic approaches. This comprehensive review explores the transformative capacity of deep learning (DL) in enhancing ALL diagnosis and classification, focusing on bone marrow image analysis. Examining ten studies conducted between 2013 and 2023 across various countries, including India, China, KSA, and Mexico, the synthesis underscores the adaptability and proficiency of DL methodologies in detecting leukemia. Innovative DL models, notably Convolutional Neural Networks (CNNs) with Cat-Boosting, XG-Boosting, and Transfer Learning techniques, demonstrate notable approaches. Some models achieve outstanding accuracy, with one CNN reaching 100% in cancer cell classification. The incorporation of novel algorithms like Cat-Swarm Optimization and specialized CNN architectures contributes to superior classification accuracy. Performance metrics highlight these achievements, with models consistently outperforming traditional diagnostic methods. For instance, a CNN with Cat-Boosting attains 100% accuracy, while others hover around 99%, showcasing DL models' robustness in ALL diagnosis. Despite acknowledged challenges, such as the need for larger and more diverse datasets, these findings underscore DL's transformative potential in reshaping leukemia diagnostics. The high numerical accuracies accentuate a promising trajectory toward more efficient and accurate ALL diagnosis in clinical settings, prompting ongoing research to address challenges and refine DL models for optimal clinical integration.

Recent advancements in machine learning and deep learning-based breast cancer detection using mammograms

OBJECTIVE

Mammogram-based automatic breast cancer detection has a primary role in accurate cancer diagnosis and treatment planning to save valuable lives. Mammography is one basic yet efficient test for screening breast cancer. Very few comprehensive surveys have been presented to briefly analyze methods for detecting breast cancer with mammograms. In this article, our objective is to give an overview of recent advancements in machine learning (ML) and deep learning (DL)-based breast cancer detection systems.

METHODS

We give a structured framework to categorize mammogram-based breast cancer detection techniques. Several publicly available mammogram databases and different performance measures are also mentioned.

RESULTS

After deliberate investigation, we find most of the works classify breast tumors either as normal-abnormal or malignant-benign rather than classifying them into three classes. Furthermore, DL-based features are more significant than hand-crafted features. However, transfer learning is preferred over others as it yields better performance in small datasets, unlike classical DL techniques.

SIGNIFICANCE AND CONCLUSION

In this article, we have made an attempt to give recent advancements in artificial intelligence (AI)-based breast cancer detection systems. Furthermore, a number of challenging issues and possible research directions are mentioned, which will help researchers in further scopes of research in this field.

Hierarchical Knowledge Propagation and Distillation for Few-Shot Learning

Recent research efforts on Few-Shot Learning (FSL) have achieved extensive progress. However, the existing efforts primarily focus on the transductive setting of FSL, which is heavily challenged by the limited quantity of the unlabeled query set. Although a few inductive-based FSL methods have been studied, most of them emphasize learning superb feature extraction networks. As a result, they may ignore the relations between sample-level and class-level representations, which are particularly crucial when labeled samples are scarce. This paper proposes an inductive FSL framework that leverages the Hierarchical Knowledge Propagation and Distillation, named HKPD. To learn more discriminative sample-level representations, HKPD first constructs a sample-level information propagation module that explores pairwise sample relations. Subsequently, a class-level information propagation module is designed to obtain and update the class-level information. Moreover, a self-distillation module is adopted to further improve the learned representations by propagating the obtained knowledge across this hierarchical architecture. Extensive experiments conducted on the commonly used few-shot benchmark datasets demonstrate the superiority of the proposed HKPD method, which outperforms the current state-of-the-art methods.

Cell segmentation in fluorescence microscopy images based on multi-scale histogram thresholding

Cell segmentation from fluorescent microscopy images plays an important role in various applications, such as disease mechanism assessment and drug discovery research. Exiting segmentation methods often adopt image binarization as the first step, through which the foreground cell is separated from the background so that the subsequent processing steps can be greatly facilitated. To pursue this goal, a histogram thresholding can be performed on the input image, which first applies a Gaussian smoothing to suppress the jaggedness of the histogram curve and then exploits Rosin's method to determine a threshold for conducting image binarization. However, an inappropriate amount of smoothing could lead to the inaccurate segmentation of cells. To address this crucial problem, a multi-scale histogram thresholding (MHT) technique is proposed in the present paper, where the scale refers to the standard deviation of the Gaussian that determines the amount of smoothing. To be specific, the image histogram is smoothed at three chosen scales first, and then the smoothed histogram curves are fused to conduct image binarization via thresholding. To further improve the segmentation accuracy and overcome the difficulty of extracting overlapping cells, our proposed MHT technique is incorporated into a multi-scale cell segmentation framework, in which a region-based ellipse fitting technique is adopted to identify overlapping cells. Extensive experimental results obtained on benchmark datasets show that the new method can deliver superior performance compared to the current state-of-the-arts.

An Optimized Deep Learning Model for Predicting Mild Cognitive Impairment Using Structural MRI

Early diagnosis of mild cognitive impairment (MCI) with magnetic resonance imaging (MRI) has been shown to positively affect patients' lives. To save time and costs associated with clinical investigation, deep learning approaches have been used widely to predict MCI. This study proposes optimized deep learning models for differentiating between MCI and normal control samples. In previous studies, the hippocampus region located in the brain is used extensively to diagnose MCI. The entorhinal cortex is a promising area for diagnosing MCI since severe atrophy is observed when diagnosing the disease before the shrinkage of the hippocampus. Due to the small size of the entorhinal cortex area relative to the hippocampus, limited research has been conducted on the entorhinal cortex brain region for predicting MCI. This study involves the construction of a dataset containing only the entorhinal cortex area to implement the classification system. To extract the features of the entorhinal cortex area, three different neural network architectures are optimized independently: VGG16, Inception-V3, and ResNet50. The best outcomes were achieved utilizing the convolution neural network classifier and the Inception-V3 architecture for feature extraction, with accuracy, sensitivity, specificity, and area under the curve scores of 70%, 90%, 54%, and 69%, respectively. Furthermore, the model has an acceptable balance between precision and recall, achieving an F1 score of 73%. The results of this study validate the effectiveness of our approach in predicting MCI and may contribute to diagnosing MCI through MRI.

Identifying the Key Components in ResNet-50 for Diabetic Retinopathy Grading from Fundus Images: A Systematic Investigation

Although deep learning-based diabetic retinopathy (DR) classification methods typically benefit from well-designed architectures of convolutional neural networks, the training setting also has a non-negligible impact on prediction performance. The training setting includes various interdependent components, such as an objective function, a data sampling strategy, and a data augmentation approach. To identify the key components in a standard deep learning framework (ResNet-50) for DR grading, we systematically analyze the impact of several major components. Extensive experiments are conducted on a publicly available dataset EyePACS. We demonstrate that (1) the DR grading framework is sensitive to input resolution, objective function, and composition of data augmentation; (2) using mean square error as the loss function can effectively improve the performance with respect to a task-specific evaluation metric, namely the quadratically weighted Kappa; (3) utilizing eye pairs boosts the performance of DR grading and; (4) using data resampling to address the problem of imbalanced data distribution in EyePACS hurts the performance. Based on these observations and an optimal combination of the investigated components, our framework, without any specialized network design, achieves a state-of-the-art result (0.8631 for Kappa) on the EyePACS test set (a total of 42,670 fundus images) with only image-level labels. We also examine the proposed training practices on other fundus datasets and other network architectures to evaluate their generalizability. Our codes and pre-trained model are available online.

Validation of a Machine Learning Expert Supporting System, ImmunoGenius, Using Immunohistochemistry Results of 3000 Patients with Lymphoid Neoplasms

(1) Background: Differential diagnosis using immunohistochemistry (IHC) panels is a crucial step in the pathological diagnosis of hematolymphoid neoplasms. In this study, we evaluated the prediction accuracy of the ImmunoGenius software using nationwide data to validate its clinical utility. (2) Methods: We collected pathologically confirmed lymphoid neoplasms and their corresponding IHC results from 25 major university hospitals in Korea between 2015 and 2016. We tested ImmunoGenius using these real IHC panel data and compared the precision hit rate with previously reported diagnoses. (3) Results: We enrolled 3052 cases of lymphoid neoplasms with an average of 8.3 IHC results. The precision hit rate was 84.5% for these cases, whereas it was 95.0% for 984 in-house cases. (4) Discussion: ImmunoGenius showed excellent results in most B-cell lymphomas and generally showed equivalent performance in T-cell lymphomas. The primary reasons for inaccurate precision were atypical IHC profiles of certain cases, lack of disease-specific markers, and overlapping IHC profiles of similar diseases. We verified that the machine-learning algorithm could be applied for diagnosis precision with a generally acceptable hit rate in a nationwide dataset. Clinical and histological features should also be taken into account for the proper use of this system in the decision-making process.

Joint Cancer Segmentation and PI-RADS Classification on Multiparametric MRI Using MiniSegCaps Network

MRI is the primary imaging approach for diagnosing prostate cancer. Prostate Imaging Reporting and Data System (PI-RADS) on multiparametric MRI (mpMRI) provides fundamental MRI interpretation guidelines but suffers from inter-reader variability. Deep learning networks show great promise in automatic lesion segmentation and classification, which help to ease the burden on radiologists and reduce inter-reader variability. In this study, we proposed a novel multi-branch network, MiniSegCaps, for prostate cancer segmentation and PI-RADS classification on mpMRI. MiniSeg branch outputted the segmentation in conjunction with PI-RADS prediction, guided by the attention map from the CapsuleNet. CapsuleNet branch exploited the relative spatial information of prostate cancer to anatomical structures, such as the zonal location of the lesion, which also reduced the sample size requirement in training due to its equivariance properties. In addition, a gated recurrent unit (GRU) is adopted to exploit spatial knowledge across slices, improving through-plane consistency. Based on the clinical reports, we established a prostate mpMRI database from 462 patients paired with radiologically estimated annotations. MiniSegCaps was trained and evaluated with fivefold cross-validation. On 93 testing cases, our model achieved a 0.712 dice coefficient on lesion segmentation, 89.18% accuracy, and 92.52% sensitivity on PI-RADS classification (PI-RADS ≥ 4) in patient-level evaluation, significantly outperforming existing methods. In addition, a graphical user interface (GUI) integrated into the clinical workflow can automatically produce diagnosis reports based on the results from MiniSegCaps.

Multiscale segmentation- and error-guided iterative convolutional neural network for cerebral neuron segmentation in microscopic images

This article uses microscopy images obtained from diverse anatomical regions of macaque brain for neuron semantic segmentation. The complex structure of brain, the large intra-class staining intensity difference within neuron class, the small inter-class staining intensity difference between neuron and tissue class, and the unbalanced dataset increase the difficulty of neuron semantic segmentation. To address this problem, we propose a multiscale segmentation- and error-guided iterative convolutional neural network (MSEG-iCNN) to improve the semantic segmentation performance in major anatomical regions of the macaque brain. After evaluating microscopic images from 17 anatomical regions, the semantic segmentation performance of neurons is improved by 10.6%, 4.0%, 1.5%, and 1.2% compared with Random Forest, FCN-8s, U-Net, and UNet++, respectively. Especially for neurons with brighter staining intensity in the anatomical regions such as lateral geniculate, globus pallidus and hypothalamus, the performance is improved by 66.1%, 23.9%, 11.2%, and 6.7%, respectively. Experiments show that our proposed method can efficiently segment neurons with a wide range of staining intensities. The semantic segmentation results are of great significance and can be further used for neuron instance segmentation, morphological analysis and disease diagnosis. Cell segmentation plays a critical role in extracting cerebral information, such as cell counting, cell morphometry and distribution analysis. Accurate automated neuron segmentation is challenging due to the complex structure of brain, the large intra-class staining intensity difference within neuron class, the small inter-class staining intensity difference between neuron and tissue class, and the unbalanced dataset. The proposed multiscale segmentation- and error-guided iterative convolutional neural network (MSEG-iCNN) improve the segmentation performance in 17 major anatomical regions of the macaque brain.