Hybrid Ant Lion Mutated Ant Colony Optimizer Technique With Particle Swarm Optimization for Leukemia Prediction Using Microarray Gene Data

IEEE Access, Год журнала: 2024, Номер 12, С. 10910 - 10919

Опубликована: Янв. 1, 2024

Leukemia refers to a type of blood malignancy that develops due certain hematological disorders. Identifying leukemia at its earlier stages through clinical operations are highly complicated task with invasive methods. Gene expression data could be collected and computational methods adopted which lead better prediction leads prevention stages. Today, feature selection has become an important step in pre-processing helps bring improvement the classification system performance is done by choosing optimal subsets means reducing or eliminating redundant irrelevant features. Particle Swarm Optimization (PSO) popular algorithm wherein solutions generated randomly move within search space obtain solutions. Another relatively new evolutionary method computation Ant Lion (ALO) lower cost compared other techniques. In this work, technique known as Hybrid Mutated Colony Optimize along was proposed for leukaemia microarray gene data. The model used identifying set features from been using Support Vector Machine (SVM) produced significant accuracy 87.88%.

Язык: Английский

---

Artificial Intelligence-Based Management of Adult Chronic Myeloid Leukemia: Where Are We and Where Are We Going?

Cancers, Год журнала: 2024, Номер 16(5), С. 848 - 848

Опубликована: Фев. 20, 2024

Artificial intelligence (AI) is emerging as a discipline capable of providing significant added value in Medicine, particular radiomic, imaging analysis, big dataset and also for generating virtual cohort patients. However, coping with chronic myeloid leukemia (CML), considered an easily managed malignancy after the introduction TKIs which strongly improved life expectancy patients, AI still its infancy. Noteworthy, findings initial trials are intriguing encouraging, both terms performance adaptability to different contexts can be applied. Indeed, improvement diagnosis prognosis by leveraging biochemical, biomolecular, imaging, clinical data crucial implementation personalized medicine paradigm or streamlining procedures services. In this review, we present state art applications field CML, describing techniques objectives, general focus that goes beyond Machine Learning (ML), but instead embraces wider field. The scooping review spans on publications reported Pubmed from 2003 2023, resulting searching “chronic leukemia” “artificial intelligence”. time frame reflects real literature production was not restricted. We take opportunity discussing main pitfalls key points must respond, especially considering critical role ‘human’ factor, remains domain.

Язык: Английский

---

Sparsity Regularization Enhances Gene Selection and Leukemia Subtype Classification via Logistic Regression

Leukemia Research, Год журнала: 2025, Номер 150, С. 107663 - 107663

Опубликована: Фев. 11, 2025

Язык: Английский

---

Gene Expression-Based Cancer Classification for Handling the Class Imbalance Problem and Curse of Dimensionality

International Journal of Molecular Sciences, Год журнала: 2024, Номер 25(4), С. 2102 - 2102

Опубликована: Фев. 9, 2024

Cancer is a leading cause of death globally. The majority cancer cases are only diagnosed in the late stages due to use conventional methods. This reduces chance survival for patients. Therefore, early detection consequently followed by diagnoses important tasks research. Gene expression microarray technology has been applied detect and diagnose most types cancers their gained encouraging results. In this paper, we address problem classifying based on gene handling class imbalance curse dimensionality. oversampling technique utilized overcome adding synthetic samples. Another common issue related dataset addressed paper applying chi-square information gain feature selection techniques. After these techniques individually, proposed method select significant genes combining those two (CHiS IG). We investigated effect individually combination. Four benchmarking biomedical datasets (Leukemia-subtypes, Leukemia-ALLAML, Colon, CuMiDa) were used. experimental results reveal that improve cases. Additionally, performance outperforms individual nearly all addition, study provides an empirical evaluating several along with ensemble-based learning. also SVM-SMOTE, random forests classifier, achieved highest results, reporting accuracy 100%. obtained surpass findings existing literature as well.

Язык: Английский

---

Leukemia Diagnosis using Machine Learning Classifiers based on MRMR Feature Selection

Engineering Technology & Applied Science Research, Год журнала: 2024, Номер 14(4), С. 15614 - 15619

Опубликована: Авг. 2, 2024

Early and accurate diagnosis of leukemia is crucial for effective treatment. Machine Learning (ML) offers promising tools classification, but the required high-dimensional datasets pose challenges. This study explores effectiveness ML algorithms disease classification investigates impact feature selection with Minimum Redundancy Maximum Relevance (MRMR ) technique. MRMR was implemented to select informative features evaluate four (Naïve Bayes (NB), K-Nearest Neighbors (KNN), Support Vector (SVM), Artificial Neural Networks (ANNs)) using subsets varying levels relevance based on scores. Our results demonstrate that effectively reduced dimensionality while maintaining even improving accuracy. KNN SVM achieved highest accuracy (100% 67, 30, 24 subsets), suggesting benefit focusing highly relevant features. NB exhibited consistent across all sets.

Язык: Английский

---

Multiclass Classification of Leukemia Cancer Subtypes using Gene Expression Data and Optimized Dueling Double Deep Q-Network

Chemometrics and Intelligent Laboratory Systems, Год журнала: 2025, Номер unknown, С. 105402 - 105402

Опубликована: Апрель 1, 2025

Язык: Английский

---

P-Glycoprotein as a Therapeutic Target in Hematological Malignancies: A Challenge to Overcome

International Journal of Molecular Sciences, Год журнала: 2025, Номер 26(10), С. 4701 - 4701

Опубликована: Май 14, 2025

P-glycoprotein (P-gp), a transmembrane efflux pump encoded by the ABCB1/MDR1 gene, is major contributor to multidrug resistance in hematological malignancies. These malignancies, arising from hematopoietic precursors at various differentiation stages, can manifest bone marrow, circulate bloodstream, or infiltrate tissues. P-gp overexpression malignant cells reduces efficacy of chemotherapeutic agents actively expelling them, decreasing intracellular drug concentrations, and promoting resistance, significant obstacle successful treatment. This review examines recent advances combating P-gp-mediated including development novel inhibitors, innovative delivery systems (e.g., nanoparticle-based delivery), strategies modulate expression activity. modulation encompass targeting relevant signaling pathways NF-κB, PI3K/Akt) exploring repurposing. While progress has been made, overcoming remains crucial for improving patient outcomes. Future research directions should prioritize potent, selective, safe inhibitors with minimal off-target effects, alongside synergistic combination therapies existing chemotherapeutics effectively circumvent

Язык: Английский

---

An efficient leukemia prediction method using machine learning and deep learning with selected features

PLoS ONE, Год журнала: 2025, Номер 20(5), С. e0320669 - e0320669

Опубликована: Май 16, 2025

Leukemia is a serious problem affecting both children and adults, leading to death if left untreated. kind of blood cancer described by the rapid proliferation abnormal cells. An early, trustworthy, precise identification leukemia important treating saving patients’ lives. Acute myelogenous lymphocytic, chronic are four kinds leukemia. Manual inspection microscopic images frequently used identify these malignant growth symptoms include fatigue, lack enthusiasm, dull appearance, recurring illnesses, easy loss. Identifying subtypes for specialized therapy one hurdles in this area. The suggested work predicts classifies gene data CuMiDa (GSE9476) using feature selection ML techniques. Curated Microarray Database (CuMiDa) collected 64 samples representing five classes genes out 22283 genes. proposed approach utilizes 25 most differentiating selected features classification machine deep learning This study has accuracy 96.15% Random Fores, 92.30 Linear Regression, SVM, 100% LSTM. Deep methods have been shown outperform traditional utilizing specific features.

Язык: Английский

---

Using Deep Learning Techniques to Enhance Blood Cell Detection in Patients with Leukemia

Information, Год журнала: 2024, Номер 15(12), С. 787 - 787

Опубликована: Дек. 8, 2024

Medical diagnosis plays a critical role in the early detection and treatment of diseases by examining symptoms supporting findings through advanced laboratory testing. Early accurate is essential for detecting medical problems then prescribing most effective strategies, especially life-threatening such as leukemia. Leukemia, blood malignancy, one prevalent cancer types affecting both adults children. It caused rapid uncontrolled growth abnormal white cells bone marrow. This accumulation interferes with production normal cells, leading to weakened immune deficiency, anemia, bleeding disorders. Conventional leukemia diagnostic methods are time-consuming, manually intensive, inefficient. research study proposes an automatic diagnostics prediction analyzing images according shape blast using digital image processing machine learning. The purpose cell precisely identify classify diverse anomalies associated cancers like supports monitoring, which leads more treatments improved results patients. To accomplish this task, we use techniques apply convolutional neural network (CNN) deep learning algorithm sample images. employs multi-stage methodology, including data preparation, preprocessing, feature extraction, classification. While our model built on typical CNN architecture, make significant advances preprocessing hyperparameter tuning. We have modified its layers combination include convolutional, pooling, fully connected that optimized characteristics. These fine-tuned better extraction classification accuracy. showed diagnosing acute based had 99% accuracy outperformed other models, DenseNet121, ResNet-50, Incep-tionv3, MobileNet, EfficientNet. comprehensive analysis reveals highest compared existing studies relevant literature.

Язык: Английский

---

Optimizing Deep Learning with Dimensionality Reduction for Analyzing the CuMiDa Brain Cancer Gene Expression Dataset

Jurnal Riset Informatika, Год журнала: 2024, Номер 6(4), С. 237 - 246

Опубликована: Сен. 15, 2024

In the digital era, machine learning and deep have become indispensable tools for bioinformatics, particularly in analyzing high-dimensional gene expression data cancer diagnosis classification. This study leverages CuMiDa brain dataset, a curated microarray database with 54,676 genes 130 samples, to evaluate effectiveness of models integrated dimensionality reduction techniques. Principal Component Analysis (PCA) Truncated Singular Value Decomposition (TruncatedSVD) were employed address challenges data, reducing noise computational complexity. Three models—DNN, MLP, TabNet—were implemented various optimizers, including ADAM, RMSprop, SGD. Results showed that TruncatedSVD outperformed PCA minimizing loss, especially MLP LBFGS achieving near-zero loss. TabNet demonstrated highest classification accuracy (96%) ADAM RMSprop. Conversely, SGD exhibited suboptimal performance across models. These findings highlight critical role optimizer selection enhancing efficiency research provides robust framework improving diagnostic oncology.

Язык: Английский

---