Machine Learning Algorithms for Breast Cancer Diagnosis: Challenges, Prospects and Future Research Directions

Journal of Oncology Research, Journal Year: 2022, Volume and Issue: 5(1)

Published: Nov. 2, 2022

Early diagnosis of breast cancer does not only increase the chances survival but also control diffusion cancerous cells in body. Previously, researchers have developed machine learning algorithms such as Support Vector Machine, K-Nearest Neighbor, Convolutional Neural Network, K-means, Fuzzy C-means, Principle Component Analysis (PCA) and Naive Bayes. Unfortunately these fall short one way or another due to high levels computational complexities. For instance, support vector employs feature elimination scheme for eradicating data ambiguity detecting tumors at initial stage. However this is expensive terms execution time. On its part, k-means algorithm Euclidean distance determine between cluster centers points. guarantee accuracy when executed different iterations. Although K-nearest Neighbor reduction, principle component analysis 10 fold cross validation methods enhancing classification accuracy, it efficient processing other hand, fuzzy c-means fuzziness value termination criteria time on datasets. However, proves be extensive several iterations measure calculations involved. Similarly, convolutional neural network employed back propagation method slow frequent retraining. In addition, achieves low predictions. Since all seem consuming, necessary integrate quantum computing principles with conventional algorithms. This because has potential accelerate computations by simultaneously carrying out calculation many inputs. paper, a review current prediction provided. Based observed shortcomings, based classifier recommended. The proposed working mechanisms are elaborated towards end paper.

Language: Английский

---

Identifying SLC2A6 as the novel protective factor in breast cancer by TP53-related genes affecting M1 macrophage infiltration

APOPTOSIS, Journal Year: 2024, Volume and Issue: 29(7-8), P. 1211 - 1231

Published: April 16, 2024

Language: Английский

---

Precision Oncology in the Era of Genomics and Artificial Intelligence

Published: Jan. 1, 2024

Abstract Cancer patient care classically represents proper diagnosis, designing appropriate therapeutics and clinical management protocols. Concept of precision medicine emerged in conjuncture to personalized when subpopulations reasonably differ disease risks, prognosis, treatment response due interpersonal differences biology. Precision oncology aims tailor medical decisions interventions optimize guidance on survival benefits or quality life for each by utilizing person’s characteristics such as clinicopathology, mutational load, biochemical test profiles, specific protein expressions, pharmacogenomics, pharmacokinetics–pharmacodynamics determine risk prediction, planning, best efficacy. Artificial intelligence (AI), i.e., the ability a machine learn recognizing patterns from multidimensional large datasets, has vast use health care, most recently been generate algorithms complex inputs improvise traditional approach cancer diagnostics therapy. AI superseding classical genetic marker panels, enabling analysis large-scale multiomic data development sophisticated predictive models, extending its applicability several aspects screening, stratification, well managements. The integration genomic profile with becomes crucial tool analyze how an individual’s unique makeup influences susceptibility outcomes. Convergence multimodal driven genomics revolutionized oncology, ultimately reshaping landscape horizon uncovering new opportunities better understanding

Language: Английский

---

Mammographic Classification of Breast Cancer Microcalcifications through Extreme Gradient Boosting

Electronics, Journal Year: 2022, Volume and Issue: 11(15), P. 2435 - 2435

Published: Aug. 4, 2022

In this paper, we proposed an effective and efficient approach to the classification of breast cancer microcalcifications evaluated mathematical model for calcification on mammography with a large medical dataset. We employed several semi-automatic segmentation algorithms extract 51 features from mammograms, including morphologic textural features. adopted extreme gradient boosting (XGBoost) classify microcalcifications. Then, compared other machine learning techniques, k-nearest neighbor (kNN), adaboostM1, decision tree, random forest (RDF), tree (GBDT), XGBoost. XGBoost showed highest accuracy (90.24%) classifying microcalcifications, kNN demonstrated lowest accuracy. This result demonstrates that it is essential microcalcification use feature engineering method selection best composition One contributions study present cancers. paper finds way select discriminative as collection improve AUC = 0.89. Moreover, highlighted performance various dataset found ideal parameters Furthermore, suitable both in theory practice calcifications mammography.

Language: Английский

---

Machine Learning-Guided Differential Gene Expression Analysis Identifies A Highly-Connected Seven-Gene Cluster in Triple-Negative Breast Cancer

Biomedicine, Journal Year: 2024, Volume and Issue: 14(4)

Published: Dec. 1, 2024

Background: One of the most challenging cancers is triple-negative breast cancer, which subdivided into many molecular subtypes. Due to high degree heterogeneity, role precision medicine remains challenging. With use machine learning (ML)-guided gene selection, differential expression analysis can be optimized, and eventually, process see great advancement through biomarker discovery.

Language: Английский

---