What type of study is this?

August 21, 2025

Machine Learning-Based Breast Cancer Prediction from Genomic Data

Key Points

The 1D CNN model achieved 72% accuracy in predicting breast cancer, outpacing other models, including XGBoost.
The core mechanism utilized was an XGBoost classifier, tested alongside SVM, KNN, Decision Tree, and others.
Assessment was done using the METABRIC dataset, showcasing the models' predictive capabilities with ROC-AUC metrics.
This accessible web interface serves healthcare professionals, enhancing timely and reliable cancer detection methodologies.

Abstract

Breast cancer remains one of the leading causes of cancer-related mortality worldwide, emphasizing the need for timely and accurate detection. This research proposes a machine learning-based framework to classify samples as cancerous or non-cancerous by leveraging high-dimensional genomic information combined with clinical data. The core model used is an XGBoost classifier, embedded within a user-friendly web interface, and benchmarked against other models such as SVM, KNN, Decision Tree, Random Forest, and a one- dimensional Convolutional Neural Network (1D CNN). These models were trained and evaluated using the METABRIC dataset. Among them, the 1D CNN achieved the highest performance, with 72% accuracy and a ROC-AUC of 0.71, while XGBoost followed closely with 68% accuracy and an AUC of 0.57. The overall system is built using the Flask framework, allows healthcare professionals to upload gene expression data and obtain instant predictions along with explanations of key contributing features. This study delivers a practical and accessible solution for breast cancer prediction, combining high reliability with clinical interpretability. Keywords: Breast Cancer, Genomic Data, Machine Learning, XGBoost, Convolutional Neural Network, Precision Medicine, Interpretability

Perguntar à IA

Bookmark