A Novel BOND-KNN Algorithm to Predict Breast Cancer Survival using Multi-modal Features.

Area of Science:

• Oncology
• Bioinformatics
• Machine Learning

Background:

• Breast cancer is the most common cancer globally, with increasing incidence.
• Despite treatment advances, precise prognostic models are still needed.
• Accurate survival prediction is crucial for effective breast cancer management.

Purpose of the Study:

• To develop an innovative approach for improving breast cancer prognosis.
• To address the challenge of accurately predicting patient survival.
• To leverage multi-modal data for enhanced predictive accuracy.

Main Methods:

• A modified K-nearest neighbors (KNN) algorithm was developed, termed BOND-KNN.
• The algorithm utilizes Bonferroni distance (BOND) to account for feature relationships.
• Multi-modal datasets including clinical, gene expression, and copy number alteration data were used.

Main Results:

• The BOND-KNN model demonstrated high accuracy, precision, recall, and F1-score.
• The methodology effectively predicted patient survival.
• Analysis of intricate relationships among sample features improved predictive performance.

Conclusions:

• The proposed BOND-KNN methodology offers a precise and effective approach for breast cancer prognosis.
• Leveraging multi-modal data with the BOND-KNN algorithm enhances survival prediction.
• This innovative method contributes to advancing breast cancer management strategies.