Multi-stage Clustering of Breast Cancer for Precision Medicine

TL;DR

This paper introduces a two-stage clustering framework that combines phenotypic and genomic data to classify breast cancer patients, aiming to enhance personalized treatment strategies in precision medicine.

Contribution

A novel hierarchical clustering method integrating phenotypic and genomic data for better patient stratification in breast cancer.

Findings

Effective identification of patient subgroups based on combined data types

Enhanced understanding of correlations among phenotypic and genomic factors

Potential to improve personalized treatment planning

Abstract

Cancer has become one of the most widespread diseases in the world. Specifically, breast cancer is diagnosed more often than any other type of cancer. However, breast cancer patients and their individual tumors are often unique. Identifying the underlying genetic phenotype can lead to precision (personalized) medicine. Tailoring medical treatment strategies to best fit the needs of individual patients can dramatically improve their health. Such an approach requires sufficient knowledge of the patients and the diseases, which is currently unavailable to practitioners. This study focuses on breast cancer and proposes a novel two-stage clustering method to partition patients into hierarchical groups. The first stage is broad grouping, which is based on phenotypes such as demographic information and clinical features. The second stage is fine grouping based on genomic characteristics, such as copy number variation and somatic mutation, of patients in a subgroup resulting from the first stage. Generally, this framework offers a mechanism to mix multiple forms of data, both phenotypic and genomic, to most effectively define individual patients for personalized predictions. This method provides the ability to detect correlation among all factors.