Ultrasound Genomics Reveals a Signature for Predicting Breast Cancer Prognosis and Therapy Response

Area of Science:

• Oncology
• Genomics
• Medical Imaging

Background:

• Breast cancer (BC) is a leading cause of cancer worldwide.
• Accurate tools for prognosis and chemotherapy response prediction in BC are lacking.
• Ultrasound imaging and transcriptome data offer potential for novel biomarker discovery.

Purpose of the Study:

• To develop an ultrasound genomics-based signature for breast cancer prognostic assessment.
• To identify key genes and ultrasound features associated with patient outcomes and treatment response.
• To establish a predictive model supporting personalized medicine in breast cancer care.

Main Methods:

• Collected ultrasound images and transcriptome data from breast cancer patients.
• Utilized Support Vector Machine (SVM) and logistic regression for feature identification.
• Employed differential expression analysis, Cox regression, and Lasso for gene signature development.
• Identified VPS28 as a potential prognostic biomarker.

Main Results:

• Identified 12 ultrasound features linked to chemotherapy response.
• Derived a five-gene signature correlating ultrasound features with prognostic genes.
• The signature effectively stratifies patients and predicts chemotherapy response.
• VPS28 showed strong association with poor prognosis and tumor invasiveness.

Conclusions:

• Ultrasound genomics provides an effective tool for breast cancer prognostic assessment.
• The developed gene signature supports personalized treatment strategies and prediction of chemotherapy response.
• VPS28 is a significant biomarker for breast cancer prognosis.