Related Experiment Video
Updated: May 10, 2026
13:19
Microarray-based Identification of Individual HERV Loci Expression: Application to Biomarker Discovery in Prostate Cancer
Published on: November 2, 2013
16.7K
Proteogenomic Biomarker Profiling for Predicting Radiolabeled Immunotherapy Response in Resistant Prostate Cancer
Benchun Yan1, Yuqiu Gao2, Yulong Zou2
1Department of Urology, Hongqi Hospital, Mudanjiang Medical University, Mudanjiang, China.
Cancer Biotherapy & Radiopharmaceuticals
|August 28, 2025
Summary
This study introduces a new machine learning approach to predict treatment response in prostate cancer (PCa) patients. It identifies ultrasound biomarkers linked to proteogenomic profiles, aiding in radiolabeled immunotherapy development for resistant cases.
Area of Science:
- Oncology
- Genomics
- Radiotherapy
- Machine Learning
Background:
- Treatment resistance is a significant challenge in prostate cancer (PCa), limiting the effectiveness of preoperative chemoradiotherapy and radiolabeled immunotherapy.
- Identifying predictive biomarkers for patient response is crucial for optimizing treatment strategies in high-risk PCa.
Purpose of the Study:
- To develop an integrative framework combining machine learning and proteogenomic profiling to identify predictive ultrasound biomarkers.
- To classify patient response to radiolabeled immunotherapy in treatment-resistant, high-risk prostate cancer.
- To establish a foundation for developing novel radiolabeled immunotherapy drugs prior to surgery.
Main Methods:
- Utilized a novel integrative framework with a deep stacked autoencoder (DSAE) and Extreme Gradient Boosting for feature refinement and classification.
- Collected multiomics data (proteomic, transcriptomic, whole-exome sequencing) from The Cancer Genome Atlas and an independent radiotherapy-treated cohort.
- Employed DSAE architecture to maintain biological variety across omics layers while reducing data dimensionality.
Main Results:
- Identified significant relationships between resistance phenotypes and proteogenomic profiles, including DNA repair pathways (BRCA2, ATM), androgen receptor (AR) signaling, and metabolic enzymes (ACLY, IDH1).
- Confirmed a panel of ultrasound biomarkers in preclinical models using patient-derived xenografts.
- Incorporated real-time phenotypic features from ultrasound imaging (perfusion, stiffness) to provide insights into the tumor microenvironment and treatment responsiveness.
Conclusions:
- The developed integrative framework successfully identifies predictive ultrasound biomarkers and classifies patient response to radiolabeled immunotherapy in treatment-resistant prostate cancer.
- Proteogenomic profiling reveals key pathways associated with treatment resistance, offering potential therapeutic targets.
- This approach provides a clinically actionable platform for advancing radiolabeled immunotherapy drug development in prostate cancer.
