PANCDR: precise medicine prediction using an adversarial network for cancer drug response.

in Briefings in bioinformatics by Juyeon Kim, Sung-Hye Park, Hyunju Lee

TLDR

  • The study is about using a computer program to predict how well a person will respond to a cancer drug based on their genetic makeup. The program is called PANCDR and it uses both preclinical and clinical data to make predictions. The study found that PANCDR outperformed other computer programs in predicting how well people responded to cancer drugs. This means that PANCDR could be used to recommend specific drugs to people based on their genetic makeup, which could improve their chances of responding to treatment. However, the study also found that more clinical data is needed to train and validate the program.

Abstract

Pharmacogenomics aims to provide personalized therapy to patients based on their genetic variability. However, accurate prediction of cancer drug response (CDR) is challenging due to genetic heterogeneity. Since clinical data are limited, most studies predicting drug response use preclinical data to train models. However, such models might not be generalizable to external clinical data due to differences between the preclinical and clinical datasets. In this study, a Precision Medicine Prediction using an Adversarial Network for Cancer Drug Response (PANCDR) model is proposed. PANCDR consists of two sub-models, an adversarial model and a CDR prediction model. The adversarial model reduces the gap between the preclinical and clinical datasets, while the CDR prediction model extracts features and predicts responses. PANCDR was trained using both preclinical data and unlabeled clinical data. Subsequently, it was tested on external clinical data, including The Cancer Genome Atlas and brain tumor patients. PANCDR outperformed other machine learning models in predicting external test data. Our results demonstrate the robustness of PANCDR and its potential in precision medicine by recommending patient-specific drug candidates. The PANCDR codes and data are available at https://github.com/DMCB-GIST/PANCDR.

Overview

  • The study aims to develop a Precision Medicine Prediction using an Adversarial Network for Cancer Drug Response (PANCDR) model to predict cancer drug response (CDR) based on genetic variability. The hypothesis being tested is that PANCDR can accurately predict CDR in external clinical data using both preclinical data and unlabeled clinical data. The methodology used for the experiment includes training the PANCDR model on preclinical and clinical data and testing it on external clinical data. The primary objective of the study is to demonstrate the robustness of PANCDR and its potential in precision medicine by recommending patient-specific drug candidates.

Comparative Analysis & Findings

  • The study compared the outcomes observed under different experimental conditions or interventions, specifically the performance of the PANCDR model in predicting CDR in external clinical data. The results showed that PANCDR outperformed other machine learning models in predicting external test data. The key findings of the study demonstrate the robustness of PANCDR and its potential in precision medicine by recommending patient-specific drug candidates.

Implications and Future Directions

  • The study's findings have significant implications for the field of research and clinical practice, as they demonstrate the potential of PANCDR in precision medicine. However, the study also identified limitations, such as the need for more clinical data to train and validate the model. Future research directions could include collecting more clinical data, exploring other machine learning models, and integrating PANCDR with other precision medicine tools.
Read Full Article