With over 200 types of cancer, each uniquely challenging, the development of precision oncology treatments is a complex endeavor. Traditional methods have focused on genetic sequencing to identify mutations in cancer driver genes and match them with potential treatments. However, many patients do not benefit from these targeted therapies.
A new study, published on April 18, 2024, in Nature Cancer, introduces PERCEPTION (PERsonalized Single-Cell Expression-Based Planning for Treatments in Oncology), an AI-based computational pipeline. Developed by Sanju Sinha, PhD, and colleagues, including senior authors Eytan Ruppin, MD, PhD, and Alejandro Schaffer, PhD, at the National Cancer Institute, PERCEPTION leverages single-cell omics to predict patient responses to cancer drugs.
PERCEPTION utilizes transcriptomics to delve into the transcription factors and messenger RNA molecules expressed by genes, offering a deeper understanding of the tumor's clonal architecture and the emergence of resistance. Sinha highlights the potential of PERCEPTION to adapt treatment strategies in response to the evolution of cancer cells.
The development of PERCEPTION faced challenges due to the limited availability of single-cell data from clinics. To overcome this, the team employed transfer learning, pre-training models with published bulk-gene expression from tumors and fine-tuning them with limited single-cell data from cell lines and patients.
PERCEPTION's effectiveness was validated through its ability to correctly stratify patients into responder and non-responder categories in clinical trials for multiple myeloma, breast, and lung cancer. Notably, it captured the development of drug resistance in lung cancer as the disease progressed.
While PERCEPTION is not yet ready for clinical use, its development marks a significant step towards creating a clinical tool that can predict individual cancer patients' treatment responses in a systematic, data-driven manner. Sinha hopes that these findings will encourage the adoption of this technology in clinics to generate more data, further refining the technology for clinical application.
The research was supported by the Intramural Research Program of the NIH, NCI, and several NIH grants, highlighting the collaborative effort behind this innovative approach to cancer treatment.
