Researchers at the A*STAR Genome Institute of Singapore (A*STAR GIS) have developed a new way of tracking circulating tumor DNA (ctDNA) in small blood samples using artificial intelligence (AI). The team reports that this new AI method, called “Fragle,” can help physicians track how well a patient is responding to cancer treatment more accurately, frequently, and at a lower cost than existing methods.
“Just as scientists tracked COVID-19 outbreaks by detecting viral particles in wastewater, Fragle analyzes DNA fragments in blood to monitor cancer treatment response and detect relapse early,” explained Anders Skanderup, PhD, Senior Principal Scientist at A*STAR GIS Laboratory of Computational Cancer Genomics and the lead author of the study.
“While existing methods have their strengths, they are often complex and expensive. We wanted to develop a simpler, more affordable, and accessible approach—one that could support accurate monitoring without adding burden to clinical workflows.”
The researchers published their findings in the journal Nature Biomedical Engineering with the title “A Deep-Learning Model for Quantifying Circulating Tumor DNA From the Density Distribution of DNA-Fragment Lengths.”
Using AI, the new model can analyze blood samples and compare the size of DNA fragments within a sample. As ctDNA usually has different sizes than healthy DNA, the AI model can quantify the ctDNA fast and reliably, and verify if and how well a patient is responding to cancer treatment.
The research team used whole-genome sequencing data from different cancers and healthy controls to validate their AI model. “In independent cohorts,” the researchers wrote in their paper, “Fragle outperformed tumor-naive methods, achieving higher accuracy and lower detection limits.”
Requiring only a small amount of DNA, Fragle can be used to monitor cancer through blood tests, costing approximately $39 (SGD $50), compared to around $780 (SGD $1000) for conventional cancer detection tests.
The new AI method can also be used alongside common and widely used DNA profiling methods in hospitals and commercial laboratories, “allowing for easy and fast adoption into existing workflows.”
This speed, versatility, and low cost mean that patients could be monitored more frequently during their cancer treatment, allowing physicians to understand if a therapy is working or not. Moreover, after surgery and/or treatment, physicians could use the new AI method to detect early signs of relapse, because Fragle is able to identify even the tiniest traces of cancer left in the blood stream.
Together with the National Cancer Centre Singapore, the A*STAR GIS researchers are currently monitoring more than 100 clinical trial patients using Fragle every two months to analyze the ctDNA levels in their blood while they undergo cancer treatment.
The research teams are also observing how changes in ctDNA may help physicians identify patients that are more or less likely to respond well to cancer therapy.
In the near future, A*STAR GIS researchers together with the researchers at the National Cancer Centre Singapore want to initiate further studies and, ultimately, introduce their new AI method for cancer screening into hospitals to improve patient care.
“We are very excited about the potential Fragle brings, to help our healthcare professionals detect and track cancer more accurately and monitor treatments more effectively, leading to better cancer care for patients,” said Wan Yue, PhD, Executive Director at A*STAR GIS.
“It is our hope that our genomic research can be translated to benefit population health not only in Singapore, but worldwide.”
