China’s new blood test achieves 90% accuracy in early gastric cancer detection

Early detection with 90% accuracy

Chinese scientists have pioneered a novel method for detecting cell-free DNA (cFDNA), which originates from both healthy and tumor cells. This cfDNA circulates in the bloodstream, primarily released through cell death and active secretion mechanisms.

The study highlights the grim reality that over 80% of gastric cancer cases in China are diagnosed at an advanced stage. Consequently, the five-year survival rate remains below 30%. However, the researchers emphasize that early detection can dramatically improve outcomes, with survival rates exceeding 90% in some cases.

Currently, diagnosis heavily relies on endoscopies and serum markers. Endoscopies, while effective, can be uncomfortable for patients and carry potential complications. Serum markers, on the other hand, have a detection rate of only 20%.

The researchers identified these loopholes in the health process. Subsequently, they sought to develop a more accurate and patient-friendly screening method, aiming for a 90% accuracy rate.

AI-powered screening method shows high accuracy

Researchers Wang Hongzhi and Nie Jinfu from the Hefei Cancer Hospital, the Chinese Academy of Sciences, and Sun Ruifang from the Shanxi Provincial Cancer Hospital led this research. They harnessed the power of AI to create a non-invasive, pain-free screening method.

According to the researchers, this method eliminates the need for intubation or imaging, offering a more comfortable and pain-free experience for patients. The research team employed AI algorithms that were in tune with the cfMeDIP (cell-free methylated DNA immunoprecipitation) technique. This technique enables the detection of minute quantities of DNA. It facilitates the development of a detection model that identifies 21 methylation biomarkers specific to gastric cancer.

A study involving 150 gastric cancer patients and 100 healthy controls demonstrated the effectiveness of this method, achieving sensitivities of 88.38% and specificities of 94.23%.

Interestingly, the study did not observe significant differences in cfDNA concentrations across various age groups. The team further validated their models using The Cancer Genome Atlas public dataset. This allowed the model to differentiate between tumor and normal samples based on the gene expressions and methylation of 21 genes, proving effective across multiple tumor types.

According to the researchers, this study is pretty important in getting to know tumor biology and could also help contribute to developing diagnostic and prognostic tools for gastric cancer.