The test uses an advanced modeling method through machine learning. Analyzes RNA sequences. Genetic data from 945 samples were used to train the model.
As a result, the test accurately determined the presence/absence of early and late stage oral cancer in patients in 90% of cases. The accuracy for squamous cell carcinoma of the oropharynx was 84.2%.
However, there is one nuance. The model only detects cancer-related changes, not the cancer itself. “It’s kind of a causal question. Do these bacteria change because of cancer? Or are there changes in the bacteria in some people who later get cancer?” added the scientists.