AI model can help improve immunotherapy efficacy.
A new AI tool can detect mutations in cancerous cells more simply and potentially less expensively while being “highly predictive” of whether immunotherapy will work for cancer patients.
Researchers at Technion-Israel Institute of Technology published a new paper that studied the use of RNA – ribonucleic acid – to measure how much of a cell has mutated, a metric known as Tumor Mutation Burden (TMB).
Immunotherapy stimulates a body’s own immune system to fight cancer cells, and TMB can help identify which cancers will respond best to this treatment. The researchers sequenced RNA, rather than the more complex DNA, and found the method to have “high precision,” the study said.
“We show that RNA-based TMB is significantly associated with patient survival, showing similar to higher significance level as compared to DNA-based TMB,” wrote authors Rotem Katzir, Noam Rudberg and Keren Yizhak.
Moreover, they did not use a patient’s “matched-normal” sample of healthy cells against which to compare a mutated cell.
Instead, they trained their AI on a robust database that included sequenced RNA from cancer patients, according to the paper, “Estimating tumor mutational burden from RNA-sequencing without a matched-normal sample,” in Nature Communications.
How immunotherapy can improve
Immunotherapy stimulates the patient’s immune system, including the white blood cells from the patient’s blood, tissues of the lymph system and organs, to fight cancer cells. It can also decrease the side effects that often occur with chemotherapy.
A body’s immune system can hunt and attack cancer cells. When a tumor has mutations, the body can distinguish it from healthy cells. When a patient has a high TMB, the tumor has more mutations, thereby making it easier for immunotherapy to work.
Evaluating RNA can be more efficient than looking at DNA because it requires a smaller amount of genetic material, eliminating one procedure for cancer patients.
RNA molecules contain small segments of the genetic codes that are copied and used as instructions within the cells. The RNA method was found to be a better predictor than traditional methods. They theorized that the RNA includes parts of the genome that are constantly used, which might be more effective in triggering an immune response.