Machine Learning Method could better Anticipate Cancer Progression

Computer-based tools have made rapid inroads into the healthcare industry. It is interesting to see how the different technologies are being used to track the progression of diseases. The evolution of tumors in cancer is responsible for how the disease progresses and spread in patient populations world over. Scientists have long been focused on finding ways to track this progression by analyzing patterns in evolutionary trends. Such efforts may also be helpful in preventing the disease altogether. Researchers from several institutes in the U.K. and the U.S. collaborated and developed a machine-learning method to track recurrent successions of genomic changes underlying repeated evolutionary processes.

The artificial intelligence (AI) technology developed is based on an algorithm that they call REVOLVER. They found that it could identify key evolutionary patterns in numerous cancer cohorts, which remain hidden. As many as 768 samples were taken from 178 patients suffering with breast, renal, colorectal cancer, and lung. The research findings are published in the peer-reviewed journal Nature Methods on August 31, 2018.

AI Algorithm could identify Repeated Tumor Evolutionary Patterns in Patient Cohorts

The scientists collected tissue samples from the patients to track tumors evolution in them. Mutation in these tumors proved to be an initial bottleneck in doing so and they sought to overcome this by using machine learning algorithms. The AI system first learned mutational patterns in the tumors responsible for their spread. The system applied the learnings—essentially consisting of multi-region sequencing datasets—to track developing tumors in new patients. They found that the algorithm was successful in identifying gene mutations in single-sample cohorts of 95 patients with colorectal cancer.

The researchers opine that the AI method could help in patient classification based on tumor evolution, which will offer insights on the cancer progression. This is likely to lead to better prognosis of cancer in patients.

Leave a Reply

%d bloggers like this: