Faced with a cancer that is difficult to treat, the Institut Curie turned to artificial intelligence. This choice enabled them to save the life of a 30-year-old young man.
Each year, we identify 20 million new people cancer patients worldwide. France, meanwhile, has more than 380,000 per year. In fact, cancer is one of the most important causes of premature deaths. Therefore, several scientists have sought to improve the understanding and treatment of this disease. The team of Sarah Watson, oncologist and researcher at the Institut Curie, is one of them. To this end, she has developed a artificial intelligence to identify the origin of cancer.
AI saves the life of a condemned patient
During his interview with The world, Sarah Watson counted the story of a former patient. In January 2020, a 30 year old man is entrusted to the biology and oncology laboratory of the Institut Curie. He presented “metastasis everywhere” et an unknown primitive cancer (CPI), the first affected organ is unknown. For simplicity, the prognosis was reserved.
This kind of cancer is not identified until it has spread to other tissues. In France, cases of unknown primary represent 7,000 patients per year. Unfortunately, doctors have very difficult to treat this kind of diseases. As a result, patients are often treated with ineffective broad-spectrum chemotherapy. This method offers less than 20% chance of surviving for a year.
To come back to the young man, he benefited from the technology developed by the Institut Curie. This is an artificial intelligence specially trained to identify the origin of cancer. So, AI identified he had kidney cancer, a clear cell carcinoma. In response, doctors applied anti-VEGF and specific immunotherapy. In the last news, the patient is in remission.
SEE ALSO: When AI improves hospital patient satisfaction
Artificial intelligence and cancer
Pour training their computer program, researchers at the Institut Curie have used a certain type of data. These come from the RNA sequencing (RNA-Seq). This technology has been used to sequence thousands of genes expressed in cancer cells.
“We said to ourselves: if we manage to teach a computer to distinguish, at the level of RNA, a tumor of the kidney, the colon or the breast, perhaps this tool will be able, if we submit to it the cancer of unknown origin, to find where it came from? »
Sarah Watson
Since the creation of this algorithm, Sarah Watson’s team has analyzed RNA-Seq data from 48 CPI. Thanks to him, she was able identify the origin of 80% of cancers. Let’s take the example of eleven newly diagnosed patients. Nine of them were able to benefit from care with 89% disease control rate. A huge difference compared to broad-spectrum chemotherapy. The results are prompting the Institut Curie to extend the adoption of this artificial intelligence to a large number of healthcare establishments.