The Curie Institute has created artificial intelligence capable of determining the origin of cancers for which only metastases have been identified.
- The Curie Institute has created a tool based on artificial intelligence which makes it possible to determine the origin of cancers from the metastatic cells found.
- Thanks to this artificial intelligence process, during the study, the origin of cancer cells could be found in more than 80% of cases.
- This artificial intelligence tool, to remain effective, must be associated with clinical examination and analyzes by medical experts.
Some cancers are only discovered at the metastatic stage, that is to say at a stage where other cancer cells have already spread elsewhere in the body (metastases). But it is necessary, to enable appropriate treatment, to know the initial cancer, called primary cancer, and therefore the organ initially affected.
The team of Dr. Sarah Watson, medical oncologist and researcher, and bioinformatics experts from the Curie institute have therefore developed a tool based on artificial intelligence which makes it possible to determine the origin of cancers from the metastatic cells found.
A tool created from RNA sequences from cancer cells
The Curie Institute, in partnership with other French centers, has developed a tool that analyzes the RNA sequences of malignant tumors recovered from samples (biopsies) of cells from metastases.
Thus, after recording more than 20,000 different samples, artificial intelligence was finally able to specify the origin of these cancer cells. But this study was only done on a small series of patients. It therefore seems necessary, according to Dr. Watson to Bfmtvto perfect it.
This tool thus makes it possible to recognize, according to gene expression profiles (a sort of cancer identity card), the origin of cancer cells with much greater speed and performance than “human” specialists.
As indicated to Bfmtv Dr. Sarah Watson: “the tool was able to indicate the origin of the cancerous tissue in more than 80% of cases“.
More specific and therefore more effective chemotherapy protocols
Cancers whose origin is not known are difficult to treat for two reasons: because the starting cancerous site is unknown but also because the stage of discovery is late (since the diagnosis is made at the metastatic stage).
And as each cancer is treated differently depending on its origin, chemotherapy protocols are specific to each type of cancer. And if the original cancer is not determined, chemotherapy is less precise and, therefore, much less effective.
This knowledge of the origin of primary cancer from cancer cells found anywhere in the body is therefore essential to implement appropriate and effective treatment. And thanks to this tool, the patient’s chances of survival are tripled.
Artificial intelligence must remain coupled with human intelligence
Dr Sarah Watson makes it clear that “the tool does not work alone, it integrates with a set of data analyzed with a meeting of experts to make the diagnosis as precise as possible.” Artificial intelligence remains a very useful tool to aid in the diagnosis and management of cancers. The research team began, after the publication of this study in the Journal of Molecular Diagnostics, to disseminate this tool to other cancer care centers.
Check out our digital video work,
only an account creation is required to access it.
