Japanese researchers have developed a technology that combines infrared imaging and artificial intelligence to detect tumors present under the mucous membranes.
- Some cancerous tumors located in the mucous membranes can be difficult to detect
- Some cancerous tumors located on the mucous membranes can be difficult to detect
- The combination of imaging and artificial intelligence makes it possible to identify those that escape biopsy
- This technique can improve cancer screening
Early detection of tumors is one of the success factors in cancer treatment. However, they can be difficult to identify: this is the case of gastrointestinal stromal tumors. The latter develop under the mucous membrane that covers the digestive walls. Even a biopsy does not guarantee their identification. Researchers from Tokyo University of Science in Japan present a new screening technique, called NIR-HSI, in the journal Nature’s Scientific Reports.
A combination of artificial intelligence and infrared imaging
“This technique is a bit like X-rays, says Dr. Hiroshi Takemura, director of the study. The idea is to use electromagnetic radiation passing through the body to generate visualizations of different organic structures..” The new method combines artificial intelligence and infrared images. To develop it, the researchers used the tissues of 12 patients who had undergone an operation to remove their gastrointestinal stromal tumor. The analysis of these organic materials made it possible to “train the artificial intelligence to recognize tumor tissue. While 10 out of 12 tumors were partially or completely covered with mucosa, the artificial intelligence was able to identify them with 86% accuracy.”This is a very exciting projectadds Dr. Takemura. Being able to diagnose different types of gastrointestinal stromal tumors quickly, accurately and with a non-invasive method, without biopsy, a technique that requires surgery, is much easier for the patient and the doctor.“According to the scientist, the emitted rays are not dangerous for the patients.
A first step, before other scientific work
With his team, the researcher is working on improving the screening technique. “We have already built a tool that attaches a NIR-HSI camera to the end of an endoscope, and we hope to be able to do such an analysis directly on a patient soon, rather than doing it on surgically removed tissue.“, indicates the researcher. “This study is the first step towards even more groundbreaking research.“, he congratulates himself. It could greatly facilitate the detection of hidden tumours.
.
