Artificial Intelligence (AI) may allow “near real-time” analysis of a brain tumor during the operation to remove it, helping surgeons more “safely and accurately,” according to a study released Monday.
The tool, tested by a team of US researchers, can tell in less than two and a half minutes whether the removed cells are cancerous or not, compared with 20 to 30 minutes from a conventional analysis, according to an article published Monday. Nature Medicine journal.
The researchers combined an innovative laser imaging technique with an algorithm based on the analysis of over 2.5 million biopsy images.
"With this tool we are better equipped to preserve healthy tissues and only remove tissues contaminated by cancer cells, which means fewer complications and better outcomes for cancer patients," said neurosurgeon Daniel Orringer of the University of New York, quoted by the AFP news agency.
According to the expert, "In neurosurgery and many other areas of cancer surgery the detection and diagnosis of tumors during the operation is essential to make the most appropriate surgical procedure."
From the sample taken, the program allows us to say what type of tumor is among the 10 most common types of brain cancer, with an efficacy comparable to that of pathologists, according to the study.
In a clinical trial involving 278 brain tumor patients, the AI tool established a correct diagnosis in 94.6% of cases, against 93.9% of human analysis. "Surprisingly, in all cases where the pathologists were mistaken our algorithm was right, and in all cases where the algorithm was mistaken the pathologists were right," noted Orringer.
According to the authors of the research, the tool could improve the relevance of human analysis and overcome the lack of experts in places where they are not enough.