Google’s algorithm with AI defines cancer rather doctors
Google provided additional evidence that AI can be a valuable source of information in the procedure of cancer.
Company researchers have developed a tool for deep learning that can identify the metastatic cancer cells of the breast more accurately than pathologists, studying the pictures. The team prepared its own algorithm (assistant for the study of lymph nodes – LYNA), which can recognize features of tumors using two sets of pathological slides, thanks it detects metastases in a variety of conditions. The result was a system of artificial intelligence that can distinguish slides with cancerous tumors in 99% of cases. Even when searching for relatively small metastases system works rather than people.
System LYNA proved to be more effective when performed as a companion of the pathologist. When performing the simulation of the diagnoses, it was discovered that the technology of deep learning made easier the work of doctors. This not only reduced the number of missed metastases, but also reduced the scan time tests.
This direct approach will be effective in the study of cancer, since he will be involved in cases with late stage of breast cancer where no known solutions. The algorithm used in the real clinics. As the researchers point out that metastases are a common factor in most forms of cancer, to tailor the system to search other cancers don’t need a lot of time.
In conclusion, it is worth noting that the system developed by Google, provides patients with a very accurate diagnoses and can serve as a huge step for mankind to free medicine.