Researchers at NYU Langone Health and the University of Glasgow have developed a smart AI tool that can diagnose lung cancer and predict if it will come back. According to Dr. Aristotelis Tsirigos, the AI was 99% accurate in telling apart two common types of lung cancer and could predict cancer recurrence with 72% accuracy, which is better than the 64% accuracy of pathologists.
The AI studied tissue samples and identified 46 key factors that link to cancer coming back or patient survival. The team hopes this tool can predict cancer outcomes for up to 5 years and be adapted for other types of cancer. More data will be added to improve the tool before it becomes available to the public.
Dr. Tsirigos explained that this AI is different because it learns by itself, without needing human input for training. This makes it more flexible and allows it to be used for different types of cancer. The researchers are focused on lung cancer for now but see potential for broader applications in the future.
Despite the promising results, the AI tool still needs further testing and fine-tuning before it can be trusted in real medical settings. The goal is for the AI to work alongside human experts, offering a second opinion and helping doctors make better decisions.
Source: targetedonc
