AI Tool Successfully Identifies Breast Cancer Missed by Human Doctors
An AI tool called Mia has been tested by an NHS hospital trust and has successfully identified signs of breast cancer in 11 women that had been missed by human doctors. The tool, piloted alongside NHS clinicians, analyzed the mammograms of over 10,000 women and flagged all those with symptoms, including the additional 11 cases that were not identified by doctors. This is a significant development in early cancer detection, as cancers can be extremely small and hard to spot at their earliest stages.
The AI tool, Mia, was able to detect tumors that were practically invisible to the human eye during a demonstration at NHS Grampian. Barbara, one of the patients whose cancer was flagged by Mia, expressed her gratitude for the less invasive treatment she received due to the early detection. Her tumor was caught early, allowing her to undergo an operation and only require five days of radiotherapy. Studies have shown that breast cancer patients with tumors smaller than 15mm at the time of detection have a 90% survival rate over the following five years.
Mia has the potential to significantly reduce waiting times for results, with claims from its developer, Kheiron, suggesting that it could provide results within three days instead of the usual 14. Currently, two radiologists review each scan, but it is hoped that one of them could be replaced by the AI tool in the future, reducing the workload for each pair.
AI tools have shown promise in spotting symptoms of specific diseases with the right training and data. However, acquiring the necessary data for training can be challenging due to patient confidentiality and privacy concerns. The development of Mia took six years and was trained on “millions” of mammograms from women all over the world.
Dr. Gerald Lip, clinical director of breast screening in the North West of Scotland, believes that tools like Mia could free up radiologists to spend more time with patients. He sees a future where AI is seen as an augmentation to medical practices rather than a replacement for doctors.
While the results from the Mia trial are encouraging, it is important to note that the research has not been peer-reviewed, and the evaluation was conducted in just one location. However, the Royal College of Radiologists acknowledges the potential of AI and believes that it will play a crucial role in patient care when used in conjunction with clinical radiologists.
Looking ahead, the increasing number of cancer cases diagnosed each year and the pressure on healthcare systems call for technological innovations like AI to improve outcomes and reduce the strain on medical staff. Further research will be essential in finding the most effective ways to utilize AI technology to benefit cancer patients.
There are other ongoing AI trials in the healthcare sector, such as the analysis of blood samples to detect signs of sepsis before symptoms emerge. These trials, although in early stages, show promise for the future of healthcare and could lead to significant advancements in disease detection and treatment.
The potential impact of AI in healthcare is immense, and it is vital that inclusivity and patient privacy are prioritized in the development and deployment of these technologies. As AI continues to evolve, its role as a valuable tool in improving patient outcomes will become increasingly apparent.