Breast cancer is the second leading cause of death from cancer in women. According to the report from Breast Cancer Research UK, Breast cancer kills roughly 11,400 people in the UK every year. But, the results have been shown to improve if spotted and treated early.
In the latest development, Google’s DeepMind, in association with Cancer Research UK Imperial Centre, Northwestern University, and Royal Surrey County Hospital, has developed a model that can spot cancer in breast screening mammograms at the earlier stage to improve health outcomes.
Titled as “International evaluation of an AI system for breast cancer screening,” the initial findings were published on journal Nature, authored by 31 scholars from Google’s Google Health unit and the Imperial College of London.
Initially, the model was trained on de-identified mammograms from 76,000 British women, using Cancer Research UK’s OPTIMAM dataset, as well as 15,000 scans from the US. Post-training, the algorithm was tested on 3000 scans in the US, and a further 25,000 in the UK.
Christopher Kelly, clinician-scientist at Google Health, and co-author of the research, which has been published in the journal Nature, said:
Our ground truth was based on biopsy results and follow-ups, so if they had a normal screen two or three years later.
As per the researchers, the AI model could predict breast cancer with the same level of accuracy as a radiology expert. In comparison with human’s output, the system witnessed a reduction in false positives – 5.7% (US) and 1.2% (UK), and in false negatives of 9.4% (US) and 2.7% (UK).
Dr. Dominic King, the health lead for Google DeepMind, said:
Our team is proud of these research findings, which suggest that we are on our way to developing a tool that can help clinicians spot breast cancer with greater accuracy. Further testing, clinical validation, and regulatory approvals are required before this could start making a difference for patients, but we’re committed to working with our partners towards this goal.
When it comes to the standard approach, it involves two assessments from a specialist, known as double-reading in the UK, in a bid to reduce the acknowledged risk of incorrect interpretations. When there is disagreement, a third will make a judgment. It’s worth mentioning that AI computer was found to be as effective as two human doctors worked together when finding the correct diagnosis.
Google Health’s Shravya Shetty and Daniel Tse note that “inaccuracies can lead to delays in detection and treatment, unnecessary stress for patients, and a higher workload for radiologists who are already in short supply.”
With this model, DeepMind hopes to use technologies like AI will spot cancer early and to ease the burden on clinicians. While the research is still in the early stages, this AI technology could be used as a second reader to give radiology experts an automated second opinion. Google Deepmind has already developed AI solutions for identifying eye disease and detecting neck cancer.
Main image credits: DeepMind
Stay tuned to Silicon Canals for more European technology news.