How Artificial Intelligence Helps Fight Breast Cancer and Save Millions of Lives
Every October for the last 40 years, Breast Cancer Awareness Month has helped draw attention to the most common cancer on Earth, which kills nearly 750,000 people every year.
Despite recorded cases dating back to Ancient Egypt, breast cancer was considered an "unbearable" disease for thousands of years. Women were expected to suffer in silence and "with dignity".
This stigma fueled academic ignorance, since just a few decades ago breast cancer remained a relatively unexplored disease. For most of the last century, women with breast cancer were offered radiation therapy and/or surgery, often radical surgical procedures that disfigured them without much benefit. Meanwhile, treatment for other cancers was advancing.
Breast cancer mortality barely changed from the 1930s through the 1970s, until joint efforts by feminist and women's liberation groups brought the study and treatment of breast cancer into their rightful place in hospitals and research institutions dominated by men. Treatment changed within a single generation.
In the 1970s, the probability that a woman diagnosed with breast cancer would survive the next 10 years was about 40%. Today that probability has nearly doubled, thanks to new drugs, advanced screening methods and more precise and effective surgical procedures. A key factor in this transformation was the emphasis on early detection. The earlier breast cancer is found, the easier it is to treat.
Artificial intelligence is playing an increasingly important role in detecting breast cancer. Although the technology is designed to augment, not replace, human doctors, it can at least help ease the shortage of radiologists, as another 2,000 people are needed to clear the backlog of scans in the NHS (National Health Service of England) amid the coronavirus pandemic.
Startups are also using artificial intelligence to tackle this problem. The British company Kheiron Medical Technologies plans to use AI to detect breast cancer in 500,000 women. The Spanish company Blue Box is developing a device that can detect breast cancer from urine samples. The Indian company Niramai is working on an inexpensive tool that can help screen large numbers of women in rural and semi-urban areas.
But equally important for improving outcomes is identifying patients at high risk of recurrence. Roughly one in ten breast cancer patients experiences a recurrence after initial treatment, which reduces their chances of survival.
“Historically, identifying them at an early stage has been difficult, but my team, working with the French Gustave Roussy Cancer Institute, has developed an AI tool that can identify 8 out of 10 patients at high risk of recurrence. AI helps patients get the necessary treatment earlier and spares patients in the low-risk group from frequent anxiety-inducing checkups. Meanwhile, pharmaceutical companies are speeding up breast cancer drug trials by enrolling high-risk patients faster”
Patient data privacy can be an understandable obstacle to rapid research. Hospitals are cautious about sending data outside their institutions, and no pharmaceutical company wants to share valuable data with competitors. But AI helps solve these problems by allowing new treatments to be developed faster, more safely and more cheaply.
Federated learning, a new form of artificial intelligence that trains on data from multiple institutions without that data leaving hospitals, is being used across Europe to give researchers access to important but previously inaccessible data. Clozel said his team will also use AI to deepen their understanding of why the most aggressive forms of breast cancer resist certain drugs. This will help scientists develop new targeted drugs that distinguish healthy cells from tumor cells better than chemotherapy does.
As the influence of artificial intelligence grows, equally important for improving outcomes is recognizing that healthcare is first and foremost a human endeavor. No algorithm will ever be able to comfort a patient in the darkest moments of their life, and no machine can instill the calm and inspire the resilience that every patient needs to beat their disease.
“I and every other doctor know that treating a disease means understanding not only the patient's illness but also the patient. Empathy is linked to higher patient satisfaction and lower levels of distress (a state of suffering), which motivates patients to continue a difficult course of treatment. Fortunately, AI technology, which increasingly helps in treating breast cancer, is designed to expand the capabilities of doctors”
Breast cancer is gradually ceasing to be "unbearable" for the millions who receive this diagnosis every year. The sea of pink ribbons that marks the start of October signals how far the world has come in the fight against one of its oldest enemies. We will never be able to completely eradicate breast cancer. However, as AI helps diagnose patients earlier and enables rapid development of treatments, it is quite possible that in a few decades we will no longer need Breast Cancer Awareness Month.
