breast cancer detection

AI used to detect breast cancer in the US

We’ve heard a lot about the dangers of artificial intelligence and the negative impact it could have on our lives, but research from the US demonstrates that it could have seriously positive effects when it comes to breast cancer detection.

The research, conducted by scientists at Harvard Medical School, the Massachusetts Computer Science and Artificial Intelligence Lab, and Massachusetts General Hospital, deployed AI through a machine learning system to predict whether breast lesions identified from a biopsy will turn out to cancerous.

The machine learning system was programmed to analyse information about breast lesions, forcing it to look for patterns among a range of data points, such as demographics, family history, biopsies and pathology reports. So far, it’s been tested on 335 high-risk lesions – and it correctly diagnosed 97% of them as malignant, which reduced the number of unnecessary surgeries by more than 30%.

False positives create fiscal negatives

Fifty-thousand women are diagnosed with breast cancer in the UK each year, but when cancers are found early enough they can often be cured. Mammograms can play a crucial role in detecting cancers early on, but the downside is that they also throw up false positives, such as ‘high-risk’ lesions that appear suspicious on mammograms and have abnormal cells when tested by needle biopsy.

Potentially, patients could undergo painful, expensive, scar-inducing surgeries to have lesions removed, even though they turn out to be benign 90% of the time.

“Because diagnostic tools are so inexact, there is an understandable tendency for doctors to over-screen for breast cancer,” claims Regina Barzilay, MIT’s Delta Electronics Professor of Electrical Engineering and Computer Science – who also happens to be a breast cancer survivor. “When there’s this much uncertainty in data, machine learning is exactly the tool that we need to improve detection and prevent over-treatment.”

Could AI breast cancer detection work over here?

According to Constance Lehman, professor at Harvard Medical School and chief of the Breast Imaging Division at MGH’s Department of Radiology, the results from the study have been very encouraging. “To our knowledge, this is the first study to apply machine learning to the task of distinguishing high-risk lesions that need surgery from those that don’t. We believe this could support women to make more informed decisions about their treatment, and that we could provide more targeted approaches to health care in general.”

However, it needs to be pointed out that this new technology may not travel well over the Atlantic. Debashis Ghosh – a consultant breast surgeon based at the Royal Free London hospital – stated that while the benefits of the technology were obvious, it may not be as effective over here. “Here we have less than 5% of patients who have these surgeries, whereas it is 30% in the US. We try to make a definite diagnosis before we operate but this technology is definitely useful where there is a lack of expertise.”

Obviously, we’re at the very beginning of the learning curve when it comes to AI as a diagnostic tool, so we await further developments with interest.