According to new research, artificial intelligence can help increase the detection of breast cancer by 20 per cent while almost halving the workload of radiologists.
Breast cancer is the most common cancer in women in Australia. The best early detection tool to date has been the mammogram, a low-dose X-ray that captures the intricacies of breast tissue to detect subtle changes that are often imperceptible during a routine physical examination. But it often eludes detection.
A new study published in The Lancet Oncology, on the other hand, has opened a new chapter in the detection of breast cancer. That’s right; AI has entered the chat, increasing breast cancer accuracy detection rates by 20 per cent more accurately than traditional radiology screenings.
Like what you see? Sign up to our bodyandsoul.com.au newsletter for more stories like this.
In a recent large-scale study in Sweden, researchers examined mammograms of 80,000 women aged 40 to 80 to highlight the potential of AI to detect breast cancer accurately. And the results? AI demonstrated its abilities by identifying 244 cancers through screening and prompting 861 recalls. In comparison, traditional screening methods detected 203 cancers and resulted in 817 recalls.
The AI detection rate was 6 per 1,000 women screened, which met the lowest safety threshold. Traditional screening methods, on the other hand, yielded a detection rate of five per 1,000 women screened. Importantly, the rate of false positives — a critical factor in determining the effectiveness of a screening technique — was 1.5 per cent for both AI and conventional methods.
Beyond the numbers, there is a more profound reassurance for anyone concerned about using AI in health screening. This study confirms that combining AI and mammography is effective and safe.
Dr Kristina Lang, the study’s lead author, told the BBC that while the preliminary results are promising, more research is needed to fully understand the role of AI in mammograms. But added that AI’s most significant potential may be its ability to alleviate radiologists‘ heavy workload.
„The excessive amount of reading“ is a constant strain for these medical professionals, she observes. If AI can shoulder some of this burden, it could be critical in relieving the strain on our healthcare system.
The study’s findings support this claim even more. According to the Radiological Society of North America, as tested in this study, AI technology reduced radiologists‘ screen-reading workload by 44 per cent. This is a significant step forward, not only in terms of technological advancement but also in terms of supporting the healthcare system.