Beware human kind!! Artificial intelligence (AI) has taken another huge step in conquering the medical field!!!
As many of you might have noticed, one of the latest topics of discussion on the newspapers, science and technology websites and social media this October has been about the use of AI in breast cancer detection. What a timely news in conjunction with Breast Cancer Awareness month.
According to the latest (2014) statistics available on Cancer Research UK, breast cancer tops the list of ten most common cancers in female in the UK. Within UK itself, it is estimated that around 150 unfortunate women will be diagnosed with breast cancer every day. And with around 31 deaths per day, breast cancer is the second most common cause of cancer death in women in the UK. While most statistics available out there are mainly on women, it is important to note that men can also become the victims of breast cancer.
Mammography has been known to be prone to false positive results, sending many poor women to go through painful and expensive surgeries each year.
Now, let us turn our attention to AI and its new role in aiding breast cancer detection. Since late 1970s, mammography has been used to screen for and detect breast cancer, and has remained the most efficient tool available to detect breast cancer. Well, at least till before AI decided that it’s time for it to rise and shine.
While many believe in the saying “old is gold”, I guess it is not fully applicable to our good-old mammography. Now, don’t get me wrong. Mammography is still a very crucial tool in breast cancer screening, but with its own flaws and maybe in need of some helps. Mammography has been known to be prone to false positive results, sending many poor women to go through painful and expensive surgeries each year. And what is worse is that 80-90% of the time, the seemingly high-risk breast cancer lesions turned out to be benign.
In order to improve the efficiency of breast cancer screening and prevent unnecessary surgeries, scientists at MIT Computer Science and Artificial Intelligence Lab have partnered up with Massachusetts General Hospital and Harvard Medical School to devise an AI model that could distinguish high-risk breast cancer lesions that are more cancerous and require surgery, from those that don’t, with highly improved accuracy.
Back in August in 2016, another group of scientists from Houston Methodist have also successfully developed an AI model that was able to, with 99% accuracy, interpret mammogram results 30 times quicker than trained medical professionals (seems like AI is doing well in breast cancer field).
AI models mainly work on mammogram results.
Putting aside the fear of AI conquering the human race, one must appreciate that indeed AI plays an important role in minimising patients’ suffering, maximising medical professionals’ time as well as cutting down money spent in cancer care. Nonetheless, mammography is still the best tool available for breast cancer screening as current AI models mainly work on mammogram results. So, it looks like AI is more like a helping nurse.
With AI getting more and more diligent and intelligent, maybe we should start to be more fearful of AI apocalypse, instead of zombie apocalypse, (or both?) in near future. What should we do if AI decided to strike humanity this coming Halloween?! (Gasp!!!)