Artificial intelligence to add breathing space into health sector

When a patient is diagnosed with breast cancer, a breast biopsy is taken and the cells checked for abnormalities, known as “high-risk breast lesions”, which could turn cancerous later down the line. Medical professionals must then decide whether the lesions should be removed, or if the patient should be put under a period of observation.

If the lesions destined to develop into cancer could be identified, the number of surgeries could be reduced minimising trauma to the patient and allowing medical professional time, and health sector funding, to be used more effectively.

This is where artificial intelligence could step in.

In the Massachusetts research, an AI model was trained to identify which lesions would develop into cancer and which would remain benign, or non-cancerous.

Over a thousand women were involved in the study. 96% had the lesions surgically removed and the rest received 2 years of follow-up imaging tests.

Excitingly, the model correctly picked 97% of the lesions-turned-cancerous so could have spared a significant number of women from going through unnecessary surgeries.

Going forward, the researchers want to see if they can train the AI to identify lesions from mammography images, perhaps removing the need for a biopsy altogether.

The application of AI in medicine is wide-reaching. It could be the solution to many common issues plaguing the medical sector such as transport, overworked staff and inadequate funding.

Transport and parking difficulties have been identified by breast cancer patients as a major barrier to receiving the healthcare and support they need. AI-capable wearables (technology worn on the body) could monitor high-risk individuals (whether that’s those with high-risk lesions, or those with the BRCA gene) for any changes and, upon detection, present the best next step without patients having to physically visit their oncologist.

Those newly diagnosed around Christmas are often put in limbo as clinics shutdown so staff can recuperate. AI provides an avenue around that – robots don’t need to eat, sleep or take leave – it could prove to be a 24/7 diagnosis and treatment pathway service accessible to all.

Inadequate medical funding is a constant struggle, as is lack of resources and subsequently overworked staff. If an AI could be trained to diagnosis and present treatment pathways, the “pathway to diagnosis” will be cut, significantly reducing the financial burden on the national health system and making high quality healthcare accessible to all.

Medical care is advancing every day but people can make mistakes and, while AI doesn’t guarantee a 100% faultless service, it can digest data in volumes outside of human capability. Millions of medical reports, patient records, clinical trials and medical journals could be accessed by an AI to inform their decisions and suggestions. It’s your oncologist on hyper drive.

Already a dynamic area in medicine, even here in Aotearoa (innovation giant, IBM, which envisages the future as “man to the power of machine” brought an AI called Watson here through a partnership with Counties Manukau DHB) AI could streamline the medical sphere hugely in the years to come, benefitting both patients and medical professionals, alike.