Michelle’s fertility struggle led to a startup that may revolutionise IVF
“I was on a whole range of fertility treatments and was booked in for IVF but in the end I got pregnant just before my appointment,” she says. “What people don’t realise is that even getting to the point you need IVF is very stressful for people who are trying to have children.”
Life Whisperer uses artificial intelligence to assess embryos for their likelihood of success in IVF and is one of a number of startups in the health sector harnessing the power of machine learning.
Pregnancy success
Michelle Perugini's startup Life Whisperer develops software-based scaleable AI technology.
CREDIT: DOMINIC LORRIMER
The Adelaide-based startup recently raised $4.5 million in funding for its product that assesses two dimensional or microscope images of embryos and provides a confidence score on how likely a pregnancy will result.
The “embryo ranking system” was built by training the artificial intelligence using thousands of images where it was known whether the pregnancy was successful or not.
Life Whisperer provides an alternative to visual assessment by embryologists, which has the potential for human error, and genetic screening through cutting away part of the embryo.
“We think it can have a huge impact on the fertility sector particularly from a patient’s perspective,” Perugini says. “Selecting the healthiest embryos first time will reduce the number of rounds patients need to go through. For the clinics it is likely to improve their success rates. From a cost perspective for the patient, if they are doing fewer IVF cycles, obviously it is a massive cost reduction for them.”
In international clinical trials conducted across 12 clinics in four countries, including Australia, Perugini says Life Whisperer improved accuracy by 25 per cent compared to world-leading embryologists.
“The actual process of IVF itself is very emotional and very financially taxing,” Perugini says. “Our whole concept around the products we develop is software-based scaleable AI technology we can deliver globally at a low cost.”
The startup is being utilised by Ovation Fertility in the United States and Ovation’s laboratory and operations director Tex VerMilyea says Life Whisperer removes the subjectivity of embryo viability.
“Life Whisperer has the potential to revolutionise the field of IVF by making embryo selection a non-invasive, dependable and affordable activity,” VerMilyea says. “AI technology has already caused a paradigm shift in the medical field and I am very pleased to see it emerge and take hold in the world of assisted reproductive technology.”
Skin cancer detection
Another startup harnessing artificial intelligence is skin cancer detection startup MetaOptima.
MetaOptima’s DermEngine uses an image recognition algorithm that has been trained on thousands of images of already-identified skin cancers.
Co-founder Maryam Sadeghi says DermEngine has begun trials in Australia and the United States.
“So many industries are using AI and healthcare is always the last to adopt but AI can provide so many solutions in this area, it’s really untapped,” she says.
“The technology is proven. We all know the potential is all there; we all know it is going to help, now it is all about implementation.”
Sadeghi says the key issues in implementation are ethics and bias.
“It is really interesting when you see in real-life settings how you deal with these problems, now that we have technology and machines to help us, we need to make sure they are in good hands,” she says.
“One simple thing is patient’s choice – one person may be comfortable with having assistance in a diagnosis, another might not.”
The Canadian-based startup has received $US6.5 million ($8.6 million) investment from Australian venture capital fund Airtree and Scott Farquhar and Kim Jackson’s investment fund Skip Capital.
AirTree co-founder Craig Blair says Australia has the ability to lead the world in several areas applying artificial intelligence to the health sector.
“Australia is in an unique position in many ways, we have really large data sets, some of the largest in the world that are labelled and ripe for machine-learning application,” he says.
“Examples of where we have seen that is melanomas, lung cancer detection, IVF embryo selection, even to gut biomes. We can create exciting world-class companies like the Cochlear and ResMed of tomorrow.”
Source: Cara Waters, The Sydney Morning Herald