US baby maker to evaluate artificial intelligence

Ovation Fertility announces study with Australian AI firm, Life Whisperer

An American IVF service is evaluating artificial intelligence technology to determine if it may help improve pregnancy outcomes for its patients.

Ovation Fertility is working with Australian AI company Life Whisperer on a retrospective study to assess whether artificial intelligence could have identified embryos that would have led to a successful pregnancy.

Tex VerMilyea, Ovation Fertility’s Austin laboratory and operations director said while the organisation currently considers preimplantation genetic screening the gold standard for selecting embryos for transfer, it is not always an option for every patient.

“Our skilled embryologists excel at visually grading embryos for transfer, and we’re interested in evaluating AI’s potential as a supplemental decision support tool for validating our embryologists’ findings on which embryos might be the most likely to result in pregnancy,” he said.

During the trial, deidentified photos of day 5, 6, and 7 embryos will be reviewed by the AI technology – the retrospective analysis will correlate the analysis of each embryo photo with its known pregnancy outcome.

Life Whisperer recently demonstrated through an Australian 600 patient study that its AI tech performed significantly better than experienced embryologists at identifying viable embryos from medical images, a critical success factor for couples undergoing IVF. The company’s tech was 30 per cent more accurate than embryologists participating in the study in identifying embryo viability.

Life Whisperer’s co-founder De Michelle Perugini said the company is using deep learning and computer vision to identify complex morphology of embryos, which cannot be seen through a microscope.

Follow CIO Australia on Twitter and Like us on Facebook… Twitter: @CIO_AustraliaFacebook: CIO Australia, or take part in the CIO conversation on LinkedIn: CIO Australia

Follow Byron Connolly on Twitter: @ByronConnolly