IVF: embryos soon to be selected by artificial intelligence



The Australian start-up, Life Whisperer, has devised a method for selecting human embryos prior to implantation, based on deep learning and artificial intelligence. Their aim is to propose a cloud application to fertility clinics to sort IVF embryos before implantation.

 

 Doctors Don and Michelle Perugini have developed this method because they believe that IVF success rates are currently too restricted and that pre-implantation selection depends too heavily on the embryologist's eye. In their opinion, “artificial intelligence could do a better job than human clinicians in selecting viable embryos".

 

 An initial commercial agreement with a fertility clinic has allowed them to access "thousands of embryo images used to train and validate imaging analysis techniques". They have thus developed algorithms capable of identifying "key embryo characteristics" that indicate its "viability": "The system recognises the results recorded with these embryos, identifies new characteristics and classifies them to put forward a relevant selection model" explained Doctor Perugini.

 

 The start-up eventually wants to sell its technology to fertility clinics "in the form of a cloud application" (…) [which would provide an instant diagnosis based on an approved and validated model". The company also is considering "using its technology for other diagnostic applications".


Sources: 

Le Monde informatique, George Nott (20/09/2017)