About us & our product
ImVitro’s SaaS platform EMBRYOLY is an AI-powered software. EMBRYOLY’s core feature provides a ranking based on the embryo’s morphokinetics as well as a subsequent transfer priority recommendation personalized to the patient for increased accuracy. See our product page for more information!
Embryo evaluation is a crucial but difficult and time-consuming task for embryologists, which AI has the potential to streamline. AI tools can automatically extract known biomarkers in embryo development, but also rank embryos based on visual information that might go unnoticed by embryologists to render the evaluation as objective as possible. Since embryologists remain the final decision maker and will always want to look at the embryos, it remains unclear how much time AI can save them, let alone whether it can save them time while also guaranteeing similar or improved performances.
Question to answer
- Can EMBRYOLY help embryologists save embryo evaluation time whilst guaranteeing comparable or improved standard of care?
The study at a glance
10 patients were randomly chosen amongst patients with cohorts of more than 5 embryos which included at least 2 blastocysts, and one embryo with known pregnancy outcome. On average, complete cohorts contained 6.7±2.4 embryos and 50±25% blastocysts. The data was collected through a MIRI® (ESCO) time-lapse system and uploaded on EMBRYOLY from an IVF center in France.
Three embryologists of different levels of seniority (junior, mid-level and senior) were asked to retrospectively evaluate all embryos in each cohort (N = 67) as they would in a laboratory context. They were asked to identify at minimum the kinetic events extracted automatically by EMBRYOLY, corresponding to: t2, t3, t4,t5, t8, tsb, tb (following the definitions of Martinez-Granados et al)*, along with any other morphological or kinetic information they felt was necessary to finalise their embryo ranking (e.g. PN number, Gardner/ASEBIR grading if blastocyst). The embryologists were then asked how confident they felt about their ranking on a scale from 1-5. The time taken to complete the evaluation was recorded for each cohort by an observer. Finally, the embryologists were shown clinical features describing the patient and their treatment, and asked to predict whether they thought the embryo that had been transferred (KID) would have led to a pregnancy or not.
After a blinding period of 1 week, the same embryologists were asked to repeat the task on the same dataset with the adjunct use of EMBRYOLY’s score and kinetic timings. The time they took to evaluate and rank all embryos of a cohort was measured by an observer. They were subsequently asked about their confidence level in their ranking and pregnancy prediction with EMBRYOLY’s help.
Conclusion at a glance
Evaluating embryos with EMBRYOLY can save significant time whilst guaranteeing a standard of care that is at least as good as that of embryologists independently of the seniority level. With EMBRYOLY, embryologists seem to be more precise in their pregnancy prediction, which helps them anticipate outcomes and adapt their treatment strategy accordingly. This demonstrates that EMBRYOLY can be used as a powerful second pair of eyes which embryologists can confidently lean on to accelerate their daily embryo evaluations, welcome more patients and focus more energy on tasks that AI cannot streamline.