A Novel Integrative Non-Invasive Embryo Selection Approach for In Vitro Fertilization based on Artificial Intelligence Enhanced Morphokinetic Analysis and Raman Spectra in Spent Culture Media

Dr CHUNG Pui Wah, Jacqueline

Department of Obstetrics and Gynaecology, CUHK

Infertility is a global problem with around 10% of couples of reproductive age having difficulty in conception and in vitro fertilization (IVF) is an efficient option for infertility treatment.  However, the clinical pregnancy rate of IVF is around 35%.  The current embryo selection approach still adopts morphology-based assessment with several obvious limitations: 1) embryo assessment under single time point ignoring the holistic morphology information during development; 2) low accuracy for aneuploidy screening unless using invasive preimplantation genetic testing (PGT-A). However, the high cost of PGT-A, prolonged duration from treatment to pregnancy and the technical challenge in biopsy limited its wide application in routine practice. To overcome this, we established non-invasive approaches in 1) advanced morphokinetic (MK) assessment under timelapse imaging with artificial intelligence (AI) enhancement and 2) embryo ploidy screening based on Raman spectra (RS) on the metabolomic profiling. Our pilot study showed that the retrospective embryo selection accuracy increased 9.36% in our AI-enhanced MK, and the aneuploidy screening sensitivity and specificity achieved 80.4% and 81.4% in the RS. Moreover, they were able to assess the embryo and give out results as early as 30 hours post fertilization and at a lower cost compared with PGT-A.