Non-invasive Metabolomic Profiling of Embryo Culture Medium Using Raman Spectroscopy With Deep Learning Model...

Area of Science:

• Reproductive biology
• Spectroscopy
• Bioinformatics

Background:

• Assessing embryo developmental potential is crucial for in vitro fertilization (IVF) success.
• Current methods often rely on morphological evaluation, which can be subjective.
• There is a need for objective, non-invasive methods to predict blastocyst development.

Purpose of the Study:

• To establish a non-invasive predictive model using Raman spectroscopy.
• To evaluate the blastocyst development potential of day 3 cleavage stage embryos.
• To differentiate between embryos capable of reaching the blastocyst stage (blastula) and those that are not (non-blastula).

Main Methods:

• Raman spectroscopy was employed to analyze the metabolic spectrum of spent day 3 embryo culture medium.
• A deep learning classification model was developed to distinguish between blastula and non-blastula embryos.
• Data from 80 blastula and 48 non-blastula samples from 34 patients were analyzed.

Main Results:

• The predictive model achieved an accuracy of 73.53%.
• Key differentiating Raman shifts identified were at 863.5, 959.5, 1,008, 1,104, 1,200, 1,360, 1,408, and 1,632 cm⁻¹.
• Specific ribose vibrations, similar to RNA, were detected, suggesting potential biomarkers.

Conclusions:

• Raman spectroscopy combined with deep learning can predict the developmental potential of day 3 embryos to the blastocyst stage.
• The non-invasive approach offers a promising tool for embryo quality assessment.
• Further research into RNA-like structures could lead to novel biomarker development for embryo selection.