On Automated Tracking and Analysis of Embryo Developmental Trajectory
We investigate the application of image processing, computer vision, and machine learning techniques to analyze time-lapse microscopy (TLM) recordings of early (preimplantation) mammalian embryogenesis. Rapid and accurate evaluation of embryo quality in the in vitro fertilization (IVF) clinic is critical for improving rates of successful post-implanation development and live births. Commercial time-lapse image capture systems to monitor embryogenesis are now entering the IVF field, however current methods to analyze embryo quality do not take full advantage of the rich information available from TLM. Recent works focus on the tracking of cells and division onsets up to the 4-cell stage (approximately 50 hours after IVF) or 8-cell stage (~70 hours after IVF). We review and discuss the techniques that have been used for this problem, both for the tracking of cells and quality evaluation based on the discovered parameters. Here we describe a method that goes beyond the 8-cell stage and considers the entire developmental trajectory of the embryo (vs. a set of parameters related with particular phases of embryo development), and we evaluate its potential application as a metric for quality evaluation for mouse and human embryos.