Multimodal Learning for Embryo Viability Prediction in Clinical IVF
This addresses the time-intensive, costly, and subjective manual embryo assessment in clinical IVF, though it appears incremental as it builds on existing multimodal techniques for a specific domain.
The paper tackled the problem of manually selecting viable embryos in IVF by developing a multimodal model using time-lapse video and EHR data to predict embryo viability, aiming to enable fast and automated predictions.
In clinical In-Vitro Fertilization (IVF), identifying the most viable embryo for transfer is important to increasing the likelihood of a successful pregnancy. Traditionally, this process involves embryologists manually assessing embryos' static morphological features at specific intervals using light microscopy. This manual evaluation is not only time-intensive and costly, due to the need for expert analysis, but also inherently subjective, leading to variability in the selection process. To address these challenges, we develop a multimodal model that leverages both time-lapse video data and Electronic Health Records (EHRs) to predict embryo viability. One of the primary challenges of our research is to effectively combine time-lapse video and EHR data, owing to their inherent differences in modality. We comprehensively analyze our multimodal model with various modality inputs and integration approaches. Our approach will enable fast and automated embryo viability predictions in scale for clinical IVF.