ABSTRACT
A major challenge in clinical In-Vitro Fertilization (IVF) is selecting the highest quality embryo to transfer to the patient in the hopes of achieving a pregnancy. Time-lapse microscopy provides clinicians with a wealth of information for selecting embryos. However, the resulting movies of embryos are currently analyzed manually, which is time consuming and subjective. Here, we automate feature extraction of time-lapse microscopy of human embryos with a machine-learning pipeline of five convolutional neural networks (CNNs). Our pipeline consists of (1) semantic segmentation of the regions of the embryo, (2) regression predictions of fragment severity, (3) classification of the developmental stage, and object instance segmentation of (4) cells and (5) pronuclei. Our approach greatly speeds up the measurement of quantitative, biologically relevant features that may aid in embryo selection.
ABSTRACT
AIMS: To compare the theoretical retinal threshold time for endoilluminators and experimental phototoxic effect using A2e laden retinal pigment epithelial (RPE) cells. METHODS: The spectral irradiances of three types of 20 gauge and 25 gauge endoilluminators, currently commercially available from two manufacturers, were evaluated in conditions where the total beam spectral power was divided by the beam spot size at an estimated use distance of 5 mm from the retina. The retinal threshold time was calculated using the guidelines recommended by the International Commission on Non-Ionizing Radiation Protection. In vitro, A2e laden cells were evenly exposed to light for 30 minutes with a standard endoilluminator positioned 1 cm above the cells and the cell viability was assessed by WST-1 assay. RESULTS: The retinal threshold times were within 1 minute for all the endoilluminators tested. A significant decrease in the viability of A2e laden RPE cells was observed after they were exposed to light from two of the three 20 gauge endoilluminators. Cell viability was not affected by the exposure to 25 gauge endoilluminators under the same conditions. There was no correlation between the theoretical threshold times and experimental data. CONCLUSIONS: Light exposure during vitrectomy can induce photochemical damage to the retina. Although the A2e laden RPE model may not correctly mimic a clinical situation, this model may be useful to estimate the possible photochemical damage to RPE cells that could not be deduced by a theoretical retinal hazard model.