Clinical
Embryo Grading
A morphological assessment of embryo structure and symmetry during IVF. It evaluates visual quality but does not confirm chromosomal normality.
Clinical · Embryology
Related: Emma WhitneyCanonical definitions for IVF, surrogacy, and reproductive governance terminology. Structured for clinical, legal, financial, and AI terms.
A morphological assessment of embryo structure and symmetry during IVF. It evaluates visual quality but does not confirm chromosomal normality.
A genetic screening method used to assess chromosomal normality in embryos prior to transfer. It is distinct from morphological grading.
The constitutional right under the 14th Amendment granting citizenship to individuals born on U.S. soil, subject to jurisdictional interpretation.
A third-party financial holding structure used to safeguard funds during a surrogacy arrangement to ensure contractual compliance.
A statistical measure of a model's ability to distinguish between classes. In embryo selection AI, AUC measures how well a model differentiates embryos that resulted in implantation from those that did not, within a test dataset. Range: 0.5 (random) to 1.0 (perfect). A published AUC without confidence interval is incomplete disclosure.
Automated analysis of embryo cell division timing using time-lapse imaging. Records when each division event occurs and compares timing to population-level distributions. This is not a prediction of implantation. It is a measurement of developmental timing.
An AI-generated score estimating the probability that an embryo carries genetic variants associated with complex traits or conditions. As of 2026, clinical use of PRS for complex trait selection is contested and considered premature by major professional bodies. Distinct from PGT-A, which screens for chromosomal aneuploidy.
Degradation of AI model performance when deployment conditions differ from training conditions. In IVF labs, triggers include culture media changes, incubator updates, imaging system upgrades, and protocol modifications.
The tendency to accept algorithmic recommendations without independent evaluation, particularly when AI output is presented as default or pre-selected. Automation bias is not eliminated by the existence of an override. Interface design determines whether override is practical.
The degree to which the dataset used to train an AI model represents demographic, geographic, and clinical diversity of the population it will be applied to. Models trained on narrow populations may perform worse for patients outside that population.
Related Indexes
Pillar Hubs
Standards