KIDScore D3

KIDScore D3 is an adjunctive decision support tool that analyzes manually annotated morphokinetic parameters from time-lapse imaging to predict the likelihood of Day 3 embryos developing to the blastocyst stage. Its methodology involves a scoring algorithm based on sequential developmental criteria, aiming to enhance embryo selection in assisted reproduction by providing statistical viability assessments.

The KIDScore D3 is a decision support tool designed as an optional software accessory for EmbryoScope and EmbryoScope+ time-lapse incubator systems, used in assisted reproduction. Its purpose is to aid embryologists in selecting embryos on Day 3 for transfer, freezing, or continued development by predicting their likelihood of reaching the blastocyst stage. The tool utilizes manually annotated morphokinetic parameters from embryo development, including the number of pronuclei and specific timings of cell division events (tPNf, t2, t3, t4, t5, t8). Based on these inputs, KIDScore D3 assigns a score from 0 to 5 by comparing the embryo's developmental parameters against model criteria; embryos are scored sequentially, and the process stops if an embryo fails a criterion or reaches the final one. Scores indicate developmental characteristics: 0 (not 2PN), 1 (too fast/direct cleavage), 2 (slow development), 3 (irregular, accelerating pace), 4 (irregular, slowing pace/unexpected cell count at 66 hours), or 5 (passed all avoidance criteria). A prospective, single-arm, multi-center clinical study involving data from 81 patients (a subset of a larger European dataset) evaluated the safety and effectiveness of KIDScore D3. The study aimed to assess if KIDScore D3, as an adjunct to traditional Day 3 morphological grading, improved the prediction of blastocyst development. Embryologists were masked to imaging data, relying on morphology and KIDScore D3 scores. The primary endpoint demonstrated that KIDScore D3 was informative for blastocyst outcome in embryos graded A, B, or C by morphology. Results showed that the adjunct use of KIDScore D3 improved embryo selection for transfer compared to morphology alone, with an odds ratio for blastocyst prediction of 4.13 (95% CI 3.48 – 4.9, p<0.0001), which was statistically significant and higher than that of the predicate device (Eeva system, OR 2.57). Secondary endpoints included various diagnostic performance measures at the embryo and treatment levels.

AI Regulatory Concerns

Powered by

Jev reads the full FDA letter and returns calibrated concern probabilities per regulatory category. These are AI estimates, not regulatory advice.

• Q1. How does the KIDScore D3 algorithm specifically weigh the different manually annotated morphokinetic parameters (e.g., tPNf, t2, t3) in its scoring model to arrive at a final score?; What was the inter-observer and intra-observer variability for the manual annotation of morphokinetic parameters required by KIDScore D3 in the clinical study, and how might this impact score consistency?; Beyond blastocyst development prediction, has the KIDScore D3 been evaluated for its correlation with subsequent clinical outcomes such as implantation rates, clinical pregnancy rates, or live birth rates?; How does the performance of KIDScore D3 (e.g., odds ratio, predictive values) vary across different patient populations, such as those with advanced maternal age, different causes of infertility, or varying oocyte quality (own vs. donor eggs)?; What are the specific definitions and thresholds for the 'avoidance criteria' in the KIDScore D3 model that lead to scores 0-4 (e.g., 'too fast,' 'slow to develop,' 'irregular development'), and how were these criteria established and validated?