Barrie, A., Homberg, R., McDowell, G., Brown, J., Kingsland, C. & Troup, S. (2017) Examining the efficacy of six published time-lapsed imaging embryo selection algorithms to predict implantation to demonstrate the need for development of specific, in-house morphokinetic selection algorithms. Fert Steril 107 pp. 613-621
Objective
To study the efficacy of five embryo selection algorithms (ESAs) at predicting implantation when applied to a large, exclusive set of known implantation embryos with the aim of demonstrating the need to develop specific, in-house ESAs.
Design
A retrospective, observational analysis.
Setting
Fertility Treatment Centre
Patients
Nine hundred and eighty embryos derived from 887 treatment cycles performed between September 2014 and September 2015. Patients undergoing treatment by either IVF or ICSI were included. Embryos were cultured using GTL™ (Vitrolife) at 5% O2, 89% N2, 6% CO2, 37°C in EmbryoScope® instruments.
Main outcome Measure
The difference in implantation rates (IR) of the categories of embryo classification in each ESA defined using specificity, sensitivity and positive predictive value, area under the receiver operating characteristic curve (AUC) and likelihood ratio. The differences in implantation rates (IR) in the categories defined by each ESA were also analysed using Fisher’s exact and Kruskall Wallis statistical tests.
Results
Each ESA specified time ranges into which embryos must fall to be identified as having the highest potential for the stated end point. The ESAs comprised a variety of observable events including time to pronuclear fading, 2, 3, 4, 5 and 8 cell, cell cycle durations and time to blastulation and blastocyst. When applied to an exclusive cohort of known implantation embryos, the PPV for IR were 42.57%, 38.14%, 44.07%, 38.79%, and 40.45%. The sensitivity was 16.70%, 51.19%, 72.94%, 98.67%, and 62.33%, respectively. Finally, the specificity was 85.90%, 48.09%, 42.12%, 2.65% and 42.62%, respectively. The AUC were 0.535, 0.512, 0.575, 0.546 and 0.583, respectively. There were no significant differences in IR between the categories in four of the five ESAs (p>0.05). One of the ESAs resulted in statistically significant differences in the embryo classifications in terms of IR (p<0.0001).
Conclusion
The results from the examination of the published ESAs examined highlight the need for the development of in house, patient, treatment and environment specific ESAs. These data suggest that currently available ESAs may not be clinically applicable and lose their diagnostic value when externally applied.
Full Text
You must be logged in to post a comment.