Advanced Imaging Techniques May Aid in Embryo Selection

Using data from continuous time-lapse (TL) monitoring of embryo culture, Cleveland Clinic researchers identified seven parameters associated with embryo chromosomal status and developed a predictive model that increases the probability of selecting chromosomally normal (euploid) blastocysts after in vitro fertilization (IVF). The work is described in an abstract presented today at the annual meeting of the American Society of Reproductive Medicine.

The model incorporates information on kinetics of embryo development and blastulation, presence of early cleavage dysmorphisms, blastocyst maturity and quality of the inner cell mass (ICM) and trophectoderm to help predict the odds of the blastocyst having a normal chromosome complement.

Embryo culture in a high tech incubator with time lapse (TL) imaging capabilities provides numerous new embryo selection criteria. With the EmbryoScope® TL chamber, researchers continuously monitor embryo development from days 1 through 6. In contrast, with traditional embryo culture methods, only a single daily observation is made in order to limit exposure of the embryo to the environment outside the incubator.

“The EmbryoScope® provides continuous imaging data on embryos, their morphology and their cell division kinetics,” says Nina Desai, PhD, HCLD, Director of Cleveland Clinic’s In Vitro Fertilization Lab. “We wanted to know whether any of the parameters seen through continuous culture and time-lapse imaging correlated with a diagnosis of euploidy on preimplantation genetic screening (PGS).”

Continuous embryo observation

This retrospective study analyzed morphokinetic data that had been prospectively collected from 2,493 zygotes subjected to aneuploidy screening at Cleveland Clinic between 2014 and 2017. During culture in the EmbryoScope® TL chamber, markers of development — such as time to two, three, four, five and eight cells; time to morula; start of blastulation; blastulation; expansion and hatching were determined by viewing captured videos. Blastocysts were scored according to maturity stage and ICM/trophectoderm quality. Trophectoderm biopsy was performed on cells on either day five or six using a laser, and excised cells were sent for genetic analysis. Blastocysts found to be euploid were then transferred in the fresh cycle or subsequent frozen cycles.

“Overall, we found that only 37% of the blastocysts were chromosomally normal, which is what we expected since the majority of patients opting for PGS in our program are older,” says Dr. Desai. “PGS can help to shorten the time to pregnancy by identifying chromosomally normal blastocysts, which is particularly important in our older patients whose reproductive potential rapidly declines with time. But PGS is a costly, labor intensive and invasive technique. These limitations make alternative noninvasive methods, which potentially enhance the likelihood of selecting a euploid, attractive. We are hopeful that models like the one described in this study will allow us to do just that.”

Based on retrospective analysis of the data captured with the EmbryoScope® TL and after adjustment for age, the authors identified seven parameters associated with a euploid outcome of aneuploidy screening.

“We found that with every additional hour it took for an embryo to start to blastulate, the odds of it being chromosomally normal decreased by 4.5%, which is really significant, “says Dr. Desai. “We also looked at cutoff points and found that if an embryo started to blastulate in less than 96.2 hours, it was 1.5 times more likely to be chromosomally normal.”

Blastocyst maturity stage, ICM grade and trophectoderm grade were also significantly associated with euploidy. Conversely, odds of euploidy were significantly decreased as the interval between tSB and time of blastocyst expansion increased. An embryo that became a blastocyst on day five rather than day six was three times as likely to be chromosomally normal. Presence of more than a single dysmorphism during early cleavage lowered the odds that a blastocyst was euploid by 49%.

The authors then used the parameters they identified to construct a logistic regression model for selection of a euploid blastocyst. A receiver operating curve analysis showed that combining all seven parameters into a model would predict euploidy with an area under the curve (AUC) of 0.70, or a 70% likelihood. In fresh embryo transfer, a model that combined cryopreservation day with trophectoderm grade had a 69% likelihood of predicting whether a euploid embryo would implant. Odds of implantation were 6-fold higher for a grade 1 trophectoderm versus a grade 3 trophectoderm and almost 2.5-fold higher for a grade 1 trophectoderm versus compared with a 2 trophectoderm.

“One of the beauties of time-lapse imaging is that you get a glimpse of events occurring at all hours — day or night. Often two embryos may look the same at a single time point, but when viewed continuously they may differ radically. For instance, some two-cell embryos divide synchronously from two to four cells, whereas others may go to three cells, remain there for several hours and further progression to four cells may take hours,” says Dr. Desai. “Irregular divisions, and other dysmorphisms may only be momentarily visible. We found an association between multinucleation in embryos and a lower rate of progressing through day four and becoming a blastocyst.”

Identifying age-adjusted parameters for embryonic development

This was the authors’ first attempt at correlating parameters from continuous embryo observation with likelihood of euploidy. While the results were promising, they cautious that the model worked well in the conditions of their specific laboratory and the data were analyzed retrospectively. They hope to apply the criteria from this research to embryo selection for transfer in patients who are not undergoing PGS to see if the parameters identified can, in fact, be used to incrementally improve pregnancy rates.

“When we see patients whose embryos are growing more slowly, we always feel that is not a good sign, and this study provides us with data to support our suspicions,” says Dr. Desai. “The age-adjusted parameters we identified appear to be independently associated with the likelihood of an embryo being chromosomally normal. Our results are dependent upon timing of embryonic development in a specific IVF laboratory. We hope that other clinics will develop their own models so that we can see if there is significant overlap with our findings.”