Clinical setting, study design and criteria for participation in the study

The present retrospective study presents embryological and clinical data from an experienced private Assisted Reproduction Unit, “IVF Athens Reproduction Center” in Athens, Greece, collected throughout October 2021 to July 2022. The study was approved by the Research and Ethics Committee of the IVF Unit (EVD1003/2022) and was conducted in accordance with the ethical standards of the National Authority for Medically Assisted Reproduction and the 1964 Helsinki Declaration and its later amendments [15].

Data were collected from 91 subsequent IVF/ICSI cycles in a matching number of infertile patients that were conducted following to infertility diagnosis of various etiologies: female/male factor or combined, unexplained infertility and repeated implantation failure following IVF/ICSI. Patients received extensive consultation throughout the stages of treatment and consented to the treatment regime. Female partners had an average age of 34.42 ± 3.34 years (min = 23 years; max = 40 years), while 429 blastocyst stage embryos were analyzed.

Exclusion criteria included IVF/ICSI cycles conducted with embryo cultures in conventional incubators, cycles with embryo transfer and/or cryopreservation at earlier stages (day 2 or day 3), early embryo arrested development and those with incomplete data that failed to follow-up. Cycles with donor gametes (donor sperm and/or donor oocyte), surrogacy or embryo biopsy for preimplantation genetic analysis for aneuploidies (PGT-A), monogenic disorders (PGT-M) or structural rearrangements (PGT-SR), were also excluded. Moreover, female participants with any endometrial or endocrinological pathology and/or any medical history of endometriosis, hydrosalpinx, or autoimmune disorders were excluded. Male partners with genitourinary infection or other reproductive pathologies, with medical history of malignancies or previous chemotherapy and/or radiotherapy were also excluded. Cycles that received adjuvant treatments, or cycles with obstructive or non-obstructive azoospermia or with absolute terazoospermia (0% typical forms in the ejaculate) according to the applied WHO strict criteria [16, 17] were not included in the present study.

Ovarian stimulation, oocyte retrieval and ICSI

For all the included cycles in the study, patients underwent an antagonist protocol for controlled ovarian stimulation initiated at day 2 to 4 of the menstrual cycle with recombinant FSH (Gonal F, MerkKGaA, Darmstadt, Germany; or Puregon (MSD, Kenilworth, NJ, USA), alone or in combination with urinary gonadotropins (hMG) (Menopur, Ferring, Saint-Prex Switzerland) and the use of antagonist Cetrotide 0.25 mg (Merck, MerkKGaA, Darmstadt, Germany) or Orgalutran 0.25 mg (Organon, Oss, Netherlands) when the leading follicle reached 14 mm and up to the day of final triggering. Ovarian response was monitored by transvaginal ultrasound with assessments of follicular growth, serum estradiol and progesterone levels every 1–3 days during stimulation. FSH and hMG dosages were adjusted accordingly to reach an optimal oocyte retrieval rate for each case.

When leading follicles reached at least 17 mm, ovulation induction was performed by administering 250 µg of choriogonadotropin alpha (Ovitrelle, Merck Serono Europe Limited, London, UK), while oocyte retrieval was performed 34–36 h after triggering under general anesthesia. Retrieved oocytes were reserved in a conventional incubator (Labotect, C200) in pre-equilibrated culture medium dishes (Universal IVF Medium, Origio a/s, Malov, Denmark) covered with mineral oil (OVOIL, Vitrolife, Sweden) at stable conditions of 5.0% O2, 6.6% CO2 and 37οC [18] until cumulus denudation and insemination by ICSI. Sperm preparation and assessment was performed according to our previous publication [19] and conformed to the WHO procedures for gamete handling [17].

Ooocyte fertilization through ICSI was performed approximately 40 h after beta-hCG administration. Following sperm injection, oocytes were transferred into the microwells of the pre-equilibrated specialized embryo culture dish (EmbryoSlide+ TM, Vitrolife A/S, Viby, Denmark) that contained single-step culture media (Sage 1-Step, Origio a/s, Malov, Denmark) covered by mineral oil (OVOIL, Vitrolife, Sweden), throughout day 5 to early day 6 of embryo development.

Embryo culture and time-lapse embryo assessment

All embryos were cultured in the FDA-approved Embryoscope+ incubator (Vitrolife, A/S, Viby, Denmark) installed with the EmbryoViewer software 7.8.2 (Vitrolife, A/S, Denmark). Collection and analysis of patient data was conducted anonymously by using specifically allocated reference codes and without any dominant of personal identification. Embryos with normal fertilization (appearance of two pronuclei) that progressed to blastocyst formation up to early day 6 were included in the analysis. Image sequences were acquired throughout the period of embryo culture via EmbryoViewer v.7.8.2 (Vitrolife, A/S, Denmark) according to the manufacturer’s settings at 10 min intervals in 11 focal planes. The accompanying software provided the implementation of time-lapse-based embryo analysis by the scoring systems of KIDScore and iDAScore® (Vitrolife, A/S, Viby, Denmark).

For KIDScore D5, all embryos were annotated by two trained and experienced embryologists in order to eliminate the intrinsic inter- and intra- reader variability, according to current guidelines [20,21,22]. For all embryos, the following information was recorded for the application of KIDScore D5: number of pronuclei (PN), timing of syngamy (tPNf), t2 (time from insemination to complete division to two cells), t3 (time from insemination to complete division to three cells), t5 (time from insemination to complete division to five cells), t8 (time from insemination to complete division to eight cells), tB (time from insemination to formation of blastocyst), ICM (Inner cell mass evaluation) and TE (Trophectoderm evaluation). Irregular morphokinetic events (such as reverse cleavage, multinucleation, abnormal pronuclei) were monitored, giving the advantage to deselect these embryos [23, 24].

iDAScore v1.2.0 software (Intelligent Data Analysis Score, Vitrolife, A/S, Denmark) provides a fully automated analysis of time-lapse sequences from the time of insemination (t0) until blastocyst stage development (108–148 h post-insemination). A higher score indicates a greater chance of achieving successful events of clinical pregnancy with positive fetal heartbeat. A score from 1 (lowest) to 9.9 (highest) is automatically generated for each embryo which is statistically correlated with its implantation potential. iDAScore provides a final grading for each embryo without being influenced by the evaluation of the operator [25], thus, eliminating the subjectivity of the conventional observatory approach. No patient data (e.g., age) or morphokinetic parameters are used as input to this model.

In freeze-all cycles or in cycles with surplus cryopreserved embryos, good quality blastocysts (GQBs) as defined by embryologists and according to Gardner’s criteria, were vitrified on day 5 if these were presented with a good quality inner cell mass (ICM) and trophectoderm (TE). Embryos not reaching adequate expansion or not meeting the above-mentioned criteria remained in culture an additional day and according to their developmental characteristics were vitrified on day 6. Blastocysts were cryopreserved by vitrification according to the protocols implementated in the IVF Unit routine practice, and according to manufacturer’s procedures for Vit Kit Freeze/Warm NX (FUJIFILM Irvine Scientific, INC, Santa Ana, CA, USA). Embryos with the highest scores were selected to be transferred first in fresh cycles or prioritized for future transfer after vitrification and warming. A maximum of three embryos per transfer was optioned if possible as allowed by the National Legislation Authority [26].

Embryo transfer and clinical outcomes

All included cycles in the analysis had either fresh or frozen embryo replacement (FER) using vitrified/warmed embryos. For the fresh embryo transfers (ET), luteal support was provided by intramuscular progesterone injection (Prolutex, IBSA Farmaceutici Italia, Lodi, Italy), whereas, for frozen embryo replacement (FER) patients were prepared through a combination of oral capsules (Utrogestan, Faran Laboratories AVEE, Attica, Greece) and intramuscular injection (Prolutex, IBSA Farmaceutici Italia, Lodi, Italy). ET was performed under trans-abdominal ultrasound guidance for adequate embryo deposition with Wallace catheters (CooperSurgical, Malov, Denmark) either on day 5 of embryonic development in fresh ET cycles or after 6 days of progesterone administration in FER by ensuring a receptive endometrium. Serum human chorionic gonadotropin levels were measured 14 days after ET to confirm biochemical pregnancy. A clinical pregnancy was assured by ultrasonographic visualization with the presence of intrauterine gestational sac/s with confirmed fetal heart activity 6 weeks following ET. The main outcome measure was the live birth prediction, as it is considered the strongest endpoint in assisted reproduction.

Data collection and analysis

Collection and analysis of patient data was conducted anonymously by using specifically allocated reference codes and without any dominant of personal identification. For all embryos at blastocyst stage, KIDScore D5 annotation and iDAScore evaluation have been recorded for statistical analysis and review in terms of reproductive outcome prediction. For the statistical analysis, descriptive statistics of the data provided an essential summary of the basic features of included population and its characteristics. Pearson correlation was subsequently carried out for the evaluation of the linear relationship between KID5 and iDA5 score. Simple logistic regression for KID5 score and iDA5 score in terms of live birth was performed. Paired t-test was applied in order to investigate the extent of KID5 and iDA5 scores difference in contrast to the probability of live birth. Multiple logistic regression was applied to explore KID5 and iDA5 scores and live birth probability, adjusted for age and number of blastocysts, followed. ROC (receiver operating characteristic) analysis was performed in order to evaluate KID5 and iDA5 score performance, with ROC curves at all possible classification thresholds designed for each score. The statistical analysis and graphical representations were carried out using SPSS version 20.0 (IBM SPSS Statistics). Outcomes were considered statistically significant when p-value was < 0.05.