Abstract Study question How is the performance of KIDScoreTM D5 version 3 for clinical pregnancy prediction in day 6 blastocysts? Summary answer Differing on Day 5 performance, KIDScoreTM D5 algorithm does not predict clinical pregnancy for day 6 blastocysts according to score subgroup grades. What is known already Embryo selection is a crucial step for a successful outcome in assisted reproductive technologies (ART). In the past years, the use of software and algorithms developed with artificial intelligence (A.I.) tools brought new parameters to be considered by the embryologists in embryo decision at the ART lab. KIDScoreTM is a commercial software (Vitrolife®, Sweden) developed and validated with a large database of known clinical outcome embryos transferred on day 3 or day 5 of development. Yet, it is unknown the performance of KIDScore D5 v3 on day 6 blastocysts, which are widely used for embryo transfer. Study design, size, duration Large retrospective cohort study including sequential single embryo transfers with blastocysts submitted or not to trophoectoderm biopsy for PGT-A, cultured in the Embryoscope® Plus incubator (Vitrolife®) in a single private IVF center between Jan/2020 and Jul/2023. Positive or negative clinical pregnancy, CP (presence/absence fetal heartbeat and gestational sac) from 738 patients/771 blastocysts developed on day 5 or day 6 were considered. Biochemical pregnancy and miscarriage were excluded from this analysis. Participants/materials, setting, methods Day 5 and day 6 embryo transfer outcomes data were compared considering three KIDScoreTM subgroups, according to the following score intervals: subgroup 1: 1.0-3.9 (n = 102), subgroup 2: 4.0-6.9 (n = 273) and subgroup 3: 7.0-9.9 (n = 396 blastocysts). Euploid (n = 517) and non-biopsied embryos (n = 254) were also analyzed in the described subgroups. For the analysis, Mann-Whitney, Chi-square and Fisher tests were used properly for statistical analysis, values of p < 0.05 were considered significant. Main results and the role of chance Maternal age were similar between positive and negative pregnancies in D5 and D6 transfers (D5:39,32±4,34 vs 39,02±4,47 and D6:38,75±3,99 vs 38,18±3,76, p = 0,35 and 0,28 respectively). In D5 transfers, CP rates significantly increased in higher score subgroups [subgroup 1:42.9% (6/14); subgroup 2:48.9% (86/176); subgroup 3:61.6% (236/383),p=0.01]. As expected, subgroup 3 showed a higher CP rate when compared to subgroup 1 + 2 (scores from 1.0 to 6.9,p=0,003). However, in D6 transfers, CP rates were similar between score subgroups 2 and 3 and lower in group 1 [subgroup 1:28.4% (25/88); subgroup 2:50.5% (49/87); subgroup 3:53.8% (7/13),p=0.005 and p = 0,001 between subgroup 1 and subgroups 2 + 3]. Similar results were seen in euploid and non-biopsied embryo transfers; in D5, CP rates were increased in higher scores subgroup (7.0-9.9) when compared to subgroups 1 + 2 (euploid:61,6% vs 42,9% p = 0,05, and non-biopsied:61,3% vs 42,6%,p=0,01 for positive and negative CP respectively in subgroup 3 vs subgroup 1 + 2). And in D6 embryo transfers CP rates were lower in subgroup 1 (1.0-3.9) and similar in subgroups 2 and 3 for both euploid and non-biopsied embryos (euploid:31,5% vs 49,4% p = 0,02, and non-biopsied:13,6% vs 56,5%,p=0,007 for positive and negative CP respectively in subgroup 1 vs subgroup 2 + 3, respectively). Limitations, reasons for caution The KIDScoreTM D5 v3 algorithm is validated for blastocyst on day 5 of development; the results showed here recognizes the limitations of evaluate the potential of CP prediction on Day 6 blastocysts and the score were not considered for embryo selection decision during the time of study. Wider implications of the findings The KIDScoreTM is effective to predict the potential of clinical pregnancy in blastocysts on day 5 of development, with higher rates on higher score subgroups, but not for blastocysts on day 6 of development, where the embryos had similar rates with score greater than 4.0. Trial registration number Not applicable.
Read full abstract