The Eeva Test combines automated time-lapse analysis with statistical modeling to predict continued embryo development. It automatically captures quantitative image features that the human eye cannot detect, and incorporates statistical modeling of dynamic parameters to maximize the test’s predictive power. A new predictive algorithm includes parameters of patient prognosis, morphology, early cleavage timings and quantitative image features to generate a 5-category score for developmental potential. The objective of this multi-center study was to validate this new algorithm against embryo implantation following blastocyst transfer. Retrospective multi-center study The study included a total of 151 patients from 9 centers who consented to use the Eeva Test. Embryos were transferred on Day 5, and embryos with known implantation data were analyzed. Images and clinical data were processed by the Eeva Test’s new Xtend algorithm, which utilizes automated image analysis software to classify embryos into five categories by incorporating key cell division timing parameters, Day 3 cell#, age and quantitative image features reflective of activity during the third cell cycle. The overall known implantation rate (IR) for the study population was 39% (84/216). The 5-category output of the extended algorithm correlated positively with blastocyst IR: 51% (36/71), 45% (33/74), 25% (10/40), 24% (5/21), 0% (0/10). In patients younger than 35yo, the overall known IR was 50% (62/123); however, the IR for “Category 1” embryos was 60% (32/53). In the same patient group, IR for the good morphology blastocysts was 56% (54/97); however, embryos with good morphology and Category 1 achieved a 61% (31/51) IR. The novel aspect of the new algorithm is including computer-extracted quantitative image attributes, patient characteristics and traditional morphology parameters in the predictive model. Although the model was built using blastocyst as outcome variables, in current study we were able to demonstrate that 5-category output in the new algorithm correlated positively with blastocyst implantation. Specifically, in young patients who are more likely in need of embryo selection, Category 1 identified a subset of embryos with 60% implantation rate, which approaches the implantation rate of euploid blastocysts (Forman et al, 2013). The algorithm also identified embryos with better implantation potential among good morphology blastocysts from young patients. Combining extended algorithm and traditional morphology may present a non-invasive approach assisting embryologists to select the best blastocyst(s) for transfer, and encourage the practice of elective SET.

Abnormal cleavage (AC) occurs when 1 cell divides to >2 daughter cells. It is reported that zygote AC is correlated with very low implantation (1%) (Rubio et al. 2012). The objectives of this study were to (1) further address the clinical relevance of zygote AC (AC1) and daughter cell AC (AC2), and (2) examine the potential of using automated time-lapse cleavage analysis to deselect AC. Retrospective cohort study. Patients undergoing D3 embryo transfer from 3 sites (Jun2011-Oct2012) consented to have their embryos monitored using the Eeva™ Test (Auxogyn), a time-lapse enabled system which provides a high or low score of developmental potential based on key P2 (2-to-3 cell) and P3 (3-to-4 cell) time intervals (Wong et al. 2010, Conaghan et al. 2013). AC1 and AC2 were manually evaluated, and clinical relevance was based on implantation and clinical pregnancy (ultrasound at 6 wks). Imaging data of AC embryos were processed using Eeva software. Statistical significance was calculated using Fisher's Exact or χ2 test. A total of 363 embryos from 43 patients were included. The overall prevalence of AC was 20% (AC1:10%, AC2:10%, Both:0.6%). The prevalence of AC among 107 embryos transferred was 29% (AC1:9%, AC2: 19%, Both:1%), because all AC embryos selected for transfer exhibited relatively good morphology (6-10 cell, <25% frag, perfect/moderate symmetry). The implantation rate of AC was 3% (1/30). In patient cohorts, the incidence of AC (0% vs. 1-25% vs. ≥25%) was inversely correlated with clinical pregnancy (50% vs. 29% vs. 16%). Eeva software categorized 87% (53/61) of AC embryos as low, suggesting that automated P2 and P3 timings may inherently capture AC. AC1 and AC2 embryos are often selected for Day 3 transfer due to their overall good morphology, but these embryos rarely implant. Automated time-lapse analysis using predictive cleavage timings may aid in the deselection of AC events to improve pregnancy rates – a prospective trial is currently underway.

Mammalian preimplantation embryo development is a complex process in which the exact timing and sequence of events are as essential as the accurate execution of the events themselves. Time-lapse microscopy (TLM) is an ideal tool to study this process since the ability to capture images over time provides a combination of morphological, dynamic and quantitative information about developmental events. Here, we systematically review the application of TLM in basic and clinical embryo research. We identified all relevant preimplantation embryo TLM studies published in English up to May 2012 using PubMed and Google Scholar. We then analysed the technical challenges involved in embryo TLM studies and how these challenges may be overcome with technological innovations. Finally, we reviewed the different types of TLM embryo studies, with a special focus on how TLM can benefit clinical assisted reproduction. Although new parameters predictive of embryo development potential may be discovered and used clinically to potentially increase the success rate of IVF, adopting TLM to routine clinical practice will require innovations in both optics and image analysis. Combined with such innovations, TLM may provide embryologists and clinicians with an important tool for making critical decisions in assisted reproduction. In this review, we perform a literature search of all published early embryo development studies that used time-lapse microscopy (TLM). From the literature, we discuss the benefits of TLM over traditional time-point analysis, as well as the technical difficulties and solutions involved in implementing TLM for embryo studies. We further discuss research that has successfully derived non-invasive markers that may increase the success rate of assisted reproductive technologies, primarily IVF. Most notably, we extend our discussion to highlight important considerations for the practical use of TLM in research and clinical settings.

ObjectiveCombined with timing of the 2nd and 3rd cytokinesis, a prolonged 1st cytokinesis duration has been reported to correlate with low blastocyst formation and reduced expression of cytokinesis genes (Wong et al, 2010). This study determined the incidence of a novel abnormal 1st cytokinesis phenotype, evaluated the duration of 1st cytokinesis, and assessed clinical relevance.DesignMultisite retrospective cohort study.Materials and MethodsPatients from 3 clinics consented to have embryos imaged using the Eeva™ Test (Auxogyn), which performs time-lapse analysis of key cell division timings (Jun2011-Oct 2012). Embryo videos were reviewed for 1st cytokinesis phenotype and duration (P1). Abnormal phenotype was defined as oolema ruffling and/or formation of pseudo cleavage furrows. P1 duration was defined as the time from appearance of the 1st cleavage furrow to completion of the 1st division. Clinical pregnancy was confirmed by ultrasound at 6 wks. Fisher's Exact test was used to assess statistical significance.ResultsA total of 638 embryos from 67 patients were categorized into groups: (1) with normal phenotype (442/638=69%) and (2) with abnormal phenotype (196/638=31%). Both groups had good morphology embryos on D3 (53% and 34% with 6-10 cells, ≤10% frag, p<0.001). Group 2 had a lower blastocyst formation rate (131/305=43% vs. 26/126=20%, p<0.0001) and trended to lower implantation rate (15/91=17% vs. 2/32=6%, p=0.2). Group 2 also exhibited prolonged P1 (0.5± 0.8 vs. 1.8±3.3 hrs, p<0.0001). Within Group 2, a subgroup of embryos with P1≥0.5 hrs had even lower blastocyst formation (6/70=9%; p<0.0001 vs. Group 1, p<0.05 vs. Group 2), and none (0/14) implanted.ConclusionEmbryos exhibiting abnormal 1st cytokinesis phenotypes represent 31% of the embryo population and have significantly lower developmental potential. Since many of these embryos have good morphology at the cleavage stage, using time-lapse to detect abnormal and prolonged 1st cytokinesis phenotypes may improve the success of embryo selection. ObjectiveCombined with timing of the 2nd and 3rd cytokinesis, a prolonged 1st cytokinesis duration has been reported to correlate with low blastocyst formation and reduced expression of cytokinesis genes (Wong et al, 2010). This study determined the incidence of a novel abnormal 1st cytokinesis phenotype, evaluated the duration of 1st cytokinesis, and assessed clinical relevance. Combined with timing of the 2nd and 3rd cytokinesis, a prolonged 1st cytokinesis duration has been reported to correlate with low blastocyst formation and reduced expression of cytokinesis genes (Wong et al, 2010). This study determined the incidence of a novel abnormal 1st cytokinesis phenotype, evaluated the duration of 1st cytokinesis, and assessed clinical relevance. DesignMultisite retrospective cohort study. Multisite retrospective cohort study. Materials and MethodsPatients from 3 clinics consented to have embryos imaged using the Eeva™ Test (Auxogyn), which performs time-lapse analysis of key cell division timings (Jun2011-Oct 2012). Embryo videos were reviewed for 1st cytokinesis phenotype and duration (P1). Abnormal phenotype was defined as oolema ruffling and/or formation of pseudo cleavage furrows. P1 duration was defined as the time from appearance of the 1st cleavage furrow to completion of the 1st division. Clinical pregnancy was confirmed by ultrasound at 6 wks. Fisher's Exact test was used to assess statistical significance. Patients from 3 clinics consented to have embryos imaged using the Eeva™ Test (Auxogyn), which performs time-lapse analysis of key cell division timings (Jun2011-Oct 2012). Embryo videos were reviewed for 1st cytokinesis phenotype and duration (P1). Abnormal phenotype was defined as oolema ruffling and/or formation of pseudo cleavage furrows. P1 duration was defined as the time from appearance of the 1st cleavage furrow to completion of the 1st division. Clinical pregnancy was confirmed by ultrasound at 6 wks. Fisher's Exact test was used to assess statistical significance. ResultsA total of 638 embryos from 67 patients were categorized into groups: (1) with normal phenotype (442/638=69%) and (2) with abnormal phenotype (196/638=31%). Both groups had good morphology embryos on D3 (53% and 34% with 6-10 cells, ≤10% frag, p<0.001). Group 2 had a lower blastocyst formation rate (131/305=43% vs. 26/126=20%, p<0.0001) and trended to lower implantation rate (15/91=17% vs. 2/32=6%, p=0.2). Group 2 also exhibited prolonged P1 (0.5± 0.8 vs. 1.8±3.3 hrs, p<0.0001). Within Group 2, a subgroup of embryos with P1≥0.5 hrs had even lower blastocyst formation (6/70=9%; p<0.0001 vs. Group 1, p<0.05 vs. Group 2), and none (0/14) implanted. A total of 638 embryos from 67 patients were categorized into groups: (1) with normal phenotype (442/638=69%) and (2) with abnormal phenotype (196/638=31%). Both groups had good morphology embryos on D3 (53% and 34% with 6-10 cells, ≤10% frag, p<0.001). Group 2 had a lower blastocyst formation rate (131/305=43% vs. 26/126=20%, p<0.0001) and trended to lower implantation rate (15/91=17% vs. 2/32=6%, p=0.2). Group 2 also exhibited prolonged P1 (0.5± 0.8 vs. 1.8±3.3 hrs, p<0.0001). Within Group 2, a subgroup of embryos with P1≥0.5 hrs had even lower blastocyst formation (6/70=9%; p<0.0001 vs. Group 1, p<0.05 vs. Group 2), and none (0/14) implanted. ConclusionEmbryos exhibiting abnormal 1st cytokinesis phenotypes represent 31% of the embryo population and have significantly lower developmental potential. Since many of these embryos have good morphology at the cleavage stage, using time-lapse to detect abnormal and prolonged 1st cytokinesis phenotypes may improve the success of embryo selection. Embryos exhibiting abnormal 1st cytokinesis phenotypes represent 31% of the embryo population and have significantly lower developmental potential. Since many of these embryos have good morphology at the cleavage stage, using time-lapse to detect abnormal and prolonged 1st cytokinesis phenotypes may improve the success of embryo selection.

Non-invasive image markers have the potential to improve selection of viable embryos. We have demonstrated that measurements of cell cycle divisions are correlated with blastocyst formation and gene expression (Wong et al, Nature Biotech 2010). The objective of this study was to develop and prospectively validate a robust model for early prediction of embryo development. Multi-site, prospective, cohort study. Patients undergoing blastocyst culture and transfer consented to have their embryos imaged using a time-lapse imaging system, Eeva (Early Embryo Viability Assessment). Embryologists who were blinded to the embryo outcome independently reviewed videos for specific cell division time intervals P2 (time between cytokinesis 1 and 2) and P3 (time between cytokinesis 2 and 3). A classification and regression tree model was developed to predict blastocysts based on P2 and P3. The prediction model was validated on an independent set of 188 embryos and assessed for performance. A total of 480 embryos from 65 patients were included. The Eeva model predicts a high probability of blastocyst development when both P2 and P3 are within specific cell division timing ranges (9.33≤P2≤11.45 and 0≤P3≤1.73). The average P2 and P3 values of blastocysts in the Development and Validation datasets were highly consistent. Prospectively using Eeva, the specificity of blastocyst prediction was significantly improved (85%) over the average prediction made by experienced embryologists using cleavage morphology (57%).Tabled 1Eeva Blastocyst Prediction Model & PerformanceDataset# EmbryosP2 (hrs)P3 (hrs)SpecificityDevelopment29210.5±2.51.1±2.586%Validation18810.7±2.41.0±1.685% Open table in a new tab We have developed and validated a model for predicting viable blastocyst formation by the cleavage stage. Eeva predictions are specific, non-invasive and easily integrated into the workflow of day 3 or 5 transfer procedures. Parallel studies are evaluating Eeva predictions for improvement to embryo selection.

Since many transferred embryos with "good morphology" fail to implant, technologies are needed to distinguish embryos with highest developmental competence. We developed and validated an integrated time-lapse and automated image analysis system which measures specific cell division timings and predicts blastocyst development by day 2. The objective of this study was to determine the degree of improvement to viable embryo selection when using our prediction model with traditional morphology. Multi-site, prospective, cohort study. The study included 43 patients (≤42 years old) undergoing in vitro fertilization at 3 clinics. Day 3 morphology and time-lapse image data captured by an integrated imaging system, Eeva (Early Embryo Viability Assessment), were prospectively collected. Eeva generated a blastocyst probability score (Low, High) based on specific cell division timings. Eeva blastocyst probability was determined to be High when all cell division timings were within the defined ranges. Five experienced embryologists made a "blastocyst" or "arrest" prediction for each embryo using Day 3 Morphology only, and then using Morphology and Eeva scores. The prediction results were compared to true blastocyst outcomes. A total of 343 embryos were evaluated. By combining Morphology+Eeva, the average blastocyst prediction accuracy significantly improved. The degree of improvement was augmented for embryos with "good morphology" on day 3.Tabled 1Blastocyst Prediction on Day 3 (% Accuracy, Mean±SD)Embryo GroupsMorphology onlyMorphology+EevaP valueTotal Embryos (n=343)62±680±2<0.0056 to 10-cell, ≤10% Frag, Perfect Symmetry (n=135)37±466±4<0.005 Open table in a new tab Adding Eeva to traditional morphology dramatically improved day 3 blastocyst prediction, particularly among "good morphology embryos". Predictions are non-invasive and available by day 2 using automated analysis. Eeva is a uniquely effective and efficient tool for viable embryo selection that may ultimately improve implantation rates.