Comparing prediction of ongoing pregnancy and live birth outcomes in patients with advanced and younger maternal age patients using KIDScore™ day 5: a large-cohort retrospective study with single vitrified-warmed blastocyst transfer

Abstract Background The KIDScore™ Day 5 (KS-D5) model, version 3, is a general morphokinetic prediction model (Vitrolife, Sweden) for fetal heartbeat prediction after embryo transfer that was developed based on a large data set that included implantation results from a range of clinics with different patient populations, culture conditions and clinical practices. However, there was no study to comparing their pregnancy and live birth prediction ability among different maternal age. The aim of this study is to analyze the performance of KS-D5 in predicting pregnancy and live birth in various maternal age groups after single vitrified-warmed blastocyst transfer (SVBT). Methods A total of 2486 single vitrified-warmed blastocyst transfer (SVBT) cycles were analyzed retrospectively. Confirmed fetal heartbeat positive (FHB+) and live birth (LB+) rates were stratified by Society for Assisted Reproductive Technology (SART) maternal age criteria (< 35, 35–37, 38–40, 41–42 and ≥ 43 years of age). Within each age group, the performance of the prediction model was calculated using the AUC, and the results were compared across the age groups. Results In all age groups, the FHB+ rates decreased as the KIDScore decreased (P < 0.05). Conversely, the AUCs increased as the maternal age increased. The AUC of the < 35 age group (0.589) was significantly lower than the AUCs of the 41–42 age group (0.673) and the ≥43 age group (0.737), respectively (P < 0.05). In all age groups, the LB+ rates decreased as the KIDScore decreased (P < 0.05). Conversely, the AUCs increased as the maternal age increased. The AUC of the ≥43 age group (0.768) was significantly higher than the AUCs of other age groups (P < 0.05; < 35 age group = 0.596, 35–37 age group = 0.640, 38–40 age group = 0.646, 41–42 age group = 0.679). Conclusions In the present study, we determined that the KIDScore model worked well for prediction of pregnancy and live birth outcomes in advanced age patients.

Download Full-text

Leave the past behind: women’s reproductive history shows no association with blastocysts’ euploidy and limited association with live birth rates after euploid embryo transfers

Human Reproduction ◽

10.1093/humrep/deab014 ◽

2021 ◽

Author(s):

Danilo Cimadomo ◽

Antonio Capalbo ◽

Lisa Dovere ◽

Luisa Tacconi ◽

Daria Soscia ◽

...

Keyword(s):

Clinical Outcomes ◽

Live Birth ◽

Maternal Age ◽

Oocyte Retrieval ◽

Reproductive History ◽

Blastocyst Transfer ◽

Blastocyst Development ◽

Live Birth Rates ◽

History Of ◽

Euploid Embryo

Abstract STUDY QUESTION Is there an association between patients’ reproductive history and the mean euploidy rates per biopsied blastocysts (m-ER) or the live birth rates (LBRs) per first single vitrified-warmed euploid blastocyst transfers? SUMMARY ANSWER Patients’ reproductive history (as annotated during counselling) showed no association with the m-ER, but a lower LBR was reported after euploid blastocyst transfer in women with a history of repeated implantation failure (RIF). WHAT IS KNOWN ALREADY Several studies have investigated the association between the m-ER and (i) patients’ basal characteristics, (ii) ovarian stimulation strategy and dosage, (iii) culture media and conditions, and (iv) embryo morphology and day of full blastocyst development. Conversely, the expected m-ER due to women’s reproductive history (previous live births (LBs), miscarriages, failed IVF cycles and transfers, and lack of euploid blastocysts among prior cohorts of biopsied embryos) still needs investigations. Yet, this information is critical to counsel new patients about a first cycle with preimplantation genetic testing for aneuploidy (PGT-A), but even more so after former adverse outcomes to prevent treatment drop-out. STUDY DESIGN, SIZE, DURATION This observational study included all patients undergoing a comprehensive chromosome testing (CCT)-based PGT-A cycle with at least one biopsied blastocyst in the period April 2013-December 2019 at a private IVF clinic (n = 2676 patients undergoing 2676 treatments and producing and 8151 blastocysts). m-ER were investigated according to women’s reproductive history of LBs: no/≥1, miscarriages: no/1/>1; failed IVF cycles: no/1/2/>2, and implantation failures after previous transfers: no/1/2/>2. Among the 2676 patients included in this study, 440 (16%) had already undergone PGT-A before the study period; the data from these patients were further clustered according to the presence or absence of euploid embryo(s) in their previous cohort of biopsied blastocysts. The clinical outcomes per first single vitrified-warmed euploid blastocyst transfers (n =1580) were investigated according to the number of patients’ previous miscarriages and implantation failures. PARTICIPANTS/MATERIALS, SETTING, METHODS The procedures involved in this study included ICSI, blastocyst culture, trophectoderm biopsy without hatching in Day 3, CCT-based PGT-A without reporting segmental and/or putative mitotic (or mosaic) aneuploidies and single vitrified-warmed euploid blastocyst transfer. For statistical analysis, Mann–Whitney U or Kruskal–Wallis tests, as well as linear regressions and generalised linear models among ranges of maternal age at oocyte retrieval were performed to identify significant differences for continuous variables. Fisher’s exact tests and multivariate logistic regression analyses were instead used for categorical variables. MAIN RESULTS AND THE ROLE OF CHANCE Maternal age at oocyte retrieval was the only variable significantly associated with the m-ER. We defined five clusters (<35 years: 66 ± 31%; 35–37 years: 58 ± 33%; 38–40 years: 43 ± 35%; 40–42 years: 28 ± 34%; and >42 years: 17 ± 31%) and all analyses were conducted among them. The m-ER did not show any association with the number of previous LBs, miscarriages, failed IVF cycles or implantation failures. Among patients who had already undergone PGT-A before the study period, the m-ER did not associate with the absence (or presence) of euploid blastocysts in their former cohort of biopsied embryos. Regarding clinical outcomes of the first single vitrified-warmed euploid blastocyst transfer, the implantation rate was 51%, the miscarriage rate was 14% and the LBR was 44%. This LBR was independent of the number of previous miscarriages, but showed a decreasing trend depending on the number of previous implantation failures, reaching statistical significance when comparing patients with >2 failures and patients with no prior failure (36% versus 47%, P < 0.01; multivariate-OR adjusted for embryo quality and day of full blastocyst development: 0.64, 95% CI 0.48–0.86, P < 0.01). No such differences were shown for previous miscarriage rates. LIMITATIONS, REASONS FOR CAUTION The sample size for treatments following a former completed PGT-A cycle should be larger in future studies. The data should be confirmed from a multicentre perspective. The analysis should be performed also in non-PGT cycles and/or including patients who did not produce blastocysts, in order to investigate a putative association between women’s reproductive history with outcomes other than euploidy and LBRs. WIDER IMPLICATIONS OF THE FINDINGS These data are critical to counsel infertile couples before, during and after a PGT-A cycle, especially to prevent treatment discontinuation due to previous adverse reproductive events. Beyond the ‘maternal age effect’, the causes of idiopathic recurrent pregnancy loss (RPL) and RIF are likely to be endometrial receptivity and selectivity issues; transferring euploid blastocysts might reduce the risk of a further miscarriage, but more information beyond euploidy are required to improve the prognosis in case of RIF. STUDY FUNDING/COMPETING INTEREST(S) No funding was received and there are no competing interests. TRIAL REGISTRATION NUMBER N/A.

Download Full-text

Intestinal parasites in swine in the Nordic countries: multilevel modelling of Ascaris suum infections in relation to production factors

Parasitology ◽

10.1017/s0031182099004941 ◽

1999 ◽

Vol 119 (5) ◽

pp. 521-534 ◽

Cited By ~ 21

Author(s):

A. ROEPSTORFF ◽

O. NILSSON ◽

C. J. O'CALLAGHAN ◽

A. OKSANEN ◽

B. GJERDE ◽

...

Keyword(s):

Intestinal Parasites ◽

Ascaris Suum ◽

Model Development ◽

Age Groups ◽

Water System ◽

Multilevel Modelling ◽

Age Group ◽

Data Set ◽

Normal Regression ◽

Weaning Age

In Denmark, Finland, Iceland, Norway and Sweden, 413 sow herds were randomly selected for sampling. Faeces from pigs of 7 age groups/categories were examined for helminth eggs (11233 individual samples), and an accompanying questionnaire was completed at each visit. In total, 1138 pigs on 230 farms were found to be positive for Ascaris suum. Considerable differences in the occurrence of A. suum could be observed directly for several of 20 independent variables at the herd or category level. However, given that univariate analyses may be severely affected by confounding of covariates resulting in spurious inference, additional multivariate analyses were undertaken. An ordinary logistic regression on Ascaris positive/negative farms showed that Denmark had the highest frequency of infected herds, while Iceland and Finland had the lowest frequencies and that herds using ‘late weaning’ and ‘Class 2’ drugs (pyrantel, levamisole) were most often infected. Because many herds were found to be totally negative for A. suum, mixed hierarchical logistic-normal regression models (both the penalized quasi-likelihood and the Markov Chain Monte Carlo methods) were developed for both a full (all herds) and a reduced (the 230 infected herds) data set using either a cut-off of >0 eggs per gram (epg) or >200 epg to counter for false-positive egg counts. Estimates for identical models, but where the animal level variance was constrained to the binomial assumption, were also calculated. Significant covariates were robust to model development with ‘Age group’, ‘Country’, ‘Weaning age’, ‘Water system’ and simple interactions between the latter two and ‘Age group’ being significantly associated with the occurrence of A. suum, while all variables concerning anthelmintic drug, anthelmintic strategy, floor type, bedding, dung removal, washing and disinfection were not. These findings are discussed in the light of the complex relationship between A. suum and its pig host.

Download Full-text

Sensitivity Analysis of the Quadratic Discriminant Function for Predicting Pregnancy Outcomes

Journal of Scientific Research and Reports ◽

10.9734/jsrr/2020/v26i230225 ◽

2020 ◽

pp. 60-69

Author(s):

Killian Asampana Asosega ◽

David Adedia ◽

Atinuke O. Adebanji

Keyword(s):

Birth Outcomes ◽

Live Birth ◽

Discriminant Function ◽

Pregnancy Outcomes ◽

Maternal Age ◽

Characteristic Curve ◽

Developed Countries ◽

Positive Effects ◽

Gestational Period ◽

Quadratic Discriminant Function

The prevalence rate of stillbirth is ten times higher in developing countries relative to developed countries with a 2016 rate of 18 percent in Ghana. This study employed the Quadratic Discriminant Function for discriminating and classifying of pregnancy outcomes based on some predictors. The study further examined the sensitivity of the Quadratic Discriminant Function in predicting pregnancy outcomes with variations in the training and test samples of deliveries recorded in a hospital in Accra, Ghana. The study considered the scenarios; 50:50, 60:40, 70:30 and 75:25 ratios of training sets to testing sets. Predictor variables on both maternal factors (maternal age, parity and gravida) and fetus variables (weight at birth and gestational period) were all statistically significant (P < .01) in discriminating between live birth and stillbirth. Results showed that maternal age had a negative effect on the live birth outcomes, while parity, gravida, gestational period and fetus weight recorded positive effects on live birth outcomes. The 75:25 ratio outperformed the other ratios in discriminating between live and stillbirth based on the Actual Error Rate of 7.28% compared to 7.81%, 12.14% and 13.79% for the 50:50, 70:30 and 60:40 ratios respectively whereas, the receiver operating characteristic curve shows the 70:30 (AUC= 0.9233) ratio outperformed the others. The study recommend the use of either the 70:30 or 75:25 training to test ratios for classification and discrimination related problems. Moreover, further research to establish the power of the respective training to test sample ratios with other statistical classification tools and more socio-economic variables can be considered.

Download Full-text

Cytoplasmic String Between ICM and mTE Is a Positive Predictor of Clinical Pregnancy and Live Birth Outcomes in Elective Frozen-thawed Single Blastocyst Transfer Cycles: a Time-lapse Study

10.21203/rs.3.rs-122470/v1 ◽

2020 ◽

Author(s):

Bing-Xin Ma ◽

Lei Jin ◽

Bo Huang

Keyword(s):

Birth Outcomes ◽

Live Birth ◽

Inner Cell Mass ◽

Cell Mass ◽

Time Lapse ◽

Clinical Pregnancy ◽

Blastocyst Transfer ◽

Single Blastocyst Transfer ◽

First Time ◽

Present Group

Abstract Background: In this study, we aim to investigate whether cytoplasmic string between inner cell mass (ICM) and mural trophectoderm (mTE) is a positive predictor of clinical pregnancy and live birth outcomes.Methods: 1,267 elective frozen-thawed single blastocyst transfer (eSBT) cycles cultured in time-lapse incubation system from January 2018 to May 2019 were involved in the study. Blastocysts were grouped according to the appearance of cytoplasmic strings between ICM and mTE cells, and identified as “Present” and “Absent” groups. In Present group, they were further categorized according to the quantity of cytoplasmic strings between ICM and mTE cells. Clinical pregnancy and live birth outcomes of blastocysts were used to evaluate the effect of cytoplasmic strings between ICM and mTE.Results: The baseline demographic and laboratory features were similar between the Present and Absent groups of cytoplasmic strings between ICM and mTE (P>0.05). According to the time-lapse analysis, cytoplasmic strings between ICM and mTE were more visible among good quality blastocysts. Furthermore, blastocysts with cytoplasmic strings showed a higher clinical pregnancy and live birth rates (P<0.05), and no significant differences were observed in abortion rate and birth weight (P>0.05).Conclusions: Although the previous conclusions of cytoplasmic strings were controversial, the present time-lapse analysis provides the evidence for the first time that cytoplasmic strings between ICM and mTE cells would be a positive predictor of clinical pregnancy and live birth outcomes in elective frozen-thawed single blastocyst transfer cycles.

Download Full-text

The Comparison of Twinning Rates Between Urban and Rural Areas in China

Twin Research and Human Genetics ◽

10.1375/twin.10.4.633 ◽

2007 ◽

Vol 10 (4) ◽

pp. 633-637 ◽

Cited By ~ 13

Author(s):

Jian-Ping Gan ◽

Zhong-Hua Wu ◽

Zhi-Ming Tu ◽

Jian Zheng

Keyword(s):

Rural Areas ◽

Urban Areas ◽

Maternal Age ◽

Age Groups ◽

Vital Statistics ◽

Differential Method ◽

Age Group ◽

Vitro Fertilization ◽

Urban And Rural Areas

AbstractBased on the birth record data from the National Vital Statistics in the 1990 Census of China, the present study analyzed the differences between urban and rural areas on monozygotic (MZ) and dizygotic (DZ) twin rates by maternal age in 1989. The twins by zygosity were calculated with Weinberg's differential method. Results show that the MZ and DZ twinning rates in China were associated with maternal age and that there were substantial differences between urban and rural areas. The MZ twinning rates in urban and rural areas were 2.36 pairs and 2.11 pairs per 1000 deliveries respectively, significantly lower than that in most studied populations. Furthermore, our analysis indicated that MZ twinning rates remained relatively constant for mothers under the age group of 36 to 38 years, but rose over this age group in both areas, albeit with a different slope. The DZ twinning rates were strikingly affected by maternal age, but the age for peak DZ rates was found within the age group of 33 to 35 years. In all maternal age groups except for 24 to 26 years, the DZ twinning rates in urban areas were higher than in rural areas. It remains unclear as to why the DZ twinning rates reversed to reach higher values within the older maternal age groups in China, but it is almost certain that the high twinning rates had nothing to do with in vitro fertilization.

Download Full-text

The impact of Data structure on classification ability of financial failure prediction model

SHS Web of Conferences ◽

10.1051/shsconf/20207405024 ◽

2020 ◽

Vol 74 ◽

pp. 05024

Author(s):

Lucia Svabova ◽

Lucia Michalkova

Keyword(s):

Data Structure ◽

Prediction Model ◽

Statistical Methods ◽

Prediction Models ◽

Global Perspective ◽

Data Set ◽

Prediction Ability ◽

Financial Difficulties ◽

The Creation ◽

The Impact

The creation of prediction models to reveal the threat of financial difficulties of the companies is realized by the application of various multivariate statistical methods. From a global perspective, prediction models serve to classify a company into a group of prosperous or non-prosperous companies, or to quantify the probability of financial difficulties in the company. In many countries around the world, real financial data about the companies are used in developing these prediction models. In Slovakia, standard data from the financial statements and annual reports of Slovak companies are used for the creation of the company’s failure model. Since in this case there are generally large data files, it is necessary to pre-process the data by the selected methods before the prediction model is constructed. A database of the companies needs to be prepared for the subsequent application of statistical methods, and it is also highly appropriate to focus globally on the detection of potential extreme and remote observations. Therefore, the article will focus on quantifying the impact of the data structure detected, for example, the occurrence of extreme and remote observations in the data set, on the resulting overall classification of the prediction ability of the models created.

Download Full-text

Maternal Age And Birth Weight Characteristics of Twins Born to Nulliparous Mothers: A Population Study

Twin Research ◽

10.1375/twin.4.1.1 ◽

2001 ◽

Vol 4 (1) ◽

pp. 1-3

Author(s):

Isaac Blickstein ◽

Ran D. Goldman ◽

Ram Mazkereth

Keyword(s):

Birth Weight ◽

Maternal Age ◽

Very Low Birth Weight ◽

Age Groups ◽

Inverse Correlation ◽

Population Based ◽

Age Group ◽

Significant Inverse Correlation ◽

Twin Birth ◽

Significant Difference

AbstractIn order to examine birth weight characteristics of twins delivered to nulliparous mothers in relation to maternal age, we used a population-based cohort of Israeli twins delivered between 1993–98 to select all 4793 (37.6%) nulliparas who delivered twins. Maternal age was subdivided as less than 20 years, 20–24, 25–29, 30–34, 35–39, and 40 years or more. We counted the frequencies of each total twin birth weight (twin A+twin B) in each of three categories (less than 3000, 3000–4999, and more than 5000 g) and the frequency of very low birth weight (VLBW, less than 1500 g) neonates in each of the six maternal age categories. There were significantly more nulliparas in the twin population at age groups less 30 years and significantly less at ages 30 years or more. We found a highly significant inverse correlation between the proportion of nulliparas and maternal age group, decreasing from 71.8% at less than 20 years to 18.6% at age 35–39 years (% nulliparas = 85 − 13.7 × age group, Pearson R2 = 0.98). However, this trend changed abruptly to the observed figure of 25.9% nulliparas aged 40 years or more instead of the expected 2.8%. We failed to reveal any significant difference in birth weight characteristics between the maternal age groups (all p > 0.05, all confidence intervals included 1.0). The more than tenfold deviation of the observed from the predicted frequency of nulliparas aged 40 years or more suggests that a different relationship between parity and age occurs at this age group. Maternal age of nulliparas is not associated with different birth weight characteristics of their twins.

Download Full-text

Cytoplasmic String Between ICM and mTE Is a Positive Predictor of Clinical Pregnancy and Live Birth Outcomes in Elective Frozen-thawed Single Blastocyst Transfer Cycles: A Time-lapse Study

10.21203/rs.3.rs-144098/v1 ◽

2021 ◽

Author(s):

Bingxin Ma ◽

Lei Jin ◽

Bo Huang

Keyword(s):

Birth Outcomes ◽

Live Birth ◽

Inner Cell Mass ◽

Cell Mass ◽

Time Lapse ◽

Clinical Pregnancy ◽

Blastocyst Transfer ◽

Single Blastocyst Transfer ◽

First Time ◽

Present Group

Download Full-text

P–179 Timing of blastocyst observation on day 5: effect on the assessment to predict live birth, and the incorporation into a blastocyst selection model

Human Reproduction ◽

10.1093/humrep/deab130.178 ◽

2021 ◽

Vol 36 (Supplement_1) ◽

Author(s):

Y Liu ◽

K Ong ◽

I Korman ◽

R Turner ◽

M Leyden ◽

...

Keyword(s):

Logistic Regression ◽

Regression Analysis ◽

Prediction Model ◽

Developmental Stage ◽

Birth Outcomes ◽

Live Birth ◽

Scoring System ◽

Roc Analysis ◽

Embryo Morphology ◽

Female Age

Abstract Study question Does variation in day 5 observation timing confound embryo-morphology-based live birth prediction, and is it possible to develop a robust comprehensive numerical prediction model. Summary answer Day 5 observation timing confounds embryo-morphology-based live birth prediction. A robust comprehensive numerical prediction model can be developed after considering a number of contributing variables. What is known already Embryo development is a dynamic process, and therefore the widely used static observations potentially lead to biased prediction of live birth outcomes. So far, little is known in regard to potential confounding impact of day 5 assessment timing on the static-morphology-based live birth prediction. In addition, the inter-observer variation in morphology-based embryo assessment requires a more robust system to improve consistency of selection. Study design, size, duration This retrospective multi-center cohort study included 8866 autologous oocyte in vitro fertilisation treatment cycles performed at 14 associated clinics within the same network during 2012–2018. Only fresh cycles with single day 5 embryo transfers were included for analysis with all pregnancies followed up until birth. Repeat cycles of same patients were excluded to avoid clustering effect in statistical analysis. Participants/materials, setting, methods Dataset was randomly split into two subsets at 60:40 ratio, with one (n = 5274) used for regression analysis and model development and the other (n = 3592) used for model testing. Multiple logistic regression was performed to evaluate live birth predicting power of several potential contributors, expressed by odds ratio (OR) and 95% confidence interval (CI). A comprehensive prediction model was subsequently developed based on calculated weights of contributing factors, then tested via receiver operating characteristics (ROC) analysis. Main results and the role of chance The timings of day 5 observation of 8866 included embryos, measured by hours post insemination (HPI), distributed in a bell shape ranging from 112.0 to 120.0 h (mean±SD 115.7±1.7 h). After taking into account female age at egg collection (grouped as < 30 yr, 30–34 yr, 35–39 yr, 40–44 yr, and 45 yr or older), whether or not the first egg collection, number of eggs collected, embryo developmental stage (grouped as pre-blastocyst, early blastocyst, expanding blastocyst, expanded blastocyst, and hatching/hatched blastocyst) and morphology score(A/B/C/D); multivariate logistic regression analysis showed significant association (OR 1.096, 95% CI 1.020–1.177, P = 0.012) between HPI groups (112–113.9 h, 114–115.9 h, 116–117.9 , and 118–120 h) and subsequent live birth outcomes. A comprehensive numerical scoring system was developed based on the statistically significant predictors including female age (OR 1.465, 95% CI 1.364–1.574, P = 0.000), embryo developmental stage (OR 1.341, 95% CI 1.244–1.445, P = 0.000), morphology score (OR 1.520, 95% CI 1.392–1.661, P = 0.000) and HPI (OR mentioned above); with a formula of Score = (Female_age_group/5)*1.465 + (Developmental_stage/5)*1.341 + (Morpho_Score/4)*1.520 + (HPI_Group/4)*1.096. ROC analysis showed statistically significant predictive power of the resulting model as expressed by area under the ROC curve using both the development (0.690, 0.675–0.704, P = 0.000) and testing (0.685, 0.667–0.703, P = 0.000) subsets. Limitations, reasons for caution The retrospective design does not allow for controlling of unknown confounders. HPI was based on static observations in this study so future time-lapse study may bring more insights with more accurate observation and measurement. Wider implications of the findings: The varying HPIs at day 5 observation were alarming as this could confound live birth prediction using embryology parameters. It is important to standardise the timing of embryo observations. The inclusion of HPI into a comprehensive numerical scoring system for live birth prediction may potentially improve its robustness Trial registration number Not applicable

Download Full-text

O-220 An annotation-free embryo scoring system (iDAScore®) based on deep learning shows high performance for pregnancy prediction after single-vitrified blastocyst transfer

Human Reproduction ◽

10.1093/humrep/deab128.044 ◽

2021 ◽

Vol 36 (Supplement_1) ◽

Author(s):

S Ueno ◽

M Ito ◽

K Uchiyama ◽

T Okimura ◽

A Yabuuchi ◽

...

Keyword(s):

Deep Learning ◽

Scoring System ◽

Prediction Models ◽

Characteristic Curve ◽

Age Groups ◽

Blastocyst Transfer ◽

In Vitro Fertilisation ◽

Age Group ◽

Grading Systems ◽

Significant Difference

Abstract Study question How is the performance of an automated embryo scoring system for pregnancy prediction after single-vitrified blastocyst transfer (SVBT) compared to other, annotation-dependent blastocyst grading systems? Summary answer Automatic embryo ranking by iDAScore shows a higher or equal performance, with regards to pregnancy prediction after SVBT, compared to manual, annotation-dependent grading systems. What is known already Blastocyst viability can be assessed by blastocyst morphology grades and/or morphokinetic parameters. However, morphological and morphokinetic embryo assessment is prone to both inter- and intra-observer variation. Recently, embryo ranking models have been developed based on artificial intelligence (AI) and deep learning. Such models rank embryos according to their potential for pregnancy only based on images and do not require any user-dependent annotation. So far, no study has independently assessed the performance of AI models compared to other embryo scoring models, including traditional morphological grading. Study design, size, duration A total of 3,014 SVBT cycles were retrospectively analysed. Embryos were stratified according to SART age groups. The quality and scoring of embryos were assessed by iDAScore v1.0 (iDAS, Vitrolife, Sweden), KIDScoreTM D5 v3 (KS; Vitrolife), and Gardner criteria. The performance of the pregnancy prediction for each embryo scoring model was compared using the area under curve (AUC) of the receiver operating characteristic curve for each maternal age group. Participants/materials, setting, methods Embryos were cultured in the EmbryoScope+ and EmbryoScopeFlex (Vitrolife). iDAS was automatically calculated using the iDAScore model running on the EmbryoViewer (Vitrolife). KS was calculated in EmbryoViewer after annotation of the required parameters. ICM and TE were annotated according to the Gardner criteria. The degree of expansion in all blastocysts was Grade 4 due to our freezing policy. Furthermore, Gardner’s scores were stratified into four grades (Excellent: AA, Good: AB BA, Fair: BB, Poor: others). Main results and the role of chance The AUCs of the < 35 years age group (n = 389) for pregnancy prediction were 0.72 for iDAS, 0.66 for KS and 0.64 for Gardner criteria. The AUC of iDAS was significantly higher (P < 0.05) compared to the other two models. For the 35–37 years age group (n = 514) the AUCs were 0.68, 0.68, and 0.65 for iDAS, KS and Gardner, respectively, and were not significantly different. The AUCs of the 38–40 years age group (n = 796) were 0.67 for iDAS, 0.65 for KS and 0.64 for Gardner criteria and where was not significantly different. The AUCs of the 41–42 years age group (n = 636) were 0.66, 0.66, and 0.63 for iDAS, KS and Gardner, respectively, and there was no significant difference among the pregnancy prediction models. For the > 42 years age group (n = 389) AUCs were 0.76 for iDAS, 0.75 for KS and 0.75 for Gardner criteria and not significantly different. Thus, for all age groups, iDAS was either highest or equal to the highest AUC, although a significant difference was only observed for the youngest age group. Limitations, reasons for caution In this study, SVBT was performed after minimal stimulation and natural cycle in vitro fertilisation (IVF). Therefore, we had only few cycles with elective blastocyst transfer. However, there was also no bias in selecting the embryos for SVBT. Wider implications of the findings Our results showed that objective embryo assessment by a completely automatic and annotation-free model, iDAScore, does perform as good or even better than more traditional embryo assessment or an annotation-dependent ranking tool. iDAS could be an optimal pregnancy prediction model after SVBT, especially in young and advanced age patients. Trial registration number not applicable

Download Full-text