BACKGROUND: Extended culture for blastocyst transfer is thought to result in embryos with high implantation potential, enabling the transfer of fewer embryos and the reduction of multiple pregnancies after in vitro fertilization (IVF). However, one major drawback of extended culture is the risk of transfer cancellation if no blastocyst develops by Day 5, despite the observation of adequate early embryo development. We set out to develop a model to predict blastocyst transfer cancellation. METHODS: The model was built from 562 consecutive first IVF cycles and tested on an independent validation cohort. Multivariable logistic regression analysis was used to test the association of patient and cycle characteristics with the presence of cultured blastocysts on Day 5 and to create a nomogram. The model's performance was quantified by discrimination and calibration and clinical utility was evaluated for various thresholds. RESULTS: Fertilization technique, number and quality ratio of the Day 3 embryos and female age were independently associated with blastocyst development. The final prediction model showed an area under the curve (AUC) of 0.75 in the training set (95% confidence interval (CI): 0.73-0.77) and was well calibrated. The AUC for the validation data set was 0.80 (95% CI: 0.78-0.83) and calibration was acceptable. Using a decision threshold of 0.55, the model showed a negative predictive value of 80.5% and a false positive rate of 17%. CONCLUSION: A cycle-based model could prove clinically-relevant for reducing the incidence of cancelled Day 5 transfers.
BACKGROUND: Extended culture for blastocyst transfer is thought to result in embryos with high implantation potential, enabling the transfer of fewer embryos and the reduction of multiple pregnancies after in vitro fertilization (IVF). However, one major drawback of extended culture is the risk of transfer cancellation if no blastocyst develops by Day 5, despite the observation of adequate early embryo development. We set out to develop a model to predict blastocyst transfer cancellation. METHODS: The model was built from 562 consecutive first IVF cycles and tested on an independent validation cohort. Multivariable logistic regression analysis was used to test the association of patient and cycle characteristics with the presence of cultured blastocysts on Day 5 and to create a nomogram. The model's performance was quantified by discrimination and calibration and clinical utility was evaluated for various thresholds. RESULTS: Fertilization technique, number and quality ratio of the Day 3 embryos and female age were independently associated with blastocyst development. The final prediction model showed an area under the curve (AUC) of 0.75 in the training set (95% confidence interval (CI): 0.73-0.77) and was well calibrated. The AUC for the validation data set was 0.80 (95% CI: 0.78-0.83) and calibration was acceptable. Using a decision threshold of 0.55, the model showed a negative predictive value of 80.5% and a false positive rate of 17%. CONCLUSION: A cycle-based model could prove clinically-relevant for reducing the incidence of cancelled Day 5 transfers.