Abstract
PURPOSE: To analyze whether the morphokinetics algorithm based on data from day 5 blastocyst transfer (KIDScoreD5 version 3) can predict the pregnancy rate of both day 5 and day 6 blastocyst transfers. METHODS: The relationship between KIDScoreD5 and clinical pregnancy rate was evaluated using the Cochran-Armitage test and receiver-operating characteristic (ROC) curve analysis. RESULTS: A positive correlation was observed between the KIDScoreD5 value and clinical pregnancy rate for both day 5 (p = 0.0003) and day 6 blastocysts (p = 0.0019) using the Cochrane-Armitage test. ROC curve analysis showed that the area under the curve (AUC) of KIDScoreD5 for clinical pregnancy was 0.627 (0.575-0.677, p < 0.0001) for day 5 blastocysts and 0.685 (0.571-0.780, p = 0.0009) for day 6 blastocysts. The combined analysis of both day 5 and day 6 blastocysts also showed an AUC of 0.680 (0.636-0.720, p < 0.0001), suggesting that it is possible to select embryos that are more likely to result in pregnancy. CONCLUSIONS: KIDScoreD5 could predict pregnancy not only in day 5 blastocysts but also in day 6 blastocysts. When both day 5 and day 6 blastocysts are vitrified, embryo selection by KIDScoreD5 is possible with a high prediction ability of pregnancy.