Enhancing IVF Success: Deep Learning for Accurate Day 3 and Day 5 Embryo Detection from Microscopic Images

Abstract

Embryo selection is a critical factor in the success of in vitro fertilization procedures, directly influencing the likelihood of a successful pregnancy. Accurate classification and detection of embryos at key developmental stages, particularly on Day 3 (cleavage stage) and Day 5 (blastocyst stage), are essential for making informed decisions regarding embryo transfer. Traditional methods of embryo assessment rely on manual morphological evaluation, which can be subjective and prone to inter-observer variability. To address these limitations, this study investigates the application of advanced deep learning models for the automated detection of embryo state using microscopic images captured on Day 3 and Day 5. We evaluated five state-of-the-art convolutional neural network (CNN) architectures—VGG19, DenseNet, ResNet50, InceptionV3, and EfficientNetV2—based on their performance in distinguishing between ‘Day 3’ and 'Day 5' embryos. The results indicate that EfficientNetV2 outperformed the other models, achieving the highest accuracy (94.34%), precision (93.29%), and recall (94.31%). This superior performance suggests that EfficientNetV2 is the most reliable model for embryo state classification, offering the potential to significantly enhance the accuracy and consistency of embryo selection in IVF clinics.

References

[1] Bódis J, Gödöny K, Várnagy Á, Kovács K, Koppán M, Nagy B, Erostyák J, Herczeg R, Szekeres-Barthó J, Gyenesei A, Kovács GL. How to reduce the potential harmful effects of light on blastocyst development during IVF. Medical Principles and Practice. 2020 May 29;29(6):558-64.

[2] Mastenbroek S, Van Der Veen F, Aflatoonian A, Shapiro B, Bossuyt P, Repping S. Embryo selection in IVF. Human Reproduction. 2011 May 1;26(5):964-6.

[3] Ziaei R, Ghasemi-Tehrani H, Movahedi M, Kalatehjari M, Vajdi M, Mokari-Yamchi A, Elyasi M, Ghavami A. The association between Diet Quality Index–International score and risk of diminished ovarian reserve: a case–control study. Frontiers in Nutrition. 2023 Nov 29;10:1277311.

[4] Allameh Z, Hajiahmadi S, Adibi A, Ebrahimi Oloun Abadi Z, Mahmoodian Dehkordi S. Diagnostic Value of Ultrasonography and MR in Antenatal Diagnosis of Placenta Accreta Spectrum. Journal of Fetal Medicine. 2020 Dec;7:275-81.

[5] Bori L, Meseguer F, Valera MA, Galan A, Remohi J, Meseguer M. The higher the score, the better the clinical outcome: retrospective evaluation of automatic embryo grading as a support tool for embryo selection in IVF laboratories. Human reproduction. 2022 Jun 1;37(6):1148-60.

[6] Sherkat R, Shahshahan Z, Kalatehjari M, Yaran M, Nasirian M, Najafi S, Zangeneh NP, Montazerin SM. Cytomegalovirus specific cell-mediated immunity status in women with preeclampsia: A case-control study. Advanced Biomedical Research. 2023 Jan 1;12(1):10.

[7] Wu C, Yan W, Li H, Li J, Wang H, Chang S, Yu T, Jin Y, Ma C, Luo Y, Yi D. A classification system of day 3 human embryos using deep learning. Biomedical Signal Processing and Control. 2021 Sep 1;70:102943.

[8] Milewski R, Milewska AJ, Czerniecki J, Leśniewska M, Wołczyński S. Analysis of the demographic profile of patients treated for infertility using assisted reproductive techniques in 2005-2010. Ginekologia Polska. 2013;84(7).

[9] Afrazeh F, Ghasemi Y, Shomalzadeh M, Rostamian S. The Role of Imaging Data from Different Radiologic Modalities During the Previous Global Pandemic. International Journal of Applied Data Science in Engineering and Health. 2024 Jul 27;1(1):9-17.

[10] Zaninovic N, Rosenwaks Z. Artificial intelligence in human in vitro fertilization and embryology. Fertility and Sterility. 2020 Nov 1;114(5):914-20.

[11] Litjens G, Kooi T, Bejnordi BE, Setio AA, Ciompi F, Ghafoorian M, Van Der Laak JA, Van Ginneken B, Sánchez CI. A survey on deep learning in medical image analysis. Medical image analysis. 2017 Dec 1;42:60-88.

[12] Shao D, Dai Y, Li N, Cao X, Zhao W, Cheng L, Rong Z, Huang L, Wang Y, Zhao J. Artificial intelligence in clinical research of cancers. Briefings in Bioinformatics. 2022 Jan;23(1):bbab523.

[13] Zadeh FS, Molani S, Orouskhani M, Rezaei M, Shafiei M, Abbasi H. Generative Adversarial Networks for Brain Images Synthesis: A Review. arXiv preprint arXiv:2305.15421. 2023 May 16.

[14] Kumar R, Kumbharkar P, Vanam S, Sharma S. Medical images classification using deep learning: a survey. Multimedia Tools and Applications. 2024 Feb;83(7):19683-728.

[15] Gong Y, Qiu H, Liu X, Yang Y, Zhu M. Research and Application of Deep Learning in Medical Image Reconstruction and Enhancement. Frontiers in Computing and Intelligent Systems. 2024 Apr 10;7(3):72-6.

[16] Yang HY, Leahy BD, Jang WD, Wei D, Kalma Y, Rahav R, Carmon A, Kopel R, Azem F, Venturas M, Kelleher CP. BlastAssist: a deep learning pipeline to measure interpretable features of human embryos. Human Reproduction. 2024 Apr 1;39(4):698-708.

[17] Sharma A, Dorobantiu A, Ali S, Iliceto M, Stensen MH, Delbarre E, Riegler MA, Hammer HL. Deep learning methods to forecasting human embryo development in time-lapse videos. bioRxiv. 2024:2024-03.

[18] Garg K, Dev A, Bansal P, Mittal H. An Efficient Deep Learning Model for Embryo Classification. In2024 14th International Conference on Cloud Computing, Data Science & Engineering (Confluence) 2024 Jan 18 (pp. 358-363). IEEE.

[19] Simonyan, K., & Zisserman, A. (2014). Very deep convolutional networks for large-scale image recognition. arXiv preprint arXiv:1409.1556.

[20] He, K., Zhang, X., Ren, S., & Sun, J. (2016). Deep residual learning for image recognition. In Proceedings of the IEEE conference on computer vision and pattern recognition (pp. 770-778).

[21] Huang G, Liu Z, Van Der Maaten L, Weinberger KQ. Densely connected convolutional networks. InProceedings of the IEEE conference on computer vision and pattern recognition 2017 (pp. 4700-4708).

[22] Xia, X., Xu, C., & Nan, B. (2017, June). Inception-v3 for flower classification. In 2017 2nd international conference on image, vision and computing (ICIVC) (pp. 783-787). IEEE.

[23] Mingxing. Tan, Quoc V. Le. Efficientnetv2: Smaller models and faster training. In International conference on machine learning, pp. 10096-10106. PMLR. 2021, July.

[24] https://www.kaggle.com/datasets/gauravduttakiit/embryo-classification-based-on-microscopic-images