Purpose To compare the in-vitro fertilization (IVF) outcomes of cancer patients who underwent oocyte retrieval and embryo/ oocyte cryopreservation prior to gonadotoxic therapy to those of age and time-matched controls with tubal factor infertility. Methods All cancer patients who underwent embryo/oocyte cryopreservation at our institution from 1997 to 2014 were reviewed. Primary outcomes were total dose of gonadotropins used, number of oocytes retrieved, and number of 2pn embryos obtained. Outcomes were compared to age-matched controls with tubal-factor infertility who underwent a fresh embryo transfer within the same relative time period as the IVF cycle of the cancer patient. Results Sixty-three cancer patients underwent 65 IVF cycles, and 21 returned for frozen embryo transfer. One hundred twenty-two age-matched controls underwent IVF cycles with fresh transfer, and 23 returned for frozen embryo transfer. No difference was seen between cancer patients and controls with respect to total ampules of gonadotropin used (38.0 vs. 35.6 respectively; p=0.28), number of oocytes retrieved (12.4 vs. 10.9 respectively; p=0.36) and number of 2pn embryos obtained (6.6 vs. 7.1 respectively; p=0.11). Cumulative pregnancy rate per transfer for cancer patients compared to controls was 37 vs. 43 % respectively (p=0.49) and cumulative live birth rate per transfer was 30 vs. 32 % respectively (p=0.85). Cancer patients had a higher likelihood of live birth resulting in twins (44 vs. 14 %; p=0.035). Conclusions Most IVF outcomes appear comparable for cancer patients and age-matched controls. Higher twin pregnancy rates in cancer patients may reflect lack of underlying infertility or need for cancer-specific transfer guidelines.
BACKGROUND: Current morphological grading methods rely on descriptive parameters to rank cleavage-stage embryos for transfer. The limited measurements and overall subjective nature of these parameters likely contribute to the low positive predictive value (PPV: 0.3-0.4) reported in the literature for identifying cleavage-stage embryos that will develop into blastocysts. Advancements in the field of artificial intelligence (AI) and convolutional neural networks (CNN) have enabled for pattern recognition in images with unprecedented accuracy.OBJECTIVE: To develop CNN using embryo images to employ AI in predicting the fate of cleavage-stage embryos.MATERIALS AND METHODS: We used the GoogleNet Inception v3 CNN architecture [1] and replaced the final classification layer with an annotated dataset of embryo images captured at day 2 (D2, 44 hours) and day 3 (D3, 68 hours). The annotation of the data was based on embryonic developmental stage recorded on day 5 (D5).A retrospective dataset of 1,282 normally fertilized embryos cultured in an EmbryoscopeÔ was used in this study. A total of 10 embryologists annotated the dataset used in this study. We divided the set into 1100 embryos for training the algorithm and 182 embryos for testing. Accuracy of predicting stage of embryo development on D5 was established for D2 and D3 time points. The accuracy of blastocyst prediction of the algorithm's top choice embryo for each patient was also evaluated.RESULTS: The deep learning algorithm was able to predict whether or not an embryo developed to the blastocyst stage with an accuracy of 68.13% using a single image captured at D2. The system predicted successful blastocyst development with sensitivity of 74.16% (95% CI 65.38-81.72) and PPV of 76.72% (95% CI 67.97-84.07). When a single captured image at D3 was used, the accuracy of predicting blastocyst development using our system was 71.42%. The sensitivity of successful blastocyst prediction using D3 data was 75.83% (67.17-83.18) with PPV of 79. 82% (71.28-86.76). When applying this algorithm to identify the top quality embryo from among a patient's entire embryo cohort (n¼20 patients), we achieved 90% accuracy of selecting an embryo that developed to the blastocyst stage on D2 and 95% on D3.CONCLUSIONS: Here we report the first AI-based mechanism for predicting the developmental fate of cleavage-stage embryos. Additionally, we demonstrated that this technology might be used to select embryos with the highest in vitro developmental potential. Utilization of AI technologies in IVF practices may allow for more objective and standardized methods for improving embryo selection.
significantly increasing the laboratory workload. Further investigation of ploidy status from Day 7 embryos may explain the lower implantation rate.References: 1. Su, Yu, et al. Aneuploidy analysis in day 7 human blastocysts produced by in vitro fertilization. Reproductive Biology and Endocrinology 14.1 (2016): 20. 2. Li, M., et al. Day 7 blastocysts, should they be discarded or cryopreserved?. Fertility and Sterility 90 (2008): S427.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.