Using Artificial Intelligence (AI) in IVF is the newest way to harness technology and enhance the reliability and success rates of ART. So, how does AI help the IVF process?
It has been more than four decades since the first baby was born using IVF, yet IVF’s success rate remains uncertain. Despite all innovations in the field, there isn’t enough data on the use of AI in IVF to influence its usage in most clinics.
Today, a few AI-equipped IVF clinics have made successful application of this technology.
The use of artificial intelligence in assisted reproductive technology has been able to improve embryologists’ capacity cycle and chances of live birth more than earlier (1). With more development in such early detection tools and more data collection, the role of AI in IVF is likely to rise in the near future.
- When to start the stimulation process?
- What shall be the timing of the egg retrieval?
- Which is the best embryo to transfer?
In the traditional approach, the embryologist makes these judgments based on the manual inspection of images in the lab using techniques like microscopic/ time-lapse imaging. However, this leaves a possibility of error due to the use of different equipment and varying image quality.
If accurate decisions are not made at these leverage points, the success rate may drop drastically with each step. This also makes the standardization of the process quite challenging.
This is exactly when AI can help to minimize the chances of human error and allow for better clinical decision-making.
At what steps in IVF is AI useful?
Once the AI is trained, it can rapidly compute and draw accurate inferences for better IVF outcomes. Supervised AI approaches are data-driven. The method uses computer algorithms or image analytics to enhance embryology and the predict the effectiveness of embryo transfer.
At present, work is going on for using:
- AI in embryo selection,
- AI in gamete selection and best oocyte and sperm combination, and
- AI in designing a personalized fertility treatment regimen.
With the upgradation and integration of this new technology in the coming times, AI in assisted reproductive technology would be used by anyone who:
- Is undergoing IVF
- Has experienced implantation failures
- Has multiple embryos available for selection
With the use of machine learning, data is collected for the different types of traits in growing embryos from women who have undergone successful pregnancies.
The AI thus enhances the embryo selection process. It compares the known components with
the sample embryo to indicate its quality and viability that may be difficult to assess by the naked human eye.
After the assessment, an Implantation Potential Score (AI score) is provided to measure the viability of a successful pregnancy through IVF.
To summarize, here are the steps used in AI for embryo selection:
Step 1 – Microscopic images of each blastocyst are uploaded in the AI software.
Step 2 – An AI score is generated for each blastocyst. The score depends on factors like implantation potential and other factors that could result in a positive pregnancy.
Step 3 – The blastocyst with the highest AI score is chosen to transfer to the uterus of the mother.
One of the best potential uses could be semen analysis, where the motility, morphology, or vitality of the sperm can be better assessed with artificial intelligence.
AI removes the subjectivity of human assessment in the traditional process, and instead objectively ranks gametes based on quality. It will be especially helpful in the selection of sperm for ICSI.
AI can be used for data mining of existing patient records and discovering novel markers to accurately predict chances of pregnancy and live birth.
Interestingly, the use of artificial intelligence applied to text mining led to the discovery of novel genes involved in the pathogenesis, development, progression, and diagnosis of endometriosis, which is one of the leading causes of infertility in females in their reproductive age (2).
Benefits of using artificial intelligence in IVF
With recent advancements, artificial intelligence is used to bridge the gap between human errors and maximize the potential success rates of IVF.
- Though there exist techniques like embryo grading that may help in the analysis of good embryos, these AI-driven methods could really take it to the next level by helping your embryologists choose the ones that have the highest potential to result in a healthy baby.
- AI can help reduce IVF costs by reducing the number of IVF cycles that might be required to achieve clinical pregnancy.
- It can help identify the best embryo from the lot
- AI can assess whether the embryo should be transferred and what is the optimal time for it
- Saves time for both the patient and the doctor
- Provides more transparency in the IVF cycle
Currently, AI algorithms are being used by a limited number of labs for enhanced embryo selection.
In July 2022, an Israeli start-up ‘Fairtility’ became the first company to receive European Conformity CE Mark to use AI for assessing embryonic vitality.
Flairtility’s CHLOE EQ is an AI design that assesses and automatically selects embryos with high accuracy and also has ~73% fair agreement with the embryologists in a study of 193 cleavage embryos in their first cycle of IVF (3).
One of our partner clinics in Malaysia also became South-East Asia’s first to achieve successful pregnancy using artificial intelligence (AI)-enhanced embryo selection technology in March, 2021.
Though there are several benefits of using IVF intelligence, most small-sized fertility clinics are yet to avail it due to several challenges like limited availability of data, integration of current practices in the workflow, limited awareness amongst the stakeholders, etc.
AI can certainly help the IVF process become more affordable and efficient and we hope that the integration of this new technology will improve IVF’s long-standing stagnant success rate and your aspirations of becoming a parent.
If you have any inputs or questions on how AI helps the IVF process, please write them below:
- Darren J X Chow, Philip Wijesinghe,Kishan Dholakia,and Kylie R Dunning. Does artificial intelligence have a role in the IVF clinic? Reprod Fertil. 2021 Jul; 2(3): C29–C34. doi: 10.1530/RAF-21-0043
- J Bouaziz, R Mashiach, S Cohen, A Kedem, A Baron, M Zajicek, I Feldman, D Seidman, D Soriano, How Artificial Intelligence Can Improve Our Understanding of the Genes Associated with Endometriosis: Natural Language Processing of the PubMed DatabaseBiomed Res Int. 2018 Mar 20;2018:6217812. DOI: 10.1155/2018/6217812
- P. Fauque, J. Frappier, J. Barberet, A. Brualla, N. Bergelson, C. Hickman, Use of CHLOE-EQ to select embryos for transfer at the Cleavage stage: a pilot study using paired sibling embryos with known implantation, Volume 45, SUPPLEMENT 1, e47-e48, October 01, 2022. DOI: https://doi.org/10.1016/j.rbmo.2022.08.082