AI Could Improve the Success of IVF Treatment

During IVF treatment, doctors use ultrasound scans to monitor the size of the follicles – small sacs in the ovaries that contain eggs – and to decide when to administer a hormone injection known as a trigger, to prepare the eggs for retrieval and ensure they are ready for fertilization with sperm to create embryos. The timing of the trigger is an important decision because it is less effective if the follicles are too small or too large at the time of administration. After the eggs are retrieved and fertilized with sperm, one embryo is selected and transferred to the uterus to hopefully establish a pregnancy.

Follicle Size and Fertility Treatment

Researchers used explainable artificial intelligence (AI) techniques – a type of AI that allows humans to understand how it works – to analyze retrospective data from more than 19,000 patients who had completed IVF treatment. They looked at which follicle sizes were associated with improved rates of mature egg retrieval, resulting in the birth of babies. They found that administering the hormone injection when a greater proportion of follicles were between 13 and 18 mm in size was associated with higher mature egg retrieval rates and improved birth rates.

Currently, doctors use ultrasound scans to measure the size of the (largest) follicles and usually give the trigger injection when a threshold of two or three follicles is reached that are larger than 17 or 18 mm. Their results suggest that maximizing the proportion of medium-sized follicles could optimize the number of mature eggs retrieved and improve birth rates.

The team believes that the results of the study demonstrate the potential of AI to support the personalization of IVF treatment to improve clinical outcomes for patients and maximize their chances of taking a baby home. The team plans to develop an AI tool that will use the results of their research to personalize IVF treatment and support clinicians’ decision-making at every stage of the IVF process. They will apply for funding to test this tool in clinical trials.

Dr. Ali Abbara, NIHR clinician scientist at Imperial College London and Consultant in Reproductive Endocrinology at Imperial College Healthcare NHS Trust, and co-senior author of the study, said: “In vitro fertilization offers help and hope to many patients who cannot get pregnant, but it is an invasive, expensive and time-consuming treatment. If it fails, it can be very painful. That is why it is important to ensure that this treatment is as effective as possible.” AI can provide a new paradigm for conducting IVF treatments and lead to better patient outcomes.

Trigger Injection

There is so much extensive data generated during an IVF treatment that it can be difficult for doctors to fully utilize it when making treatment decisions for their patients. The researchers’ study showed that AI methods are well suited to analyzing complex IVF data. In the future, AI could be used to provide accurate recommendations to improve decision-making and support the personalization of treatment, so that we can give every couple the best possible chance of having a baby. This study is the first to analyze a large data set to show that AI can more accurately identify the specific follicle sizes most likely to yield mature eggs than current methods.

Infertility affects one in six couples, and in vitro fertilization has proven to be a valuable measure to help patients conceive. An important decision in IVF treatment is the timing of the “trigger” injection of hormones such as human chorionic gonadotropin (hCG) to mature the eggs for retrieval. The timing of the triggering injection affects the number of mature eggs retrieved and the success of the treatment. Doctors use ultrasound scans to measure the size of the (largest) follicles. They usually administer the triggering shot when a threshold of two or three follicles with a diameter of more than 17 or 18 mm is reached. However, this method is not precise enough and does not take into account the size of each individual follicle and the likelihood of each follicle producing a mature egg.

How AI is Helping to Improve the Efficiency of IVF Treatment

In the retrospective study, the team used AI techniques on data from 19,082 patients aged between 18 and 49 years who were treated between 2005 and 2023 at 11 clinics in the UK – including IVF clinics at Imperial College Healthcare NHS Trust – and two in Poland. They examined the size of individual follicles on the days leading up to and on the day of trigger administration. The researchers found that follicles measuring 13–18 mm were more likely to result in the later retrieval of more mature eggs. The data suggested that a greater number of follicles in this range on the day of triggering was associated with better clinical outcomes.

They also found that ovarian stimulation that was too long, in which there were a greater number of larger follicles (over 18 mm) present on the day of trigger administration, can lead to a premature increase in the hormone progesterone. This can negatively impact IVF outcomes by interfering with the proper development of the endometrium – the tissue that lines the uterus and is important for allowing implantation of the embryo to result in a pregnancy. This reduces the likelihood of an embryo implanting and subsequently resulting in a live birth. These insights gained from AI could help the team develop evidence-based IVF protocols that are guided by data and designed to improve treatment efficiency.