AI has the potential to improve in vitro fertilisation treatments, offering new hope in the journey towards parenthood. — Photos: ART CHEN/The Star
In a fertility centre in Kota Damansara, Selangor, chief embryologist Tee Sze Tian meticulously examines images of embryos, a critical step in the in vitro fertilisation (IVF) process that helps couples struggling with infertility.
Statistics from the World Health Organization indicate that infertility affects one in every six adults, prompting many to consider IVF, where eggs and sperm are combined in a lab to produce embryos.
“My team and I will observe a single cell developing into a structured cluster, eventually forming a blastocyst – a crucial stage in IVF, indicating readiness for implantation in the mother’s uterus,” says Tee.
This is where life begins, from a single cell invisible to the naked eye. Tee illustrates this by making a tiny dot on paper, saying, “See, that was you.”
“The process has the potential to produce embryos – usually about five to eight – that can later be implanted into the uterus to help couples struggling with infertility conceive,” explains Tee, who has been working at TMC Fertility for more than 15 years.
Before implantation, embryos are placed on a small culture dish, roughly the size of a palm. Tee carefully positions the dish in a specially designed incubator equipped with a time-lapse imaging system, which includes a built-in camera and microscope. Over the next five days, the camera captures images of the embryos’ development every 10 minutes.
Tee and her team will meticulously analyse hundreds of thousands of images, observing each embryo’s shape and movements. Their goal is to identify the embryo with the highest potential to develop into a healthy pregnancy to give couples the best chance of achieving their dream of having a child.
Tee (standing) notes that AI scores embryos for viability, assisting the clinical team in making informed decisions without having to manually check each image. — ART CHEN/The Star
“We would have to click on every image one by one, run through videos, and hope that your eyes won’t feel tired because you don’t want to miss out on anything as we have to manually score the embryos on their potential,” Tee explains.
This is where artificial intelligence (AI) becomes an invaluable tool for embryologists like Tee.
Guided by data
Private fertility centres in Malaysia have been integrating AI technology into IVF treatments, using it to make data-driven decisions aimed at achieving the best possible pregnancy outcomes for hopeful couples.
According to the Malaysian Society of Assisted Reproduction Technology (MSART) immediate past president and fertility specialist Dr Navdeep Singh Pannu, AI is most commonly employed in embryo selection, where it is incorporated into time-lapse incubation systems.
Dr Navdeep says AI reduces human bias, making IVF embryo selection more objective and less reliant on subjective human judgment. — DR NAVDEEP SINGH PANNU
“The system helps to monitor the growth of the embryo, and the AI will then assist in determining which is the best embryo for eventual transfer,” he says.
Lakshmi Menon, CEO of TMC Fertility, says that the centre, which began exploring the use of AI in 2019, is also using the technology to assist in selecting the most viable embryos for implantation.
“The AI systems are trained on extensive datasets, including time-lapse imaging of embryos from previous successful pregnancies,” she explains.
“By analysing these historical records, the AI identifies subtle developmental patterns and characteristics that indicate embryo viability and potential for successful implantation.”
Tee explains that the system helps “score” the embryos in the incubator to indicate their viability, aiding the clinical team in making more informed decisions while reducing the strain of having to check every image manually.
“The AI is trained on a dataset of about 10,000 to 20,000 videos of embryos that have already resulted in pregnancy. For example, if we’re looking at this embryo here (on the screen), the AI has assigned a score of 9.1, which indicates it’s likely to result in a successful pregnancy,” she says.
Dr Navdeep claims that the use of AI helps reduce human bias: “This makes things more objective rather than leaving them to subjective methods because humans tend to have their own ways of looking at things and producing results.”
Tee agrees, saying that before the use of AI, she might feel more optimistic about certain images on a good day, while on bad days, she might view some images less positively or with more uncertainty.
“AI removes this variability, ensuring consistency and objectivity in the evaluation process. This also makes it easier for us to explain to patients about the embryos that are being recommended for implantation,” she adds.
Tee notes that patients are also becoming increasingly open to the role of AI in their IVF journey, with some even enquiring about its applications before it’s integrated into the assessment processes.
“They love it. We’ve met patients who have learned about the use of AI-assisted software through reproductive tech conventions and are mostly updated about upcoming technology mainly used abroad,” she adds.
The human touch
Despite the exciting prospects that technology holds, Lakshmi says embryologists remain integral in the embryo selection process, with them making the final decision.
While AI helps in choosing embryos with the best chance of success, Lakshmi notes that other factors — such as patient conditions and treatment plans — play a role in achieving a live birth. — ART CHEN/The Star
“Many IVF laboratory techniques still rely heavily on human expertise and experience, ensuring that the quality of care remains high. The AI also depends on accurate data entry by our team, reinforcing the importance of human oversight over these technologies,” she says.
She says the centre has strict protocols to ensure precise data entry through human verifications at critical checkpoints and that the team is trained to maintain meticulous records to minimise the risks of errors.
While AI helps in choosing embryos with the best chance of success, Lakshmi notes that many other factors play a role in achieving a live birth.
To improve outcomes, it’s crucial to address individual patient conditions with personalised treatment plans crafted through detailed evaluations.
Melaka-based fertility specialist Dr Pravin Peraba of Metro IVF says that apart from embryo quality, the uterus and hormone levels play a huge role in helping in achieving a successful pregnancy.
Dr Pravin points out that IVF success rates of over 80% usually refer to couples under 30, with reduced rates for those over 40. — DR PRAVIN PERABA
He says Metro IVF is currently in the process of vetting and evaluating the use of AI as part of fertility treatments.
“The real question is: Can AI help to increase pregnancy rates for IVF treatments? There have been small-scale studies done in Europe that showed a slight increase. But we can’t say if it has happened in Malaysia because no one has published their data,” he says.
Dr Pravin says the overall IVF success rate in Malaysia was about 30%, as announced by Health Minister Datuk Seri Dr Dzulkefly Ahmad during an event in April.
Private centres, however, may report higher success rates due to the use of more advanced equipment and medication, with his centre claiming a success rate of around 40% to 45%.
Dr Pravin notes that centres claiming a success rate of around 80% may be referring to couples under the age of 30.
When including statistics for couples over the age of 40, the overall success rate is likely to be lower, he adds.
If the first IVF cycle fails, couples may be advised to go through another cycle. In government hospitals, the cost for one IVF cycle is about RM3,000, while private clinics charge between RM10,000 and RM30,000. The emotional burden and high cost of treatment may result in patients having unrealistic expectations over the potential of AI-assisted technology in fertility treatments.
“I think doctors have to think carefully on this, especially when you consider the patient’s own IVF journey. If you can’t explain exactly how the AI is helping them in the process, there is a chance here that you’d end up building the patient’s hope too far,” he says.
Advancing through AI
Experts like Lakshmi feel that AI could play a broader role in the field of reproductive technology in Malaysia.
Statistics from the World Health Organization indicate that infertility affects one in every six adults, prompting many to consider IVF, where eggs and sperm are combined in a lab to produce embryos. — Image by freepik
She shares that TMC Fertility is exploring more ways to integrate AI into fertility treatment, including for improving personalised treatment planning.
“For instance, AI could optimise stimulation protocols that are used in IVF to stimulate the ovaries to produce multiple eggs by providing more precise dosages for injections and medications based on data-driven analysis of blood test results,” she explains.
Meanwhile, Dr Navdeep says AI could even help to assist in sperm selection through a tool called SiD.
“It’s a computer program that helps to assess sperm based on factors like movement or shape. It helps identify the most suitable to be fertilised with the egg, potentially leading to higher-quality embryos,” he adds.
Tee says she has heard such a feature is currently under development but remains cautious about its potential.
“Before we plan on integrating the feature into our process, we have to make sure that it is actually useful. Companies developing such a feature could make all sorts of claims.
“If I am not seeing an AI that could give me an accuracy of at least between 70% to 80%, then what is the point of charging patients extra for that?” she adds.
Tee encourages patients to be inquisitive about AI: “Patients should ask if the data has been validated and how. What is the accuracy rate? Ask about the model it was trained on. For example, if the dataset is from Europe, is it applicable for patients here? Regarding privacy, will your data be anonymised to ensure there are no identifying features?”
According to both Dr Navdeep and Dr Pravin, there are currently no local regulations on the use of AI in the reproductive field.
“The field is ever-evolving, and there are no guidelines regarding the use of AI yet. At the moment, humans are still very much involved in the process, and AI is being seen as helping to facilitate some processes, not replacing anyone,” Dr Navdeep says.
According to Dr Pravin, his patients in Melaka are also concerned if AI will affect them negatively or even physically.
“You can debunk that immediately by saying it’s not invasive at all because it’s just software,” he says.
Privacy is another concern they have shared, but the ultimate question is always about the effectiveness of the AI.
“At the end of the day, a patient’s goal is to go home with a baby. Though AI is going to be in every facet of our lives, the truth is, we are still vetting its potential,” he adds.
In retrospect, Dr Pravin says there is no denying how technology has improved the IVF process and success rate for patients over the years.
“The first IVF baby was born in 1978, and at the time, the success rate was said to be one in 100. Now, we are able to freeze embryos and perform more targeted procedures with better success rates than before,” he says, adding that AI could be particularly helpful in improving the success rates of patients with repeated failures.
Lakshmi says there are key factors to consider when it comes to the future of AI in the field: “Continuous research, robust data security measures, and patient education will be crucial to expanding AI applications in fertility treatments, ultimately driving better success rates and more positive patient experiences.”
Having personally gone through IVF with his wife, Dr Navdeep understands the disappointment faced by patients who don’t achieve their desired results.
“When we fail some patients, it’s a moment for me to find out how we can do better. I do believe that if we can use more technology in this field, then we will be able to improve outcomes for more patients,” he says.
