This Israeli Startup is Using Deep Learning and Computer Vision to Revolutionize IVF Treatments

IVF, In-Vitro Fertilization, the process of combining a female egg and male sperm outside the human body and implanting the resulting embryo into the uterus, is the most common method today to tackle fertility issues. Despite that, it's still not the ideal solution to the world's growing reproductive woes. IVF treatment's global success rates range between 35% and 40%, and the cost can often reach $100,000 to go through the stressful conception cycles, exhausting couples both emotionally and financially. This low success rate is largely attributed to the inherent challenge involved in optimal embryo selection: choosing which embryo has the highest likelihood to be implanted successfully. Embryologists make these decisions based on subjective visual criteria which are subject to a great degree of human error. A new Israeli startup, AIVF, was founded last year to bridge this gap by developing an artificial intelligence (AI) based system for the embryo selection process. [caption id="attachment_53514" align="alignnone" width="2000"] AIVF's team at their offices located in Tel Aviv.[/caption] AIVF was founded by Daniella Gilboa, Prof. Daniel Seidman, and Prof. Eyal Schiff in 2018 to provide embryologists with an end-to-end decision support tool using the power of deep learning and computer vision algorithms. Gilboa and her team at AIVF are developing a system to optimize embryo selection. [caption id="attachment_53515" align="alignnone" width="2000"] AIVF's co-founders Prof. Daniel Seidman and Daniella Gilboa (left to right).[/caption] “AIVF’s decision support system won’t replace embryologists, but rather automate one of the processes in their workflow” explained Gilboa. This would increase success rates, reduce associated costs of the IVF lab, and increase lab capacity. With over 2.5 million yearly IVF treatments administered globally, and over $15 billion market size, IVF treatment poses a major opportunity for the AI-Healthcare revolution. In fact, “IVF treatment is becoming a common practice in today’s reality, where 12% of the world population struggle to conceive naturally” explained Prof. Seidman. [caption id="attachment_53517" align="alignnone" width="2000"] Prof. Daniel Seidman operating the embryo incubation machines with time-lapse functionality video recording.[/caption] The intersection of AI and IVF originates back to 2010 when the amount of visual data on embryos skyrocketed. Hospitals adopted new incubation equipment equipped with a special camera capable of recording time-lapse videos of the developing embryo. More importantly, the new time-lapse videos maintained records of the embryo’s success or failure in generating pregnancy and eventually a liveborn. Moreover, in the case of Israel, the government subsidizes IVF for its citizens, turning Israel into an IVF empire of treatments, and consequently has an abundance of data. It is through a partnership with Israeli private hospital Assuta, as well as eight other clinics worldwide, that AIVF currently holds the largest dataset of over 250,000 time-lapse videos of developing embryos and expects to reach over 300,000 videos by the end of the year. [caption id="attachment_53512" align="alignnone" width="1966"] AIVF's deep learning and computer vision algorithms applied to time-lapse videos and stills of embryo development with proprietary markers and identifiers. Photo: Courtesy.[/caption] With these time-lapse videos, the company is developing AI algorithms using a combination of deep learning and computer vision methodologies. Using a combination of AI methodologies, like convolutional neural networks to proprietary image processing systems designed for the unique characteristics of these relatively new time-lapse images, AIVF is at the forefront of the application of technology to revolutionize IVF process. The startup is planning on starting clinical trials in 2020, with plans to receive regulatory approval for usage of the system in clinics in 2021. [caption id="attachment_53518" align="alignnone" width="2000"] Embryo storage units.[/caption] Clinicians around the world are realizing the potential of AI in improving medical outcomes, and raising IVF success rates is one of the most promising applications. Over 400,000 children are born through IVF each year and the number is constantly growing. “We aim to make IVF treatment more affordable, more successful and more accessible to all couples seeking to conceive” said Gilboa.