Nitiz Murdia, Director and Co-founder, Indira IVF highlights the role of technology in treating reproductive disorders
Artificial Intelligence (AI) refers to a computer programme’s capacity to accomplish activities or thinking processes that we normally associate with human intellect. AI, which is based on automation, has the potential to transform healthcare and help address some major health issues. It has the potential to enhance patient outcomes as well as the productivity and efficiency of healthcare delivery.
In the diagnosis and treatment of reproductive disorders, AI can be used in different ways from analysing data on patient visits to the clinic, medication prescribed, lab tests, and procedures performed. It has proven to be a highly effective tool for finding the best methods and outcomes, and determination of the optimal course of treatment for a patient.
Analysis of sperm quality and motility is a vital part of Assisted Reproductive Techniques (ART) pregnancies which helps in determination of the best approach for that couple. Because of the lack of objectivity in the manual evaluation of sperm morphology and the different levels of laboratory competence, the computer-aided sperm analysis (CASA) systems now in use for this purpose are unable to perform optimally. Even in situations of idiopathic male infertility, which the existing framework of examination fails to assess, having an AI-enabled platform can aid in obtaining more exact and objective results.
While women under 35 have a better than 20 per cent probability of having a full term, normal birth weight, and singleton live birth every cycle, women 35 to 37 years old have a 17 per cent chance, and the percentage diminishes with age.[1] One of the most common ways of infertility treatment is in vitro fertilisation (IVF). The most crucial element of the IVF procedure is embryo assessment and selection. Its goal is to choose the best embryos from a larger group of fertilised oocytes, the majority of which will be found to be unviable owing to aberrant development or genetic abnormalities. In such circumstances, embryo selection becomes critical, and AI simplifies the process.
Automatic annotation of embryo development, embryo grading, and embryo selection for implantation are the three categories of AI applications in embryology. Usually, the embryo selection procedure is based on the morphology and photos that embryologists record of that specific embryo, and the embryos are graded based on the morphology. Due to manual intervention, these grades will vary from individual to individual, laboratory to laboratory, and place to place, and errors in the process are inevitable. This is where standardising and automating the process with AI becomes critical.
An embryologist with the experience of a particular number of years has limited knowledge but the AI predictor contains lakhs of images in its database. It can make more accurate predictions and has more experience than any doctor. One person may have lived through it for 10 years, but the AI will have 150 years of cumulative experience. AI has the potential to improve clinical effectiveness and efficiency, therefore improving the ART treatment cycle.
Machine learning needs a lot of data and processing large amounts of data is now feasible because to the fast development of graphics processing units (GPUs). AI and machine learning can substantially accelerate the development of reproductive medicine in the near future because of the steady improvement of hardware and software.
For ovary and uterus evaluation by feeding an ultrasound picture into an AI system, AI can assist anticipate the right gonadotropin levels in the ovary. In the same way that cancer therapy predicts the volume of the ovary, it suggests a certain dose of gonadotropin injection. The procedure stimulates the ovaries to generate eggs in the patient’s desired quantity.
AI is also helpful to a great extent in prevention of miscarriages. By reversing the data input before implantation, AI models are utilised to find the embryo with the highest risk of miscarriage. Because it is done on a limited set, the method has a 77 percent accuracy in preventing faulty embryos. Despite having previously unsuccessful implantation circumstances, the technique of Preimplantation Genetic testing (PGT) allows people to produce a healthy child. It is a cutting-edge technique that examines chromosomes and removes defective embryos.
In the realm of reproductive science, AI is progressively becoming a useful complement to traditional methods of assessing and treatment of reproductive disorders. It has been utilised from clinical monitoring, determination of best timings, to prediction of pregnancy outcomes; although the adaptation of such technology is quite slow. However, adaptation of AI has the potential to overcome the constraints of cost, access, and low success rates which can prove to be highly beneficial for patients in a country like India.
Reference
[1] https://www.pennmedicine.org/updates/blogs/fertility-blog/2018/march/ivf-by-the-numbers