AI helps with an unfulfilled desire to have children
IVF treatment involves a great deal of effort and expense as well as physical and psychological stress for patients. Eggs are fertilized in a laboratory environment and the embryos are implanted into the uterus at the blastocyst stage (day 5 of development).
BirthAI is an AI-supported procedure for the quality assessment and classification of blastocysts. Embryologists can use it to insert the most suitable blastocyst into the uterus. This significantly increases the probability of pregnancy. [nbsp]
Together with the KUK Fertility Center at Kepler University Hospital, a new procedure has been developed to support IVF treatments using AI. The novelty of the procedure is the use of AI methods to assess the quality characteristics of the blastocysts, the so-called Gardner criteria. This is an evaluation scheme consisting of 3 criteria (cell expansion, inner cell mass, quality of the epithelial cell layer covering the blastocyst) which makes it possible to select the most suitable blastocyst for transfer. This increases the objectivity of the assessment and standardization can be achieved, regardless of the person making the assessment or the level of training. In contrast to other research approaches, which determine a direct ranking of blastocysts, the decision-making process remains interpretable on the basis of the Gardner criteria. This interpretability is additionally supported in our method by visualizations of the image regions used for the decision making of the AI method as well as by the calculation of confidence intervals that reflect the certainty of the decision. In order to evaluate and validate our method, a comprehensive image dataset was created. On the one hand, a training dataset was created and annotated by an experienced expert. On the other hand, a test dataset was generated and evaluated by an international panel of experts (12 experts from Europe and Egypt). This allowed the AI methods to be compared with the expert assessment, with the result that the developed method clearly outperformed the average expert assessment in all 3 Gardner criteria.
Services / Products:
Automated blastocyst classification for physicians, biologists and embryologists (prototype available)
The benefits of the new AI procedure for couples undergoing IVF treatment and for embryologists are clear:
Benefits for doctors:
- Optimal decision support for the medical team. The use of AI reduces subjectivity. Previously, blastocysts were analyzed under a microscope and the average consensus of a group of specialist embryologists decided on "suitable" and "less suitable" blastocysts with a certain degree of subjectivity.
- Smart artificial intelligence will also be a kind of sparring partner for embryologists in the future. Above all, the time needed to train the experts could be significantly reduced.
- Increase in training data for more accurate forecasts
Benefits for the couples:
- Increased chances of a successful pregnancy
- Cost minimization for the couple, as fewer transfers are necessary and the pregnancy occurs more quickly.
- Less psychological stress for the woman/couple. Every attempt without a resulting pregnancy is a high psychological burden
- Less physical stress for the woman. (hormone administration, etc.)
Founder:
- Markus Manz
Awards:
- Science Corner at the Vienna Museum of Technology
- eAward
- Upper Austrian Innovation Award
