The “Dr. Fertilidad” podcast video presents a discussion between the host Jorge and biologist Isabel Puerta from Future Fertility about the application of artificial intelligence (AI), specifically “deep learning,” in assessing oocyte quality in assisted reproduction.
The key points of the conversation include:
Subjectivity in oocyte evaluation: It is mentioned that the assessment of oocyte quality used to be based on the subjective criteria of biologists and doctors, which resulted in imprecise classifications.
Application of Deep Learning: Isabel explains that “deep learning” allows a model to learn to identify patterns in thousands of oocyte images to predict their potential to develop into a blastocyst. This model is trained with data from various clinics to ensure its generalizability.
Degree of certainty: The current model has an “area under the curve” (AUC) of 0.69, which represents a 20% improvement compared to human evaluation.
Practical applications:
Social vitrification: It is useful for young patients freezing their eggs, providing an objective assessment to manage expectations and plan treatments.
Egg banks: It allows banks to assess the quality of donor eggs to ensure consistency and high quality.
Diagnosis of cycle failures: It helps identify if poor oocyte quality is the cause of a failed cycle.
Group culture: It facilitates the grouping of higher-scoring oocytes for cultivation.
Euploidy prediction: Oocyte quality scores correlate with a higher probability of euploidy in the blastocyst.
Patient reaction: Patients react positively to the report, as it provides them with a tangible and objective understanding of their egg quality.
Future of Future Fertility: The company aims to expand the report to include the probability of euploidy and explore other AI applications in assisted reproduction.
