Artificial intelligence is increasingly being used to refine various aspects of medical care, including assisted reproductive technology (ART). One researcher, Akriti Sharma, is at the forefront of this wave, exploring how AI can improve the selection of embryos during fertility treatments.

Akriti Sharma’s doctoral research at OsloMet,leveraging data from Volvat Spiren fertility clinic in Oslo,focuses on creating AI models to aid embryologists. These models are designed to assess and select the most viable embryos for transfer to the uterus.

“The main goal was to develop AI models that can help embryologists in their daily tasks and provide better assessments of embryos”, explains Sharma.

An embryo represents the earliest stage of human development, lasting from fertilization to approximately eight weeks, after which it is indeed considered a fetus.

Current AI Embryo Tools: High Cost,Limited Insight

Existing AI tools for embryo assessment are often costly due to licensing fees and provide limited insight into the AI’s decision-making process,according to Sharma.

“AI-driven tools can automate embryo assessment but are too expensive for many fertility clinics”, she notes.

Many medical professionals are hesitant to adopt these technologies because of the lack of transparency in how the AI arrives at its conclusions.

The proprietary nature of these commercial AI tools means that the rationale behind their assessments is often hidden, hindering trust and understanding.

Sharma aims to develop AI models that are more accessible, transparent, and easily understood by medical professionals.

“AI can automate low-risk tasks, enabling embryologists to focus on more complex decisions. This can make the fertility process faster and more efficient.”

How AI Enhances Assisted Reproductive technology

Sharma provides a straightforward description of how AI integrates with assisted reproductive technology:

“In assisted reproductive technology, eggs and sperm are fused outside the body in a controlled environment. AI can assist medical experts by providing an automated assessment of embryo quality.”

AI’s ability to analyze large datasets allows it to identify patterns that might be imperceptible to the human eye.This is achieved through:

• Analyzing videos or images to evaluate an embryo’s quality based on its changes in shape and structure over time.
• Identifying patterns that typically lead to successful implantations and pregnancies.
• Combining patient data, such as age and lifestyle habits, with embryo morphokinetics, like cell division timing.

The Advantages of Explainable Artificial Intelligence (XAI)

Sharma utilizes explainable artificial intelligence (XAI) to ensure that AI models are more easily understood by embryologists.

“The intention of explainable AI is to clarify how AI arrives at decisions. Explainable AI was found to function as a ‘debugging’ tool, allowing us to validate whether the AI models consider correct patterns.”

Explainable AI clarifies the decision-making process of AI models, making it easier to identify errors or incorrect patterns. It also ensures that the model focuses on biologically relevant data.

When embryologists understand how an AI model makes decisions, they can adjust the model to improve its accuracy and reliability.

Building Trust Through Transparency

Explainable AI enhances trust in AI technology by making it easier for developers to optimize models and for embryologists to trust AI-driven decisions. This trust is essential for the effective use of AI in critical areas.