Page 14 - GTM-2-2

Global Translational Medicine Clinical algorithms in ART

predict which embryos are most likely to develop treatment recommendations, optimizing treatment
successfully, allowing embryologists to prioritize protocols, and predicting responses to treatment.
these embryos for transfer and reducing the time However, these approaches are still in the early stages of
required for manual embryo selection . development, and further research is needed to validate
[39]
ii. Embryo scoring algorithms: Embryo scoring their effectiveness in clinical practice.
algorithms use machine learning techniques These new instruments are improving the experiences
to predict the likelihood of successful embryo of both clinicians and patients, with a larger applicative
implantation based on a range of factors, such as in gamete, embryo selection, and egg/embryo storage.
morphological characteristics and developmental Nevertheless, the application of AI in the IVF laboratory
stage. These algorithms can prioritize embryos for to streamline patient care is a growing but not yet fully
transfer and reduce the time required for manual realized concept. This is why we still believe that, in
embryo selection [39,40] . clinical rather than laboratory, AI applications are more
iii. Fertilization prediction algorithms: Fertilization immediately useful in medical practice for ART. The
prediction algorithms use patient-specific data, question of accurately estimating the overall probability of
such as age, hormonal levels, and sperm quality, a medical outcome resulting from two independent events
to predict the likelihood of successful fertilization. still remains for clinicians. While it is possible to do so in
These algorithms can optimize the timing of certain cases, such diagnostic and prognostic decisions
procedures and reduce the time required for often require the consideration of multiple probabilities or
manual monitoring and intervention . steps. However, in cases where multiple independent events
[8]
iv. Quality control algorithms: Quality control are involved, the misestimation of the overall probability of
algorithms use statistical techniques to monitor success (known as the conjunction fallacy) is likely to lead
laboratory performance and ensure the safety to diagnostic and prognostic errors .
[32]
and accuracy of procedures. These algorithms can
detect anomalies and deviations from established That happens in the diagnosis and treatment of RIF
protocols and alert laboratory staff to potential where several factors are concurring to the failure, one
[43]
issues before they affect outcomes . independent from the other .
[12]
v. Cryopreservation algorithms: Cryopreservation The diagnostic tool by questionnaire of Chapron
algorithms use machine learning techniques to et al. [21] is a real improvement of the clinical approach to
predict the likelihood of successful embryo or gamete this disease because it is able to shorten the first diagnosis
cryopreservation based on a range of factors, such and accelerate the possible treatment with an improved
as age, hormonal levels, and clinical history. These prognostic value for prospective fertility treatment results
algorithms can optimize the timing and methods of in those women (Table 2). The assessment of male and
[25]
cryopreservation and reduce the time required for female [26-29] gametes with specific automatic prognostic
manual monitoring and intervention [41,42] . grading allocation according to the production and/or
However, even if these algorithms could still improve reservoir is a substantial help in the clinical management
the laboratory efficiency of IVF programs, reducing the of the infertile couple without adding some new potential
time required for manual procedures, optimizing the use improvement in pregnancy outcome (Table 2). The starting
of resources, and improving outcomes for patients, they dose of gonadotropins in the COS represents a great help in
still do not exhibit evidence of an improvement in the the same management, improving the clinical effort in that
procedure’s efficiency. By integrating these algorithms management [28,29] , as well as the trigger timing calculation
into clinical practice, IVF programs could achieve that avoids mistakes and delays or anticipation leading
higher success rates and provide safer and more effective to reduced oocyte pick up (OPU) efficiency (Table 2).
treatment. But still, randomized and controlled trials are There is still discussion about the efficiency of picking up
needed for full validations. and fertilizing the maximal oocytes as possible for each
aspiration or performing mild stimulation and partial
The use of AI in infertility treatment can improve the fertilization of oocytes. Thus, the model for calculating the
accuracy, efficiency, and success rates of reproductive magic number to maximize results and reduce the number
medicine, leading to better patient outcomes and increased of oocytes to pick up remains to be established for efficiency
access to fertility care. We reported available tools and contribution and the use of this algorithm, but it is useful
algorithms focused to personalize medical efforts. to follow the concern of clinicians that refuse to collect
Specific algorithms have the potential to improve the as many embryos as possible in the perspective of better
effectiveness and safety of COS by providing personalized efficiency of this procedure (Table 2). The ART calculator
[44]

Volume 2 Issue 2 (2023) 8 https://doi.org/10.36922/gtm.0308