"Transforming Reproductive Medicine with AI"
Brief overview of the impact of AI on various fields, leading into its applications in reproductive medicine.
2. What is AI
Systems or machines that
mimic human intelligence to
perform tasks and can
iteratively improve
themselves based on the
information they collect.
3. What is ML
Machine learning (ML)
is a type of artificial
intelligence (AI) that
allows software
applications to become
more accurate at
predicting outcomes
without being explicitly
programmed to do so
4. What is DL
A type of machine learning
based on artificial neural
networks in which multiple
layers of processing are used
to extract progressively higher
level features from data.
5.
6.
7. Drawbacks of Human
Intelligence
• No one person is the
same
• Education
• Experience
• Interpretation
• Intuition
• Bias
• Physical State
• Mental State
• Emotional State
8. How AI is
Impacting our
Lives
Smart Phones
Security & Surveillance
Social Media Platforms
Navigation
E- Commerce
Banking & Finance Sector
Autonomous Vehicles
Smart House
9. Role of AI in Healthcare
Better
Prediction
Early
Detection of
Ailments
Improve
Decision
Making
Improve
Access to
Medical Care
Training Research
13. ART Software
• Covers the entire workflow of IVF treatments,
from the time of first appointment of the patient
until the final discharge summary
• Keeps the verification and matching data
• Matching of sperm and egg samples of patient
Concern about data protection and security
14. How can AI Help?
Access Treatment Protocol Gametes
15. Algorithm for Treatment Decision
Age AMH
Semen
Parameters
Duration of
Trying
Associated
Factors
Treatment
availed so far
TVS findings
Laws of the
Country
• Wait and Watch
• Medical
• Laparoscopy
• IUI
• IVF
• ICSI
• Third Party
• Stop Rx/Adoption
16. Access to
Treatment-
Algorithm
• Location
• Nature of work
• Time available
• Transportation
available
• Whether able to
travel alone/not
• Purely Online
• Some Offline visits
Needed
• How many visits for
treatment
• Nurse for Injections-
visits
17. Optimizing
ART Protocol
Formulation of an
IVF treatment
regimen
Dose Formulation
Follicular monitoring
and Endometrial
Receptivity
measurement
Egg/Sperm selection
Embryo Selection
18. ART Software for IVF Treatment Regimen
Age
AMH
BMI
Previous Treatment Protocol
FORT
No.of oocytes
OMR
• Which Protocol
• Which Gonadotrophin
• Dose of Gonadotrophin
• Adjuvants
• How long
• Which trigger
19. Imaging& AI in
RM
• Stored USG scan data
• Reduced examination time
• Reduces operator’s influence on scan
interpretation and objectivity
• Reduced interobserver variability
AFC
Uterine Anomalies
Follicle Monitoring
20. Endometrial
Receptivity
Measurement & AI
• Segmentation of region of
endometrium
• Classification of
endometrial pattern
• Estimating the accurate
motion of endometrium
• Assessing the blood supply
of endometrium
quantitatively
24. AI can help in TOS
Oocyte shape
Oocyte size
Ooplasm
characteristics
Structure of
PVS
Zona Pellucida
Polar body
25. Sperm Selection and AI
• Synchronization of CASA
systems with AI ART
software to permit to
select sperm for ART
• Provide faster and more
Precise Results
26. Male Infertility & AI possibilities
• Predict the likelihood of sperm extraction in
azoospermia
• Identification of sperm cells in microsurgical
testicular samples of patients
• Require a massive number of sperm images
for machine training in order to correctly
differentiate sperm from other tissue cells
• Rapid and applicable in real time.
27. Embryo
Selection & AI-
Embryoscope
• Non-invasive objective assessment of time-lapse
imaging of embryos and image data to predict
embryo development and implantation potential
29. • Khosravi et al. (2019)
• Analyzed more than
10,000 embryos
• Predict the blastocyst
quality with an
AUC>0.98
• Tran et al. (2019)
• Development of a deep
learning model for
automatically
recording
morphokinetic videos.
• Analyzed more than
10,000 videos
• Could identify images
of blastocysts that
yielded a fetal
heartbeat, with an
AUC> 0.90.
30. Zaninovic et al. (2018a)
• They evaluated 50,392
images of 10,148 embryos
grown in a Time Lapse
system, and found 97.52%
accuracy to classify good
and bad blastocysts.
Bori et al. (2020)
• KIDScoreD5 algorithm
classifies embryos into
categories based on the
points of cleavage times and
blastocyst appearance
• 22,461 embryos were
analysed
• Ability to distinguish
between embryos with
similar morphological
characteristics
31. PGTa & AI
• Pre-screen embryos and identify those with a
low probability of being genetically altered
• Therefore,only a few embryos would go to
PGT-A, preventing all embryos from having to
be biopsied
• The AI system was able to extract 94
features using a deep neuronal network to
predict euploid versus aneuploid BLs.
32. Cryopreservation & AI
• Quality control tool in
thawing embryos
• Monitoring embryo
culture systems
throughout the year.
33. AI in
Reproductive
Medicine
Important future in improving IVF success
Provides a single solution for the whole IVF
journey
Could improve success, reduce errors and lead to
faster, cheaper and more accessible results.
Help in research and industry to help understand
differences in embryo quality.
But AI cannot and should not replace
embryologists and clinicians
34.
35.
36. Thank You • Until reproductive specialists
adopt a common clinical language
and standard data acquisition
criteria, data mining cannot occur
to the degree required for off-
the-shelf ART applications.
• Reproductive specialists can begin
to standardize their systems as
collective knowledge grows.
• Comprehensive note-taking,
detailed outcomes reporting, and
routine collection of high quality
imaging, can accelerate this
innovation.
Gupta, Fertil Steril Dialogue 2020