Empower your decisions
iDAScore - Intelligent Data Analysis for embryo evaluation
Automated analysis at your fingertips
iDAScore makes use of deep learning to provide fully automated analyses of developing embryos. Now you can get an objective ranking based on the likelihood of implantation of each of a patient´s embryos at once - at the touch of a button. The future of embryo evaluation begins here.
Traditional embryo evaluation methods are prone to subjectivity and external factors. Experience level, time constraints and lab staffing may all affect evaluation of embryos. It has been shown that an embryologist may score the same blastocyst differently on separate occasions. 1,2 iDAScore objectively “compares” a given embryo with other embryos that have similar development patterns, and generates a score correlating with the likelihood of implantation.
AI scoring is not affected by
Simplicity in action
iDAScore is simple and intuitive, with no manual export or import of images or videos required.
Get a glimpse of the iDAScore interface by watching the trailer.
"With just a touch of a button, it provides us with scores instantly. It is impressive and this helps us improve our workflow and efficiency dramatically as there will be no need for annotations. It is a great tool for improving consistency in our evaluations as well.”
-Hana Watanabe, Embryologist, Hiroshima HART Clinic, Japan
Empower your decisions
The iDAScore algorithm analyses each of the patient’s embryos and provides a score correlating with the likelihood of implantation. Output scores are automatically generated for all embryos in a consistent and objective manner.
Consistent approach to embryo
evaluation. Day to day differences
in staffing, or introduction of new
embryologists will not affect the
way embryos are scored.
Objectively analyses the full embryo
development sequence, without
inherent human bias.
Analyses time-lapse sequences
continuously, without the need
for manual processing of data or
Provides a reliable ranking of a
patient’s cohort of embryos. Embryos
can be evaluated adjunctively with
EmbryoViewer software or other well
established scoring schemes.
How iDAScore works
“With a clear easy to use interface iDAScore makes choosing between multiple blastocysts an easier job. The score matches well with the decisions of experienced embryologists, and KIDScore D5. We think that in time, iDAScore will become an invaluable tool in helping to make our day to day clinical decisions. In addition, we are hoping to use it to assess already frozen embryos which were cultured in the EmbryoScope+. ”
-Alan Birks, Senior Clinical Embryologist
Manchester Fertility, UK
"iDAScore provides an impressive ranking of embryos. We see that the correlation with implantation, shows the same positive trend as with the KIDScore D5 decision support tool. This could be a game changer in the future of embryo evaluation, as we gain more trust in using AI for embryo evaluation.”
-Tony Price, Embryology Manager
Wessex Fertility, UK
“iDAScore is an ideal tool for providing a fast prioritization of embryos for further examination. This is especially useful for guiding new embryologists and helping them to evaluate embryos in a way that is more consistent with evaluation by senior staff with many years of experience.”
-Kirsten Simonson, Clinical embryologist
Maigaard fertility, Denmark
BUILDING ON THE POWER OF AI
iDAScore was trained on 115,000 full time-lapse sequences of embryos. Of these 14,644 were transferred 4,337 were Fetal Heart Beat positive. Data was from 18 clinics and included a wide range of ages as well as including both fresh and frozen transfers. Validation was performed by having hold out data sets that were not included in the training and were used for validation
* KID Embryos: Embryos with Known Implantation Data
A robust algorithm
iDAScore was trained and validated on a large and diverse dataset, resulting in a robust AI model that generalises well to clinics whose data was not included in the training process. Automatic embryo scoring by AI shows convincing results for all provided subgroups of data and can be expected to generalise well to all clinics.
To test how well the model generalises, a clinic hold out test* was performed3. Data from 18 clinics was used, where data from 1 clinic was excluded from training and the AUC* for the hold out clinic was calculated. Data from the 12 clinics contributing the most data is shown. AUC* when considering all embryos exceeded 0.9 for the hold out clinics.
The algorithm was tested on data across age groups3. We see that iDAScore performs equally well at ranking all embryos in all of the age groups tested. AUC* in the 40+ category is slightly lower, likely due to patient factors.
Different clinics may have different distributions of the proportion of IVF or ICSI insemination. In this study3, all data with associated labelling of insemination type was analysed. This shows equal performance regardless of insemination method.
Time of culture was also shown to provide similar results3. Data was subdivided into 4 subgroups, 1<15 hpi, 115-117 hpi, 117-119 hpi and >119 hpi to get a more detailed view. This confirms that clinics have a flexible time window to perform analysis.
Presentations from our scientific symposiums:
'The Future is here: Starting the journey with AI based embryo evaluation' at IVF WW Congress 2021 &
'AI based embryo evaluation: Empower your decisions' at ESHRE 2020
The Future is here: Starting the journey with AI based embryo evaluation
- Simon Cooke, Scientific Director of IVF Australia
- Jens Rimestad, Deep Learning Specialist at Vitrolife
- Anthony Price, Embryology Manager at Wessex Fertilty Clinic
From AI to Bedside
Presented by Dr. Aengus Tran, MD, Harrison.AI (Sydney, Australia).
iDAScore – development of robust AI based embryo evaluation
Presented by Jens Rimestad, Deep Learning Specialist (Aarhus, Denmark).
Webinar: 'iDAScore - Changing the paradigm of embryo evaluation'
In this recorded webinar, Jørgen Berntsen, Data Science Manager at Vitrolife, presents data showing the performance of a fully automated AI-based embryo evaluation algorithm. We will also look at how the AI scores correlate with currently accepted embryo evaluation parameters.Watch webinar
'The future of AI methods to automate embryo evaluation'
Watch this recorded webinar, where Dr. Mikkel Fly Kragh presents new developments based on time-lapse and artificial intelligence.
'AI in time-lapse - automatic grading of human blastocysts'
In this recorded webinar, Dr. Mikkel Fly Kragh will go through how he and the AI team at Vitrolife have developed an AI method to analyse time-lapse sequences for automatic grading of blastocysts.
What is the potential of AI in ART?
The potential of artificial intelligence to improve processes in IVF is immense.
Vitrolife has worked with artificial intelligence to improve processes in IVF for a long time. Find out more about how Vitrolife is utilising AI in this blog post.
Learn more about the future of AI in IVF
iDAScore® - Intelligent Data Analysis for embryo evaluation
Contact us to learn more
How can you benefit from time-lapse in IVF?
Dr. Markus Montag has written a white paper where he discusses the clinically proven results with time-lapse, as well as how time-lapse can improve the workflow in the lab and facilitate communication.
EmbryoScope+, making time-lapse standard of care
EmbryoScope+ is designed to meet the needs of clinics wishing to implement time-lapse as a standard of care to more of their patients.Learn more
EmbryoScope 8 - optimal capacity to suit your clinics need
EmbryoScope 8 time-lapse system is designed to meet the needs of smaller clinics who want to enjoy the same great benefits of the EmbryoScope+ family of time-lapse systems.Learn more
EmbryoScope Flex, ideal for mild stimulation & low responder patients
Expanding your possibilities to offer time-lapse as a standard of care to more patients.Learn more
- Bormann et. al (2020), Fertil Steril 113(4): 781-787.e1.
- Storr et. al (2017), Hum Reprod 32(2): 307-314.
- Rimestad et al., (2020) Robust embryo scoring model based on artificial intelligence (AI) applied to a large time-lapse dataset. Human Reproduction, Volume 35, Issue Supplement_1, July 2020, Pages i213-i214.