Blog · By IVF Automation team · 22 Oct 2026

The FDA has now cleared four embryo-scoring tools. Here's what that does and doesn't mean for your lab.

On 15 May 2026 the US Food and Drug Administration cleared Embryo Predict, an AI tool from Alife Health that scores blastocyst images. Its maker announced the clearance on 28 May, and AI embryo selection was back in the headlines. If you run an IVF lab in India, you will hear about it from vendors soon, if you haven't already.

Before anyone signs a purchase order, it helps to know what an FDA clearance in this category actually covers, what the research says, and which Indian rules apply. This guide explains all three in plain terms. It is operational information only. Any decision to use embryo assessment software belongs to your lab director and clinical team.

The FDA category these tools sit in

The FDA groups these products under one device type, the "Assisted Reproduction Embryo Image Assessment System" (21 CFR 884.6195, product code PBH). The regulation describes a prescription device that analyses images of developing embryos and "provides information to aid in the selection of embryo(s) for transfer when there are multiple embryos deemed suitable for transfer or freezing."

It is a Class II device with special controls. Among other things, makers must show clinical performance data (sensitivity, specificity and predictive values), software validation and hazard analysis, and the reproducibility of the prediction algorithm. The category was created in 2014 through a De Novo request for Auxogyn's Eeva test.

The FDA's 510(k) database lists four later clearances under this code:

Read the "indications for use", not the headline

Each clearance is tied to a narrow intended use, and the details differ a lot.

According to its 510(k) summary, CHLOE BLAST gives adjunctive information on whether a Day 3 embryo is likely to reach the blastocyst stage on Day 5. It is cleared only for images captured by the EmbryoScope version D incubator.

The Embryo Predict summary describes a different job. It analyses images of Day 5, 6 and 7 blastocysts that the embryologist has already judged suitable for transfer, combines them with patient age and day of culture, and provides adjunctive information when several blastocysts are suitable. The company says the clearance was supported by a randomised trial of 440 patients at seven US clinics, and that the product also carries a CE mark in Europe.

Two words keep appearing: adjunctive and suitable. These tools don't replace morphological grading by your embryologists. They rank or add information to embryos your team has already graded as usable. A clearance tells you the FDA accepted that the device does that specific job as claimed. It does not mean the device was cleared for other microscopes, other culture days or other uses.

What the research says so far

The largest randomised trial so far is a useful reality check. Published in Nature Medicine in 2024, it compared Vitrolife's deep-learning score, iDAScore, with standard morphological selection across 14 IVF clinics in Australia and Europe, with 1,066 patients. Clinical pregnancy rates were 46.5% in the AI arm and 48.2% in the morphology arm, and the study could not show that the AI approach was non-inferior. In a timing substudy, though, evaluating embryos with the AI score took about 21 seconds on average, against about 208 seconds for manual assessment.

ESHRE's 2023 good practice recommendations on add-ons take a similarly cautious line. They do not recommend time-lapse imaging, with or without embryo selection software, as a tool to improve live birth rates. They also note that many selection algorithms appear to be laboratory-dependent, probably because culture conditions differ between labs. At the same time, ESHRE recognises time-lapse as a tool for teaching, standardising assessment, lab workflows and quality control.

For an Indian lab, the practical lesson is simple. The strongest current case for these tools is operational: more consistent grading between embryologists, a digital record of every assessment, and less time spent at the microscope. Be wary of any marketing that promises better results for patients. That is also a legal risk for clinics. Check claims with our free NMC post checker before you publish anything about the technology you use.

What applies in India

India has no embryo-specific AI rule, but two documents matter.

CDSCO's guidance on medical device software. The Central Drugs Standard Control Organisation's Guidance Document on Medical Device Software (CDSCO/MD/GD/MDSW/01/2026) explains how software is regulated under the Medical Devices Rules 2017. It lists image-analysis software and AI/ML-based tools among its examples of software that can be a medical device. It says the risk class (A to D) depends on the intended use, and that the intended-use statement should make the software's clinical role clear, for example whether it is decision-support software. It also describes an "Algorithm Change Protocol" so that updates to the software don't compromise its safety or intended use.

ICMR's AI ethics guidelines. The Indian Council of Medical Research's Ethical Guidelines for Application of AI in Biomedical Research and Healthcare (2023) cover accountability and liability, and set out guiding principles for validating and deploying AI tools.

Add the record-keeping duties in the ART Act and Rules, and the fact that embryo images linked to patient records are personal data under India's Digital Personal Data Protection Act. Our free DPDPA readiness check can help with that part.

Seven questions to ask any AI embryo-scoring vendor

  1. What is the exact intended use, and can I see it in writing? Which culture days, which embryo stages, and which incubators or microscopes?
  2. What is the product's regulatory status in India under the Medical Devices Rules 2017, and what is its risk class? Ask for the licence, not a slide.
  3. Which FDA, CE or other clearances does it hold, and do they cover the same version and hardware you are selling me?
  4. On which patient populations was the model trained and validated? Has it been tested on data from Indian labs?
  5. How do you handle algorithm updates? Will I be told when the model changes, and is there a change protocol?
  6. Where are embryo images and patient data stored and processed, and how does that fit the DPDP Act?
  7. Who supports it locally, what does the AMC cover, and can I export every score and audit log if we stop using it? See our AMC questions for IVF lab vendors for more.

Keep the answers on file. They help with your own validation and with inspections. If you already run time-lapse incubators, start by checking how scores are logged alongside your daily lab QC logs and witnessing records.

Sources

Your next step

Comparing time-lapse or embryo assessment suppliers? Browse our free IVF vendor directory and take the seven questions above to every demo. Suppliers can claim a free listing there too.

For more on automation in the lab, see our cryo-storage monitoring guide and the workflow library.

General operational information only; not medical, clinical or legal advice. Product names are mentioned for information and are not endorsements. See our disclaimer.

Get IVF Automation Weekly

Compliance, automation and opportunities for India's IVF labs and clinics, every week in five minutes. Free, unsubscribe anytime.

Also: open IVF lab jobs · vendor directory · featured listings

Get IVF Automation Weekly

Compliance, automation and opportunities for India's IVF labs, every week in five minutes. Subscribe from the box at the bottom of our events page, browse open IVF lab jobs, or find a vendor.