Static images or time-lapse? Choosing an embryo assessment workflow

Every IVF laboratory assesses embryos, but not every laboratory does it the same way. Some photograph embryos at set time points under a standard microscope; others culture them in time-lapse incubators that record images continuously. Both approaches are used around the world, and AI tools now exist for each. This article compares what they capture and what they ask of the lab, so you can choose with your own constraints in mind.

What a static image captures

A static image is a snapshot of the embryo at one moment, typically at the blastocyst stage on day 5 or 6. From it, embryologists grade expansion, inner cell mass and trophectoderm — the familiar Gardner-style grading — and record what they see.

What it asks of the lab: a microscope with a camera, a consistent imaging protocol (focal plane, magnification, timing) and a way to store images with the case. Most labs already have this.

What it cannot show: events between observations, such as the timing of cleavage divisions, or transient changes that resolve before the image is taken.

What time-lapse captures

Time-lapse systems take images at short intervals inside the incubator, so the embryo's development can be reviewed as a sequence. This adds morphokinetic information — when divisions happen — and lets embryos stay in culture without being taken out for each assessment.

What it asks of the lab: dedicated incubators and culture dishes, the capacity for the number of patients you treat, and time to annotate or review the sequences.

What it cannot show: anything outside the system's field and focal planes, and it does not by itself tell you how to weigh the extra information.

Where AI fits in each

AI models can read either kind of input, but they are trained for one. A model built for static images expects a single, well-taken image; a model built for time-lapse expects a sequence. What matters is that the model's intended purpose matches your workflow — and that it is regulated for that purpose where you practise.

ERICA works from a single static image per embryo, so it fits labs with or without time-lapse incubators. It orders the embryos of a procedure by estimated euploid potential to support the embryologist's selection; it does not predict pregnancy, and the decision stays with the team.

Questions that help choose

1. What does your current workflow look like? Changing culture systems is a bigger step than adding software to existing images.
2. What volume do you handle, and how? Incubator capacity and annotation time scale with cycles.
3. What evidence exists for the purpose you have in mind? Professional bodies such as ESHRE publish good-practice recommendations on laboratory technologies and add-ons; read them for the specific claim you are evaluating.
4. How will you see the results in your own lab? Whatever the approach, recording outcomes consistently is what lets you judge it with your own data.

There is no single right answer for every laboratory. The best workflow is the one your team runs consistently, documents well and can review.

ERICA is a Class I medical device under Regulation (EU) 2017/745 for clinical decision support. It does not predict pregnancy and does not replace clinical judgement.