AI-powered endometrial diagnostics to support personalised IVF decisions

MetaEndo is an AI-enabled diagnostic service designed for fertility clinics and IVF centres. It combines digital histology, deep neural networks and multi-parameter patient data to help clinicians better understand endometrial receptivity and optimise embryo transfer timing within the individual "implantation window".

Why endometrium and why now?

IVF success is influenced by many factors: oocyte and sperm quality, embryo viability, endometrial receptivity and other uterine and systemic conditions. Even in well-run clinics, a significant proportion of cycles fail despite good-quality embryos, leaving clinicians and patients with limited explanations.

MetaEndo is being developed to address this gap by:

• providing a structured, data-driven assessment of the endometrium,
• improving interpretation of the individual implantation window,
• and offering decision support for embryo transfer planning rather than relying solely on empirical timing.

What MetaEndo analyses

MetaEndo integrates multiple layers of information into one diagnostic framework:

• Digital histological images of endometrial biopsies (whole-slide images).
• Morphological features extracted using computer vision and deep neural networks (endometrial glands, stroma, structural patterns).
• Patient-level clinical data, such as:
• age and infertility history (cause, duration, type),
• hormonal profile (blood tests),
• ultrasound findings of the endometrium,
• progesterone concentration and other relevant biomarkers,
• selected laboratory and imaging parameters.

Using these inputs, the system builds a personalised decision-making model that characterises the patient's endometrial receptivity state and its position relative to the expected implantation window.

How it fits into your workflow

MetaEndo is designed as a remote service and SaaS platform, so clinics can use it without investing in new physical equipment.

A typical workflow:

• Biopsy and routine histology:
• The clinic performs an endometrial biopsy at the chosen cycle time using its usual protocol.
• A local pathology lab prepares the slide as in standard practice.
• Slide digitisation:
• The histology slide is scanned into a digital whole-slide image (WSI).
• Existing slide scanners and pathology infrastructure can be used.
• Secure upload to MetaEndo:
• The clinic (or pathology partner) uploads the digital image to the MetaEndo web platform.
• The clinician enters or confirms the key clinical and laboratory parameters in a structured form.
• AI-driven analysis:
• MetaEndo applies deep neural networks and other machine-learning algorithms to the histology image and integrates it with the patient’s clinical profile.
• Digital report back to the clinic:
• A structured report is generated and made available through the secure online portal.
• Reports can be downloaded as PDFs and, where needed, integrated into local systems.

The service is entirely remote, requires only one image analysis per biopsy with accompanying patient information, and keeps sample logistics simple, while providing secure digital reports and robust data protection.

What the MetaEndo report provides

The MetaEndo report is being designed to deliver clear, clinically useful output rather than raw algorithm scores. It aims to provide:

• An assessment of endometrial receptivity status relative to the expected implantation window.
• Interpretation of key histological and algorithm-derived features.
• Contextualisation with patient-specific clinical data (age, infertility factors, hormonal environment, ultrasound characteristics).
• Guidance to support the timing of embryo transfer, always as decision support for the treating physician, not as a replacement for clinical judgement.

The goal is to help IVF teams reduce repeated cycles, shorten time to pregnancy and provide more transparent explanations to patients when adjusting treatment plans.

Benefits for your clinic and patients

According to the MetaEndo project documentation, the intended benefits for clinics include:

• Improved IVF diagnostics and performance. More precise assessment of endometrial receptivity, supporting better cycle planning.
• More objective and predictive decision support. Deep-learning-based interpretation of histology and clinical data, providing an additional layer of objectivity versus purely visual, subjective evaluation.
• Efficient resource allocation. Potential reduction in repeated treatments and unnecessary procedures, better use of clinical and lab capacity.
• Cost-effectiveness. By improving diagnostic accuracy and reducing the number of failed cycles, MetaEndo aims to lower overall treatment costs for patients and the healthcare system.
• Personalised treatment options. More consistent and individualised treatment pathways, tailored to each patient's endometrial and clinical profile.
• Remote access and digital delivery. Clinics of any size can access the service, regardless of geography, as long as they can provide digital slides and basic clinical data.

Scientific and clinical collaboration

MetaEndo is being developed in collaboration with the Lithuanian University of Health Sciences (LSMU) and specialists from its Reproductive Medicine Centre. These partners contribute clinical expertise, sample data and outcome information for algorithm training and validation.

The project builds on earlier work where a computer-vision algorithm for detecting endometrial glands in histology images was developed and tested on digitalised slides, with results submitted as a scientific publication ("Detection model of endometrial glands in histology images"). This model forms part of the technological foundation for MetaEndo.

Development stage and responsible use

MetaEndo is an R&D-intensive diagnostic system and is currently in the development and validation phase. The project documentation explicitly recognises potential challenges, including:

• limited or heterogeneous research and clinical data,
• possible accuracy limitations of machine-learning algorithms,
• integration challenges between AI outputs and clinical workflows,
• the need for continuous data collection, algorithm refinement and usability improvements.

For this reason:

• MetaEndo is intended to be used as decision support, not as a standalone diagnostic or a substitute for a clinician's professional judgement.
• Any clinical use must comply with local regulations, ethics requirements and applicable medical device or diagnostic rules in the relevant jurisdiction.

Work with us

We are actively seeking fertility clinics, IVF centres and diagnostic laboratories interested in:

• participating in pilot projects and validation studies,
• providing anonymised data under appropriate ethical and legal frameworks,
• co-developing clinical workflows and report formats that are genuinely useful for IVF teams,
• and exploring how AI-enabled endometrial diagnostics can be integrated into existing protocols (including recurrent implantation failure, repeated IVF failures and other complex cases).

If you would like to discuss a potential collaboration or pilot, please contact our team. We will be happy to arrange a detailed technical and clinical discussion.