Atlas H&E-TME: A Self-Service Tool for Comprehensive Analysis of the Tumor Microenvironment (TME)

Aignostics recently announced early access to Atlas H&E-TME. Read on to learn more about the tool and sign up for a free trial!

Ryan Sargent
June 20, 2025

Atlas H&E-TME is a self-service tool for comprehensive analysis of the tumor microenvironment (TME) in H&E images at single-cell resolution. The tool provides deep insights about spatial biology across 16 cancer subtypes. The product is built using Atlas – the foundation model we co-developed with Mayo Clinic and Charité Berlin – under ISO 13485-certified design control quality standards. Whole-slide images are fully analyzed in just a few hours, with a range of readouts including:

  • Qualitative polygon overlays for tissue QC, tissue segmentation, cell detection, and cell classification
  • Over 5,000 quantitative spatial metrics, including cell, tissue, and neighborhood-level features

Model performance has been extensively validated for each cancer type and our Performance Assurance Guarantee ensures that results meet your standards. Available via our easy-to-use API, Atlas H&E-TME empowers researchers with deeper, more scalable insights from every H&E image. 

Robust Outputs

Atlas H&E-TME provides robust qualitative and quantitative outputs for a range of common cancer types including bladder, breast, colorectal, liver (incl. HCC & CCA), and lung (incl. NSCLC & SCLC). 

For each H&E image, Atlas H&E-TME applies robust quality control models to help researchers isolate their analyses to valid tissue regions. The product then applies a series of cell- and tissue-level models, with results delivered in just a few hours including:

  1. Qualitative polygon outputs for tissue QC, tissue segmentation, cell detection, and cell classification that can be integrated and viewed within your IMS of choice.
  2. Over 5,000 quantitative spatial outputs at the cell, tissue, and neighborhood-level, provided in CSV format for straightforward analysis.

You Control Your Data

Atlas H&E-TME complies with GDPR and was developed following ISO 27001 information security standards and ISO 13485-certified quality standards for design control. Your data and results remain secure while in transit and during processing and are automatically deleted from our network. 

This means that Aignostics does not retain any of your data for model training purposes.

Performance Validation

During development of Atlas H&E-TME, we established stringent performance requirements that ensured results are precise and consistent. To confirm we met these requirements, our team conducted extensive benchmarking that measured performance across a variety of real-world conditions.

Performance was benchmarked against pathologist annotations for:

  • Different morphological subtypes
  • Multiple scanners types
  • Data from diverse labs and biobanks

In total we assessed and validated over 1500 metrics to verify that Atlas H&E-TME met our performance standards. Detailed benchmarking results are available upon request. 

Performance Assurance Guarantee

Atlas H&E-TME is backed by our Performance Assurance Guarantee. At Aignostics, our model reliability is engineered through extensive validation by board-certified pathologists. Our models consistently deliver precise readouts across diverse clinical datasets.

Your satisfaction is our commitment - for edge cases where performance is not up to standard, we will fine-tune our models at zero cost and provide you with improved results, ensuring continuous performance without compromise.

Get Started

Early access is open for Atlas H&E-TME! As part of the program, all users will get access to a free trial to try out the product and see the results for themselves. 

Sign up here to get started!

Related Articles

Meet Neelay Shah, Machine Learning Engineer on our Data Processing team!

29.4.2025

Neelay and the Data Processing team develop pipelines to process data at scale for use-cases such as training foundation models. They ensure that data and models can be processed in a distributed fashion, efficiently spreading computational work across multiple machines to allow for scaling when large data sets are involved. They also partner closely with the Data Science team who use the infrastructure developed by the Machine Learning Engineers to perform experiments for client projects.

Meet PD Dr. med. Julika Ribbat-Idel, IFCAP, Principal Pathologist on our Pathology team!

30.5.2025

Julika and the Pathology team serve as the medical backbone within Aignostics. Their primary responsibility is bringing specialized medical expertise to a tech-focused company, making sure Aignostics' AI can accurately analyze tissue samples across various applications.

Enhancing Prognostic Precision in Bladder Cancer: AI-Driven Tumor Microenvironment Analysis from H&E Images

10.6.2025

Our TME-enhanced model outperformed traditional UICC staging, achieving higher predictive accuracy and clearer separation of prognostic risk groups. This demonstrates that integrating TME data from routine H&E slides with UICC staging improves risk stratification, helping pinpoint high-risk bladder cancer patients more precisely than anatomical staging alone.