Aignotics Atlas H&E-TME
A Berlin-based artificial intelligence company is commercializing technology that promises to extract comprehensive tumor microenvironment insights from routine pathology slides in hours rather than weeks, addressing a persistent bottleneck in cancer drug development and translational research.
Aignostics announced general availability of Atlas H&E-TME, an application that analyzes the tumor microenvironment in hematoxylin and eosin-stained images without specialized staining or multiplexing. According to the company, the platform comprehensively analyzes whole-slide images and delivers detailed readouts for quality control, seven tissue types, nine cell classes, and over 5,000 quantitative metrics per image—all at single-cell resolution.
The application was built using Atlas, a foundation model co-developed by Aignostics, Mayo Clinic, and Charité Berlin. Following successful early access pilots with biopharma partners, the platform is now available as a self-service application, with integrations to major image management systems planned for future releases.
Understanding the tumor microenvironment—the complex ecosystem of immune cells, blood vessels, connective tissue, and signaling molecules surrounding cancer cells—is critical for developing effective therapies, particularly immunotherapies that depend on immune system interactions with tumors. However, traditional TME analysis approaches require specialized immunohistochemistry staining, multiplexed imaging panels, or spatial transcriptomics, all of which are costly, time-consuming, and difficult to scale across large clinical trial cohorts.
From Routine Slides to Single-Cell Insights
Atlas H&E-TME addresses this limitation by extracting deep analytical insights from H&E slides—the most common staining method in pathology, used routinely in clinical diagnosis. According to Aignostics, this enables researchers to maximize value from existing tissue samples without additional processing or specialized equipment.
The platform has been extensively validated across primary and metastatic sites from various laboratories and scanner types, according to the company, to ensure generalizability to real-world conditions. Validation metrics are available to prospective partners upon request. The company emphasizes that validation covered diverse technical environments to demonstrate robustness beyond single-institution datasets.
“This is the first time I have seen a model that demonstrates consistently high accuracy across multiple cancer indications with exceptional robustness,” said Frederick Klauschen, MD, Co-Founder of Aignostics and Director of the Institute of Pathology at Ludwig-Maximilians-Universität München.
The application supports diverse research use cases, from rapidly identifying tumors with immune infiltration to integrating H&E data with spatial transcriptomics for refined cell niche analyses. Initial availability focuses on breast, bladder, colorectal, liver, and lung cancer for biopharma partners, with academic access and additional cancer types planned for coming months.
“Understanding the TME is essential for developing effective cancer therapies, but traditional approaches are costly, time-consuming, and often limited in scale,” said Viktor Matyas, Co-Founder and CEO of Aignostics. “Atlas H&E-TME changes this paradigm. Our pilots with biopharma have confirmed we can deliver high-quality results with unparalleled accuracy, speed, and scale, helping researchers get maximum value from every H&E image.”
Competitive Landscape: Foundation Models and TME Analysis
Aignostics enters a rapidly evolving digital pathology market where multiple companies are leveraging AI and foundation models to extract insights from pathology images. The competitive dynamics center on data scale, clinical validation, and the specific problems each platform addresses.
PathAI, a Boston-based leader in AI-powered pathology, offers PathExplore, a tumor microenvironment analysis platform that delivers over 600 human-interpretable features from H&E whole-slide images. PathExplore has been validated across multiple cancer types and integrated with real-world clinical data through partnerships with companies like ConcertAI. PathAI’s approach emphasizes linking quantitative pathology data with clinical outcomes, enabling biomarker discovery and patient stratification for clinical trials.
PathAI operates a CAP/CLIA-certified laboratory and provides comprehensive services from wet lab processing to algorithm deployment, positioning itself across the full pathology value chain. The company has secured regulatory clearances and extensive partnerships with biopharma companies, giving it significant market presence in drug development applications.
Paige, which spun out of Memorial Sloan Kettering Cancer Center, has developed Virchow, a foundation model trained on over one million digitized pathology slides. Paige’s approach focuses on pan-cancer detection and diagnosis, with FDA-cleared applications for prostate cancer and expanding capabilities across 17 tissue types. The company recently released PRISM2, a whole-slide foundation model designed to connect pathology images with clinical language used by pathologists, enabling multimodal AI applications.
Paige’s competitive advantage stems from its access to Memorial Sloan Kettering’s extensive pathology archives and its collaboration with Microsoft Research, which provided computational resources for training large-scale models. The company’s emphasis on regulatory clearance and clinical-grade applications positions it strongly for diagnostic use cases, though its TME analysis capabilities appear less specialized than dedicated platforms.
What distinguishes Aignostics’ Atlas H&E-TME in this landscape is its focused specialization on tumor microenvironment profiling specifically, rather than broader cancer detection or general pathology analysis. By delivering over 5,000 quantitative metrics per image—significantly more granular than PathAI’s 600 features—Aignostics aims to provide deeper TME characterization for researchers developing immunotherapies and other microenvironment-targeted treatments.
The platform’s collaboration with Mayo Clinic and Charité Berlin for foundation model development also provides clinical credibility, though the extent of validation data and peer-reviewed publications will be critical factors in adoption decisions by biopharma partners who require rigorous evidence before integrating new analytical tools into drug development programs.
Strategic Implications for Drug Development
For pharmaceutical companies developing cancer therapies, particularly immunotherapies and targeted agents that depend on specific tumor microenvironment characteristics, Atlas H&E-TME offers potential advantages in patient stratification, biomarker discovery, and understanding mechanisms of treatment response or resistance.
Traditional TME analysis using immunohistochemistry or multiplexed imaging requires planning staining panels upfront, limiting exploratory analysis. By extracting comprehensive data from routine H&E slides retrospectively, researchers can analyze archived clinical trial samples without additional tissue consumption—a significant advantage when tissue availability is limited.
The platform’s speed—delivering results in hours rather than weeks—could accelerate timelines for correlative studies that link treatment outcomes with TME characteristics. Faster turnaround enables more iterative hypothesis testing and could compress drug development cycles, particularly in early-phase trials where rapid decision-making about patient selection criteria or dose optimization is valuable.
However, adoption will depend on validation that Atlas H&E-TME’s computational predictions correlate meaningfully with gold-standard measurements and clinical outcomes. Biopharma companies have been burned by analytical tools that generate impressive-looking data without translating to actionable insights or regulatory acceptance. The company’s validation claims will require peer-reviewed publication and prospective confirmation in clinical trials to gain broad acceptance.
The self-service model could accelerate adoption by reducing implementation friction, though integration with existing laboratory information management systems and digital pathology infrastructure remains complex. Organizations with mature digital pathology operations will find integration easier than those still transitioning from analog workflows.
Adoption Barriers and Market Realities
Several factors will influence whether Atlas H&E-TME gains traction beyond early-adopter biopharma partners. Clinical validation depth matters enormously—regulators and payers increasingly demand evidence that AI-derived biomarkers predict outcomes reliably across diverse patient populations, scanner types, and tissue preparation protocols.
While Aignostics states the platform has been validated across various labs and scanners, the company will need to publish validation datasets, performance metrics, and evidence of clinical utility in peer-reviewed journals. Biopharma companies evaluating the platform will scrutinize sensitivity, specificity, reproducibility, and concordance with established reference standards.
Regulatory pathway considerations also apply. If Atlas H&E-TME-derived metrics will be used for patient selection in clinical trials or as companion diagnostics, the platform may require FDA clearance or CE marking, adding time and cost to commercialization. The self-service research use model may initially avoid these requirements, but any transition to clinical use will require regulatory engagement.
Pricing and business model transparency matter. Digital pathology and AI analysis platforms vary widely in pricing structures—per-slide fees, subscription models, enterprise licenses—and total cost of ownership includes not just software licenses but also computational infrastructure, data storage, and staff training. Aignostics will need to demonstrate clear return on investment for biopharma partners evaluating competing platforms.
Intellectual property landscapes in computational pathology are complex, with overlapping patents on image analysis methods, specific algorithms, and applications. Companies commercializing TME analysis tools must navigate potential infringement risks, though foundation models trained on large public datasets may offer some freedom to operate compared to narrower task-specific algorithms.
Looking Ahead: Foundation Models Transform Pathology Research
Aignostics’ launch of Atlas H&E-TME reflects broader momentum toward foundation models in medical imaging—large-scale AI systems trained on massive datasets that can be adapted to multiple downstream tasks. This approach mirrors developments in natural language processing and computer vision, where foundation models have demonstrated superior performance and generalizability compared to task-specific models.
For cancer researchers, foundation model-based pathology analysis promises to unlock insights from the billions of existing H&E slides in hospital archives and clinical trial biobanks. Rather than requiring prospective sample collection with specialized staining protocols, researchers could retrospectively analyze samples to discover new biomarkers, understand treatment mechanisms, or identify patient subgroups who benefit most from specific therapies.
The democratization potential is significant. Academic researchers without access to expensive multiplexed imaging platforms or spatial transcriptomics equipment could leverage H&E-based TME analysis to pursue competitive research programs. Aignostics’ planned academic access program could accelerate discovery if priced accessibly and integrated with public datasets like The Cancer Genome Atlas.
For patients, more sophisticated tumor microenvironment understanding could enable more precise treatment selection—matching individuals to therapies most likely to work based on their tumor’s immune landscape, vascular characteristics, and stromal composition. This level of precision moves beyond simple biomarkers like PD-L1 expression toward comprehensive microenvironment profiling that captures the complexity driving treatment response.
Ultimately, the success of platforms like Atlas H&E-TME will be measured not by technical sophistication or feature counts, but by whether they generate insights that translate into better therapies, more accurate patient stratification, and improved outcomes for people living with cancer. In a field where computational promises often exceed clinical delivery, rigorous validation and demonstrated clinical utility will separate transformative tools from incremental innovations.
– This original article was created with AI support.
