elea AI: Modernizing Pathology Labs with AI

AI Investment in Healthcare and Elea's Innovative Approach

Venture capital investment in artificial intelligence tools for the healthcare industry was anticipated to reach $11 billion in the previous year. This substantial figure underscores the growing belief that AI will fundamentally change this vital sector.

A significant number of healthcare-focused AI startups are concentrating on improving operational efficiency through the automation of administrative tasks surrounding patient care. Elea, headquartered in Hamburg, aligns with this trend.

However, Elea distinguishes itself by focusing on a frequently neglected area: pathology laboratories. These labs analyze patient samples to detect diseases.

Focus on Pathology Labs

Elea believes that by addressing the needs of pathology labs, it can effectively scale its AI-powered, voice-based workflow system. The goal is to significantly enhance lab productivity and achieve a global impact.

The company also intends to extend its workflow-centric methodology to other healthcare departments, further accelerating their output.

Revolutionizing Lab Workflows

Elea’s primary AI tool is engineered to completely transform the working methods of clinicians and lab personnel. It aims to supersede outdated information systems and conventional practices.

This includes replacing tools like Microsoft Office for report writing with an “AI operating system.” This system utilizes speech-to-text technology and other automation features to dramatically reduce the time required to deliver a diagnosis.

Demonstrated Results

Following approximately six months of operation with initial users, Elea reports a substantial reduction in report turnaround time.

Specifically, the system has enabled the lab to produce roughly half of its reports in just two days – a significant improvement in efficiency.

Revolutionizing Pathology with Automated Workflows

According to Elea’s CEO and co-founder, Dr. Christoph Schröder, pathology laboratories, traditionally reliant on manual processes, stand to gain significant productivity improvements through the implementation of AI. He explains that Elea fundamentally alters the existing paradigm, introducing a far more automated approach to all operational stages.

The system functions as a central hub, facilitating communication between physicians and medical technical assistants (MTAs). Staff members interact with Elea, conveying observations and desired actions, which the AI then executes.

“Elea acts as the agent, completing all tasks within the system,” Dr. Schröder clarifies. “This includes slide preparation, staining procedures, and other essential steps, resulting in a considerably faster and more streamlined workflow.”

Rather than simply enhancing current processes, Elea aims to replace the entire existing infrastructure. The cloud-based software is designed to supersede legacy systems and their fragmented approach, which often relies on separate applications for individual tasks. The core concept is to provide a unified platform capable of orchestrating all laboratory operations.

Leveraging Large Language Models

The startup is developing its platform by fine-tuning various large language models (LLMs) with specialized pathology information and data. This enables key functionalities within the laboratory environment.

A crucial feature is the integration of speech-to-text technology, which transcribes staff voice notes. Furthermore, the system incorporates “text-to-structure” capabilities, converting these transcriptions into actionable instructions that drive the AI agent’s activities. These actions can include directing laboratory equipment to maintain a consistent workflow.

Elea also intends to create its own foundational model specifically for analyzing slide images. This development is a step towards incorporating diagnostic capabilities into the platform, though the immediate focus remains on expanding its current offerings.

Significant Time Savings and Efficiency Gains

Elea’s proposition to laboratories centers on dramatically reducing turnaround times. Processes that traditionally require two to three weeks can potentially be completed in hours or days. The integrated system compounds productivity gains by eliminating inefficiencies such as manual report typing, which is prone to human error and workflow disruptions.

Laboratory personnel can access the system through an iPad app, a Mac app, or a web application, providing flexible access points tailored to different user preferences.

Company Background and Expansion Plans

Founded in early 2024, Elea launched its first laboratory implementation in October, following a period of development throughout 2023. Dr. Schröder brings extensive experience in AI application, having previously worked on autonomous driving projects at Bosch, Luminar, and Mercedes.

Co-founder Dr. Sebastian Casu, the startup’s CMO, contributes a clinical perspective, with over a decade of experience in intensive care, anesthesiology, and emergency medicine. He also previously served as a medical director for a large hospital network.

Early Adoption and Future Growth

Elea has established a partnership with a prominent German hospital group – details of which remain undisclosed – that processes approximately 70,000 cases annually. Currently, the system supports hundreds of users.

Additional customer launches are planned in the near future, and the company is actively exploring international expansion, with a particular interest in entering the U.S. market.

Seed Funding for Elea

Elea is now publicly announcing a €4 million seed funding round secured last year. The investment was spearheaded by Fly Ventures and Giant Ventures, and has been instrumental in expanding the engineering team and initiating product deployment within initial laboratory settings.

While this amount is comparatively modest against the multi-billion dollar investments currently being seen in the AI sector, Schröder posits that success for AI startups doesn't necessarily require massive engineering teams or substantial capital. He suggests that strategic resource allocation is paramount.

Specifically within the healthcare domain, a focused, departmental approach and thorough maturation of the intended application are key before expanding into new areas.

Future Funding Plans

The company is preparing to initiate a Series A funding round, anticipated to occur this summer. This next phase will see Elea transition from a word-of-mouth marketing strategy to proactive outreach, aiming to increase adoption among laboratories.

Schröder highlights a key differentiator between Elea and its competitors: “We offer a focused solution, rather than a fully vertically integrated system.”

A Different Approach to AI in Healthcare

Many existing AI tools function as add-ons to established systems, like Electronic Health Records (EHRs). This often requires users to navigate multiple interfaces and tools, presenting challenges for those less comfortable with digital technology.

“This complexity inherently limits the potential of these solutions,” Schröder explains.

Elea, conversely, has been deeply integrated into its own laboratory information system – termed a “pathology operating system.” This integration ensures a seamless user experience, eliminating the need for separate tools or interfaces.

Users can interact with Elea directly within their existing workflow, simply stating their observations and desired actions.

Efficiency Through Focused Expertise

Schröder emphasizes that achieving significant results doesn't necessitate a large workforce. “You don’t need an enormous number of engineers – a team of a dozen or two dozen highly skilled individuals is sufficient.”

Elea currently employs approximately two dozen engineers who are capable of delivering exceptional results. This approach aligns with the trend of rapidly growing companies that prioritize a small team of experts over a large, generalized workforce.

“These experts can accomplish remarkable things,” he states. “This is the core of our philosophy, and why we don’t initially require hundreds of millions in funding.”

“It represents a genuine shift in the conventional model of company building.”

• Seed Funding: €4 million
• Lead Investors: Fly Ventures and Giant Ventures
• Team Size: Approximately 24 engineers

Cultivating a Workflow-Centric Approach

Elea’s decision to initially focus on pathology laboratories was a deliberate one. According to Schröder, the potential market is valued at several billion dollars. He further characterizes the pathology sector as possessing a distinctly “global” nature.

This global reach facilitates scalability for their Software as a Service (SaaS) offering, particularly when contrasted with the more decentralized landscape of hospital supply. Global laboratory organizations and suppliers are actively seeking solutions to enhance scalability.

“A single application can be developed and deployed across multiple regions – from Germany to the U.K. and the U.S.,” Schröder explains. “Workflow processes remain remarkably consistent across these locations.”

He notes that solving a workflow challenge in one language, such as German, can be readily adapted to others, including English and Spanish, leveraging current Large Language Models (LLMs). This expands the scope of potential applications considerably.

Schröder also highlights pathology labs as experiencing “one of the most rapid growth phases within medicine.” Advances in fields like molecular pathology and DNA sequencing are driving increased demand for diverse and frequent analyses.

This heightened demand translates to increased workloads for laboratories, and consequently, greater pressure to optimize productivity. Efficiency is paramount in this evolving landscape.

Following the refinement of the laboratory application, Elea intends to explore areas where AI is more commonly utilized in healthcare. This includes assisting hospital physicians with documenting patient encounters.

However, any future applications will maintain a strong emphasis on workflow optimization. The core principle will remain consistent.

“Our aim is to introduce a workflow mindset, treating each task as a component within a larger workflow, culminating in a final report,” Schröder states. “This report then requires distribution.”

He clarifies that within a hospital setting, Elea would not venture into diagnostics, but instead concentrate on “operationalizing the workflow” itself. Workflow efficiency is the key differentiator.

Image processing represents another area of interest for Elea’s future healthcare applications. Specifically, they are exploring ways to accelerate data analysis in radiology.

Challenges in AI-Powered Healthcare Transcription

A key consideration revolves around the issue of accuracy. Given the sensitive nature of healthcare applications, inaccuracies within AI transcriptions – particularly concerning critical assessments like biopsy results for cancerous tissues – could have severe repercussions if discrepancies arise between a physician’s statements and Elea’s reported data to other healthcare professionals involved in patient care.

Currently, accuracy is being assessed by monitoring the extent of user edits made to reports generated by the AI. Schröder indicates that between 5% and 10% of reports require some level of manual correction, potentially signaling errors. However, he acknowledges that modifications may also be necessary for reasons beyond error correction, and efforts are underway to minimize the need for manual intervention.

He emphasizes that ultimate responsibility rests with the physicians and staff who review and validate the AI’s outputs. This suggests Elea’s workflow isn’t fundamentally different from traditional methods, where human typists transcribing doctors’ voice notes could also introduce errors; the difference lies in the initial creation being performed by Elea AI rather than a human typist.

The potential for increased throughput through automation could, however, place greater pressure on these review processes, as staff may need to manage a significantly larger volume of data and reports compared to previous workflows.

Schröder concedes that such risks exist, but highlights a built-in “safety net” feature. This feature prompts physicians to re-examine potentially problematic areas, acting as a “second pair of eyes.” The AI analyzes current findings reports in relation to the doctor’s recent statements, providing comments and suggestions.

Patient confidentiality is another crucial concern, particularly with agentic AI utilizing cloud-based processing, as opposed to maintaining data on-premise under direct lab control. Schröder asserts that Elea addresses these “data privacy” concerns through a process of separating patient identities from diagnostic outputs, relying on pseudonymization to ensure data protection compliance.

“Data remains anonymous throughout the process – each step performs a single function – and data is combined only on the device where the physician views it,” he explains. “We utilize temporary pseudo IDs in all processing stages, which are subsequently deleted, but are combined when the doctor accesses patient information.”

The company also utilizes servers located in Europe and ensures full compliance with data privacy regulations. Their primary client, a publicly owned hospital network considered critical infrastructure in Germany, has validated the security of their data handling practices.

“We likely exceeded the necessary security measures, but it’s always prudent to prioritize safety, especially when dealing with sensitive medical data,” Schröder concludes.