Joining Forces for Pathology Diagnostics with AI Assistance: The EMPAIA Initiative

Norman Zerbe,Lars Ole Schwen,Christian Geißler,Katja Wiesemann,Tom Bisson,Peter Boor,Rita Carvalho,Michael Franz,Christoph Jansen,Tim-Rasmus Kiehl,Björn Lindequist,Nora Charlotte Pohlan,Sarah Schmell,Klaus Strohmenger,Falk Zakrzewski,Markus Plass,Michael Takla,Tobias Küster,André Homeyer,Peter Hufnagl

2024-04-16

Abstract:Over the past decade, artificial intelligence (AI) methods in pathology have advanced substantially. However, integration into routine clinical practice has been slow due to numerous challenges, including technical and regulatory hurdles in translating research results into clinical diagnostic products and the lack of standardized interfaces. The open and vendor-neutral EMPAIA initiative addresses these challenges. Here, we provide an overview of EMPAIA's achievements and lessons learned. EMPAIA integrates various stakeholders of the pathology AI ecosystem, i.e., pathologists, computer scientists, and industry. In close collaboration, we developed technical interoperability standards, recommendations for AI testing and product development, and explainability methods. We implemented the modular and open-source EMPAIA platform and successfully integrated 14 AI-based image analysis apps from 8 different vendors, demonstrating how different apps can use a single standardized interface. We prioritized requirements and evaluated the use of AI in real clinical settings with 14 different pathology laboratories in Europe and Asia. In addition to technical developments, we created a forum for all stakeholders to share information and experiences on digital pathology and AI. Commercial, clinical, and academic stakeholders can now adopt EMPAIA's common open-source interfaces, providing a unique opportunity for large-scale standardization and streamlining of processes. Further efforts are needed to effectively and broadly establish AI assistance in routine laboratory use. To this end, a sustainable infrastructure, the non-profit association EMPAIA International, has been established to continue standardization and support broad implementation and advocacy for an AI-assisted digital pathology future.

Computers and Society,Artificial Intelligence,Computer Vision and Pattern Recognition,Human-Computer Interaction

What problem does this paper attempt to address?

The main problem that this paper attempts to solve is the standardization and interoperability of artificial intelligence (AI) technology in pathological diagnosis. Specifically, the paper discusses the following key challenges:

1. **Standardization and Interoperability**:
- Current AI solutions face serious interoperability problems when integrated into different image management and laboratory information systems. Software systems of different manufacturers use proprietary application programming interfaces (APIs) and file formats, causing AI developers to need to adapt their solutions separately for each system, increasing the workload and cost.
- The paper proposes the EMPAIA platform and its open, vendor - neutral standard interface (EMPAIA App Interface), aiming to simplify the integration process of AI applications and improve the level of interoperability and standardization.

2. **Practical Application in Clinical Practice**:
- Although AI has made significant progress in the field of pathology, it is still very difficult to widely apply it in clinical practice. Through cooperation with multiple pathology laboratories, the paper evaluates and tests the practical application effect of the EMPAIA platform, and collects feedback from clinical and technical perspectives to improve the function and user experience of the platform.

3. **Explainability**:
- AI algorithms are usually considered as "black boxes", and it is difficult to understand their decision - making processes. Especially in the application scenarios such as medical diagnosis that may endanger lives, it is crucial to control and understand AI solutions. The paper explores how to provide customized explainability methods for different stakeholders (such as pathologists, developers, decision - makers, etc.) to enhance users' trust and acceptance of AI systems.

4. **Regulations and Verification**:
- The commercialization and clinical implementation of AI - assisted diagnosis tools still need to overcome legal and regulatory barriers. The paper mentions the guidelines and support services released by the EMPAIA team to help AI suppliers meet these challenges and ensure that their products meet the certification requirements of medical products.

In summary, the core objective of this paper is to promote the wide application and effective integration of AI technology in pathological diagnosis by establishing standardized interfaces, promoting cross - vendor interoperability, improving the explainability of AI systems, and supporting regulatory verification. This will not only help improve the work efficiency of pathologists, but also improve the diagnostic accuracy and treatment effect of patients.

### Formula Examples

When discussing the performance of AI models, the paper may involve some evaluation metrics, such as the accuracy of classification tasks and the Intersection over Union (IoU) of segmentation tasks. Here are the Markdown representations of these formulas:

- **Classification Accuracy (Accuracy)**:
\[
\text{Accuracy}=\frac{\text{TP}+\text{TN}}{\text{TP}+\text{TN}+\text{FP}+\text{FN}}
\]
where TP is True Positive, TN is True Negative, FP is False Positive, and FN is False Negative.

- **Intersection over Union (IoU)**:
\[
\text{IoU}=\frac{\text{Area of Overlap}}{\text{Area of Union}}=\frac{|A\cap B|}{|A\cup B|}
\]
where A and B are the predicted area and the real area respectively.

These formulas are used to evaluate the performance of AI models in pathological image analysis.