What role does AI and digital pathology play in drug discovery and pharmaceutical diagnostics? Following her panel discussion at DP&AI: Europe 2025, we caught up with Monika Lamba Saini, Global Translational Pathology Leader at ADC Therapeutics, to discuss this often forgotten area of diagnostics.
From a diagnostic perspective, where do you see AI and digital pathology adding the most value within pharma today?
Their greatest value is in accelerating drug development. AI-powered digital pathology enables faster compound screening and lead optimization through predictive modeling and virtual screening, reducing reliance on costly physical testing. By analyzing multimodal data – imaging, genomic, and proteomic – AI can identify novel morphologic features and spatial tissue patterns that serve as predictive or prognostic biomarkers, offering insights beyond what is visible to the human eye.
AI also supports clinical trials through central review, improved patient selection, remote collaboration, and standardized training, enhancing data quality and reliability. The shift to digital workflows, combined with AI-driven automated analysis, improves laboratory efficiency, quality assurance, and consistency.
Integrating AI-derived pathology data with molecular and clinical information further advances precision medicine, enabling more accurate prognostication and tailored treatment strategies based on a comprehensive understanding of each patient’s disease.
Your panel discussion at DP&AI: Europe 2025 highlighted the importance of building biology models, not just pathology models. What does this shift mean for how pathologists generate and interpret tissue data?
Pathologists play a central role in linking scientific data with cellular and tissue features, disease processes, and the microenvironment. Their morphologic expertise supports the development and validation of complex in vitro systems, organoids, humanized animal models, and AI tools, ensuring translational relevance through integration of multi-omics data.
Drawing on comparative medicine, pathologists connect findings from animal and in vitro studies to human disease, guiding evaluation of drug efficacy, safety, and broader applicability. They also define study endpoints and establish evidence-based, clinically meaningful biomarkers.
Today, the pathologist is not only an interpreter of tissue but also a critical link between clinical insight and AI-driven biological models, ensuring scientific and semantic coherence.
How is pharma approaching diagnostic innovation using AI-enabled digital pathology?
Pharmaceutical companies are using AI-powered digital pathology to advance biomarker discovery and patient stratification, driven by the need for more efficient, objective, and personalized drug development.
In biomarker development, AI analyzes large digital pathology datasets that exceed human capacity, enabling automated quantification of histologic features and identification of novel predictive biomarkers through deep learning. Multiplex imaging further reveals spatial relationships within the tumor microenvironment, which is particularly relevant for immunotherapy. Rather than assessing PD-L1 expression alone, AI can evaluate the proximity of immune and tumor cells to better predict treatment response. Integrating pathology with genomic and proteomic data through multimodal models further improves biomarker precision.
For patient stratification, AI supports more efficient clinical trial design by identifying patients most likely to benefit from targeted therapies, improving success rates while reducing costs and timelines. For example, Roche is developing the VENTANA TROP2 (EPR20043) RxDx Assay using AstraZeneca’s QCS algorithm to quantify TROP2 expression in non–small cell lung cancer and guide selection for antibody-drug conjugates. GSK is collaborating with PathAI to develop algorithms for nonalcoholic steatohepatitis to quantify inflammation and fibrosis in clinical trials.
Pharmaceutical companies are also prioritizing data standardization, validation, and explainable AI to support regulatory submissions and build confidence in companion diagnostics.
What role do pathologists play in ensuring that AI models developed in pharma remain diagnostically relevant and interpretable?
Pathologists ensure that AI models in pharmaceutical development are clinically relevant, accurate, and interpretable by contributing essential domain expertise. Their key roles include:
Data annotation and curation: Pathologists label complex datasets, such as tissue slides, establishing the ground truth for AI training and ensuring data quality and diagnostic accuracy.
Defining clinical relevance: They shape the diagnostic questions AI models are designed to address, ensuring outputs reflect meaningful clinical needs rather than isolated statistical patterns.
Model validation: As the clinical gold standard, pathologists compare AI results with expert diagnoses to assess real-world performance and identify biases or errors that automated metrics may miss.
Interpretability: By reviewing the features and patterns AI systems use, pathologists help confirm that model reasoning aligns with established medical knowledge, improving transparency and trust in clinical application.
How do you see digital pathology platforms evolving to better support collaboration between diagnostic labs and pharma research teams?
I believe the key areas of evolution would be:
AI-based quantitative analysis: Delivering more reliable, reproducible tissue data to support drug development.
Integrated data ecosystems: Seamless links with LIS, EHRs, and genomic platforms to enable a more comprehensive understanding of disease.
Secure cloud networks: Global access to images and datasets, supporting large-scale, real-world evidence generation.
Enhanced collaboration: Real-time interaction across research teams and institutions.
Remote analysis: Facilitating decentralized clinical trials and distributed expertise.
Co-development with pharma: Expanding partnerships in companion diagnostics and creating new revenue streams for laboratories.
As digital pathology platforms evolve, greater use of AI analytics, data integration, and secure cloud infrastructure will enable more efficient drug discovery, streamlined clinical trials, and broader implementation of precision medicine.
Looking ahead, how do you expect AI-driven digital pathology to shape diagnostics in pharma over the next five to ten years?
Digital pathology is increasingly incorporating AI-driven analytics, enabling advanced diagnostic decision-making through deep learning and supporting the development of AI-integrated companion diagnostics. Large-scale, secure cloud networks will enhance collaboration, accelerate clinical trials, and expand global access to precision medicine.
Future platforms will integrate data from laboratory information systems, electronic health records, and genomic databases. Combining pathology with clinical, radiologic, and molecular data will deepen understanding of disease and support discovery of novel biomarkers and therapeutic targets.
Scalable cloud storage will reduce reliance on physical slide transport, while value-based health care models and trained laboratory personnel will support digital and AI-enabled workflows.
As a result, the pathologist’s role will become increasingly multidisciplinary, bridging experiential expertise with data-driven insights in a more collaborative environment.
Newsletters
Receive the latest pathologist news, personalities, education, and career development – weekly to your inbox.
