In this interview, we engage with PathAI leaders: Eric Walk, MD, Chief Medical Officer; Ella Stewart, Senior Product Manager; and Christina Jayson, PhD, Head of Inflammation & Immunology AI Product. We discussed PathAI’s AISight platform, which is revolutionizing clinical trials with advanced AI integration. The conversation explored the platform’s capabilities, its impact on pathology workflows, and the future of AI in drug development, offering insights into how AI is set to transform the field.
Moe: How does AISight integrate with lab systems, and what challenges have you addressed to maintain workflow continuity?
Ella Stewart: Integration is crucial for maintaining workflow continuity, especially given the diversity of lab information systems (LIS) across different labs. For our clinical trials platform, we manage samples from our biopharma lab in Memphis, which is integrated with our LIS to streamline data reception. This integration allows us to handle data from trials efficiently. For external sites, we facilitate a bulk upload of images with a data manifest, ensuring that data from various sources is consolidated effectively.
Eric Walk, MD: In the pathology lab space, we’ve integrated with multiple different LIS systems, each with unique requirements. Our dedicated engineering team leads these bi-directional integrations, ensuring our platform remains adaptable and efficient across different lab environments. This adaptability is key to maintaining seamless operations and ensuring that our AI tools can be effectively utilized in diverse settings.
Moe: How has AISight transformed pathologist workflows, and what feedback have you received on its usability and impact on accuracy?
Christina Jayson: Our entire workflow, including the AI tools integrated into the AISight Clinical Trials Platform is designed to enhance reproducibility, accuracy, and efficiency in pathology clinical trials. Traditionally, pathologists rely on manual scoring, which can be subjective and variable. Using the AIM-MASH algorithm as an example, our AI algorithms allow a single pathologist to perform at the level of a consensus panel of three pathologists and, with a regulatory approval, could replace the current gold standard of multiple pathologist consensus reads. This transformation reduces complexity and variability, as the AI provides a consistent anchor for pathologists.
Ella Stewart: Feedback from pathologists on the AISight Clinical Trials Platform has been overwhelmingly positive. They appreciate the platform’s ability to streamline workflows and enable precise and accurate AI assisted scoring . For example, our AIM-MASH algorithm work published in Nature Medicine last year demonstrated that AI-assisted workflows can significantly enhance reproducibility with the same precision as a consensus panel of pathologists, reducing the need for multiple pathologists and expediting the trial process. This efficiency is crucial in clinical trials, where time, reproducibility and accuracy are of the essence.
Moe: Could you elaborate on the data sets and methodologies used to train and validate the AI algorithms within AISight?
Christina Jayson: Developing robust AI algorithms requires diverse and comprehensive data sets. We train our algorithms using data from multiple clinical trials, ensuring they encounter various disease severities and patient populations. This diversity is crucial for creating reliable algorithms across different conditions. Our training involves annotations from specialized pathologists, providing a solid ground truth for the models.
Eric Walk, MD: Validation is conducted against a ground truth panel, ensuring that our algorithms meet high standards of accuracy and reliability. This rigorous approach ensures that our AI tools are effective and adaptable to various pathology laboratory scenarios, enhancing their utility in real-world applications. By maintaining this high standard, we ensure that our tools can be trusted by pathologists and researchers alike.
Moe: How has PathAI collaborated with pharmaceutical enterprises and academic organizations to enhance AISight capabilities?
Christina Jayson: Collaboration is at the heart of our approach to enhancing AISightcapabilities, especially for building our AI algorithms that are integrated into AISight. We’ve adopted a consortium model for various indications, partnering with biopharma companies and academic centers to develop standardized tools for clinical trials. In the inflammatory bowel disease space, for instance, we were selected by the Foundation for the National Institutes of Health Biomarkers Consortium Mucosal Healing in ulcerative colitis project team to develop a gold standard histology endpoint assessment tool.
Eric Walk, MD: This collaboration involved eight leading biopharma partners, several nonprofits, and academic medical centers, providing diverse perspectives and data. These partnerships have accelerated the development of our algorithms, allowing us to integrate feedback from various stakeholders and enhance the tools’ effectiveness in clinical trials. By working together, we can ensure that our tools meet the needs of the entire research community.
Moe: How does AISight support real-time data analysis and reporting, and how has this influenced trial decision-making?
Ella Stewart: AISight’s digital platform automates adjudication, ensuring consistency and efficiency. The platform automatically triggers secondary reviews or consensus tasks when disagreements occur, streamlining the process. Our AISight Live tool facilitates real-time collaboration, allowing pathologists to review slides online.
Eric Walk, MD: This capability enhances decision-making by providing a seamless platform for adjudication and consensus-building. The ability to conduct real-time analysis and reporting has the potential to significantly improve trial outcomes, reducing delays and improving the accuracy of results. This real-time capability is a game-changer in clinical trials, where timely and accurate data is critical.
Moe: Looking ahead, how does PathAI envision AI’s role evolving in pathology, and what innovations are planned for AISight?
Ella Stewart: The future of AI in pathology is promising, with numerous innovations on the horizon. We’re expanding into new areas like celiac disease with tools such as AI-assisted counting for intra-epithelial lymphocyte counts. This tool provides immediate results, enhancing the efficiency of clinical trials.
Christina Jayson: We’re also excited about integrating new AI algorithms into AISight, such as our Liver Explore and IBD Explore tools. These tools extract comprehensive data from biopsies, providing insights into drug mechanisms of action and potential biomarkers. By continuing to innovate and expand our capabilities, PathAI aims to remain at the forefront of AI in pathology, driving advancements in drug development and clinical research.
Moe Alsumidaie is Chief Editor of The Clinical Trial Vanguard. Moe holds decades of experience in the clinical trials industry. Moe also serves as Head of Research at CliniBiz and Chief Data Scientist at Annex Clinical Corporation.
