Article originally posted on The Clinical Trial Vanguard by Moe Alsumidaie on May 30, 2025
Article originally posted on The Clinical Trial Vanguard by Moe Alsumidaie on May 30, 2025
In this conversation with Jeffrey Sorenson, CEO at Yunu, we explore the transformative potential of treating imaging data as a first-class citizen in clinical trials. Sorenson, who spearheads innovative imaging solutions, discusses the necessary structural shifts in clinical trial protocols, the integration of advanced data systems, and the future of imaging in oncology studies. This interview highlights the challenges and opportunities in modernizing imaging workflows to enhance data integrity, trial efficiency, and patient outcomes.
Moe: Imaging is often siloed in trials. What shifts are needed to treat imaging as a first-class data citizen?
Jeffrey Sorenson: To treat imaging as a first-class data citizen, we need a new imaging source data system for cancer centers. Current PACS systems aren’t equipped to handle the specific measurements required by clinical trial protocols, resulting in a disconnect between source imaging data and textual results. This often results in data that can’t be traced back to the original measurements. By creating a unified system across trial sites, we can ensure high-quality readouts at the site level, rather than waiting for central review, which is often too late.
For instance, our platform includes a new eCRF builder that aligns with sponsor requirements, ensuring all necessary data elements are captured accurately and efficiently. This approach streamlines the workflow and enhances the traceability and auditability of imaging data, which is crucial for maintaining data integrity and improving trial outcomes.
Moe: With faster timelines and decentralization, what are the common data integrity risks with imaging?
Jeffrey Sorenson: One of the most significant risks is the profound error rate and backlogs in imaging reads at cancer centers. Sponsors often cannot audit beyond the eCRF values they receive, which means they cannot effectively manage quality across sites. To mitigate these risks, we need to use a common set of tools, ensure high-quality trained readers, and produce real-time data for site assessments. This approach will help maintain data integrity and improve trial outcomes. For example, during our onboarding of 12 cancer centers, we observed significant error rates and backlogs, highlighting the need for a system that provides real-time observability and traceability. By implementing a robust imaging data system, sponsors can gain insights into workflow processes, identify discrepancies, and address them proactively, ultimately enhancing the quality and reliability of trial data.
Moe: Imaging endpoints are central in trials. How should sponsors and regulators standardize these measurements?
Jeffrey Sorenson: Sponsors should provide the necessary technology to sites, similar to how they provide Contract Research Organizations (CROs) with specific Clinical Trial Management Systems (CTMS) or Electronic Data Capture (EDC) systems. Centralization should focus on qualified personnel who can contribute to trials remotely through secure, web-based tools. In oncology trials, it’s crucial to involve the treating investigator in decision-making, as the context of each therapy cycle is vital for accurate measurements and patient care. For example, in cancer care, the selection of target and non-target lesions is critical for determining the overall response and treatment plan. By facilitating real-time collaboration between oncologists and radiologists, we can ensure that imaging endpoints are consistently and accurately measured, leading to more reliable trial outcomes.
Moe: What are the long-term consequences of not modernizing imaging practices, and what’s the cost to sponsors and patients?
Jeffrey Sorenson: Allowing imaging data to remain paper-based means losing the connection between source imaging data and measurements. This disconnect results in delayed signals, transposition errors, and increased site burden. By digitizing these processes, we can achieve real-time signal, error-free data, and maintain the context between imaging data and measurements. This modernization is essential for leveraging AI and advanced processing to extract more value from clinical trial data. For instance, our platform’s ability to integrate AI and radiomics into the workflow enables the extraction of additional insights, such as body fat composition and its impact on drug metabolism, which can optimize patient eligibility and trial outcomes. The cost of not modernizing is significant, as it limits the innovation potential and hinders the ability to make data-driven decisions that could improve patient care and trial success.
Moe: With growing imaging metadata, what new analytics are emerging, and how might they reshape trial success?
Jeffrey Sorenson: There are two exciting fields: pragmatic solutions and future-forward analytics. Pragmatically, we can extract simple yet powerful insights, such as the impact of body fat composition on drug metabolism, directly from images. These insights can optimize patient eligibility and trial outcomes. Future-forward, while radiomics holds promise, it requires extensive data collaboration across pharmaceutical companies and CROs. Our platform enables in-flight and post-hoc analysis, allowing sponsors to extract value from trial data, including even unsuccessful trials. This approach is crucial for advancing imaging analytics in clinical trials. For example, by integrating advanced processing capabilities into our system, sponsors can gain real-time insights into patient progression, identify trends across trials, and make informed decisions about safety and efficacy. This level of analytics not only enhances trial success but also paves the way for more personalized and effective treatments.
