ICH E6(R3) Raises the Bar on Imaging in Clinical Trials—Are You Ready?

Lori
03.09.25 06:08 PM

Article originally posted on The Clinical Trial Vanguard by Moe Alsumidaie on September 3, 2025

When the FDA inspector arrived, no one expected imaging to be the issue.

The scans had been read. The assessments were submitted. But there were no audit trails—just static PDFs, manual uploads, and spreadsheets emailed between radiologists, sites, and CROs. What once passed with a nod and a note now triggered a formal inspection finding.


With ICH E6(R3), that’s no longer a fluke. It’s the future.

Sponsors are expected to work with their clinical research sites to ensure they have processes and technology in place to extend quality programs where it matters most, near the patient. Sites can expect formal inspection findings related to these new guidelines, if not fully prepared.

Imaging, the primary endpoint in over 90% of oncology trials, has been one of the most error-prone and loosely governed data types in clinical research. But the new ICH E6(R3) guidance changes that. Imaging data must now meet the same Good Clinical Practice (GCP) standards as EDC-captured lab results, ePROs, and adverse event reports.

The shift is timely. According to Enhancing Data Integrity: New Clinical Trial Imaging Techniques, legacy processes are responsible for a sizable portion of inspection findings related to data traceability and measurement discrepancies. When radiologists submit read results outside the parameters required in the protocols, or site staff input lesion measurements based on protocol non-compliant PACS reading workflows manually into homegrown spreadsheets, trials inevitably inherit a fragmented data trail that cannot be defended during audits or monitor inspections.

Meanwhile, sponsors are shouldering the enormous downstream costs of these failures. As explored in The Unseen Financial and Human Toll of Clinical Trial Imaging Errors, we reported error rates of 25%, 30%, and 50% at three NCI-designated Comprehensive Cancer Centers before implementing a protocol-compliant imaging informatics platform—rates that were reduced to under 3% after adoption. Central review cannot fix this problem because site reads determine which patients are eligible to participate on trials. Thus, the chaos of site reads leads to secondary central reviews censoring over 10% of patients who were inappropriately enrolled and treated on trials they did not qualify for. This is a travesty for patients who were not afforded their chance at another line of therapy that could have been their true best hope.

For sponsors who ubiquitously rely on non-traceable site reads and outdated workflows, the compliance clock is ticking. As R3 puts it,

“Procedures should be in place to cover the full data life cycle… for review of trial-specific data, audit trails and other relevant metadata should be in place. It should be a planned activity, and the extent and nature should be risk-based, adapted to the individual trial and adjusted based on experience during the trial.” – ICH E6(R3)

Source Data Must Be Traceable—That Includes Imaging

ICH E6(R3) defines source data as:

“Original documents or data (which includes relevant metadata) or certified copies of the original documents or data, irrespective of the media used.” – ICH E6(R3)

This presents a major challenge for trials where imaging assessments are scattered across PDFs, local drives, PACS systems, or third-party portals that lack integration or audit-ready infrastructure.

Computerised systems maintain logs of user account creation, changes to user roles and permissions and user access; … Systems are designed to permit data changes in such a way that the initial data entry and any subsequent changes or deletions are documented, including, where appropriate, the reason for the change. – ICH E6(R3)

Legacy imaging workflows often depend on radiology reports emailed as attachments, scans uploaded via manual portals, and assessment results copied into spreadsheets by CRCs. These workflows result in fragmented, non-standardized records that are difficult to audit, nearly impossible to reconcile against the database, and prone to delays or transcription errors.

This is especially dangerous in oncology, where imaging is often the primary endpoint. Unlike lab values, which are structured and timestamped in EDCs, imaging readouts may never be tied back to the originating DICOM files and anatomic measurements—or worse, may vary based on subjective interpretations or an inconsistent application of required tumor response criteria measurements.

Yunu, a cloud-based imaging workflow platform, addresses this gap directly. It ensures the accurate conduct of radiology reads in real time across all trials at the site, as well as all sites in each trial. The workflow-driven approach maintains full audit trails, ties each read results and image annotations to the original source images, and offers harmonized, granular, timestamped traceability across all participating sites and reader groups—meeting ICH expectations for source data verification, inspection readiness, and GCP-aligned data integrity requirements in real time.

Protocol Deviations in Radiology Reads on Trials = Compliance Failures

Imaging deviations often go unnoticed—until it’s too late: lesions measured inconsistently across visits, RECIST 1.1 misapplied, or the wrong criteria used entirely. These aren’t just operational slip-ups. Under ICH E6(R3), they undermine trial endpoints and trigger compliance risk.

In panel discussions at SCDM and audits across leading NCI-designated cancer centers, imaging protocol deviations were found in up to 50% of reads in early-phase oncology trials—driven by inconsistent criteria, missed measurements, and subjective lesion tracking.


Adoption of structured imaging informatics platforms like Yunu’s reduced these error rates to under 3%, enforcing real-time protocol adherence and traceability.

If imaging deviations alter or obscure endpoint data—and sponsors often only detect them after database lock—then these foundational expectations for GCP and data quality and integrity have already been breached.


Jeffrey Sorenson, CEO of Yunu

“The updated ICH E6(R3) guidance places greater emphasis on real-time protocol adherence and risk-based monitoring to address issues exactly like these. But most imaging workflows are still built for retrospective cleanup—not proactive compliance.“

Traditional workflows often rely on static imaging data files manually uploaded after patient visits, subjective assessments written in free-text fields, and emails or spreadsheets shared between sites and radiologists.

Procedures for review of trial-specific data, audit trails and other relevant metadata should be in place. It should be a planned activity, and the extent and nature should be risk-based. – ICH E6(R3)

As detailed in the article Addressing the Accuracy Crisis in Clinical Trial Imaging, this fragmentation leads to “a delayed feedback loop where protocol violations aren’t caught until several cycles later—if at all.” Without structured, real-time capture, imaging data becomes vulnerable to inconsistencies, misreads, and delayed recognition of critical deviations—all of which compromise endpoint reliability and breach the proactive quality management expectations outlined in ICH E6(R3).

At the point of data capture, automated data validation checks to raise data queries should be considered as required based on risk, and their implementations should be controlled and documented. – ICH E6(R3)

Real-Time Oversight Is No Longer Optional

ICH E6(R3) places significant weight on sponsor oversight and RBQM (Risk-Based Quality Management) and imaging can no longer be overlooked. Busy radiologists in the sites may no longer have the time required. Almost none have the local systems required. Most are not trained or experienced with the measurement criteria.

Traditionally, imaging oversight has relied on retrospective review—where monitors sift through image files and reports weeks or months after acquisition. There’s little to no in-flight visibility into whether site assessments were done on time, per protocol, or using consistent response criteria. As noted in Addressing the Accuracy Crisis in Clinical Trial Imaging, this delay creates a blind spot: “reviewing data long after the patient has completed treatment limits the ability to identify deviations when they happen—and makes real-time correction impossible.”

Thes updated guidance in ICH E6(R3) reinforce that real-time visibility is not just beneficial—it’s a GCP expectation

The sponsor should ensure that the investigator has timely access to data … including … centrally read imaging data. – ICH E6(R3)

Where does this leave sites and sponsors? Thankfully, Yunu is a solution available to addresses these requirements directly. The platform offers real-time dashboards that give sponsors, CROs, and imaging stakeholders live visibility and access to:

  • Read status
  • Site compliance
  • Lesion tracking
  • Imaging queries and resolution times
  • De-identified source DICOM images
  • Data lock ready results


This allows site staff and sponsor study teams to intervene quickly when issues arise—with a codified solution that works across any number of sites and trials simultaneously, preventing protocol deviations, accelerating timelines, and ensuring that imaging reads contribute reliably to efficacy signals.


It’s not just about speed—it’s about integrity. In one trial, Yunu helped reduce site monitor image handling across the entire trial by up to 80%, freeing resources to focus on patients and compliance while aligning imaging activities with ICH E6(R3)’s vision for continuous quality and real-time control.


Decentralized Imaging Requires Centralized Standards

Decentralized clinical trials (DCTs) are on the rise—and ICH E6(R3) formally recognizes their role in expanding access and enhancing patient centricity. However, this flexibility comes with a clear caveat: data integrity must be preserved. That means centralized standards must apply, regardless of where or how the data is captured.

In imaging, this poses a major challenge. Decentralized trials with primary imaging endpoints may have radiology reads performed at local hospitals, community centers, or via tele-radiology, often using different systems, protocols, and tumor response interpretations. Without harmonization of techniques and processes, the result is inconsistent reads, subjective measurements, and conflicting endpoint data.

Simply stated, decentralization presents an even greater challenge for GCP compliance related to imaging endpoints.

Yunu solves this problem at scale. The platform enforces uniformity through:

  • Built-in response criteria (RECIST, Lugano, RANO, etc.) covering all major measurement types and possible modifications
  • Configurable workflow templates for academic, community, and outsourced readers that support site, central, BICR, and hybrid reads
  • Web accessible dashboards that monitor in-flight operational and clinical results across large numbers of trials, sites, and subjects


In one real-world deployment, Yunu enabled a major NCI-designated cancer center to connect its imaging protocol to a regional hospital network. This allowed patients in underserved areas to participate in oncology trials locally while expert radiologists hundreds of miles and a mountain range away work in complete harmony with their treating oncologists.


This was not only a technical milestone but a model for equity in action. As decentralization expands, the ability to harmonize reads across diverse settings becomes more critical—not because decentralization is flawed, but because consistency is foundational to data integrity.


Platforms like Yunu don’t replace decentralization—they enable it at scale by ensuring that imaging data remains consistent, compliant, and audit-ready across all environments.


Imaging Has Reached a Compliance Crossroads

For years, imaging in clinical trials operated in a regulatory gray zone—mission-critical, yet loosely managed. That era is over.

With ICH E6(R3), imaging data must be traceable, auditable, and monitored in real time. The bar has been raised. The FDA is watching. And the burden is no longer just on central labs, it’s on sites, sponsors, CROs, and tech partners to get imaging right.

Yunu is already enabling that shift at scale, harmonizing imaging data across thousands of trials, reducing error rates, and enabling decentralized reads without compliance risk.


Avatar   
Website   |   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.

Lori