by Tim Bolus, Software Product Manager, Molecular Devices
As cell-based assays become increasingly central to pharmaceutical quality control, laboratories face an important challenge: how to integrate imaging workflows into validated GMP environments while maintaining data integrity standards. While regulations such as FDA 21 CFR Part 11 and EU Annex 11 have long governed electronic records in pharmaceutical settings, the expansion of imaging technologies into QC applications is revealing gaps that the industry has to address.
The Regulatory Landscape
The core regulatory framework governing electronic records and data integrity in pharmaceutical laboratories has remained relatively stable. FDA 21 CFR Part 11 continues to focus primarily on the data integrity of electronic records and electronic signature requirements, ensuring that any data used to make decisions about products for human use is captured, stored, and protected appropriately.
In the EU, Annex 11 serves a similar function and is currently undergoing revision. Perhaps more significantly, regulators are developing Annex 22, a new framework specifically addressing artificial intelligence (AI) and machine learning applications in pharmaceutical contexts.
This emerging regulation reflects the growing role of AI-driven analysis in imaging workflows, where algorithms increasingly support decisions about cell counting, confluence assessment, and viability testing. The challenge for QC laboratories is not that regulations have dramatically changed, but rather that imaging technologies are now being applied in contexts where compliance requirements were originally designed for simpler electronic systems.
The Validation Barrier
The most significant barrier to adopting imaging systems in GMP QC laboratories is instrument validation. Unlike research environments where flexibility is important, QC labs cannot use any instrument that has not undergone formal installation qualification (IQ) and operational qualification (OQ) procedures.
This requirement, outlined in EU Annex 15, applies not only to instrumentation but also to the test methods themselves.
For imaging systems, this creates a particular challenge. Many cytometers and imaging platforms were developed primarily for research applications, where the emphasis is on capability and throughput rather than reproducibility and traceability. Until recently, cytometers have been largely absent from QC labs precisely because the validation infrastructure did not exist.
The Workflow Fragmentation Problem
The workflow fragmentation problem compounds these difficulties. In a typical cell-based potency assay, laboratory personnel may need to count cells manually using a hemocytometer and microscope, then transfer samples to a plate reader for the actual assay, and finally move data into a laboratory information management system (LIMS) for analysis and reporting.
Each transition point represents a potential source of error and a gap in the audit trail. When assays are developed in R&D laboratories and then transferred to QC, the cell verification step often becomes a bottleneck. The assay itself may be validated, and the cell line qualified, but without a validated system, there is no way to verify that instruments—such as microplate readers and imagers—are performing within their specifications.
This gap can introduce variability that obscures whether potency differences reflect genuine manufacturing issues or simply inconsistent cell preparation.
Why Data Integrity Demands Extend to Imaging
The fundamental purpose of data integrity requirements is traceability. Audit trails must tell a complete story: who performed each action, when they performed it, what the original data showed, and what changes were made.
This narrative capability serves both scientific and regulatory purposes.
From a scientific perspective, robust audit trails enable laboratories to investigate unexpected results. If a potency assay yields an anomalous value, investigators can trace back through the entire workflow to identify potential causes. Was the cell confluence within specification? Did the instrument pass its most recent calibration? Were any parameters modified during analysis? Did a user modify results? If so, were there proper change controls in place?
From a regulatory perspective, audit trails protect against both inadvertent errors and deliberate manipulation. When electronic records can be modified without detection, results can be adjusted to meet specifications—an issue known as “testing into compliance.” Comprehensive audit trails make such modifications visible, creating accountability that protects product quality and patient safety.
For imaging workflows, these requirements mean that every image acquisition, every algorithmic analysis, and every cell count must be captured in a tamper-evident record. The same rigor that applies to plate reader data must extend to the imaging systems that increasingly support QC decision-making.
Algorithm Reproducibility
One often-overlooked aspect of imaging validation is algorithm reproducibility. When a laboratory reads the same plate five times, intuition suggests the results should be identical. In practice, many imaging algorithms produce slightly different results on each run because of how they identify and segment cellular objects.
For research applications, this variability may be acceptable. For QC applications, where specifications are tight and out-of-specification results trigger investigations, algorithm precision becomes critical. Laboratories implementing imaging in validated environments should evaluate not only whether an algorithm produces accurate results on average, but whether it produces consistent results across repeated measurements.
Preparing for AI and Automation
The pharmaceutical industry is moving steadily toward greater automation, driven by the need to reduce costs, improve throughput, and minimize human error. Imaging workflows sit at the intersection of this trend with emerging AI capabilities, making them a focal point for regulatory attention.
Current regulatory thinking emphasizes that AI and machine learning systems must be validated with the same rigor as any other system that influences product quality decisions. When an algorithm determines whether cells meet acceptance criteria, that determination must be defensible in an audit. Laboratories should anticipate that regulators will expect documentation of algorithm training, validation datasets, and ongoing performance monitoring.
The development of Annex 22 in the EU signals that regulators are actively working to establish frameworks for AI governance in pharmaceutical contexts. Laboratories that begin building validation strategies for AI-enabled imaging now will be better positioned when formal requirements emerge.
Guidance for Laboratory Leaders
For QC leaders evaluating imaging solutions, several principles should guide decision-making.
First, prioritize systems designed for regulated environments.
Research-grade instruments may offer impressive capabilities, but if they cannot support audit trails, electronic signatures, and instrument qualification protocols, they will not meet GMP requirements.
Second, consider workflow integration.
Systems that combine imaging and plate reading under a single validated software environment reduce the number of separate validations required and eliminate data transfer gaps that can compromise traceability.
Third, engage technology providers as compliance partners rather than simply equipment suppliers.
The responsibility for validation ultimately rests with the laboratory, but technology providers who offer qualification protocols, validation guidance, and regulatory expertise can significantly reduce the burden.
Finally, understand that validation strategy is a business decision as well as a technical one.
Regulations require that systems be validated, but do not prescribe how extensively. Laboratories are at liberty to balance the cost of validation activities against the risks of inadequate documentation, making deliberate choices about validation depth that they can defend to auditors.
Last Thoughts
Integrating imaging workflows across GLP environments—which include pre-clinical and clinical development phases—and GMP QC laboratories—involving manufacturing product used in clinical trials through commercialization—presents both an opportunity and a challenge. The technology exists to automate cell counting, improve assay consistency, and generate richer data about cell-based processes. Realizing these benefits in regulated environments requires careful attention to validation, data integrity, and algorithm reproducibility.
As regulators develop new frameworks for AI and machine learning, and as vendors deliver solutions specifically designed for GMP applications, the barriers to imaging adoption in QC will continue to fall. Laboratories that invest now in understanding compliance requirements and building validation capabilities will be best positioned to capture the efficiency gains that imaging technologies promise.
About the author: Tim Bolus is Software Product Manager at Molecular Devices, with more than 20 years of experience in biotechnology and pharmaceutical compliance, spanning GMP and GLP requirements across research, development, and quality control environments.