Comprehending the spate of recent rejections in the cell and gene therapy space may require looking no further than early-stage clinical trials of candidates from REGENXBIO, Excision BioTherapeutics and Intellia Therapeutics.
The FDA’s recent rejection of a high-profile gene therapy application from REGENXBIO has reignited a familiar debate across the biotech industry. The therapy had generated substantial optimism, supported by encouraging preclinical data and years of regulatory engagement. Yet when the agency evaluated the full evidentiary package, the data ultimately failed to meet the threshold required for approval.
This decision is not an isolated event. Across advanced therapeutics, the field is seeing a widening range of regulatory outcomes. Some programs achieve regulatory success, as seen with recent approvals such as Sarepta’s Elevidys, reflecting alignment between clinical data and regulatory expectations. Others, such as BioMarin’s Roctavian, Astellas’ AT132 and Solid Biosciences’ SGT-001, stall in late-stage trials or fail to secure approval despite years of promising development signals.
These mixed outcomes raise an important question for the industry. Are the setbacks primarily the result of regulatory caution, or do they reflect deeper weaknesses in how evidence is generated and evaluated long before therapies reach the clinic? In my role as cofounder and chief operating officer of Revalia Bio, whose platform helps to enable human data trials, I’ve observed that increasingly, the answer appears to involve both.
The industry’s growing concern: Are early signals reliable?
Drug development has always involved uncertainty. Historically, roughly 90% of therapies entering clinical trials fail to achieve approval. But the recent pattern of setbacks in advanced therapeutics has intensified scrutiny of the earliest stages of development, where foundational evidence is generated. Regulators, investors and developers I’ve spoken to are asking a similar question: How reliable are the signals that shape early decision-making?
Early biological data influences nearly every aspect of a program’s trajectory. It informs trial design, patient selection, dosing strategies and regulatory engagement. When the underlying evidence is robust, development strategies can proceed with confidence. When it is less predictive, uncertainty accumulates throughout the pipeline.
Late-stage setbacks do not always arise because a clinical trial fails in the conventional sense. In some cases, as seen with REGENXBIO’s RGX-121, the core issue lies in how efficacy is defined and measured. The FDA’s rejection did not hinge on a clear lack of safety or biological activity, but rather on concerns around the primary endpoint itself, the reliance on a biomarker not yet validated to predict clinical benefit, and the use of an external natural history control.
In such cases, early data may appear compelling, but ultimately, it still falls short of demonstrating meaningful clinical impact. As a result, the limitation is not simply that a trial fails, but that the evidence used to justify advancing into late-stage development may be insufficiently anchored to outcomes that regulators consider clinically reliable.
Gene and cell therapies have amplified this issue because of their inherent complexity. Unlike traditional small-molecule drugs, many advanced therapies aim to modify biological systems at a fundamental level—editing genes, delivering viral vectors or reprogramming immune responses. These interventions can have durable or even permanent effects, making safety and efficacy signals particularly critical.
The biological mechanisms involved are often highly sensitive to factors such as tissue specificity, immune responses, delivery efficiency and patient variability. Small gaps in understanding during early development can translate into significant uncertainty once therapies reach human trials. In other words, advanced therapeutics do not necessarily create new translational challenges. They simply make existing ones harder to ignore.
The structural challenge: Evidence built on incomplete biology
Much of today’s development infrastructure still relies on experimental systems that were designed decades ago. Animal models and simplified laboratory systems have played an indispensable role in drug discovery and preclinical safety testing. They remain essential tools for understanding biological mechanisms and identifying potential risks. But when therapies depend on highly specific interactions within human biology, these systems can struggle to capture the full picture.
Certain immune responses, disease pathways and tissue dynamics differ substantially between species. Even closely related organisms may respond differently to gene delivery systems or long-term protein expression. As a result, therapies that appear promising in preclinical models can behave differently when tested in human patients. These limitations do not invalidate traditional approaches. However, they do introduce uncertainty into the early evidence base that developers rely on when advancing therapies into clinical trials.
When early biological insight is incomplete, developers often attempt to manage uncertainty through trial design. Adaptive trials, expanded cohorts, biomarker-driven endpoints and accelerated regulatory pathways are all tools that can help gather data more efficiently. These strategies have become increasingly common in advanced therapeutics, particularly in areas with high unmet medical need. But trial design can only do so much.
Clinical trials are ultimately designed to test hypotheses about safety and efficacy. If those hypotheses are built on incomplete biological understanding, even well-designed studies may struggle to generate clear answers. In such cases, the burden of resolving fundamental uncertainty shifts to regulators, who must evaluate whether the available data sufficiently support approval.
Regulatory pressure is increasing
As advanced therapies become more powerful the stakes of regulatory decisions grow correspondingly higher. Regulatory agencies must balance two competing priorities: accelerating access to innovative treatments and ensuring that therapies entering the market are safe and effective.
For interventions with long-lasting biological effects, the evidentiary bar may rise accordingly. Recent decisions such as the FDA’s initial rejection of BioMarin’s Roctavian, the complete response letter for REGENXBIO’s RGX-121 and ongoing scrutiny of Sarepta’s Elevidys suggest that regulators are examining the strength of early biological evidence more closely, particularly when therapies rely on complex mechanisms or when long-term outcomes remain uncertain. This scrutiny is not necessarily a sign of resistance to innovation. Rather, it reflects the responsibility regulators carry when evaluating therapies that may permanently alter biological systems.
In response, many researchers and biotechnology companies are exploring ways to strengthen the evidence pipeline before clinical trials begin. Across academia and industry, including at Revalia Bio, there is growing interest in integrating human biological data earlier in development. Advances in organoid systems, organ-on-a-chip technologies and ex vivo human tissue models are providing new ways to study disease mechanisms in human-relevant contexts.
At the same time, increasingly sophisticated computational models and large-scale biological datasets are helping translate complex biological signals into more predictive insights. The goal is not to replace traditional experimental systems but to complement them with additional sources of evidence that better reflect human biology. By combining multiple forms of human-derived data, developers may be able to reduce uncertainty earlier in the development process.
Rethinking the evidence pipeline
The recent wave of gene therapy setbacks may ultimately reveal something broader than isolated program failures. They may reflect structural challenges in how the industry generates and evaluates evidence long before therapies reach patients. If early biological signals are incomplete or insufficiently predictive, uncertainty accumulates across every stage of development. Strengthening the evidence pipeline earlier could reduce the likelihood of late-stage surprises and improve confidence across the ecosystem.
Gene therapy remains one of modern medicine’s most promising frontiers. But as therapies become more sophisticated, so must the frameworks used to evaluate them. The next phase of innovation may depend not only on new technologies but on better ways of understanding human biology before clinical trials begin.
