Design and Run Smarter Gynaecological Oncology Clinical Trials

Gynaecological oncology clinical trials are entering a more precise phase, shaped by biomarker-led recruitment, adaptive methodologies, richer data, and new approaches to patient participation.

The opportunity is not simply to run trials faster, but to design studies that produce more relevant evidence and translate biological insight into better clinical decisions.

Image

The Oncology Clinical Trial Landscape Is Changing

Gynaecological oncology has become an increasingly important testing ground for new approaches to clinical development.

Ovarian, endometrial, cervical, vulvar, and vaginal cancers differ substantially in biology, disease course, treatment options, and patient populations. A single trial model cannot address all of these variables effectively.

An analysis of more than 2,000 gynaecological oncology trials registered on ClinicalTrials.gov between 2007 and 2020 illustrates the scale and uneven distribution of this research.

The analysis identified 2,152 trials classified as gynaecological oncology, with cervical cancer accounting for 350 studies. Cervical oncology trials also enrolled a disproportionately large share of participants, reflecting the scale of the disease burden in some regions. 

Yet volume does not necessarily mean optimal evidence.

A more recent analysis of 279 NCI-sponsored interventional gynaecological cancer trials found that 45% focused on ovarian cancer, 27% on uterine cancer, and 20% on cervical cancer. Pharmaceutical agents accounted for 85% of interventions

 The analysis also found substantial variation in eligibility criteria, raising questions about who can realistically participate in research. 

Clearly, the next stage of gynaecology oncology clinical development needs to optimise the trial itself, not simply increase the number of trials.

That means asking better questions before recruitment begins.

Does the study identify the right population? Are eligibility criteria clinically realistic? Can the trial recruit from the population that will ultimately use the treatment? Are endpoints meaningful beyond statistical significance? And can emerging evidence influence the study without compromising its integrity?

These questions are moving trial design away from a fixed protocol towards a more integrated development model.

Patient Recruitment Starts With Trial Design

Patient recruitment is often described as an operational problem. However, the evidence suggests it is also a design problem.

The 2025 PENTAGON study examined factors associated with enrolment in gynaecological cancer trials at a multi-site academic and community practice. Of 230 patients assessed for eligibility, 151 screened positive for at least one trial, and 57 ultimately enrolled. 

The study found that certain trial characteristics were associated with higher enrolment.

Trials that did not exclude patients with a previous cancer history had a higher enrolment rate than those that did. Additionally, studies allowing previous chemotherapy also showed a trend towards stronger enrolment. 

That makes a strong argument, and now regulatory focus, that eligibility criteria should be considered alongside patient characteristics when designing a clinical trial portfolio. 

This is very important as highly selective eligibility criteria can improve scientific precision while reducing the pool of patients who can participate.

That creates a tension between trial purity and clinical relevance.

A patient population can be biologically attractive on paper but difficult to recruit in practice. If the eventual treatment is intended for a broader population, an excessively narrow development programme can also leave important evidence gaps.

The PENTAGON findings are exploratory and came from a single clinical practice, so they should not be generalised to every trial setting. But they reinforce a wider principle that trial feasibility should be evaluated from the patient's perspective as well as the sponsor’s. 

That means considering factors such as:

• Previous treatments and comorbidities
• Distance from trial sites
• Biopsy and testing requirements
• Frequency of in-person visits
• Disease stage
• Availability of molecular testing
• The practical burden placed on patients and caregivers

The question is not whether a patient meets the protocol. It’s whether the protocol has been designed around the patients the trial needs to reach.

View interview How Should Pharma Companies Audit AI Vendors?

How Should Pharma Companies Audit AI Vendors?

“This is one of the riskiest industries there is because you're dealing with human lives, you're dealing with experimental protocols, and you're dealing with regulatory bodies where you might not get another shot at that clinical trial."

Image

Biomarkers Are Reshaping Patient Selection

The growing use of biomarkers is changing what “the right patient” means in oncology.

A 2026 systematic review of gynecologic oncology trials in 2025 synthesised 23 important clinical studies reported, including 17 Phase III randomised controlled trials and six earlier-phase or non-Phase III studies. The review identified biomarker-driven treatment as one of the major themes shaping contemporary gynaecological oncology. 

Across ovarian, endometrial, cervical, and vulvar cancers, clinical development is increasingly incorporating biological characteristics into treatment decisions.

That includes markers such as BRCA, homologous recombination deficiency, mismatch repair status, PD-L1, and circulating tumour DNA.

The purpose is not simply to collect more molecular data, but to establish whether a biological characteristic can meaningfully inform treatment selection, prognosis, response, or resistance. 

This distinction is critical.

A biomarker does not automatically become a useful clinical biomarker because it correlates with disease. It needs to demonstrate sufficient analytical and clinical validity for its intended use.

For trial sponsors, this introduces another layer of complexity.

The trial needs to identify eligible patients through molecular testing before randomisation. Testing must be available at participating sites and samples need to be collected and handled consistently. The clinical endpoint must be appropriate for the decision being made and the statistical analysis must reflect the biological hypothesis.

This is where clinical development increasingly connects with translational research.

The biomarker is no longer simply an additional data point, and it must become part of the trial architecture.

The shift is visible in recent gynaecological oncology research. The 2026 review highlights the growing use of molecularly defined treatment strategies, including differential responses according to mismatch repair status in endometrial cancer and biomarker analyses in cervical and ovarian cancer.

This creates both opportunity and risk.

Better patient selection can make a treatment effect easier to detect. But increasingly narrow populations can also make recruitment more difficult and reduce the size of the eventual evidence base.

Precision therefore needs to be balanced with representativeness.

Image

Adaptive and Pragmatic Trials Make Development More Responsive

The traditional clinical trial is deliberately controlled. Its protocol is defined before recruitment, and changes are tightly governed.

That remains essential for generating reliable evidence. But some clinical questions benefit from a design that can respond to accumulating information.

Adaptive methodologies allow pre-specified elements of a trial to change in response to interim data. These may include sample size, treatment allocation, continuation decisions, or the dropping of ineffective treatment arms.

In gynaecological oncology, this principle is already being explored.

The ACTOv study, for example, is a multicentre Phase II randomised trial investigating an adaptive treatment strategy for relapsed platinum-sensitive ovarian cancer. The protocol used changes in the serum biomarker CA125 to inform carboplatin dose modulation, while maintaining a conventional randomised comparison.

The importance of such designs extends beyond one trial.

If evidence can inform pre-planned decisions during development, sponsors may be able to reduce unnecessary exposure to ineffective interventions and make more efficient use of patient populations.

But adaptive does not mean uncontrolled.

Statistical discipline is the foundation of adaptive development.

Decision rules must be established in advance and interim analyses need appropriate statistical controls. Data must arrive quickly enough to support decisions.

Operational teams need to understand how changes will be implemented and regulators need confidence that the resulting evidence remains interpretable.

A related development is the growing interest in pragmatic clinical trials.

The 2026 roadmap from the Gynecologic Cancer InterGroup argues for pragmatic approaches that better reflect routine clinical populations and settings. It highlights the potential role of real-world data, external or synthetic control arms, and AI and machine learning where conventional randomised approaches are difficult to execute, including rare cancers and narrow biomarker populations. 

This is particularly relevant to gynaecological oncology.

Some cancers have relatively small patient populations. Some molecular subgroups are smaller still. Recruiting sufficient numbers into conventional trials can therefore become a limiting factor.

Pragmatic design offers another route. The objective is not to weaken evidence standards but to reduce unnecessary separation between the trial environment and the clinical environment.

View interview AI in Pharma: Why the Future of Healthcare Starts With Patients, Not Tech

AI in Pharma: Why the Future of Healthcare Starts With Patients, Not Tech

Kate O’Reilly, President & Chair of the Healthcare Businesswomen’s Association (HBA) Dublin-Ireland Chapter & Healthcare Transformation Partner at Roche, discusses AI in pharma, patient engagement, and the future of healthcare innovation.

Image

AI Connects Data, Design, and Decision-Making in Smart Gynaecological Oncology Clinical Trials

AI adds another dimension to this evolution in designing smarter gynaecological oncology clinical trials.

The potential value of AI in clinical development goes beyond automated recruitment. Models can support feasibility assessment, identify potential participants, analyse historical data, model patient flow, and assist with interim analysis.

AI-assisted modelling could be a very important potential tool for predicting trial outcomes and improving recruitment, while data harmonisation can reduce variability across multi-centre studies.

But the strongest opportunity may lie in connecting these capabilities.

Consider a trial in which historical clinical data informs site selection. EHR systems then identify potential participants; molecular testing determines eligibility; an adaptive statistical model evaluates interim evidence; a central data environment harmonises information across sites, and patient-reported outcomes provide additional context.

Each component already exists in isolation. The strategic opportunity is to connect them into one evidence system.

Pharmatica’s recent Insights into EHR-supported patient recruitment makes a similar point. EHR systems can help identify potential participants, but technical matching alone does not solve the recruitment problem. Data quality, interoperability, workflow integration, and clinical acceptance remain essential.

The same principle applies to AI.

A model can identify a likely trial candidate. It cannot, by itself, determine whether the patient should enrol.

A prediction can inform a decision. It should not replace clinical judgement, informed consent, statistical governance, or regulatory accountability.

This is particularly important as AI moves deeper into regulated development.

The value of AI therefore depends on data quality, validation, transparency, governance, and appropriate human oversight.

What Smarter Gynaecological Oncology Trials Look Like

The future of gynaecological cancer clinical trials is unlikely to come from one new statistical method or one AI platform.

It will come from better-connected trial architecture.

This does not mean traditional randomised trials are becoming obsolete. They remain the foundation for establishing efficacy and safety. The opportunity is to make the surrounding development system more intelligent.

That means designing eligibility criteria with real-world populations in mind. It also means integrating biomarkers when they have a clear clinical purpose.

Adaptive approaches should be used where the scientific question supports them and pragmatic designs should be considered when conventional recruitment is unrealistic.

And it means treating data infrastructure as part of clinical development rather than a downstream technical function.

The 2026 pragmatic trials roadmap makes this broader shift explicit, including the potential for AI and machine learning to improve recruitment efficiency, align trial populations more closely with real-world demographics, and reduce infrastructure burdens, particularly in lower-resource settings. 

Image

The Clinical Compass for Smarter Gynaecological Oncology Clinical Trials

Gynaecological oncology offers a useful case study for the wider transformation of clinical development.

The science is becoming more precise, but that precision creates new operational challenges.

Molecularly defined populations can improve biological relevance while making recruitment harder. More data can improve decision-making while increasing the need for harmonisation and governance. Adaptive methodologies can increase responsiveness while demanding greater statistical and operational discipline.

Smarter trials therefore do not mean less rigour. They mean applying rigour more intelligently.

The question is whether clinical development organisations can bring these capabilities together.

The most valuable trial may not be the one that generates the most data.

It may be the one that generates the right evidence, from the right patients, at the right time, in a form that can support the next decision.

At Pharmatica, we examine the systems, strategies, and technologies shaping the future of Clinical Development, connecting trial design, data, biomarkers, AI, and patient participation to the decisions that determine whether promising science can translate into meaningful clinical impact.

Pharmatica: Insight. Connection. Impact.