--- title: "Designed for Approval, Judged by Adoption: Where Oncology IVD Clinical Studies Slip" description: "Discover why oncology IVD clinical performance studies often fail beyond regulatory approval, and how better study design supports reimbursement and clinical adoption." url: https://qbdgroup.com/en/blog/oncology-ivd-clinical-study-design-mistakes type: "Blog post" language: en published: 2026-07-10 author: "Annelies Rotthier" category: "Clinical" publisher: "QbD Group" citation: "QbD Group, \"Designed for Approval, Judged by Adoption: Where Oncology IVD Clinical Studies Slip\", https://qbdgroup.com/en/blog/oncology-ivd-clinical-study-design-mistakes" --- # Designed for Approval, Judged by Adoption: Where Oncology IVD Clinical Studies Slip > Discover why oncology IVD clinical performance studies often fail beyond regulatory approval, and how better study design supports reimbursement and clinical adoption. For: Founders, R&D leads, and clinical strategy teams at early-stage oncology IVD companies planning or running clinical performance studies In oncology in vitro diagnostics, the delays that stretch timelines and drain capital rarely originate in the science. The science is usually the part the team is best at. They originate in clinical performance study design. Samples that seemed available prove insufficiently annotated. Recruitment runs at a fraction of the projected rate. A study built to satisfy a Notified Body turns out to be insufficient for the Health Technology Assessment body. Each of these is preventable. Each is a planning gap, not a scientific one. Clinical performance studies for oncology devices are especially challenging because these problems compound. Patient populations are often small and biomarker-stratified. Sample material is scarce and precious. Meaningful clinical endpoints can require years of follow-up. Tumour heterogeneity complicates both design and interpretation. And regulatory and reimbursement evidence requirements rarely align. In a recent QbD Group webinar, **Annelies Rotthier, PhD** (Business Unit Manager, IVD CRO and Clinical Evidence Services at QbD Group) and **Egbert Smit, PhD MBA** (CEO, MLA Diagnostics) examined where clinical performance studies most often go wrong in oncology, and shared good practices to prevent the kind of rework that pushes programmes off plan. The points below draw on that conversation. The full session is available on demand. ## Designed for Approval, Judged by Adoption A clinical performance study under IVDR demonstrates that the device performs as claimed and as intended. A study built solely to satisfy a Notified Body will produce that data. The difficulty is that reimbursement decisions turn on a different question, whether the result changes patient management in ways that justify cost, and clinical adoption turns on a third: whether a clinician would act on the result, in this context, at this point in the pathway. A CE mark can therefore be obtained with evidence that fails at reimbursement. The outcome is not regulatory failure; it is approved technology that is not used. Closing the gap is best achieved by designing the study, from the outset, to address the questions every audience will ask. > Manufacturers sometimes lock the study design before they have a good view of their intended purpose. Your intended purpose should be your anchor, your compass. It drives you in the right direction on how to set up your studies. > > — Annelies Rotthier, PhD, QbD Group ## Four Pressures on the Study Design The webinar walks through four areas where clinical performance studies in oncology most consistently break down. The disciplines that prevent each are addressed in detail in the session. The framing below indicates where the leverage points sit. ### Study Design Designs locked before the intended purpose is settled. Retrospective designs chosen for speed that prove insufficient for clinical utility claims at reimbursement. Overly broad designs that dilute effect size and leave the primary endpoint underpowered. In many oncology indications there is no agreed gold standard, forcing composite reference approaches that are valid but methodologically demanding and more heavily scrutinised by reviewers. ### Endpoints Regulatory endpoints and clinical utility endpoints are not the same construct. Sensitivity and specificity against a reference standard answer the Notified Body's question; treatment-decision change, disease-free survival, or impact on overall survival answer the HTA body's. Selecting endpoints by what is easiest to measure, rather than by what each downstream audience requires, is among the more expensive errors a sponsor can make. ### Samples Sample availability is the most consistently underestimated bottleneck in oncology IVD development. Biobanked material is fast to access but often missing the longitudinal clinical annotation that meaningful endpoints require. Prospective collection delivers the annotation but at significant cost. Low biomarker prevalence, tumour heterogeneity, and small biomarker-positive populations compound the difficulty. ### Technical and Statistical Complexity Complex workflows limit reproducibility across sites. Multiple endpoints, analytes, and subgroup analyses demand a pre-specified statistical analysis plan, not an afterthought. Insufficient statistical power is the most common reason studies are rejected or required to generate additional data. It is almost always identifiable in advance, and almost always too late once the study has run. > I would try to avoid commercial sample providers. They charge high prices, and the clinical annotation is not always reliable. Work directly with investigators. This gives you much better insight into sample quality and the data that comes with it. > > — Egbert Smit, PhD MBA, CEO, MLA Diagnostics ## A Stakeholder Discipline, Not a Stakeholder Slide Multi-stakeholder evidence planning is not a section in the protocol. It is a discipline applied at the outset, mapping the audiences whose decisions will determine the eventual fate of the test: the Notified Body; the HTA bodies and payers; the clinicians and key opinion leaders; the laboratory directors who will operate the workflow; and the procurement leads who control purchasing. Each has a different question. A well-designed study answers all of them from a single dataset. The webinar discussion includes the stakeholder mapping approach applied at MLA Diagnostics, the relationship-building disciplines that turn clinical sites and KOLs into eventual advocates for adoption, and the sample-feasibility framework used to pressure-test a protocol before it is locked. > You really need to understand all of their perspectives. What every stakeholder needs, and what they have to gain or lose from a new test, will influence how things are purchased and how things are reimbursed. > > — Egbert Smit, PhD MBA, CEO, MLA Diagnostics ## Key Takeaway A clinical performance study in oncology IVD is one of the largest investments a sponsor will make before commercial launch. The decisions that determine whether that investment produces a credible, defensible, adoption-ready evidence base are taken before the protocol is signed. The discipline is to take them deliberately, with all the audiences in view, and with the intended purpose as the anchor. ## Before You Lock Your Protocol: A Practical Checklist Use this to review your study plan before committing:

Clinical performance study readiness checklist

## Avoid These Mistakes. Start With the Right Foundation Watch the full on-demand webinar featuring Egbert Smit (MLA Diagnostics) and QbD Group's clinical experts for practical, experience-based guidance on getting your oncology IVD to market faster. **[Watch the free on-demand webinar →](/en/webinars/oncology-ivd-market-faster-what-founders-underestimate)** --- Source: https://qbdgroup.com/en/blog/oncology-ivd-clinical-study-design-mistakes — © QbD Group. Quote freely with attribution and a link back.