Oncology Clinical Trial Imaging: When Does “Worse” Become Progression?

A tumor gets bigger. A new lesion appears. A scan looks worse.

It may seem straightforward to call that disease progression. But in an oncology clinical trial, such a change is not necessarily the same as progression. Determining whether a patient’s disease has progressed requires more than reviewing a single image. It depends on a structured sequence of evidence:

Image → Measurement → Criteria → Time → Context → Assessment

Each component helps transform an imaging observation into a clinical trial event that can influence treatment decisions, endpoint evaluation, and the evidence generated for a therapy.

What Is Disease Progression in an Oncology Clinical Trial?

In an oncology clinical trial, disease progression is determined by applying prespecified response criteria to imaging findings over time. Changes in tumor size, new lesions and other imaging findings must be evaluated according to the criteria defined for the study, such as RECIST 1.1, iRECIST, mRECIST, RANO or PCWG, before they contribute to an assessment of progressive disease.

From Image Change to Tumor Response Assessment

The image is the starting point. It shows what has changed since a previous scan, such as an increase in tumor size, the appearance of a new lesion, or a change in the characteristics of existing disease.

Measurement then helps quantify the observed change, but the result can be influenced by lesion selection, measurement technique, and variability between assessments. Even small differences in how a lesion is selected, measured, or outlined can affect the resulting assessment, particularly when imaging findings are close to a prespecified threshold.

The applicable response criteria, such as RECIST 1.1, iRECIST, mRECIST, RANO, and PCWG provide the rules for interpreting those measurements. Choosing and implementing the appropriate criteria is essential to producing endpoint assessments that support the objectives of the trial.

Time also matters. Progression assessment is inherently longitudinal, with the current scan interpreted in relation to relevant prior imaging. Prior imaging provides critical context for understanding whether a finding represents meaningful disease change, a borderline variation, or a pattern that requires confirmation. However, the specific requirements for determining progression and confirming an assessment depend on the applicable response criteria and trial protocol.

One particularly important aspect of longitudinal assessment is the nadir, the smallest recorded tumor measurement during treatment. A tumor may initially shrink substantially, for example, from the size of a grapefruit to a pea, but subsequently grow to the size of a plum. Although it remains much smaller than at baseline, this regrowth may represent progression because the disease is growing despite ongoing treatment. This is why the current assessment must be evaluated against nadir, typically the smallest tumor burden recorded since treatment began, rather than simply the immediately preceding scan, which may not represent the appropriate reference point for determining progression. Without tracking this reference, gradual increases across successive scans could go unrecognized.

Beyond these measurements and their evolution over time, the findings must also be interpreted within the disease-specific and treatment-related context addressed by the applicable assessment framework. Treatment-related effects, new lesions, disease biology, and other factors can complicate what readers see and how criteria should be applied. This makes clear, prespecified assessment rules particularly important in clinical trials, where findings must be interpreted consistently within the defined review process.

Reader Variability in Oncology Clinical Trial Imaging

Even when readers evaluate the same images using the same response criteria, they may not always arrive at the same assessment.

Variability can arise from:

• Differences in tumor measurements
• Selection of different target or non-target lesions
• Findings that fall near response thresholds
• Treatment-related imaging effects
• Differences in criteria interpretation
• Interpretation in context of prior imaging

These differences do not necessarily mean that one reader is simply right and another is wrong. They may reveal areas in which the imaging data or application of the response criteria require closer examination.

Central review with adjudication provides a structured process for resolving discordant assessments between central readers. Examining the reasons behind these discordant assessments can help identify recurring sources of variability and opportunities to improve review requirements and assessment rules, thereby supporting greater consistency in trial data.

How Disease Progression Affects Oncology Clinical Trial Endpoints

In an oncology clinical trial, progression is not simply a radiological assessment; it is a defined trial event that contributes to endpoint evaluation, with implications for progression-free survival, treatment decisions, statistical analyses, and ultimately the evidence used to evaluate a therapy.

That is why rigor across the entire assessment pathway matters. Clearly defined imaging requirements and response criteria, appropriate reader training, consistent longitudinal assessment, and a prespecified process for resolving discordance all contribute to more reliable endpoint evaluation.

The key distinction is therefore not simply whether an image looks worse, but whether the observed change meets the prespecified criteria for progression when evaluated within the complete longitudinal assessment framework.