In colorectal cancer (CRC), biomarkers have long been used to decide if patients are eligible for a particular therapy or clinical trial. The underlying logic is straightforward: match the target, select the patient.
CRC remains a major global health challenge, affecting around 1.9 million people worldwide and ranking among the leading causes of cancer mortality, with many patients diagnosed at an advanced stage where treatment options are limited.
Access deeper industry intelligence
Experience unmatched clarity with a single platform that combines unique data, AI, and human expertise.
As the number of targeted therapies has expanded, applying this model has become more difficult. A single biomarker result does not necessarily capture the underlying tumour heterogeneity or how a tumour may behave and respond to treatment.
This is starting to challenge the assumption that just one biomarker test can reliably guide development or treatment decisions in CRC. In this setting, the question is no longer whether a biomarker is present, but whether it is enough.
Why single-marker approaches are not always sufficient
Biomarker-driven development has delivered important advances in oncology. However, as understanding of tumour biology has deepened, it has become clear that CRC tumours often rely on multiple biological drivers to overcome resistance mechanisms rather than a single dominant pathway, particularly as resistance develops.
Multiple signalling pathways can contribute to disease progression at the same time, meaning tumours may depend on more than one mechanism to grow and survive.
Rather than appearing in isolation, these biomarkers may be present together within the same tumour. A patient may therefore test positive for more than one relevant biomarker within the same tumour – indicating that more than one biological pathway may be driving the cancer.
This has direct implications for patient selection. Restricting eligibility to a single marker can exclude individuals whose disease is influenced by additional drivers, particularly where trial criteria require patients to meet a single, specific biomarker definition. However, expanding inclusion criteria may introduce a different challenge, making trial outcomes harder to interpret when responses vary across biologically distinct groups.
How biomarker overlap is reshaping development decisions
Once patient populations are selected based on multiple biomarker criteria, trial design becomes more complex.
Evidence from across multiple solid tumours shows that biomarkers often coexist rather than define clearly separated groups. In CRC, this affects decisions around inclusion criteria, cohort structure, and statistical analysis.
From an early development perspective, isolating clear efficacy signals becomes more difficult when populations are heterogeneous. From a clinical perspective, patients rarely fit neatly into a single biomarker-defined category, highlighting a gap between trial design and routine care.
Development and medical teams are therefore balancing two competing needs: maintaining enough precision to detect meaningful outcomes, while ensuring studies reflect the diversity of patients seen in practice. Addressing this requires trial strategies that can accommodate overlapping biology from the outset, rather than treating it as an exception. This is increasingly being considered earlier in development, as teams look to anticipate how overlapping biology may affect both response, resistance and survival.
Why emerging therapies are challenging existing frameworks
The way newer therapies work adds another layer of complexity. Many are not confined to a single biological pathway.
Some treatments are designed to both target tumour cells directly and influence the surrounding immune microenvironment.
This creates a mismatch between the therapy mechanism of action and traditional patient selection frameworks. A therapy developed for a specific target may still show activity in tumours driven by multiple overlapping mechanisms.
Predicting response becomes less straightforward under these conditions. Biomarker results may point in different directions, requiring more nuanced decision-making, including consideration of sequencing and combination strategies, rather than a single binary choice. This suggests that patient selection strategies will need to evolve alongside therapy design, rather than relying on historical biomarker frameworks.
How trial design is starting to evolve
These developments are already influencing how clinical trials are structured.
Rather than relying solely on fixed inclusion criteria, there is increasing interest in designs that better reflect the variability seen in patients. The aim is to adapt biomarker-driven approaches to a more complex biological landscape.
Key shifts include:
More flexible cohort structures
Trials are increasingly including patients with more than one biomarker, rather than restricting enrolment to narrowly defined groups.
Greater use of adaptive trial designs with novel endpoints
Basket, umbrella, and adaptive trials – which allow multiple patient groups or biomarkers to be studied within a single protocol and modified as data emerge – are becoming more widely used. Novel effect measures such as ctDNA and MRD are becoming more dynamic and are emerging as regulatory acceptable endpoints, enabling identification of patients at risk for early progression and carrying predictive value with conventional endpoints in oncology clinical trials.
Integration of real-world evidence
Randomised controlled trials remain essential, but they do not always reflect the full range of patients seen in clinical practice. Real-world data (RWD), such as information from electronic health records or registries, provide insight into how therapies perform across broader populations.
From a medical affairs perspective, combining RWD with clinical trial data supports a more complete understanding of how findings translate into routine care.
Moving beyond single biomarkers
Combining genomic, clinical, and real-world data allows for a more detailed view of disease biology. Advances in data analysis are helping identify patterns that would not be visible using a single biomarker alone.
Digital pathology is also poised to operationalise this shift beyond single biomarkers in CRC by transforming routine histology into a high-dimensional, spatially resolved readout of tumour biology. When integrated with genomics and clinical data, these composite signatures can be prospectively embedded in trials to stratify patients, enrich for responders, and help link complex biological patterns to clinical outcomes.
Organisations such as AbbVie are exploring these approaches, integrating different data sources and trial designs to better reflect how the disease behaves outside controlled study settings. For development and medical teams, this represents a shift toward more flexible and integrated trial models.
How sequencing and resistance are influencing strategy
Patient selection is also shaped by how the disease evolves over time. CRC treatment typically involves multiple lines of therapy.
For example, cancers may initially respond to treatment but later progress by activating alternative pathways that allow them to continue growing despite initial therapy, for instance, through RAS/RAF bypass signalling or MET amplification. These dynamics influence not only which therapies are used, but also their timing.
As a result, sequencing treatment strategies are increasingly being considered earlier in oncology development. Combination approaches are also gaining attention, particularly where tumours are driven by more than one biological mechanism, including novel combination approaches.
What emerging data may help clarify
Data from upcoming scientific meetings, including the American Society of Clinical Oncology (ASCO), the European Hematology Association (EHA) Congress and the European Society for Medical Oncology (ESMO) Congress, are expected to provide further insight into how these challenges are being addressed.
Ongoing studies are examining how therapies perform in patients with overlapping biomarkers, and whether broader or more flexible selection frameworks can support more consistent outcomes.
These findings are likely to shape how future trials are designed and how patient selection strategies evolve.
What needs to change in patient selection
CRC remains an area of significant unmet need, particularly in advanced disease where treatment options are limited and outcomes remain poor.
Addressing this requires more than identifying new targets. It also depends on rethinking how patient selection is approached.
Continuing to rely on single-marker frameworks risks oversimplifying the disease and limiting how new therapies are evaluated. This means a more integrated strategy is needed – one that considers multiple biological signals and how they interact within individual patients, rather than relying on a single test result to guide decisions.
In this setting, designing trials around a single biomarker risk overlooking the biological complexity that ultimately determines how patients respond to treatment. Across the industry, including at AbbVie, there is increasing focus on aligning early development, clinical evidence generation, precision medicine approaches and real-world insights to better understand how therapies can be applied across complex patient populations.
Designing trials that reflect real-world biology while producing interpretable results remains a key priority. From a clinical and medical affairs perspective, there is also a need to ensure that emerging evidence translates into practical guidance for treatment decisions. Ultimately, improving outcomes in CRC will depend not only on scientific advances, but on developing patient selection strategies that better reflect how the disease presents and evolves in practice.
