Beyond colorectal cancer: how deep molecular profiling uncovers the biology that actually drives patient outcomes
Colorectal cancer (CRC) is not a uniform disease. Patients may share the same diagnosis, but the biological mechanisms driving their tumors can differ fundamentally - that is why treatment responses vary so widely. We need to start understanding the complexity of the disease to develop targeted therapies responding to the right patient population.
CRC remains one of the most complex tumor types in oncology. Despite major advances in targeted therapies and immuno-oncology, many new programs fail in clinical trials. Why? A major reason is that tumor biology is still not characterized deeply enough to identify the right patients for a given therapy upfront. This leads to trial cohorts that are too heterogeneous, with targets that generate a biological effect in some patients but not others, and to late‑stage failures that could have been anticipated much earlier. Progress requires a multidimensional understanding of what is actually driving each individual tumor – one that goes beyond a single genomic or transcriptomic snapshot.
Everything starts with the quality of the biological material. Capturing cancers in their native state is essential. Minimal ischemia times during resection, combined with highly standardized collection procedures in the OR, ensure comparability across patients. The resulting samples truly reflect tumor biology. Together with pre‑clinical and longitudinal follow‑up data, deep molecular profiling enables us to understand the true complexity and heterogeneity of CRC.
This enables better matching of patient groups to therapy from the get-go, de‑risking clinical trials and driving smarter development pathways.
Four studies. Four dimensions of CRC biology.
Through continuous collaboration with our Global Clinical Network, we continue to sharpen our understanding of cancer biology. These close partnerships have enabled insights that translate into high‑quality scientific publications. The studies below illustrate what becomes visible when this level of biological and clinical detail is applied to CRC.
)
1. Exon-Skipping: making molecular subtyping clinically feasible
What we found: In collaboration with the Wilmot Cancer Institute, we showed that just 29 exon-skipping RNA splicing events can reliably classify CRC tumors into clinically meaningful consensus molecular subtypes (CMS) – the four major CRC subtypes defined by distinct biological and clinical characteristics. This splicing-based approach outperformed traditional gene-expression classifiers while dramatically simplifying the assay design.
Why it matters: One of the main barriers to molecular subtyping in routine clinical practice has been assay complexity. A classifier based on 29 RNA splicing events is faster and more scalable than conventional transcriptional panels, making CRC subtype information genuinely accessible at the point of diagnosis. In drug development, that translates to more refined patient selection and stronger trial design from the outset. For patients, it points toward a future where subtype information is available earlier, enabling more tailored treatment decisions sooner.
)
2. Left vs. right-sided CRC: a tractable molecular classifier for stratification
What we found: Left- and right-sided colorectal tumors display fundamentally different genomic and signaling profiles, as well as different responses to targeted therapies - a clinical distinction that has been recognized for years. We identified a pair of transcription factors capable of nearly perfectly distinguishing left- from right-sided tumors at the molecular level, providing a precise molecular basis for a difference that previously relied on anatomy alone.
Why it matters: This transcription factor signature creates a tractable classifier for patient stratification. Tumor location moves from a descriptive label to a molecularly defined subtype with predictive value, one that could be embedded into diagnostic workflows to improve therapy design and sharpen cohort selection in trials.
)
3. Immune signatures: refining patient selection beyond MSI status
What we found: MSI-H status is the current standard for selecting CRC patients for immune checkpoint inhibitor (ICI) therapy, but it does not fully predict who will respond. Our pan-cancer multi-omics analysis showed that immune context matters. Immune signatures, further dissected based on their immune cell type of origin, reveal meaningfully different immune profiles across tumors, including within MSI-H tumors where response rates still vary considerably.
Why it matters: MSI-H alone is not sufficient for patient selection. Immune cell-specific molecular markers can refine that selection beyond current diagnostic labels, identifying which patients are most likely to benefit from ICI therapy and potentially flagging patients who may respond despite not meeting current MSI-H criteria. The result is a sharper basis for treatment decisions than a single biomarker can provide.
)
4. Chromosomal instability: structural genome architecture as a disease-defining feature
What we found: Much oncology research focuses on driver mutations, but large-scale structural rearrangements have also been shown to be biologically relevant. Our analysis of chromosomal instability (CIN) patterns across CMS subtypes showed that CMS2 and CMS4, both classified as CIN-high, share broad global similarities but differ substantially in specific large genomic events. These differences in structural organization are likely to underlie the distinct clinical behavior of each subtype and help explain why patients within these groups do not respond uniformly to the same therapies.
Why it matters: CIN patterns may reveal subtype-specific vulnerabilities that mutation-focused approaches would not identify. Structural genome architecture distinguishes clinically relevant CRC subtypes in ways that go beyond driver mutations, opening new avenues for target identification and more biologically precise trial stratification.
)
What this means for the field – and for patients
CRC is entering a new era. Progress increasingly depends on pushing molecular resolution to a level where the underlying biology can be understood at an individual level.
Through deep molecular profiling, we can uncover shared tumor behaviors, resistance mechanisms and molecular signatures that make the disease’s complexity more interpretable. This depth of stratification allows therapies to be developed and applied in smaller, more precisely defined patient populations - resulting in greater accuracy and a higher likelihood of clinical success.
Indivumed connects high‑quality data and precise biological understanding with deep molecular insight to reach a milestone where the right patient, the right target, and the right therapeutic modality align.