Protein blood tests may be better predictors of bowel cancer relapse and survival rates: study

Analysis from Moffitt Cancer Center has shown that DNA-derived protein epiScores measured in blood prior to treatment can enhance risk stratification in colorectal cancer, with certain markers linked to a far higher likelihood of recurrence and death

Researchers at the Moffitt Cancer Center, Tampa, Florida, USA, have reported that the blood-based DNA markers, protein epiScores, can help to identify patients with colorectal cancer who face a significantly higher risk of recurrence or death. The findings add to a growing body of evidence that epigenetic profiling may refine prognostic assessment beyond conventional clinicopathological criteria.

Protein epiScores are derived from patterns of DNA methylation, a chemical modification to DNA that influences gene activity without altering the genetic code itself. These methylation signatures can act as molecular surrogates for circulating protein levels. By analysing blood samples collected before treatment, the investigators sought to determine whether such signatures might predict outcomes in patients with colorectal cancer more accurately than current standard of care.

The research team identified four protein epiScores that showed strong associations with adverse clinical outcomes. Patients with higher levels of these markers experienced a greater risk of between 60 and 70 per cent of chance of cancer recurrence during follow-up compared with those with lower levels. One marker in particular – LGALS3BP – also demonstrated a clear association with overall survival. Patients with elevated LGALS3BP epiScores faced an 80 per cent increased risk of death during the study period.

When the researchers incorporated the protein epiScores into existing prognostic models – such as tumour stage and patient age – predictive performance was seen to improve. The accuracy of recurrence prediction rose from 64 to 70 per cent. For overall survival, predictive accuracy increased from 70 to 75 per cent. Although these gains may appear modest, even incremental improvements in risk discrimination can influence clinical decision-making, particularly in a disease where adjuvant therapy carries substantial toxicity and cost.

Colorectal cancer remains one of the most common malignancies worldwide and a leading cause of cancer-related mortality. Current risk assessment relies heavily on histopathological staging and selected molecular features identified within tumour tissue. However, such approaches do not always capture the systemic biological processes that influence disease progression, including host immune response, angiogenesis and coagulation pathways.

The investigators reported that the protein epiScores appeared to reflect biological pathways that conventional tools do not measure directly. These included aspects of immune function, blood vessel growth and blood clotting, all of which can contribute to tumour spread and metastatic potential. A blood-based assay therefore offers a window into tumour–host interactions that tissue-based staging systems may miss.

The study has indicated that a single blood draw taken before treatment may provide clinically relevant information to refine risk stratification. Such an approach would offer practical advantages. Blood sampling is minimally invasive, inexpensive and already integrated into routine oncology workflows. If validated in broader and more diverse patient cohorts, protein epiScores could support personalised treatment planning by identifying patients who may benefit from intensified surveillance or adjuvant therapy.

The authors acknowledged that further validation in independent populations is essential before clinical implementation can proceed. Biomarker development requires careful replication to confirm robustness across ethnic groups, healthcare systems and treatment regimens. Nonetheless, the present findings have demonstrated the potential for epigenetic signatures in blood to complement established protocols.

The study contributes to an expanding field that seeks to harness epigenetic information for management of cancers. DNA methylation patterns provide a stable and quantifiable molecular record of gene regulation. By translating these patterns into protein-level proxies, researchers have begun to bridge genomic and proteomic information without the need for complex protein assays.

If subsequent research confirms these associations, protein epiScores could form part of a more nuanced framework for colorectal cancer prognosis. Such tools would not replace tumour staging but rather augment it, in order to deliver a more precise estimate of individual risk.

Q&A with Dr. Jacob Kresovich, lead author and assistant member in the Cancer Epidemiology Programme at Moffitt. 

What is the biggest takeaway from this research? 

“We wanted to know whether a blood-based test could better identify colorectal cancer patients at higher risk for recurrence or death. Doctors typically rely on tumour and patient characteristics but those factors do not fully explain why patients with similar profiles can have very different outcomes. We tested whether protein epiScores could add useful information beyond standard clinical factors. Better prediction could help tailor follow-up care and treatment intensity while avoiding unnecessary interventions for lower-risk patients.”

What are protein epiScores and why use them instead of traditional blood protein tests? 

“Protein epiScores are based on DNA methylation patterns in blood cells. They act like a fingerprint of the immune cell’s state related to certain protein signals, rather than measuring protein levels directly. We used them because protein levels can fluctuate from day to day, while DNA methylation patterns tend to be more stable over time. These scores may also capture broader immune-related processes that could be more relevant to cancer progression.”

Which findings stood out most for predicting patient outcomes? 

“Four protein epiScores were strongly linked to cancer recurrence: HCII, VEGFA, CCL17 and LGALS3BP. Patients with higher scores were about 60 per cent to 70 per cent more likely to experience recurrence. The LGALS3BP score was also strongly associated with overall survival, with higher levels linked to an 80 per cent greater risk of death. These markers reflect different biological systems, including blood clotting, tumour blood vessel growth and immune regulation, suggesting that multiple pathways influence cancer outcomes.”

How did protein epiScores improve prediction beyond standard clinical factors like stage and age? 

“Using only traditional [standard of care] clinical factors, our model predicted recurrence with about 64 per cent accuracy. Adding the four key protein epiScores increased accuracy to 70 per cent. For overall survival, accuracy improved from 70 to 75 per cent when the LGALS3BP score was added. We also saw improved risk classification, with more accurate prediction for 34 per cent of patients for recurrence and 16 per cent for mortality. Unfortunately, we were lacking information from circulating tumour DNA, so we will need to account for that in future studies, but even the small improvements shown here can have meaningful clinical impact.”

What do these results suggest about future colorectal cancer care or research? 

“These findings suggest protein epiScores could eventually be incorporated into tools that help guide treatment and follow-up decisions. More broadly, the study shows that useful prognostic information can be obtained from a simple blood [test] taken before treatment begins. Because these scores reflect relatively stable immune characteristics, they may help explain why some patients have poorer outcomes. However, the findings must be validated in other populations before clinical use.”

For further reading please visit: 10.1186/s13148-026-02059-3