Letter to the Editor: Role of perioperative carcinoembryonic Antigen in guiding adjuvant chemotherapy for colorectal cancer

Abstract

Du et al reported that postoperative carcinoembryonic antigen (CEA) may serve as a cost-effective marker to guide adjuvant chemotherapy in colorectal cancer. Although this study provides clinically relevant findings, several issues need further discussion. First, the retrospective design limits causal interpretation and remains susceptible to selection bias and unmeasured confounding, particularly in non-randomized treatment allocation. Second, the definition and timing of postoperative CEA measurement may influence its predictive value, given its short half-life and potential interference from perioperative factors. Third, the clinical application of CEA-guided treatment requires more rigorous evaluation, including formal interaction testing and consideration of treatment toxicity. Integrating postoperative CEA with emerging biomarkers such as circulating tumor DNA could improve risk stratification. Further prospective studies are needed to validate these findings and clarify their clinical utility.

Key Words: Postoperative carcinoembryonic antigen; Colorectal cancer; Adjuvant chemotherapy; Circulating tumor DNA; Risk stratification

Core Tip: Postoperative carcinoembryonic antigen (CEA) is a simple and inexpensive marker, but its use for guiding adjuvant chemotherapy in colorectal cancer still requires caution. The retrospective design, variable timing of postoperative CEA testing, and absence of treatment–biomarker interaction analysis may limit the strength of the conclusion. Future studies should standardize CEA measurement and combine it circulating tumor DNA to improve postoperative risk stratification.

TO THE EDITOR

We read with great interest the article by Du et al[1] published in the World Journal of Gastroenterology, which reported that perioperative serum carcinoembryonic antigen (CEA) is a powerful marker for prognostic prediction and adjuvant chemotherapy decision-making in stage II and III colorectal cancer. The authors demonstrated that postoperative CEA levels decrease significantly compared to preoperative levels, and that elevated postoperative CEA may identify patients who benefit from adjuvant chemotherapy, potentially offering a more cost-effective alternative to expensive molecular biomarkers such as circulating tumor DNA (ctDNA)[2]. We appreciate the authors' significant contribution to this field and would like to offer some methodological and clinical perspectives for further consideration.

Firstly, due to its retrospective characteristic that cannot establish a cause-and-effect relationship with the effects of postoperative CEA values and adjuvant chemotherapy results. Although there has a large sample size and includes a validation cohort in this study, selection bias from a retrospective design cannot be completely eliminated. Although the author included a validation cohort and performed Cox regression analyses, patients were not randomly divided into treatment groups for adjuvant chemotherapy. Unobservable confounders such as patients' perceptions of their own diseases, treatment choices, subtle differences in surgical quality not captured by American Joint Committee on Cancer staging can affect whether they receive chemotherapy and clinical results differently when observed separately. In addition, the comparison of XELOX vs FOLFOX among patients with elevated postoperative CEA was not randomized. Therefore, it might be affected by institutional preference or toxicity-induced selection, leading to bias in survival benefits.

Secondly, the definition and application scope of CEA differences (CEA-diff) in this paper also require more verification. The authors pointed out that there were some shortcomings in the predictive ability of dynamic changes in CEA. It is possible that this has been affected by a limited definition period for measuring postoperative CEA. Given the 5-day half-life of CEA, measurements taken too early, such as within the first week after surgery, may be influenced by residual surgical inflammation, thereby obscuring the elevation caused by minimal residual disease, while delaying measurements too long may compromise timely treatment. We suggest that postoperative CEA measurement be standardized at approximately 2-4 weeks after surgery to balance the resolution of perioperative inflammation and the timely initiation of adjuvant therapy[3]. Moreover, a single postoperative measurement without a standardized, protocolbased time point introduces substantial nondifferential misclassification bias, likely weakening the predictive value of CEA-diff[4]. Therefore, we suggest prospective studies that standardize postoperative CEA timing based on its established 5day halflife kinetics to reduce these biases.

Thirdly, we believe the study should more explicitly address clinical decision-making, concluding that perioperative CEA can guide adjuvant chemotherapy without balancing treatment toxicity and healtheconomic considerations may be unconvincing. The authors propose that patients with elevated postoperative CEA benefit significantly from adjuvant chemotherapy, whereas those with normal CEA do not. For better clinical translation, a formal interaction test between CEA status and treatment effect is needed, because subgroup Kaplan-Meier curves may overestimate treatment benefit[5]. In addition, while XELOX showed a longer overall survival compared to FOLFOX among those with elevated postoperative CEA levels, the high-risk incidence of oxaliplatin-related neurotoxicity needs to be considered alongside potential gains when selecting regimens in real-world practice[6].

We suggest that we should incorporate the postoperative CEA level and other technologies, such as ctDNA, to link back to existing practices based on retrospective data in a dynamic riskstratification model. Given that ctDNA provides different biological information regarding minimal residual disease, postoperative CEA may serve as a complementary and initial screening tool to select patients for more expensive ctDNA testing[2]. Additionally, a randomized controlled trial stratified by postoperative CEA levels could confirm the causality of the association demonstrated in this study. Compared to universal chemotherapy or ctDNAguided therapy, a costeffectiveness analysis of CEAguided therapy would also be valuable to guide that this approach not only improves survival but also reduces the economic burden on patients.

CONCLUSION

In summary, the study by Du et al[1] has provided valuable insights and confirmed the potential clinical applicability of postoperative CEA in this area. However, we recommend that postoperative CEA be used cautiously as a low-cost complement to, rather than a replacement for, emerging biomarkers like ctDNA, and that its clinical implementation be supported by prospective validation, standardized measurement protocols, and integrated risk-stratification models to ultimately ease the burden on patients with colorectal cancer.