Letter to the Editor: Unmeasured Confounding and Clinical Decision Making in T1N0M0 Esophageal Cancer

We read with great interest the article by Wang et al. [1], titled “Chemoradiotherapy Versus Radiotherapy in Patients With Stage T1N0M0 Esophageal Cancer: A Retrospective Cohort Study.” The authors addressed a clinically significant question regarding the optimal nonsurgical management for this early-stage disease, a population for which high-level evidence is scarce. We commend the authors for their valuable contribution. However, we have concerns regarding the interpretation of the overall survival (OS) benefit in the T1b subgroup and believe a more nuanced discussion of patient selection is warranted. One of the most thought-provoking findings is that for the T1b subgroup, chemoradiotherapy was an independent prognostic factor for OS in multivariable analysis, yet it had no significant effect on cancer-specific survival (CSS). An OS benefit that cannot be explained by an improvement in CSS strongly suggests that the survival advantage may not be driven by the antitumor effect of chemotherapy itself, but rather by unmeasured confounders, particularly significant selection bias. Our inference is based on two key observations. First, as shown in table 1 of the original article by Wang et al., the chemoradiotherapy group, which received more intensive treatment, paradoxically had a higher proportion of Grade III/IV tumors (34.1% vs. 12.2%). This counterintuitive finding suggests that the treatment decision was likely influenced not just by tumor characteristics but by the patient's ability to “tolerate” chemotherapy. Second, the SEER database lacks data on performance status (PS) and comorbidities, both of which are powerful predictors of OS. A more plausible explanation, therefore, is that patients triaged to radiotherapy alone were in poorer health for unrecorded reasons, leading to higher non-cancer-related mortality and, consequently, worse OS. This methodological concern has profound clinical implications. If the observed OS benefit is indeed an artifact of selection bias, it would be hazardous to adopt “chemoradiotherapy is superior for T1b patients” as a generalizable conclusion. Rather than simplifying the clinical decision, the study's results highlight the paramount importance of individualized patient stratification within this heterogeneous population. For a fit T1b patient, adding chemotherapy to address the higher risk of micrometastases may be justified; for a frail or elderly patient, the toxicity of chemotherapy could negate any potential benefit. In conclusion, we believe the observed association with OS should be interpreted with great caution as hypothesis-generating, not as a definitive conclusion. This work impels us to recognize that future research should move beyond simply comparing two treatment modalities. Instead, it should focus on identifying which biological or clinical features (such as molecular markers and lymphovascular invasion) can distinguish those patients who truly need and will benefit from intensified therapy. Only then can we make the most appropriate treatment decisions for each nonsurgical patient with early-stage disease. The authors have nothing to report. The authors have nothing to report. The authors declare no conflicts of interest. No new data were generated in support of this letter. All data and figures cited herein were taken directly from Wang et al. [1].