Abstract
We read with great interest the recent article by Geng et al., 1 which investigates the role of adjuvant chemotherapy in patients with esophageal squamous cell carcinoma (ESCC) following neoadjuvant chemoradiotherapy (nCRT) and surgery. This real-world study addresses an important and timely clinical question; however, several methodological and analytical aspects warrant further clarification to strengthen the robustness and interpretability of the findings.
The observed survival difference in the overall population narrowly missed statistical significance (p = 0.058), and a similarly nonsignificant result was reported in the pathological complete response (pCR) subgroup (p = 0.110). Considering the limited sample size (n = 218, including only 95 patients with pCR), these findings may be attributed to insufficient statistical power rather than true therapeutic equivalence. A post hoc power analysis would be helpful to assess the probability of detecting clinically meaningful differences—such as a hazard ratio (HR) of 0.7. Furthermore, given the wide confidence intervals associated with small sample sizes, caution should be exercised in interpreting the absence of statistical significance as evidence of no benefit.
In addition, the subgroup analyses of treatment effect by pCR status appear to rely solely on univariate log-rank tests. The absence of multivariable modeling limits the capacity to account for key prognostic factors. We encourage the authors to consider separate multivariable survival analyses for the pCR and non-pCR groups, along with interaction testing between pCR status and receipt of adjuvant therapy. Incorporating established pathological covariates—such as post-treatment pathological TNM (ypTNM) stage, lymphovascular invasion, perineural invasion, and resection margin status—into Cox proportional hazards models would substantially strengthen the validity and clinical interpretability of the conclusions.2,3
Another methodological concern lies in the use of clinical TNM (cTNM) rather than ypTNM staging as the primary stratification factor for survival analysis. This approach diverges from current standard practices in post-nCRT evaluation and may explain several unexpected findings in the study, including the lack of a survival association with tumor stage and the counterintuitive result that a lower clinical T-stage was associated with worse prognosis (HR = 1.19). Clarification regarding the rationale for this choice, along with sensitivity analyses using ypTNM staging, would enhance the transparency and interpretability of the results.
Finally, the report that 16.2% of patients in the adjuvant chemotherapy group experienced recurrence within 6 months of surgery—some within as early as 3 months—raises concerns about potential confounding by early relapse. It remains unclear whether these cases were excluded from eligibility for adjuvant therapy per study protocol, or whether recurrence occurred despite the planned administration of chemotherapy. A detailed comparison of recurrence timing relative to the actual initiation of adjuvant chemotherapy would be helpful in clarifying this point.
In summary, while this study offers important insights into the potential role of adjuvant therapy following nCRT in ESCC, its clinical relevance would be considerably enhanced by addressing the above concerns. Rigorous multivariate analyses and transparent methodological justification are essential to accurately define the value of adjuvant treatment in this setting and to inform future prospective trials.
