Comment on "Chemoradiotherapy Followed by Active Surveillance Versus Standard Esophagectomy for Esophageal Cancer: A Systemic Review and Individual Patient Data Meta-Analysis".

In the March issue of the Annals of Surgery, van der Wilk et al1 published a meta-analysis examining patients with esophageal cancer who had a complete clinical response following chemoradiotherapy. The authors concluded that there was no difference in overall or progression-free survival when comparing active surveillance and esophagectomy for patients who were deemed to have a complete clinical response. Surveillance following complete clinical response is undoubtedly a hot-button topic among clinicians who treat esophageal malignancies. Although we applaud the authors for going the extra mile to obtain individual study data to compile for a larger analysis, we had several important questions regarding the results and conclusions of this study. The main question we should ask ourselves when interpreting these results is whether they are generalizable to the general population of patients with esophageal carcinoma receiving neoadjuvant chemoradiation? Notably, the 7 studies within this analysis included patients treated over a 26-year period and, obviously, this represents a very heterogenous group. There was very little consistency in the chemoradiotherapy treatments utilized among the different studies, and only 1 study utilized a more modern CROSS-trial regimen. Baseline patient and tumor characteristics were not provided for either cohort, making it impossible for us to determine similarities with our own patient population. In addition, the authors did not present any subgroup analysis for overall or progression-free survival regarding tumor histology. We know from the CROSS trial that patients with squamous cell carcinoma typically have greater rates of pathologic complete response compared to those with adenocarcinoma.2 Without stratification, we are unable to tell if the results are similar for both histologic subtypes. The authors do take time within the discussion to point out the bias inherent within their study. However, we feel there should be a greater emphasis regarding this bias before “discuss[ing] an active surveillance strategy with patients” as the authors have suggested. Selection bias cannot be ignored in the current meta-analysis as only one of the 7 included studies involved randomized, nonretrospective data. Clearly there are reasons some patients were deemed unfit or refused surgical resection. Given that the I2 of the meta-analysis for overall survival was 55% (P = 0.037), we can conclude that there was substantial heterogeneity between the included studies. Further analysis of the individual studies included in the meta-analysis does reveal some conflicting results. Jeong et al retrospectively analyzed 154 patients with locally advanced esophageal squamous cell carcinoma.3 They found that trimodal therapy, including chemoradiotherapy and esophagectomy, was associated with improved local recurrence-free survival and disease-free survival when compared to definitive chemoradiation, even in patients with a complete metabolic response on FDG-PET. The lone source of nonretrospective data, a randomized trial published by Park et al, also showed a trend toward improved disease-free survival for patients undergoing esophagectomy opposed to those randomized to observation alone.4 Ultimately, at the crux of any discussion of active surveillance, is our ability to define and detect a complete clinical response. Within the current meta-analysis, the definition of clinical complete response varied between studies as different combinations of radiologic, endoscopic, and pathologic methods were employed. Are we to the point where we can reliably and accurately detect clinical complete response? At best, current diagnostic approaches still miss at least 10% of cases of residual carcinoma following neoadjuvant chemoradiation.5 In the current study, 40% of patients in the active surveillance group developed a locoregional occurrence compared to only 7% in the standard esophagectomy arm. Of the 40% that developed local recurrence, only half underwent salvage esophagectomy. Survival data were not presented for these patients. This is not an inconsequential group when you consider that a majority may have been able to undergo a curative intent operation at the outset. In addition, no data were provided concerning the additional interventions required in patients with local recurrences which we know from existing literature can be substantial (ie, EGDs, dilations, stents, etc). We too are eagerly awaiting the results of the ongoing clinical trials comparing standard surgical resection to active surveillance.6,7 Nevertheless, we are wary of suggesting that the results of retrospective, heavily biased studies should guide clinical practice. There ultimately may be some role for active surveillance among patients with an excellent response to neoadjuvant chemoradiation and who are unfit for an operation. However, it should be stressed that at the present time, “watch and wait” for esophageal cancer is not ready day to day clinical practice outside of the context of a clinical trial. ACKNOWLEDGMENT Conceptualization and planning: Z.E.S., K.M.A., M.K.; Analysis and literature search: Z.E.S.; Drafting of manuscript: Z.E.S., K.M.A., M.K.; Critical revision: Z.E.S., K.M.A., M.K.