Advertisement

Disease-free survival as a surrogate endpoint for overall survival in adults with resectable esophageal or gastroesophageal junction cancer: A correlation meta-analysis

Open AccessPublished:May 20, 2022DOI:https://doi.org/10.1016/j.ejca.2022.04.027

      Highlights

      • First published DFS/OS surrogacy analysis spanning all treatment settings in EC/GEJ.
      • Meta-analysis of 26 trials suggests correlated hazard ratios for DFS and OS.
      • DFS can be a valid and useful surrogate for OS in resectable EC/GEJ.

      Abstract

      Objectives

      To evaluate disease-free survival (DFS) as a surrogate endpoint for overall survival (OS) using aggregate-level data from resectable esophageal or gastroesophageal junction cancer (EC/GEJC) trials assessing therapies in (neo)adjuvant and perioperative settings.

      Methods

      A systematic literature review was conducted to identify trials reporting OS and DFS, or compatible progression-free survival (PFS). Bivariate random-effects meta-analysis was used to estimate correlation between the treatment effects on DFS/PFS and OS, and weighted linear regression models assuming trial sample sizes as weights were used to estimate surrogacy equations. The primary analysis consisted of trials across all treatment settings, and secondary analysis consisted of trials only in the adjuvant setting. Leave-one-out cross-validation (LOOCV) was performed to measure the stability and predictive accuracy of the surrogacy equations while surrogate threshold effects (STE)—the minimum treatment effect on DFS/PFS that would translate into a positive OS benefit—were derived to measure their usefulness.

      Results

      The primary analysis included 26 trials. The estimated correlation coefficient between the hazard ratio (HR) of DFS/PFS (HRDFS/PFS) and HR of OS (HROS) was 0.83 (95% confidence interval [CI]: 0.70–0.90). The estimated surrogacy equation was log(HROS) = 0.80 × log(HRDFS/PFS) with a corresponding STE of 0.82. Reported HROS was within the 95% prediction interval of the predicted HROS from the model for more than 95% of the trials in the LOOCV, indicating a valid model. Secondary analysis included 7 trials with an estimated correlation coefficient of 0.76 (95% CI: 0.18–0.95). Through LOOCV, the surrogacy equation in the adjuvant setting was deemed valid.

      Conclusions

      Our meta-analysis suggests that HRDFS/PFS —where DFS/PFS is defined as time from resection to disease recurrence (local, locoregional, or distant) or death—is correlated to HROS, and a valid and useful surrogate predictor for HROS in the neoadjuvant, perioperative, or adjuvant settings.

      Keywords

      1. Introduction

      Esophageal cancer (EC) ranks seventh in terms of incidence, and sixth in mortality, across all cancers globally [
      • Bray F.
      • Ferlay J.
      • Soerjomataram I.
      • Siegel R.L.
      • Torre L.A.
      • Jemal A.
      Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries.
      ]. The mainstay of treatment for locoregional EC and gastroesophageal junction cancer (GEJC) across histologies is preoperative chemoradiotherapy, although pre- or perioperative chemotherapy is also used [
      • Lordick F.
      • Mariette C.
      • Haustermans K.
      • Obermannová R.
      • Arnold D.
      Oesophageal cancer: ESMO Clinical Practice Guidelines for diagnosis, treatment and follow-up.
      ,
      • Ajani J.A.
      • D'Amico T.A.
      • Bentrem D.J.
      • Chao J.
      • Corvera C.
      • Das P.
      • et al.
      Esophageal and esophagogastric junction cancers, version 2.2019, NCCN clinical practice guidelines in oncology.
      ]. Despite these multimodal strategies, the risk of recurrence remains high as ECs are inherently aggressive tumors. To improve efficacy and safety outcomes, systemic adjuvant therapies have been explored [
      • Cartwright E.
      • Keane F.K.
      • Enzinger P.C.
      • Hong T.
      • Chau I.
      Is there a precise adjuvant therapy for esophagogastric carcinoma?.
      ,
      • Kelly R.J.
      • Ajani J.A.
      • Kuzdzal J.
      • Zander T.
      • Van Cutsem E.
      • Piessen G.
      • et al.
      Adjuvant nivolumab in resected esophageal or gastroesophageal junction cancer.
      ].
      Overall survival (OS) remains as the gold-standard endpoint in most oncology trials due to its objectivity (i.e., death can be measured without prejudice or favor [
      • Evans S.R.
      Fundamentals of clinical trial design.
      ]) and its clinical meaningfulness. Despite these advantages, collection of OS data poses challenges such as requiring longer times and larger trial populations to detect statistically significant and clinically meaningful benefits. Moreover, for earlier stage cancers, rapidly changing treatment landscape in later stages can introduce confounding effects on OS. A common way to address these challenges is to use statistically appropriate and clinically relevant surrogate endpoints (SEs) that can alleviate the uncertainty in long-term survival. Establishing surrogacy relationships may help expedite acceptance of treatments that are under investigation by regulatory authorities and payers [
      • Ronellenfitsch U.
      • Jensen K.
      • Seide S.
      • Kieser M.
      • Schwarzbach M.
      • Slanger T.E.
      • et al.
      Disease-free survival as a surrogate for overall survival in neoadjuvant trials of gastroesophageal adenocarcinoma: pooled analysis of individual patient data from randomised controlled trials.
      ,
      • Kataoka K.
      • Nakamura K.
      • Mizusawa J.
      • Kato K.
      • Eba J.
      • Katayama H.
      • et al.
      Surrogacy of progression-free survival (PFS) for overall survival (OS) in esophageal cancer trials with preoperative therapy: literature-based meta-analysis.
      ]. Prentice et al., 1989 defined SEs as well as a set of criteria to validate an SE [
      • Prentice R.L.
      Surrogate endpoints in clinical trials: definition and operational criteria.
      ]. Extensions of methodologies to validate surrogacy relationships have since been proposed, such as Daniels and Hughes 1997, which uses results from clinical trials as the evidence base [
      • Daniels M.J.
      • Hughes M.D.
      Meta-analysis for the evaluation of potential surrogate markers.
      ].
      Surrogacy relationships between intermediate efficacy outcomes and OS are dependent on the oncological settings (e.g., treatment settings and line of therapy) [
      • Xie W.
      • Halabi S.
      • Tierney J.F.
      • Sydes M.R.
      • Collette L.
      • Dignam J.J.
      • et al.
      A systematic review and recommendation for reporting of surrogate endpoint evaluation using meta-analyses.
      ]. Despite the large number of analyses exploring SEs for OS in other tumor areas (e.g., colorectal, breast, and lung cancers), a recent systematic review of surrogacy analyses by Xie et al., 2019 suggests that there is a paucity of surrogacy work in the adjuvant setting for EC and GEJC [
      • Xie W.
      • Halabi S.
      • Tierney J.F.
      • Sydes M.R.
      • Collette L.
      • Dignam J.J.
      • et al.
      A systematic review and recommendation for reporting of surrogate endpoint evaluation using meta-analyses.
      ]. A prior meta-analysis by Oba et al., 2013 using individual patient data (IPD) from 14 randomized clinical trials (RCT) of gastric cancer (GC) in the adjuvant setting concluded that disease-free survival (DFS) is an appropriate SE for OS [
      • Oba K.
      • Paoletti X.
      • Alberts S.
      • Bang Y.-J.
      • Benedetti J.
      • Bleiberg H.
      • et al.
      Disease-free survival as a surrogate for overall survival in adjuvant trials of gastric cancer: a meta-analysis.
      ]. SEs for OS, specifically for EC and GEJC, have been previously explored only in the neoadjuvant setting [
      • Ronellenfitsch U.
      • Jensen K.
      • Seide S.
      • Kieser M.
      • Schwarzbach M.
      • Slanger T.E.
      • et al.
      Disease-free survival as a surrogate for overall survival in neoadjuvant trials of gastroesophageal adenocarcinoma: pooled analysis of individual patient data from randomised controlled trials.
      ,
      • Kataoka K.
      • Nakamura K.
      • Mizusawa J.
      • Kato K.
      • Eba J.
      • Katayama H.
      • et al.
      Surrogacy of progression-free survival (PFS) for overall survival (OS) in esophageal cancer trials with preoperative therapy: literature-based meta-analysis.
      ]. In an analysis using IPD from 8 trials, Ronellenfitsch et al., 2019 detected a strong individual-level surrogacy (Spearman rank correlation coefficient of 0.89), and a strong trial–level association (coefficient of determination R2 of 0.912, with 95% confidence interval [CI] of 0.75–1.0) between DFS and OS [
      • Ronellenfitsch U.
      • Jensen K.
      • Seide S.
      • Kieser M.
      • Schwarzbach M.
      • Slanger T.E.
      • et al.
      Disease-free survival as a surrogate for overall survival in neoadjuvant trials of gastroesophageal adenocarcinoma: pooled analysis of individual patient data from randomised controlled trials.
      ]. In the other neoadjuvant study, Kataoka et al., 2017, assessed the correlation between treatment effects on progression-free survival (PFS) and OS from 10 trials in EC. PFS did not appear to be an appropriate SE for OS (R2 of 0.283 [95% CI: 0.00–0.90]) [
      • Kataoka K.
      • Nakamura K.
      • Mizusawa J.
      • Kato K.
      • Eba J.
      • Katayama H.
      • et al.
      Surrogacy of progression-free survival (PFS) for overall survival (OS) in esophageal cancer trials with preoperative therapy: literature-based meta-analysis.
      ]. However, Kataoka et al., 2017 acknowledged the inconsistent PFS definitions used in their included studies potentially affecting this result.
      The purpose of this study was to determine, through a correlation meta-analysis utilizing aggregate-level data from RCTs, whether the hazard ratio (HR) of DFS/PFS (HRDFS/PFS) can serve as a valid SE for the HR of OS (HROS) in resectable EC/GEJC. We considered trials across neoadjuvant, perioperative, or adjuvant settings to study the surrogacy relationship between DFS and OS in early-stage EC/GEJC.

      2. Materials and methods

      2.1 Systematic literature review (SLR)

      We identified RCTs reporting relative treatment effects on OS and at least one of DFS or PFS for adults with resectable EC/GEJC by searching Embase®, MEDLINE®, and CENTRAL on 10 June 2020 with no publication date restrictions using predefined search strategies (Appendix Table A1). The search was inclusive of the neoadjuvant, adjuvant, and perioperative settings.
      Study selection and data extraction were performed by two independent investigators. Our target SE was DFS whose definition is time from randomization to recurrence (local, locoregional, or distant) or death, with censoring at last date of disease assessment. Upon completion of the SLR, the definitions reported in the evidence base were reviewed (Appendix Table A2). Considering the limited availability of data from the literature with a DFS definition consistent with our initial one, we decided to lift the constraints for our correlation meta-analysis to be inclusive of more broadened definitions, including alternative DFS endpoints such as PFS. Specifically, when DFS was missing, compatible PFS was used instead as these endpoints were used interchangeably in early-stage EC/GEJC trials. The inclusion of PFS as a suitable candidate SE was encouraged by a similar surrogacy study [
      • Kataoka K.
      • Nakamura K.
      • Mizusawa J.
      • Kato K.
      • Eba J.
      • Katayama H.
      • et al.
      Surrogacy of progression-free survival (PFS) for overall survival (OS) in esophageal cancer trials with preoperative therapy: literature-based meta-analysis.
      ]. In the adjuvant setting, across our evidence base, the definitions between the PFS and DFS outcomes were consistent in terms of event descriptions.

      2.2 Data and outcomes

      Adjusted HRs and corresponding 95% CI reported from the trials were preferred over their unadjusted counterparts, although differences between them were negligible given the randomized study designs. If HR estimates were not reported, Kaplan–Meier (KM) curves were extracted via the Grab-It XP tool (v10.0, DataTrend Software, https://download.cnet.com/Grab-It-XP/3000-2053_4-41084.html) to generate digitized pseudo patient-level data via Guyot algorithm, which were later used in a Cox proportional hazards model to calculate the underlying HRs [
      • Guyot P.
      • Ades A.E.
      • Ouwens M.J.N.M.
      • Welton N.J.
      Enhanced secondary analysis of survival data: reconstructing the data from published Kaplan-Meier survival curves.
      ]. Datasets were stored in Excel worksheets.

      2.3 Statistical methods

      Two independent steps were taken to assess the overall surrogacy: 1) Assessment of the strength of association between treatment effects on DFS/PFS and OS; and 2) Estimation of the surrogacy equation to estimate treatment effect on OS as a linear function of treatment effect on DFS/PFS. HRs were used as measure of treatment effect estimates on DFS/PFS and OS only when the proportional hazards assumption was satisfied. Since our methodologies were not suitable to address non-stationary hazards over time, we tested proportionality assumption using the global Schoenfeld test using digitized KM curves if they were reported in the publications. Studies not meeting the proportionality assumption were excluded from the analyses.
      Association between DFS/PFS and OS was assessed using a bivariate random-effects meta-analyses (BRMA) [
      • Riley R.D.
      • Thompson J.R.
      • Abrams K.R.
      An alternative model for bivariate random-effects meta-analysis when the within-study correlations are unknown.
      ] to estimate correlation [
      • Daniels M.J.
      • Hughes M.D.
      Meta-analysis for the evaluation of potential surrogate markers.
      ,
      • van Houwelingen H.C.
      • Arends L.R.
      • Stijnen T.
      Advanced methods in meta-analysis: multivariate approach and meta-regression.
      ]. BRMA maintains the individual weighting of each study but removes the need to know the within-study correlations, which were not available for this evidence base due to lack individual patient level data. Weighted linear regression (WLR) models using the sample sizes of trials as their weights were then used to estimate a linear surrogacy equation between the log-transformed treatment effects (i.e., the HRs) on DFS/PFS and OS. In WLR models, consistent with the linearity assumption for the relationship between the treatment effects, logarithmic transformations on HRs were conducted to linearize the HRs in terms of the log-transformed hazard rates in treatment and control arms as well as to convert their scales and the corresponding confidence intervals to a more symmetric spectrum. The log-transformation of HRs was also easier to interpret with an intuitive appeal since they attain the value 0 when there was no treatment effect (i.e., HR = 1).
      All analyses were performed using R (version 4.0.2). For the BRMA, the “metamisc” package was used [

      Debray T. metamisc: diagnostic and prognostic meta-analysis, https://CRAN.R-project.org/package=metamisc (accessed 16 May 2022).

      ]. The R “wCor” function was used to calculate the Pearson correlation coefficients in weighted form.

      2.4 Data sets for primary, secondary, and sensitivity analyses

      In addition to the primary analysis set including all trials in the neoadjuvant, perioperative, and adjuvant settings, we also conducted multiple sensitivity and subgroup analyses. These included: 1) removal of studies with misaligned index dates on DFS and OS; 2) removal of studies containing GC patients; and 3) removal of Zhao et al., 2020 due to high risk of bias in this study due to large cross-over [
      • Zhao Y.
      • Han L.
      • Zhang W.
      • Shan L.
      • Wang Y.
      • Song P.
      • et al.
      Preoperative chemotherapy compared with postoperative adjuvant chemotherapy for squamous cell carcinoma of the thoracic oesophagus with the detection of circulating tumour cells randomized controlled trial.
      ]. Additionally, a secondary analysis was conducted in the adjuvant setting only.

      2.5 Assessment and validation of the surrogacy equation

      Leave-one-out cross-validation (LOOCV) method was used to validate the surrogacy equation obtained from the WLR [
      • Armitage P.
      • Berry G.
      • Matthews J.
      Statistical methods in medical research.
      ,
      • Buyse M.
      • Molenberghs G.
      • Burzykowski T.
      • Renard D.
      • Geys H.
      The validation of surrogate endpoints in meta-analyses of randomized experiments.
      ,
      • Efron B.
      Estimating the error rate of a prediction rule: improvement on cross-validation.
      ]. LOOCV estimated a predicted HROS for each trial in the evidence base using the same model with WLR structure, but with the remaining trials. Validity is described as the predictive value of the SE on the treatment effect on the true endpoint [
      • Mariette C.
      • Dahan L.
      • Mornex F.
      • Maillard E.
      • Thomas P.A.
      • Meunier B.
      • et al.
      Surgery alone versus chemoradiotherapy followed by surgery for stage I and II esophageal cancer: final analysis of randomized controlled phase III trial FFCD 9901.
      ]. To assess whether DFS/PFS can be used as a valid surrogate for OS, we referred to the National Institute for Health and Care Excellence (NICE) DSU TSD 20 [
      • Bujkiewicz S.
      • Achana F.
      • Papanikos T.
      • Riley R.D.
      • Abrams K.R.
      NICE DSU Technical Support Document 20: multivariate meta-analysis of summary data for combining treatment effects on correlated outcomes and evaluating surrogate endpoints.
      ], which recommends that at least 95% of the trials in the respective evidence to have their reported HROS to be within the 95% prediction interval (PI) of the predicted HROS (i.e., coverage is ≥95%) at the end of cross-validation.
      To assess the usefulness of the surrogacy equations, we calculated the surrogate threshold effect (STE). The STE is defined as the minimum treatment effect on the SE (i.e. the maximum value of HRDFS/PFS) to predict a positive treatment effect on the HROS (null hypothesis of no significant treatment difference at the 95% confidence level) [
      • Burzykowski T.
      • Buyse M.
      Surrogate threshold effect: an alternative measure for meta-analytic surrogate endpoint validation.
      ]. It is worth to note that STE calculation can also be customized for a specific trial with missing OS data using its own sample size and the 95% PI of the WLR models [
      • Baker S.G.
      Five criteria for using a surrogate endpoint to predict treatment effect based on data from multiple previous trials.
      ].

      3. Results

      A total of 49 publications corresponding to 33 unique trials [
      • Zhao Y.
      • Han L.
      • Zhang W.
      • Shan L.
      • Wang Y.
      • Song P.
      • et al.
      Preoperative chemotherapy compared with postoperative adjuvant chemotherapy for squamous cell carcinoma of the thoracic oesophagus with the detection of circulating tumour cells randomized controlled trial.
      ,
      • Deng W.
      • Yang J.
      • Ni W.
      • Li C.
      • Chang X.
      • Han W.
      • et al.
      Postoperative radiotherapy in pathological T2-3N0M0 thoracic esophageal squamous cell carcinoma: interim report of a prospective, phase III, randomized controlled study.
      ,
      • Mariette C.
      • Dahan L.
      • Mornex F.
      • Maillard E.
      • Thomas P.A.
      • Meunier B.
      • et al.
      Surgery alone versus chemoradiotherapy followed by surgery for stage I and II esophageal cancer: final analysis of randomized controlled phase III trial FFCD 9901.
      ,
      • Al-Batran S.E.
      • Homann N.
      • Pauligk C.
      • Goetze T.O.
      • Meiler J.
      • Kasper S.
      • et al.
      Perioperative chemotherapy with fluorouracil plus leucovorin, oxaliplatin, and docetaxel versus fluorouracil or capecitabine plus cisplatin and epirubicin for locally advanced, resectable gastric or gastro-oesophageal junction adenocarcinoma (FLOT4): a randomised, phase 2/3 trial.
      ,
      • Alderson D.
      • Cunningham D.
      • Nankivell M.
      • Blazeby J.M.
      • Griffin S.M.
      • Crellin A.
      • et al.
      Neoadjuvant cisplatin and fluorouracil versus epirubicin, cisplatin, and capecitabine followed by resection in patients with oesophageal adenocarcinoma (UK MRC OE05): an open-label, randomised phase 3 trial.
      ,
      • Ando N.
      • Iizuka T.
      • Ide H.
      • Ishida K.
      • Shinoda M.
      • Nishimaki T.
      • et al.
      Surgery plus chemotherapy compared with surgery alone for localized squamous cell carcinoma of the thoracic esophagus: a Japan Clinical Oncology Group Study--JCOG9204.
      ,
      • Ando N.
      • Kato H.
      • Igaki H.
      • Shinoda M.
      • Ozawa S.
      • Shimizu H.
      • et al.
      A randomized trial comparing postoperative adjuvant chemotherapy with cisplatin and 5-fluorouracil versus preoperative chemotherapy for localized advanced squamous cell carcinoma of the thoracic esophagus (JCOG9907).
      ,
      • Barbour A.P.
      • Walpole E.T.
      • Mai G.T.
      • Barnes E.H.
      • Watson D.I.
      • Ackland S.P.
      • et al.
      Preoperative cisplatin, fluorouracil, and docetaxel with or without radiotherapy after poor early response to cisplatin and fluorouracil for resectable oesophageal adenocarcinoma (AGITG DOCTOR): results from a multicentre, randomised controlled phase II trial.
      ,
      • Burmeister B.H.
      • Smithers B.M.
      • Gebski V.
      • Fitzgerald L.
      • Simes R.J.
      • Devitt P.
      • et al.
      Surgery alone versus chemoradiotherapy followed by surgery for resectable cancer of the oesophagus: a randomised controlled phase III trial.
      ,
      • Burmeister B.H.
      • Thomas J.M.
      • Burmeister E.A.
      • Walpole E.T.
      • Harvey J.A.
      • Thomson D.B.
      • et al.
      Is concurrent radiation therapy required in patients receiving preoperative chemotherapy for adenocarcinoma of the oesophagus? A randomised phase II trial.
      ,
      • Cunningham D.
      • Stenning S.P.
      • Smyth E.C.
      • Okines A.F.
      • Allum W.H.
      • Rowley S.
      • et al.
      Peri-operative chemotherapy with or without bevacizumab in operable oesophagogastric adenocarcinoma (UK Medical Research Council ST03): primary analysis results of a multicentre, open-label, randomised phase 2-3 trial.
      ,
      • Kleinberg L.R.
      • Catalano P.J.
      • Forastiere A.A.
      • Keller S.M.
      • Mitchel E.P.
      • Anne P.R.
      • et al.
      Eastern cooperative oncology group and American college of radiology imaging network randomized phase 2 trial of neoadjuvant preoperative paclitaxel/cisplatin/radiation therapy (RT) or irinotecan/cisplatin/RT in esophageal adenocarcinoma: long-term outcome and implications for trial design.
      ,
      • Le Prise E.
      • Etienne P.L.
      • Meunier B.
      • Maddern G.
      • Ben Hassel M.
      • Gedouin D.
      • et al.
      A randomized study of chemotherapy, radiation therapy, and surgery versus surgery for localized squamous cell carcinoma of the esophagus.
      ,
      • Lee J.L.
      • Park S.I.
      • Kim S.B.
      • Jung H.Y.
      • Lee G.H.
      • Kim J.H.
      • et al.
      A single institutional phase III trial of preoperative chemotherapy with hyperfractionation radiotherapy plus surgery versus surgery alone for resectable esophageal squamous cell carcinoma.
      ,
      • Lim S.H.
      • Shim Y.M.
      • Park S.H.
      • Kim H.K.
      • Choi Y.S.
      • Ahn M.J.
      • et al.
      A randomized phase II study of leucovorin/5-fluorouracil with or without oxaliplatin (LV5FU2 vs. FOLFOX) for curatively-resected, node-positive esophageal squamous cell carcinoma.
      ,
      • Lv J.
      • Cao X.F.
      • Zhu B.
      • Ji L.
      • Tao L.
      • Wang D.D.
      Long-term efficacy of perioperative chemoradiotherapy on esophageal squamous cell carcinoma.
      ,
      • Mukherjee S.
      • Hurt C.
      • Cox C.
      • Radhakrishna G.
      • Gwynne S.
      • Bateman A.R.
      • et al.
      Induction oxaliplatin capecitabine followed by switch to carboplatin-paclitaxel based RT versus continuing oxaliplatin capecitabine RT in operable esophageal adenocarcinoma: survival analysis of the randomized phase II neoscope trial.
      ,
      • Ruhstaller T.
      • Thuss-Patience P.
      • Hayoz S.
      • Schacher S.
      • Knorrenschild J.R.
      • Schnider A.
      • et al.
      Neoadjuvant chemotherapy followed by chemoradiation and surgery with and without cetuximab in patients with resectable esophageal cancer: a randomized, open-label, phase III trial (SAKK 75/08).
      ,
      • Shapiro J.
      • van Lanschot J.J.B.
      • Hulshof M.
      • van Hagen P.
      • van Berge Henegouwen M.I.
      • Wijnhoven B.P.L.
      • et al.
      Neoadjuvant chemoradiotherapy plus surgery versus surgery alone for oesophageal or junctional cancer (CROSS): long-term results of a randomised controlled trial.
      ,
      • Stahl M.
      • Walz M.K.
      • Riera-Knorrenschild J.
      • Stuschke M.
      • Sandermann A.
      • Bitzer M.
      • et al.
      Preoperative chemotherapy versus chemoradiotherapy in locally advanced adenocarcinomas of the oesophagogastric junction (POET): long-term results of a controlled randomised trial.
      ,
      • Tepper J.
      • Krasna M.J.
      • Niedzwiecki D.
      • Hollis D.
      • Reed C.E.
      • Goldberg R.
      • et al.
      Phase III trial of trimodality therapy with cisplatin, fluorouracil, radiotherapy, and surgery compared with surgery alone for esophageal cancer: CALGB 9781.
      ,
      • Thomas A.
      • Virdee P.S.
      • Eatock M.
      • Lord S.R.
      • Falk S.
      • Anthoney D.A.
      • et al.
      Dual Erb B Inhibition in Oesophago-gastric Cancer (DEBIOC): A phase I dose escalating safety study and randomised dose expansion of AZD8931 in combination with oxaliplatin and capecitabine chemotherapy in patients with oesophagogastric adenocarcinoma.
      ,
      • Urba S.G.
      • Orringer M.B.
      • Turrisi A.
      • Iannettoni M.
      • Forastiere A.
      • Strawderman M.
      Randomized trial of preoperative chemoradiation versus surgery alone in patients with locoregional esophageal carcinoma.
      ,
      • von Döbeln G.A.
      • Klevebro F.
      • Jacobsen A.B.
      • Johannessen H.O.
      • Nielsen N.H.
      • Johnsen G.
      • et al.
      Neoadjuvant chemotherapy versus neoadjuvant chemoradiotherapy for cancer of the esophagus or gastroesophageal junction: long-term results of a randomized clinical trial.
      ,
      • Yamasaki M.
      • Yasuda T.
      • Yano M.
      • Hirao M.
      • Kobayashi K.
      • Fujitani K.
      • et al.
      Multicenter randomized phase II study of cisplatin and fluorouracil plus docetaxel (DCF) compared with cisplatin and fluorouracil plus Adriamycin (ACF) as preoperative chemotherapy for resectable esophageal squamous cell carcinoma (OGSG1003).
      ,
      • Zieren H.U.
      • Müller J.M.
      • Jacobi C.A.
      • Pichlmaier H.
      • Müller R.P.
      • Staar S.
      Adjuvant postoperative radiation therapy after curative resection of squamous cell carcinoma of the thoracic esophagus: a prospective randomized study.
      ,
      • Baba M.
      • Natsugoe S.
      • Shimada M.
      • Nakano S.
      • Kusano C.
      • Fukumoto T.
      • et al.
      Prospective evaluation of preoperative chemotherapy in resectable squamous cell carcinoma of the thoracic esophagus.
      ,
      • Boonstra J.J.
      • Kok T.C.
      • Wijnhoven B.P.
      • van Heijl M.
      • van Berge Henegouwen M.I.
      • Ten Kate F.J.
      • et al.
      Chemotherapy followed by surgery versus surgery alone in patients with resectable oesophageal squamous cell carcinoma: long-term results of a randomized controlled trial.
      ,
      • Bosset J.F.
      • Gignoux M.
      • Triboulet J.P.
      • Tiret E.
      • Mantion G.
      • Elias D.
      • et al.
      Chemoradiotherapy followed by surgery compared with surgery alone in squamous-cell cancer of the esophagus.
      ,
      • Cai X.W.
      • Zeng Y.
      • Feng W.
      • Liu M.N.
      • Yu W.
      • Zhang Q.
      • et al.
      Randomized phase II trial comparing tumor bed alone with tumor bed and elective nodal postoperative radiotherapy in patients with locoregionally advanced thoracic esophageal squamous cell carcinoma.
      ,
      • Kelsen D.P.
      • Winter K.A.
      • Gunderson L.L.
      • Mortimer J.
      • Estes N.C.
      • Haller D.G.
      • et al.
      Long-term results of RTOG trial 8911 (USA Intergroup 113): a random assignment trial comparison of chemotherapy followed by surgery compared with surgery alone for esophageal cancer.
      ,
      • Ténière P.
      • Hay J.M.
      • Fingerhut A.
      • Fagniez P.L.
      Postoperative radiation therapy does not increase survival after curative resection for squamous cell carcinoma of the middle and lower esophagus as shown by a multicenter controlled trial. French University Association for Surgical Research.
      ,
      • Yang H.
      • Liu H.
      • Chen Y.
      • Zhu C.
      • Fang W.
      • Yu Z.
      • et al.
      Neoadjuvant chemoradiotherapy followed by surgery versus surgery alone for locally advanced squamous cell carcinoma of the esophagus (NEOCRTEC5010): a phase III multicenter, randomized, open-label clinical trial.
      ] were identified through the SLR (Fig. 1). Seven trials were excluded from the analyses due to evidence of non-proportional hazards [
      • Baba M.
      • Natsugoe S.
      • Shimada M.
      • Nakano S.
      • Kusano C.
      • Fukumoto T.
      • et al.
      Prospective evaluation of preoperative chemotherapy in resectable squamous cell carcinoma of the thoracic esophagus.
      ,
      • Boonstra J.J.
      • Kok T.C.
      • Wijnhoven B.P.
      • van Heijl M.
      • van Berge Henegouwen M.I.
      • Ten Kate F.J.
      • et al.
      Chemotherapy followed by surgery versus surgery alone in patients with resectable oesophageal squamous cell carcinoma: long-term results of a randomized controlled trial.
      ,
      • Bosset J.F.
      • Gignoux M.
      • Triboulet J.P.
      • Tiret E.
      • Mantion G.
      • Elias D.
      • et al.
      Chemoradiotherapy followed by surgery compared with surgery alone in squamous-cell cancer of the esophagus.
      ,
      • Cai X.W.
      • Zeng Y.
      • Feng W.
      • Liu M.N.
      • Yu W.
      • Zhang Q.
      • et al.
      Randomized phase II trial comparing tumor bed alone with tumor bed and elective nodal postoperative radiotherapy in patients with locoregionally advanced thoracic esophageal squamous cell carcinoma.
      ,
      • Kelsen D.P.
      • Winter K.A.
      • Gunderson L.L.
      • Mortimer J.
      • Estes N.C.
      • Haller D.G.
      • et al.
      Long-term results of RTOG trial 8911 (USA Intergroup 113): a random assignment trial comparison of chemotherapy followed by surgery compared with surgery alone for esophageal cancer.
      ,
      • Ténière P.
      • Hay J.M.
      • Fingerhut A.
      • Fagniez P.L.
      Postoperative radiation therapy does not increase survival after curative resection for squamous cell carcinoma of the middle and lower esophagus as shown by a multicenter controlled trial. French University Association for Surgical Research.
      ,
      • Yang H.
      • Liu H.
      • Chen Y.
      • Zhu C.
      • Fang W.
      • Yu Z.
      • et al.
      Neoadjuvant chemoradiotherapy followed by surgery versus surgery alone for locally advanced squamous cell carcinoma of the esophagus (NEOCRTEC5010): a phase III multicenter, randomized, open-label clinical trial.
      ]. Consequently, 26 trials remained in the primary analysis-set. One trial contained three treatment arms [
      • Lv J.
      • Cao X.F.
      • Zhu B.
      • Ji L.
      • Tao L.
      • Wang D.D.
      Long-term efficacy of perioperative chemoradiotherapy on esophageal squamous cell carcinoma.
      ], hence there were a total of 27 comparisons in the meta-analysis. The study characteristics are described in Table 1. Sample sizes ranged from 30 to 1063 patients (median: 144). Fifteen trials were in the neoadjuvant setting [
      • Mariette C.
      • Dahan L.
      • Mornex F.
      • Maillard E.
      • Thomas P.A.
      • Meunier B.
      • et al.
      Surgery alone versus chemoradiotherapy followed by surgery for stage I and II esophageal cancer: final analysis of randomized controlled phase III trial FFCD 9901.
      ,
      • Alderson D.
      • Cunningham D.
      • Nankivell M.
      • Blazeby J.M.
      • Griffin S.M.
      • Crellin A.
      • et al.
      Neoadjuvant cisplatin and fluorouracil versus epirubicin, cisplatin, and capecitabine followed by resection in patients with oesophageal adenocarcinoma (UK MRC OE05): an open-label, randomised phase 3 trial.
      ,
      • Barbour A.P.
      • Walpole E.T.
      • Mai G.T.
      • Barnes E.H.
      • Watson D.I.
      • Ackland S.P.
      • et al.
      Preoperative cisplatin, fluorouracil, and docetaxel with or without radiotherapy after poor early response to cisplatin and fluorouracil for resectable oesophageal adenocarcinoma (AGITG DOCTOR): results from a multicentre, randomised controlled phase II trial.
      ,
      • Burmeister B.H.
      • Smithers B.M.
      • Gebski V.
      • Fitzgerald L.
      • Simes R.J.
      • Devitt P.
      • et al.
      Surgery alone versus chemoradiotherapy followed by surgery for resectable cancer of the oesophagus: a randomised controlled phase III trial.
      ,
      • Burmeister B.H.
      • Thomas J.M.
      • Burmeister E.A.
      • Walpole E.T.
      • Harvey J.A.
      • Thomson D.B.
      • et al.
      Is concurrent radiation therapy required in patients receiving preoperative chemotherapy for adenocarcinoma of the oesophagus? A randomised phase II trial.
      ,
      • Le Prise E.
      • Etienne P.L.
      • Meunier B.
      • Maddern G.
      • Ben Hassel M.
      • Gedouin D.
      • et al.
      A randomized study of chemotherapy, radiation therapy, and surgery versus surgery for localized squamous cell carcinoma of the esophagus.
      ,
      • Lee J.L.
      • Park S.I.
      • Kim S.B.
      • Jung H.Y.
      • Lee G.H.
      • Kim J.H.
      • et al.
      A single institutional phase III trial of preoperative chemotherapy with hyperfractionation radiotherapy plus surgery versus surgery alone for resectable esophageal squamous cell carcinoma.
      ,
      • Mukherjee S.
      • Hurt C.
      • Cox C.
      • Radhakrishna G.
      • Gwynne S.
      • Bateman A.R.
      • et al.
      Induction oxaliplatin capecitabine followed by switch to carboplatin-paclitaxel based RT versus continuing oxaliplatin capecitabine RT in operable esophageal adenocarcinoma: survival analysis of the randomized phase II neoscope trial.
      ,
      • Shapiro J.
      • van Lanschot J.J.B.
      • Hulshof M.
      • van Hagen P.
      • van Berge Henegouwen M.I.
      • Wijnhoven B.P.L.
      • et al.
      Neoadjuvant chemoradiotherapy plus surgery versus surgery alone for oesophageal or junctional cancer (CROSS): long-term results of a randomised controlled trial.
      ,
      • Stahl M.
      • Walz M.K.
      • Riera-Knorrenschild J.
      • Stuschke M.
      • Sandermann A.
      • Bitzer M.
      • et al.
      Preoperative chemotherapy versus chemoradiotherapy in locally advanced adenocarcinomas of the oesophagogastric junction (POET): long-term results of a controlled randomised trial.
      ,
      • Tepper J.
      • Krasna M.J.
      • Niedzwiecki D.
      • Hollis D.
      • Reed C.E.
      • Goldberg R.
      • et al.
      Phase III trial of trimodality therapy with cisplatin, fluorouracil, radiotherapy, and surgery compared with surgery alone for esophageal cancer: CALGB 9781.
      ,
      • Thomas A.
      • Virdee P.S.
      • Eatock M.
      • Lord S.R.
      • Falk S.
      • Anthoney D.A.
      • et al.
      Dual Erb B Inhibition in Oesophago-gastric Cancer (DEBIOC): A phase I dose escalating safety study and randomised dose expansion of AZD8931 in combination with oxaliplatin and capecitabine chemotherapy in patients with oesophagogastric adenocarcinoma.
      ,
      • Urba S.G.
      • Orringer M.B.
      • Turrisi A.
      • Iannettoni M.
      • Forastiere A.
      • Strawderman M.
      Randomized trial of preoperative chemoradiation versus surgery alone in patients with locoregional esophageal carcinoma.
      ,
      • von Döbeln G.A.
      • Klevebro F.
      • Jacobsen A.B.
      • Johannessen H.O.
      • Nielsen N.H.
      • Johnsen G.
      • et al.
      Neoadjuvant chemotherapy versus neoadjuvant chemoradiotherapy for cancer of the esophagus or gastroesophageal junction: long-term results of a randomized clinical trial.
      ,
      • Yamasaki M.
      • Yasuda T.
      • Yano M.
      • Hirao M.
      • Kobayashi K.
      • Fujitani K.
      • et al.
      Multicenter randomized phase II study of cisplatin and fluorouracil plus docetaxel (DCF) compared with cisplatin and fluorouracil plus Adriamycin (ACF) as preoperative chemotherapy for resectable esophageal squamous cell carcinoma (OGSG1003).
      ]; 3 were in the perioperative setting [
      • Al-Batran S.E.
      • Homann N.
      • Pauligk C.
      • Goetze T.O.
      • Meiler J.
      • Kasper S.
      • et al.
      Perioperative chemotherapy with fluorouracil plus leucovorin, oxaliplatin, and docetaxel versus fluorouracil or capecitabine plus cisplatin and epirubicin for locally advanced, resectable gastric or gastro-oesophageal junction adenocarcinoma (FLOT4): a randomised, phase 2/3 trial.
      ,
      • Cunningham D.
      • Stenning S.P.
      • Smyth E.C.
      • Okines A.F.
      • Allum W.H.
      • Rowley S.
      • et al.
      Peri-operative chemotherapy with or without bevacizumab in operable oesophagogastric adenocarcinoma (UK Medical Research Council ST03): primary analysis results of a multicentre, open-label, randomised phase 2-3 trial.
      ,
      • Kleinberg L.R.
      • Catalano P.J.
      • Forastiere A.A.
      • Keller S.M.
      • Mitchel E.P.
      • Anne P.R.
      • et al.
      Eastern cooperative oncology group and American college of radiology imaging network randomized phase 2 trial of neoadjuvant preoperative paclitaxel/cisplatin/radiation therapy (RT) or irinotecan/cisplatin/RT in esophageal adenocarcinoma: long-term outcome and implications for trial design.
      ]; 1 compared a neoadjuvant treatment to a perioperative treatment [
      • Ruhstaller T.
      • Thuss-Patience P.
      • Hayoz S.
      • Schacher S.
      • Knorrenschild J.R.
      • Schnider A.
      • et al.
      Neoadjuvant chemotherapy followed by chemoradiation and surgery with and without cetuximab in patients with resectable esophageal cancer: a randomized, open-label, phase III trial (SAKK 75/08).
      ]; 4 were in the adjuvant setting [
      • Deng W.
      • Yang J.
      • Ni W.
      • Li C.
      • Chang X.
      • Han W.
      • et al.
      Postoperative radiotherapy in pathological T2-3N0M0 thoracic esophageal squamous cell carcinoma: interim report of a prospective, phase III, randomized controlled study.
      ,
      • Ando N.
      • Iizuka T.
      • Ide H.
      • Ishida K.
      • Shinoda M.
      • Nishimaki T.
      • et al.
      Surgery plus chemotherapy compared with surgery alone for localized squamous cell carcinoma of the thoracic esophagus: a Japan Clinical Oncology Group Study--JCOG9204.
      ,
      • Lim S.H.
      • Shim Y.M.
      • Park S.H.
      • Kim H.K.
      • Choi Y.S.
      • Ahn M.J.
      • et al.
      A randomized phase II study of leucovorin/5-fluorouracil with or without oxaliplatin (LV5FU2 vs. FOLFOX) for curatively-resected, node-positive esophageal squamous cell carcinoma.
      ,
      • Zieren H.U.
      • Müller J.M.
      • Jacobi C.A.
      • Pichlmaier H.
      • Müller R.P.
      • Staar S.
      Adjuvant postoperative radiation therapy after curative resection of squamous cell carcinoma of the thoracic esophagus: a prospective randomized study.
      ]; and 3 compared a neoadjuvant treatment with an adjuvant treatment [
      • Zhao Y.
      • Han L.
      • Zhang W.
      • Shan L.
      • Wang Y.
      • Song P.
      • et al.
      Preoperative chemotherapy compared with postoperative adjuvant chemotherapy for squamous cell carcinoma of the thoracic oesophagus with the detection of circulating tumour cells randomized controlled trial.
      ,
      • Ando N.
      • Kato H.
      • Igaki H.
      • Shinoda M.
      • Ozawa S.
      • Shimizu H.
      • et al.
      A randomized trial comparing postoperative adjuvant chemotherapy with cisplatin and 5-fluorouracil versus preoperative chemotherapy for localized advanced squamous cell carcinoma of the thoracic esophagus (JCOG9907).
      ,
      • Lv J.
      • Cao X.F.
      • Zhu B.
      • Ji L.
      • Tao L.
      • Wang D.D.
      Long-term efficacy of perioperative chemoradiotherapy on esophageal squamous cell carcinoma.
      ]. The set of trials including at least one arm studying an adjuvant treatment formed the secondary analysis-set.
      Table 1Summary of RCTs identified through SLR.
      Author Year (Trial name)Sample sizeInterventionComparatorFollow-Up (months)
      Estimated follow-up time based on maximum time-to-event data from digitized pseudo patient-level data. Abbreviations: 5-FU – 5-fluorouracil; CDDP – Cisplatin; CRT – Chemoradiotherapy; CT – Chemotherapy; HR – Hazard ratio; KM – Kaplan–Meier curve; NACRT – Neoadjuvant chemoradiotherapy; NR – Not reported; NACT – Neoadjuvant chemotherapy; RT – Radiotherapy; S – Surgery; TXT – Docetaxel.
      Alderson et al., 2017 (UK MRC OE05) [
      • Alderson D.
      • Cunningham D.
      • Nankivell M.
      • Blazeby J.M.
      • Griffin S.M.
      • Crellin A.
      • et al.
      Neoadjuvant cisplatin and fluorouracil versus epirubicin, cisplatin, and capecitabine followed by resection in patients with oesophageal adenocarcinoma (UK MRC OE05): an open-label, randomised phase 3 trial.
      ]
      897NACT (epirubicin, CDDP and capecitabine) + SNACT (CDDP and 5-FU) + S72
      Barbour et al., 2020 (AGITG DOCTOR) [
      • Barbour A.P.
      • Walpole E.T.
      • Mai G.T.
      • Barnes E.H.
      • Watson D.I.
      • Ackland S.P.
      • et al.
      Preoperative cisplatin, fluorouracil, and docetaxel with or without radiotherapy after poor early response to cisplatin and fluorouracil for resectable oesophageal adenocarcinoma (AGITG DOCTOR): results from a multicentre, randomised controlled phase II trial.
      ]
      66NACRT (RT + CDDP + 5-FU + TXT)NACT (CDDP + 5-FU + TXT)60
      Burmeister et al., 2005 [
      • Burmeister B.H.
      • Smithers B.M.
      • Gebski V.
      • Fitzgerald L.
      • Simes R.J.
      • Devitt P.
      • et al.
      Surgery alone versus chemoradiotherapy followed by surgery for resectable cancer of the oesophagus: a randomised controlled phase III trial.
      ]
      257NACRT (CDDP, 5-FU) + SS alone72
      Burmeister et al., 2011 [
      • Burmeister B.H.
      • Thomas J.M.
      • Burmeister E.A.
      • Walpole E.T.
      • Harvey J.A.
      • Thomson D.B.
      • et al.
      Is concurrent radiation therapy required in patients receiving preoperative chemotherapy for adenocarcinoma of the oesophagus? A randomised phase II trial.
      ]
      75NACRT (CDDP, 5-FU) + SNACT (CDDP, 5-FU) + S71
      Cunningham et al., 2017 (ST03) [
      • Cunningham D.
      • Stenning S.P.
      • Smyth E.C.
      • Okines A.F.
      • Allum W.H.
      • Rowley S.
      • et al.
      Peri-operative chemotherapy with or without bevacizumab in operable oesophagogastric adenocarcinoma (UK Medical Research Council ST03): primary analysis results of a multicentre, open-label, randomised phase 2-3 trial.
      ]
      1063Bevacizumab + capecitabine + CDDP + epirubicin + SCapecitabine + CDDP + epirubicin + S36.2 median
      Deng et al., 2020 (IMRT) [
      • Deng W.
      • Yang J.
      • Ni W.
      • Li C.
      • Chang X.
      • Han W.
      • et al.
      Postoperative radiotherapy in pathological T2-3N0M0 thoracic esophageal squamous cell carcinoma: interim report of a prospective, phase III, randomized controlled study.
      ]
      167S + postoperative RTS alone60
      Al-Batran et al., 2019 (FLOT4) [
      • Al-Batran S.E.
      • Homann N.
      • Pauligk C.
      • Goetze T.O.
      • Meiler J.
      • Kasper S.
      • et al.
      Perioperative chemotherapy with fluorouracil plus leucovorin, oxaliplatin, and docetaxel versus fluorouracil or capecitabine plus cisplatin and epirubicin for locally advanced, resectable gastric or gastro-oesophageal junction adenocarcinoma (FLOT4): a randomised, phase 2/3 trial.
      ]
      7165-FU + leucovorin + oxaliplatin + TXT (FLOT)5-FU/capecitabine + CDDP + epirubicin (ECF/ECX)43 median
      Ando et al., 2003 (JCOG9204) [
      • Ando N.
      • Iizuka T.
      • Ide H.
      • Ishida K.
      • Shinoda M.
      • Nishimaki T.
      • et al.
      Surgery plus chemotherapy compared with surgery alone for localized squamous cell carcinoma of the thoracic esophagus: a Japan Clinical Oncology Group Study--JCOG9204.
      ]
      242S + CDDP + 5-FUS alone72
      Ando et al., 2012 (JCOG9907) [
      • Ando N.
      • Kato H.
      • Igaki H.
      • Shinoda M.
      • Ozawa S.
      • Shimizu H.
      • et al.
      A randomized trial comparing postoperative adjuvant chemotherapy with cisplatin and 5-fluorouracil versus preoperative chemotherapy for localized advanced squamous cell carcinoma of the thoracic esophagus (JCOG9907).
      ]
      330CDDP + 5-FU (neoadjuvant)CDDP + 5-FU (adjuvant)72
      Kleinberg et al., 2016 [
      • Kleinberg L.R.
      • Catalano P.J.
      • Forastiere A.A.
      • Keller S.M.
      • Mitchel E.P.
      • Anne P.R.
      • et al.
      Eastern cooperative oncology group and American college of radiology imaging network randomized phase 2 trial of neoadjuvant preoperative paclitaxel/cisplatin/radiation therapy (RT) or irinotecan/cisplatin/RT in esophageal adenocarcinoma: long-term outcome and implications for trial design.
      ]
      97NACRT (CDDP, irinotecan) + S + adjuvant CT (CDDP, irinotecan)NACRT (CDDP, paclitaxel) + S + adjuvant CT (CDDP, paclitaxel)72
      Le Prise et al., 1994 [
      • Le Prise E.
      • Etienne P.L.
      • Meunier B.
      • Maddern G.
      • Ben Hassel M.
      • Gedouin D.
      • et al.
      A randomized study of chemotherapy, radiation therapy, and surgery versus surgery for localized squamous cell carcinoma of the esophagus.
      ]
      86NACRT (CDDP, 5-FU) + SS alone27
      Lee et al., 2004 [
      • Lee J.L.
      • Park S.I.
      • Kim S.B.
      • Jung H.Y.
      • Lee G.H.
      • Kim J.H.
      • et al.
      A single institutional phase III trial of preoperative chemotherapy with hyperfractionation radiotherapy plus surgery versus surgery alone for resectable esophageal squamous cell carcinoma.
      ]
      101NACRT (CDDP, 5-FU) + SS alone36
      Lim et al., 2017 [
      • Lim S.H.
      • Shim Y.M.
      • Park S.H.
      • Kim H.K.
      • Choi Y.S.
      • Ahn M.J.
      • et al.
      A randomized phase II study of leucovorin/5-fluorouracil with or without oxaliplatin (LV5FU2 vs. FOLFOX) for curatively-resected, node-positive esophageal squamous cell carcinoma.
      ]
      625-FU + leucovorin + oxaliplatin5-FU + leucovorin58
      Lv et al., 2010 [
      • Lv J.
      • Cao X.F.
      • Zhu B.
      • Ji L.
      • Tao L.
      • Wang D.D.
      Long-term efficacy of perioperative chemoradiotherapy on esophageal squamous cell carcinoma.
      ]
      238CDDP + paclitaxel + radiation (neoadjuvant/adjuvant)S alone72
      Mariette et al., 2014 (FFCD 9901) [
      • Mariette C.
      • Dahan L.
      • Mornex F.
      • Maillard E.
      • Thomas P.A.
      • Meunier B.
      • et al.
      Surgery alone versus chemoradiotherapy followed by surgery for stage I and II esophageal cancer: final analysis of randomized controlled phase III trial FFCD 9901.
      ]
      195NACRT (CDDP, 5-FU) + SS alone72
      Mukherjee et al., 2020 (NeoSCOPE) [
      • Mukherjee S.
      • Hurt C.
      • Cox C.
      • Radhakrishna G.
      • Gwynne S.
      • Bateman A.R.
      • et al.
      Induction oxaliplatin capecitabine followed by switch to carboplatin-paclitaxel based RT versus continuing oxaliplatin capecitabine RT in operable esophageal adenocarcinoma: survival analysis of the randomized phase II neoscope trial.
      ]
      85NACRT (RT + carboplatin + paclitaxel) + SNACRT (RT + oxaliplatin + capecitabine) + S40.7 median
      Ruhstaller et al., 2018 (SAKK7508) [
      • Ruhstaller T.
      • Thuss-Patience P.
      • Hayoz S.
      • Schacher S.
      • Knorrenschild J.R.
      • Schnider A.
      • et al.
      Neoadjuvant chemotherapy followed by chemoradiation and surgery with and without cetuximab in patients with resectable esophageal cancer: a randomized, open-label, phase III trial (SAKK 75/08).
      ]
      300Cetuximab + multimodal therapyMultimodal therapy48 median
      Shapiro et al., 2015 (CROSS) [
      • Shapiro J.
      • van Lanschot J.J.B.
      • Hulshof M.
      • van Hagen P.
      • van Berge Henegouwen M.I.
      • Wijnhoven B.P.L.
      • et al.
      Neoadjuvant chemoradiotherapy plus surgery versus surgery alone for oesophageal or junctional cancer (CROSS): long-term results of a randomised controlled trial.
      ]
      368NACRT (carboplatin, paclitaxel) + SS alone72
      Stahl et al., 2017 (POET) [
      • Stahl M.
      • Walz M.K.
      • Riera-Knorrenschild J.
      • Stuschke M.
      • Sandermann A.
      • Bitzer M.
      • et al.
      Preoperative chemotherapy versus chemoradiotherapy in locally advanced adenocarcinomas of the oesophagogastric junction (POET): long-term results of a controlled randomised trial.
      ]
      126S + NACRT (RT + 5-FU + folinic acid + CDDP + etoposide)S + NACT (5-FU + folinic acid + CDDP)72
      Tepper et al., 2008 (CALGB 9781) [
      • Tepper J.
      • Krasna M.J.
      • Niedzwiecki D.
      • Hollis D.
      • Reed C.E.
      • Goldberg R.
      • et al.
      Phase III trial of trimodality therapy with cisplatin, fluorouracil, radiotherapy, and surgery compared with surgery alone for esophageal cancer: CALGB 9781.
      ]
      56S + NACRT (RT + CDDP + 5-FU)S alone71
      Thomas et al., 2020 (DEBIOC) [
      • Thomas A.
      • Virdee P.S.
      • Eatock M.
      • Lord S.R.
      • Falk S.
      • Anthoney D.A.
      • et al.
      Dual Erb B Inhibition in Oesophago-gastric Cancer (DEBIOC): A phase I dose escalating safety study and randomised dose expansion of AZD8931 in combination with oxaliplatin and capecitabine chemotherapy in patients with oesophagogastric adenocarcinoma.
      ]
      30Neoadjuvant AZD8931 + Xelox CT (oxaliplatin + capecitabine)Neoadjuvant Xelox CT (oxaliplatin + capecitabine)32
      Urba et al., 2001 [
      • Urba S.G.
      • Orringer M.B.
      • Turrisi A.
      • Iannettoni M.
      • Forastiere A.
      • Strawderman M.
      Randomized trial of preoperative chemoradiation versus surgery alone in patients with locoregional esophageal carcinoma.
      ]
      100S + NACRT (RT + CDDP + 5-FU + vinblastine)S alone72
      Von Dobeln et al., 2019 (NeoRes) [
      • von Döbeln G.A.
      • Klevebro F.
      • Jacobsen A.B.
      • Johannessen H.O.
      • Nielsen N.H.
      • Johnsen G.
      • et al.
      Neoadjuvant chemotherapy versus neoadjuvant chemoradiotherapy for cancer of the esophagus or gastroesophageal junction: long-term results of a randomized clinical trial.
      ]
      181NACRT (RT + CDDP + 5-FU)NACT (CDDP + 5-FU)59
      Yamasaki et al., 2017 [
      • Yamasaki M.
      • Yasuda T.
      • Yano M.
      • Hirao M.
      • Kobayashi K.
      • Fujitani K.
      • et al.
      Multicenter randomized phase II study of cisplatin and fluorouracil plus docetaxel (DCF) compared with cisplatin and fluorouracil plus Adriamycin (ACF) as preoperative chemotherapy for resectable esophageal squamous cell carcinoma (OGSG1003).
      ]
      162S + NACT (adriamycin + CDDP + 5-FU)S + NACT (TXT + CDDP + 5-FU)38
      Zhao et al., 2020 [
      • Zhao Y.
      • Han L.
      • Zhang W.
      • Shan L.
      • Wang Y.
      • Song P.
      • et al.
      Preoperative chemotherapy compared with postoperative adjuvant chemotherapy for squamous cell carcinoma of the thoracic oesophagus with the detection of circulating tumour cells randomized controlled trial.
      ]
      115NACTAdjuvant CT34
      Zieren et al., 1995 [
      • Zieren H.U.
      • Müller J.M.
      • Jacobi C.A.
      • Pichlmaier H.
      • Müller R.P.
      • Staar S.
      Adjuvant postoperative radiation therapy after curative resection of squamous cell carcinoma of the thoracic esophagus: a prospective randomized study.
      ]
      68RT + SS alone35
      a Estimated follow-up time based on maximum time-to-event data from digitized pseudo patient-level data. Abbreviations: 5-FU – 5-fluorouracil; CDDP – Cisplatin; CRT – Chemoradiotherapy; CT – Chemotherapy; HR – Hazard ratio; KM – Kaplan–Meier curve; NACRT – Neoadjuvant chemoradiotherapy; NR – Not reported; NACT – Neoadjuvant chemotherapy; RT – Radiotherapy; S – Surgery; TXT – Docetaxel.
      The BRMA correlation estimate between log-transformed HRDFS/PFS and log-transformed HROS for the primary analysis was 0.83 (95% CI: 0.70–0.90). Using the WLR, the estimate increased to 0.89 (95% CI: 0.76–0.95). Fig. 2 displays the resulting surrogacy equation between the log-transformed treatment effects for DFS/PFS and OS from the WLR. The estimated surrogacy equation was log (HROS) = 0.80 × log(HRDFS/PFS) and the resulting STE was 0.82, indicating that a reported HRDFS/PFS of 0.82 or lower from a trial would lead to a statistically significant HROS at the 95% confidence level favoring the intervention.
      Fig. 2
      Fig. 2WLR to estimate the association between the log-transformed HRDFS/PFS and HROS using sample sizes as weights (primary analysis). Legend: The predictive surrogacy equation is graphed as a solid straight line with its corresponding predictive interval boundaries as dotted curved lines. The red dots are plotted using measures from the reported log(HRDFS/PFS) on the x-axis against measures of log(HROS) on y-axis for each treatment comparison per trial. Sizes of the dots correspond to the weights associated within the surrogacy equation. For this particular regression model, the STEs are crossed on the predictive line at 0.82. Abbreviations: HR = Hazard ratio; OS = Overall survival; DFS/PFS = Disease-free survival or progression-free survival.
      The results of the LOOCV are presented in Fig. 3. Since observed HROS fell outside of 95% PI of the predicted HROS only for one trial [
      • Le Prise E.
      • Etienne P.L.
      • Meunier B.
      • Maddern G.
      • Ben Hassel M.
      • Gedouin D.
      • et al.
      A randomized study of chemotherapy, radiation therapy, and surgery versus surgery for localized squamous cell carcinoma of the esophagus.
      ], the SE was deemed to be valid [
      • Mariette C.
      • Dahan L.
      • Mornex F.
      • Maillard E.
      • Thomas P.A.
      • Meunier B.
      • et al.
      Surgery alone versus chemoradiotherapy followed by surgery for stage I and II esophageal cancer: final analysis of randomized controlled phase III trial FFCD 9901.
      ]. Across the sensitivity analyses there were minimal changes to the correlation estimate, STE, and overall results of LOOCV (Table 2).
      Fig. 3
      Fig. 3Results from the LOOCV in the primary analysis. Legend: The overall results of the LOOCV are presented in the above graph. In this figure, the red dots are the reported HROS measures and the black dots are the predicted HROS measures. Corresponding to the predicted HROS measures, the 95% PIs are presented as black solid lines per each trial. To assess the cross-validation results, reported log(HROS) measures should fall within the plotted interval. Abbreviations: HR – Hazard ratio; OS – Overall survival.
      Table 2Pearson's correlation coefficients from the sensitivity analyses.
      Analysis setModelEstimate (95% CI)
      Removing trials with misaligned outcome definitions with other trials in the evidence base (24 trials remaining):

      Lim et al., 2017 [
      • Lim S.H.
      • Shim Y.M.
      • Park S.H.
      • Kim H.K.
      • Choi Y.S.
      • Ahn M.J.
      • et al.
      A randomized phase II study of leucovorin/5-fluorouracil with or without oxaliplatin (LV5FU2 vs. FOLFOX) for curatively-resected, node-positive esophageal squamous cell carcinoma.
      ]

      Lv et al., 2010 [
      • Lv J.
      • Cao X.F.
      • Zhu B.
      • Ji L.
      • Tao L.
      • Wang D.D.
      Long-term efficacy of perioperative chemoradiotherapy on esophageal squamous cell carcinoma.
      ]
      BRMA0.83 (0.69–0.91)
      WLR
      Weighted linear regression uses trials' sample sizes as weights. Abbreviations: CI – Confidence interval; BRMA – Bivariate random effect meta-analyses; GC – Gastric cancer; WLR – Weighted linear regression.
      0.90 (0.74–0.97)
      Removing trial with mixed GC patients (25 trials remaining):

      Cunningham et al., 2017 (ST03) [
      • Cunningham D.
      • Stenning S.P.
      • Smyth E.C.
      • Okines A.F.
      • Allum W.H.
      • Rowley S.
      • et al.
      Peri-operative chemotherapy with or without bevacizumab in operable oesophagogastric adenocarcinoma (UK Medical Research Council ST03): primary analysis results of a multicentre, open-label, randomised phase 2-3 trial.
      ]
      BRMA0.80 (0.66–0.88)
      WLR
      Weighted linear regression uses trials' sample sizes as weights. Abbreviations: CI – Confidence interval; BRMA – Bivariate random effect meta-analyses; GC – Gastric cancer; WLR – Weighted linear regression.
      0.88 (0.72–0.96)
      Removing trial with low participant adherence (25 trials remaining):

      Zhao et al., 2020 [
      • Zhao Y.
      • Han L.
      • Zhang W.
      • Shan L.
      • Wang Y.
      • Song P.
      • et al.
      Preoperative chemotherapy compared with postoperative adjuvant chemotherapy for squamous cell carcinoma of the thoracic oesophagus with the detection of circulating tumour cells randomized controlled trial.
      ]
      BRMA0.83 (0.7–0.91)
      WLR
      Weighted linear regression uses trials' sample sizes as weights. Abbreviations: CI – Confidence interval; BRMA – Bivariate random effect meta-analyses; GC – Gastric cancer; WLR – Weighted linear regression.
      0.89 (0.73–0.96)
      a Weighted linear regression uses trials' sample sizes as weights. Abbreviations: CI – Confidence interval; BRMA – Bivariate random effect meta-analyses; GC – Gastric cancer; WLR – Weighted linear regression.
      The predictive performance of the surrogacy equation was tested by comparing the model predictions to the reported values of HROS in two pivotal trials: CROSS and FLOT4 (Table 3). In both trials the observed HROS estimates were not only contained within the 95% PI but also were close to their observed counterparts. Table 3 also presents the predicted HROS for the ongoing CheckMate 577 trial, the first phase 3 trial assessing adjuvant nivolumab versus placebo [
      • Kelly R.J.
      • Ajani J.A.
      • Kuzdzal J.
      • Zander T.
      • Van Cutsem E.
      • Piessen G.
      • et al.
      Adjuvant nivolumab in resected esophageal or gastroesophageal junction cancer.
      ]. The predicted HROS in this trial was statistically significant at the 95% confidence level. Since OS data from this study continue to mature and have not been publicly reported yet, the observed HROS is currently unavailable for comparison with model predictions.
      Table 3Observed and predicted HROS obtained by using observed HRDFS/PFS in the surrogacy equation from the WLR model.
      TrialObserved HRDFS/PFS (95% CI)Observed HROS (95% CI)Predicted HROS (95% PI)
      CROSS
      Observed hazard ratios were obtained from the multivariable estimates in Table 2 of the Shapiro et al. 2015 publication [41].
      0.64 (0.49–0.82)0.69 (0.53–0.89)0.70 (0.58–0.85)
      FLOT4
      Observed hazard ratio for OS in the GEJC subgroup was obtained from Fig. 3 of the Al-Batran et al. 2019 publication [26]. For the observed PFS (from randomization), the hazard ratio estimate was obtained directly from the study investigators.
      0.74 (0.58–0.95)0.76 (not reported)0.79 (0.66–0.94)
      CheckMate 577
      Observed estimate for DFS was obtained from the Kelly et al. 2021 publication [5].
      0.69 (0.56–0.86)Not available0.75 (0.65–0.85)
      Abbreviations: CI – Confidence interval; DFS/PFS – Disease-free survival or progression-free survival; HR – Hazard ratio; OS – Overall survival; PI – Prediction interval.
      a Observed hazard ratios were obtained from the multivariable estimates in Table 2 of the Shapiro et al. 2015 publication [
      • Shapiro J.
      • van Lanschot J.J.B.
      • Hulshof M.
      • van Hagen P.
      • van Berge Henegouwen M.I.
      • Wijnhoven B.P.L.
      • et al.
      Neoadjuvant chemoradiotherapy plus surgery versus surgery alone for oesophageal or junctional cancer (CROSS): long-term results of a randomised controlled trial.
      ].
      b Observed hazard ratio for OS in the GEJC subgroup was obtained from Fig. 3 of the Al-Batran et al. 2019 publication [
      • Al-Batran S.E.
      • Homann N.
      • Pauligk C.
      • Goetze T.O.
      • Meiler J.
      • Kasper S.
      • et al.
      Perioperative chemotherapy with fluorouracil plus leucovorin, oxaliplatin, and docetaxel versus fluorouracil or capecitabine plus cisplatin and epirubicin for locally advanced, resectable gastric or gastro-oesophageal junction adenocarcinoma (FLOT4): a randomised, phase 2/3 trial.
      ]. For the observed PFS (from randomization), the hazard ratio estimate was obtained directly from the study investigators.
      c Observed estimate for DFS was obtained from the Kelly et al. 2021 publication [
      • Kelly R.J.
      • Ajani J.A.
      • Kuzdzal J.
      • Zander T.
      • Van Cutsem E.
      • Piessen G.
      • et al.
      Adjuvant nivolumab in resected esophageal or gastroesophageal junction cancer.
      ].
      Considering the close familial link (particularly with regard to etiologies and treatment) between GC and EC/GEJC, we also assessed the predictive accuracy of the estimated surrogacy equations on select adjuvant GC trials, namely ACTS-GC, CLASSIC, and ARTIST (Table 4) [
      • Park S.H.
      • Sohn T.S.
      • Lee J.
      • Lim D.H.
      • Hong M.E.
      • Kim K.M.
      • et al.
      Phase III trial to compare adjuvant chemotherapy with capecitabine and cisplatin versus concurrent chemoradiotherapy in gastric cancer: final report of the adjuvant chemoradiotherapy in stomach tumors trial, including survival and subset analyses.
      ,
      • Bang Y.J.
      • Kim Y.W.
      • Yang H.K.
      • Chung H.C.
      • Park Y.K.
      • Lee K.H.
      • et al.
      Adjuvant capecitabine and oxaliplatin for gastric cancer after D2 gastrectomy (CLASSIC): a phase 3 open-label, randomised controlled trial.
      ,
      • Sakuramoto S.
      • Sasako M.
      • Yamaguchi T.
      • Kinoshita T.
      • Fujii M.
      • Nashimoto A.
      • et al.
      Adjuvant chemotherapy for gastric cancer with S-1, an oral fluoropyrimidine.
      ]. The surrogacy equation's predictions were fairly close to the reported HROS for ACTS-GC and CLASSIC, but not for ARTIST which could be explained with the following reasoning: Most of the trials (25/27) in our evidence base had concordant results, meaning treatment effects were either favorable, or harmful, for both endpoints (i.e., OS and DFS/PFS). Therefore, the resulting surrogacy equation had the tendency to characterize a positive relationship between the treatment effects of OS and DFS. On the other hand, ARTIST trial had a discordant result and investigated earlier stage tumors (I/II) [
      • Park S.H.
      • Sohn T.S.
      • Lee J.
      • Lim D.H.
      • Hong M.E.
      • Kim K.M.
      • et al.
      Phase III trial to compare adjuvant chemotherapy with capecitabine and cisplatin versus concurrent chemoradiotherapy in gastric cancer: final report of the adjuvant chemoradiotherapy in stomach tumors trial, including survival and subset analyses.
      ].
      Table 4Observed and predicted HROS for select GC trials obtained by using observed HRDFS/PFS in the surrogacy equation from the WLR model.
      TrialObserved HRDFS/PFS (95% CI)Observed HROS (95% CI)Predicted HROS (95% PI)
      ACTS-GC [
      • Sakuramoto S.
      • Sasako M.
      • Yamaguchi T.
      • Kinoshita T.
      • Fujii M.
      • Nashimoto A.
      • et al.
      Adjuvant chemotherapy for gastric cancer with S-1, an oral fluoropyrimidine.
      ]
      0.62 (0.50–0.77)0.68 (0.52–0.87)0.69 (0.61–0.78)
      CLASSIC [
      • Bang Y.J.
      • Kim Y.W.
      • Yang H.K.
      • Chung H.C.
      • Park Y.K.
      • Lee K.H.
      • et al.
      Adjuvant capecitabine and oxaliplatin for gastric cancer after D2 gastrectomy (CLASSIC): a phase 3 open-label, randomised controlled trial.
      ]
      0.56 (0.44–0.72)0.72 (0.52–1.00)0.63 (0.55–0.72)
      ARTIST [
      • Park S.H.
      • Sohn T.S.
      • Lee J.
      • Lim D.H.
      • Hong M.E.
      • Kim K.M.
      • et al.
      Phase III trial to compare adjuvant chemotherapy with capecitabine and cisplatin versus concurrent chemoradiotherapy in gastric cancer: final report of the adjuvant chemoradiotherapy in stomach tumors trial, including survival and subset analyses.
      ]
      0.74 (0.52–1.05)1.13 (0.78–1.65)0.79 (0.67–0.93)
      Abbreviations: CI – Confidence interval; DFS/PFS – Disease-free survival or progression-free survival; GC – Gastric cancer; HR – Hazard ratio; OS – Overall survival; PI – Prediction interval.
      The secondary analysis in the adjuvant setting showed an overall correlation estimate of 0.76 (95% CI: 0.18–0.95) between DFS/PFS and OS using the BRMA, which increased to 0.82 (95% CI: 0.19–0.97) when using WLR. The resulting 95% CIs were wider compared to the primary analysis due to the substantially low number of trials in this analysis set. As shown in Fig. 4, the WLR estimated the surrogacy equation to be log(HROS) = −0.03 + 0.65 × log(HRDFS/PFS). Using this equation, for the ongoing adjuvant CheckMate 577 trial, the HROS was predicted to be 0.76 (95% PI: 0.58–0.99). This HROS prediction was similar to that from our primary analysis albeit with a larger PI. Relative to the primary analysis, a lower STE of 0.64 was estimated due to increased uncertainty borne by relatively sparse evidence base. During LOOCV, the reported HROS estimates for all seven trials in the secondary analysis fell within the 95% PI of the predicted HROS, supporting the validity of the surrogacy model per NICE TSD 20.
      Fig. 4
      Fig. 4WLR to estimate the association between the log-transformed HRDFS/PFS and HROS using sample sizes as weights (secondary analysis). Legend: The predictive surrogacy equation is graphed as a solid straight line with its corresponding predictive interval boundaries as dotted curved lines. The red dots are plotted using measures from the reported log(HRDFS/PFS) on the x-axis against measures of log(HROS) on y-axis for each treatment comparison per trial. Sizes of the dots correspond to the weights associated within the surrogacy equation. For this particular regression model, the STEs are crossed on the predictive line at 0.64. Abbreviations: HR – Hazard ratio; OS – Overall survival; DFS/PFS – Disease-free survival or progression-free survival.
      To assess the degree of strength for the correlation estimates, we followed the German Institute of Quality and Efficiency in Health Care (IQWiG) guidelines [
      ]. High correlation is defined as the lower limit of the 95% CI of the correlation statistic being ≥ 0.85, and low correlation is defined as the upper limit of the 95% CI of the correlation statistic being ≤ 0.7 [
      ]. Correlation is considered moderate outside of these two scenarios. Under this guidance, for both the primary and secondary analyses, the correlation between the treatment effects on DFS/PFS and OS was moderate regardless of the approach it was derived from (e.g., BRMA and WLR).

      4. Discussion

      Our surrogacy analysis shows that within the neoadjuvant, perioperative, and adjuvant settings for EC and GEJC, DFS/PFS is a useful and valid SE to OS. Its usefulness as a SE was seen through its high STE, while its validity was established through the high coverage of the LOOCV. While these two criteria emphasize the validity of the model in alignment with the guidance outlined by NICE [
      • Mariette C.
      • Dahan L.
      • Mornex F.
      • Maillard E.
      • Thomas P.A.
      • Meunier B.
      • et al.
      Surgery alone versus chemoradiotherapy followed by surgery for stage I and II esophageal cancer: final analysis of randomized controlled phase III trial FFCD 9901.
      ], with respect to the IQWiG criteria, the correlation between the endpoints was moderate [
      ]. The finding of the high STE in primary analysis did not carry through to the secondary analysis of trials in the adjuvant setting only as there were limited number of trials, leading to greater model uncertainty. Nonetheless, the rest of the results were robust with the estimates being consistent across a variety of sensitivity analyses.
      The growing interest in surrogacy analyses has also led to improved guidance for both manufacturers and the authorities granting regulatory and reimbursement approval [
      • Mariette C.
      • Dahan L.
      • Mornex F.
      • Maillard E.
      • Thomas P.A.
      • Meunier B.
      • et al.
      Surgery alone versus chemoradiotherapy followed by surgery for stage I and II esophageal cancer: final analysis of randomized controlled phase III trial FFCD 9901.
      ]. However, there are no universally accepted criteria to classify the strength of the surrogacy relationship based on a commonly used standard correlation measure (e.g., Pearson's correlation coefficient, or Spearman's or Kendall's rank correlation coefficient). For a correlation to be classified as “strong,” the IQWiG criteria require the lower limit of the 95% CI of the correlation coefficient to be above 0.85 with a lack of clarity on the applicability of these criteria for classifying trial level correlation [
      ]. The NICE DSU argues that strength of association may not be the best measure of classifying surrogacy [
      • Mariette C.
      • Dahan L.
      • Mornex F.
      • Maillard E.
      • Thomas P.A.
      • Meunier B.
      • et al.
      Surgery alone versus chemoradiotherapy followed by surgery for stage I and II esophageal cancer: final analysis of randomized controlled phase III trial FFCD 9901.
      ], rather it is the ability to make meaningful predictions with a valid model that matters most.
      One of the unique strengths of our evidence base is that it includes two key trials: CROSS, which set the neo-adjuvant treatment standard of care, and FLOT4, which expanded the available options for the treatment of locally advanced, resectable GC and GEJC [
      • Al-Batran S.E.
      • Homann N.
      • Pauligk C.
      • Goetze T.O.
      • Meiler J.
      • Kasper S.
      • et al.
      Perioperative chemotherapy with fluorouracil plus leucovorin, oxaliplatin, and docetaxel versus fluorouracil or capecitabine plus cisplatin and epirubicin for locally advanced, resectable gastric or gastro-oesophageal junction adenocarcinoma (FLOT4): a randomised, phase 2/3 trial.
      ,
      • Shapiro J.
      • van Lanschot J.J.B.
      • Hulshof M.
      • van Hagen P.
      • van Berge Henegouwen M.I.
      • Wijnhoven B.P.L.
      • et al.
      Neoadjuvant chemoradiotherapy plus surgery versus surgery alone for oesophageal or junctional cancer (CROSS): long-term results of a randomised controlled trial.
      ]. CROSS evaluated chemoradiotherapy followed by surgery as the intervention versus surgery alone as control, finding that the intervention significantly improved 5-year OS rates over the control arm. FLOT4 had a similar finding for its treatments, finding that perioperative combination chemotherapy of fluorouracil plus leucovorin, oxaliplatin, and docetaxel significantly improved OS as compared with perioperative combination therapies epirubicin, plus cisplatin and 5-fluorouracil (5-FU) or capecitabine. It is important to recognize that our modeled estimates and their statistical significance were similar to the reported values from the study publications.
      To our knowledge, this study is the only recent surrogacy analysis in the literature that spans all treatment settings in early EC/GEJC. The evidence base contains a large number of studies (n = 26), supporting generalizability of the results and creating smaller margin of errors for the surrogacy equation. Previous surrogacy studies in EC/GEJC have only considered 8 or 10 trials due to the limited scope of their treatment settings [
      • Kataoka K.
      • Nakamura K.
      • Mizusawa J.
      • Kato K.
      • Eba J.
      • Katayama H.
      • et al.
      Surrogacy of progression-free survival (PFS) for overall survival (OS) in esophageal cancer trials with preoperative therapy: literature-based meta-analysis.
      ]. Additionally, neither of these past studies employed BRMA while ours explored it in addition to the WLR. Lastly, proportional hazards assumption was rigorously tested so that all contributing studies contained meaningful and statistically valid HR estimates. To this end, digitization of KM curves provided the opportunity to verify the statistical appropriateness of using HRs for relative treatment effects in the analyses which helped eliminate any potential source of bias.
      Our study has limitations in terms of the utilization of available data. First, we acknowledge that the two-stage approach described by Burzykowski et al., 2006, which assesses both endpoint and treatment–effect correlations, would provide more robust conclusions on the appropriateness of DFS as a surrogate to OS [
      • Burzykowski T.
      • Buyse M.
      Surrogate threshold effect: an alternative measure for meta-analytic surrogate endpoint validation.
      ]. If provided access to IPD, individual endpoint correlations could also be explored further to complement our findings. Second, the studies included in this analysis evaluated chemotherapy-based regimens, and therefore the results and conclusions from our meta-analyses may not be generalizable to other therapies with varying mechanisms of actions. Third, our analyses assume that prognostic factors for DFS are generally similar to those of OS. In the event that a prognostic factor affected DFS, but not OS, then differences in the prognostic factor between studies would affect the estimated surrogacy equation. On the other hand, if the prognostic factors were similar, as assumed, then differences across studies would only serve to strengthen the analysis by increasing the heterogeneity in treatment effects and providing better support for a more reliable estimation of the surrogacy equation. Lastly, due to limited availability of consistent DFS definitions in the evidence base we decided to be inclusive of more broadened definitions, which may pose as a potential source of bias. For example, taking DFS from the time of surgery, or a week or 6 months post-surgery, may create a bias in that patients are theoretically balanced at randomization but after surgery may no longer be balanced with respect to important prognostic factors (e.g., quality of the preoperative radiation if given, individual surgeons definitions of resectability, the extent and type of resection and approach, Intensive Therapy Unit facilities, and anesthetics between the treatment groups).

      5. Conclusions

      Based on a relatively large evidence base in resectable EC and GEJC, our correlation meta-analysis suggests that HRDFS/PFS can be used as a SE for HROS in the neoadjuvant, perioperative, or adjuvant settings. LOOCV confirmed the validity of the surrogacy equation, and the high STE of 0.82 suggested that the equation has strong predictive utility.

      Ethics approval and consent to participate

      Not applicable.

      Consent for publication

      Not applicable.

      Availability of data and material

      Not applicable. The data for this literature review was retrieved from published studies listed in the manuscript.

      Source of funding

      This study was funded by Bristol Myers Squibb.

      Authors’ contributions

      J.A.A., P.S., M.K., and I.K. contributed to study conceptualization and design. L.L., M.P., and S.K. were responsible for the development of the methodology and data acquisition, and L.L. and S.K. performed the statistical analysis. All authors performed interpretation of data and were responsible for the writing, review, and/or revisions of the article. All authors read and approved the final article.

      Conflict of interest statement

      The authors declare the following financial interests/personal relationships that may be considered as potential competing interests: J.A.A. is an ad hoc advisor to Bristol Myers Squibb and his institution has also received research grants from Bristol Myers Squibb. P.S., M.K., and I.K. report employment by Bristol Myers Squibb. M.P. and L.L. report employment, and S.K. reports contract, by Evidinno Outcomes Research Inc., which was commissioned by Bristol Myers Squibb to conduct this study. Authors report no other conflicts of interest.

      Acknowledgments

      The authors received medical writing support for the preparation of this manuscript provided by Evidinno Outcomes Research Inc. (Vancouver, BC, Canada), funded by Bristol Myers Squibb.

      Appendix A. Supplementary data

      The following is the Supplementary data to this article:

      References

        • Bray F.
        • Ferlay J.
        • Soerjomataram I.
        • Siegel R.L.
        • Torre L.A.
        • Jemal A.
        Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries.
        Ca - Cancer J Clin. 2018; 68: 394-424
        • Lordick F.
        • Mariette C.
        • Haustermans K.
        • Obermannová R.
        • Arnold D.
        Oesophageal cancer: ESMO Clinical Practice Guidelines for diagnosis, treatment and follow-up.
        Ann Oncol. 2016; 27: v50-v57
        • Ajani J.A.
        • D'Amico T.A.
        • Bentrem D.J.
        • Chao J.
        • Corvera C.
        • Das P.
        • et al.
        Esophageal and esophagogastric junction cancers, version 2.2019, NCCN clinical practice guidelines in oncology.
        J Natl Compr Cancer Netw. 2019; 17: 855-883
        • Cartwright E.
        • Keane F.K.
        • Enzinger P.C.
        • Hong T.
        • Chau I.
        Is there a precise adjuvant therapy for esophagogastric carcinoma?.
        Am Soc Clin Oncol Educ Book. 2018; 38: 280-291
        • Kelly R.J.
        • Ajani J.A.
        • Kuzdzal J.
        • Zander T.
        • Van Cutsem E.
        • Piessen G.
        • et al.
        Adjuvant nivolumab in resected esophageal or gastroesophageal junction cancer.
        N Engl J Med. 2021; 384: 1191-1203
        • Evans S.R.
        Fundamentals of clinical trial design.
        J Exp Stroke Transl Med. 2010; 3: 19-27
        • Ronellenfitsch U.
        • Jensen K.
        • Seide S.
        • Kieser M.
        • Schwarzbach M.
        • Slanger T.E.
        • et al.
        Disease-free survival as a surrogate for overall survival in neoadjuvant trials of gastroesophageal adenocarcinoma: pooled analysis of individual patient data from randomised controlled trials.
        Eur J Cancer. 2019; 123: 101-111
        • Kataoka K.
        • Nakamura K.
        • Mizusawa J.
        • Kato K.
        • Eba J.
        • Katayama H.
        • et al.
        Surrogacy of progression-free survival (PFS) for overall survival (OS) in esophageal cancer trials with preoperative therapy: literature-based meta-analysis.
        Eur J Surg Oncol. 2017; 43: 1956-1961
        • Prentice R.L.
        Surrogate endpoints in clinical trials: definition and operational criteria.
        Stat Med. 1989; 8: 431-440
        • Daniels M.J.
        • Hughes M.D.
        Meta-analysis for the evaluation of potential surrogate markers.
        Stat Med. 1997; 16: 1965-1982
        • Xie W.
        • Halabi S.
        • Tierney J.F.
        • Sydes M.R.
        • Collette L.
        • Dignam J.J.
        • et al.
        A systematic review and recommendation for reporting of surrogate endpoint evaluation using meta-analyses.
        JNCI Cancer Spectr. 2019; 3: pkz002
        • Oba K.
        • Paoletti X.
        • Alberts S.
        • Bang Y.-J.
        • Benedetti J.
        • Bleiberg H.
        • et al.
        Disease-free survival as a surrogate for overall survival in adjuvant trials of gastric cancer: a meta-analysis.
        J Natl Cancer Inst. 2013; 105: 1600-1607
        • Guyot P.
        • Ades A.E.
        • Ouwens M.J.N.M.
        • Welton N.J.
        Enhanced secondary analysis of survival data: reconstructing the data from published Kaplan-Meier survival curves.
        BMC Med Res Methodol. 2012; 12: 9
        • Riley R.D.
        • Thompson J.R.
        • Abrams K.R.
        An alternative model for bivariate random-effects meta-analysis when the within-study correlations are unknown.
        Biostatistics. 2008; 9: 172-186
        • van Houwelingen H.C.
        • Arends L.R.
        • Stijnen T.
        Advanced methods in meta-analysis: multivariate approach and meta-regression.
        Stat Med. 2002; 21: 589-624
      1. Debray T. metamisc: diagnostic and prognostic meta-analysis, https://CRAN.R-project.org/package=metamisc (accessed 16 May 2022).

        • Zhao Y.
        • Han L.
        • Zhang W.
        • Shan L.
        • Wang Y.
        • Song P.
        • et al.
        Preoperative chemotherapy compared with postoperative adjuvant chemotherapy for squamous cell carcinoma of the thoracic oesophagus with the detection of circulating tumour cells randomized controlled trial.
        Int J Surg. 2020; 73: 1-8
        • Deng W.
        • Yang J.
        • Ni W.
        • Li C.
        • Chang X.
        • Han W.
        • et al.
        Postoperative radiotherapy in pathological T2-3N0M0 thoracic esophageal squamous cell carcinoma: interim report of a prospective, phase III, randomized controlled study.
        Oncol. 2020; 25: e701-e708
        • Armitage P.
        • Berry G.
        • Matthews J.
        Statistical methods in medical research.
        4th ed. 2001 (Malden, MA)
        • Buyse M.
        • Molenberghs G.
        • Burzykowski T.
        • Renard D.
        • Geys H.
        The validation of surrogate endpoints in meta-analyses of randomized experiments.
        Biostatistics. 2000; 1: 49-67
        • Efron B.
        Estimating the error rate of a prediction rule: improvement on cross-validation.
        J Am Stat Assoc. 1983; 78: 316-331
        • Mariette C.
        • Dahan L.
        • Mornex F.
        • Maillard E.
        • Thomas P.A.
        • Meunier B.
        • et al.
        Surgery alone versus chemoradiotherapy followed by surgery for stage I and II esophageal cancer: final analysis of randomized controlled phase III trial FFCD 9901.
        J Clin Oncol. 2014; 32: 2416-2422
        • Bujkiewicz S.
        • Achana F.
        • Papanikos T.
        • Riley R.D.
        • Abrams K.R.
        NICE DSU Technical Support Document 20: multivariate meta-analysis of summary data for combining treatment effects on correlated outcomes and evaluating surrogate endpoints.
        (available from:)
        • Burzykowski T.
        • Buyse M.
        Surrogate threshold effect: an alternative measure for meta-analytic surrogate endpoint validation.
        Pharmaceut Stat. 2006; 5: 173-186
        • Baker S.G.
        Five criteria for using a surrogate endpoint to predict treatment effect based on data from multiple previous trials.
        Stat Med. 2018; 37: 507-518
        • Al-Batran S.E.
        • Homann N.
        • Pauligk C.
        • Goetze T.O.
        • Meiler J.
        • Kasper S.
        • et al.
        Perioperative chemotherapy with fluorouracil plus leucovorin, oxaliplatin, and docetaxel versus fluorouracil or capecitabine plus cisplatin and epirubicin for locally advanced, resectable gastric or gastro-oesophageal junction adenocarcinoma (FLOT4): a randomised, phase 2/3 trial.
        Lancet. 2019; 393: 1948-1957
        • Alderson D.
        • Cunningham D.
        • Nankivell M.
        • Blazeby J.M.
        • Griffin S.M.
        • Crellin A.
        • et al.
        Neoadjuvant cisplatin and fluorouracil versus epirubicin, cisplatin, and capecitabine followed by resection in patients with oesophageal adenocarcinoma (UK MRC OE05): an open-label, randomised phase 3 trial.
        Lancet Oncol. 2017; 18: 1249-1260
        • Ando N.
        • Iizuka T.
        • Ide H.
        • Ishida K.
        • Shinoda M.
        • Nishimaki T.
        • et al.
        Surgery plus chemotherapy compared with surgery alone for localized squamous cell carcinoma of the thoracic esophagus: a Japan Clinical Oncology Group Study--JCOG9204.
        J Clin Oncol. 2003; 21: 4592-4596
        • Ando N.
        • Kato H.
        • Igaki H.
        • Shinoda M.
        • Ozawa S.
        • Shimizu H.
        • et al.
        A randomized trial comparing postoperative adjuvant chemotherapy with cisplatin and 5-fluorouracil versus preoperative chemotherapy for localized advanced squamous cell carcinoma of the thoracic esophagus (JCOG9907).
        Ann Surg Oncol. 2012; 19: 68-74
        • Barbour A.P.
        • Walpole E.T.
        • Mai G.T.
        • Barnes E.H.
        • Watson D.I.
        • Ackland S.P.
        • et al.
        Preoperative cisplatin, fluorouracil, and docetaxel with or without radiotherapy after poor early response to cisplatin and fluorouracil for resectable oesophageal adenocarcinoma (AGITG DOCTOR): results from a multicentre, randomised controlled phase II trial.
        Ann Oncol. 2020; 31: 236-245
        • Burmeister B.H.
        • Smithers B.M.
        • Gebski V.
        • Fitzgerald L.
        • Simes R.J.
        • Devitt P.
        • et al.
        Surgery alone versus chemoradiotherapy followed by surgery for resectable cancer of the oesophagus: a randomised controlled phase III trial.
        Lancet Oncol. 2005; 6: 659-668
        • Burmeister B.H.
        • Thomas J.M.
        • Burmeister E.A.
        • Walpole E.T.
        • Harvey J.A.
        • Thomson D.B.
        • et al.
        Is concurrent radiation therapy required in patients receiving preoperative chemotherapy for adenocarcinoma of the oesophagus? A randomised phase II trial.
        Eur J Cancer. 2011; 47: 354-360
        • Cunningham D.
        • Stenning S.P.
        • Smyth E.C.
        • Okines A.F.
        • Allum W.H.
        • Rowley S.
        • et al.
        Peri-operative chemotherapy with or without bevacizumab in operable oesophagogastric adenocarcinoma (UK Medical Research Council ST03): primary analysis results of a multicentre, open-label, randomised phase 2-3 trial.
        Lancet Oncol. 2017; 18: 357-370
        • Kleinberg L.R.
        • Catalano P.J.
        • Forastiere A.A.
        • Keller S.M.
        • Mitchel E.P.
        • Anne P.R.
        • et al.
        Eastern cooperative oncology group and American college of radiology imaging network randomized phase 2 trial of neoadjuvant preoperative paclitaxel/cisplatin/radiation therapy (RT) or irinotecan/cisplatin/RT in esophageal adenocarcinoma: long-term outcome and implications for trial design.
        Int J Radiat Oncol Biol Phys. 2016; 94: 738-746
        • Le Prise E.
        • Etienne P.L.
        • Meunier B.
        • Maddern G.
        • Ben Hassel M.
        • Gedouin D.
        • et al.
        A randomized study of chemotherapy, radiation therapy, and surgery versus surgery for localized squamous cell carcinoma of the esophagus.
        Cancer. 1994; 73: 1779-1784
        • Lee J.L.
        • Park S.I.
        • Kim S.B.
        • Jung H.Y.
        • Lee G.H.
        • Kim J.H.
        • et al.
        A single institutional phase III trial of preoperative chemotherapy with hyperfractionation radiotherapy plus surgery versus surgery alone for resectable esophageal squamous cell carcinoma.
        Ann Oncol. 2004; 15: 947-954
        • Lim S.H.
        • Shim Y.M.
        • Park S.H.
        • Kim H.K.
        • Choi Y.S.
        • Ahn M.J.
        • et al.
        A randomized phase II study of leucovorin/5-fluorouracil with or without oxaliplatin (LV5FU2 vs. FOLFOX) for curatively-resected, node-positive esophageal squamous cell carcinoma.
        Cancer Res Treat. 2017; 49: 816-823
        • Lv J.
        • Cao X.F.
        • Zhu B.
        • Ji L.
        • Tao L.
        • Wang D.D.
        Long-term efficacy of perioperative chemoradiotherapy on esophageal squamous cell carcinoma.
        World J Gastroenterol. 2010; 16: 1649-1654
        • Mukherjee S.
        • Hurt C.
        • Cox C.
        • Radhakrishna G.
        • Gwynne S.
        • Bateman A.R.
        • et al.
        Induction oxaliplatin capecitabine followed by switch to carboplatin-paclitaxel based RT versus continuing oxaliplatin capecitabine RT in operable esophageal adenocarcinoma: survival analysis of the randomized phase II neoscope trial.
        J Clin Oncol. 2020; 38: 373
        • Ruhstaller T.
        • Thuss-Patience P.
        • Hayoz S.
        • Schacher S.
        • Knorrenschild J.R.
        • Schnider A.
        • et al.
        Neoadjuvant chemotherapy followed by chemoradiation and surgery with and without cetuximab in patients with resectable esophageal cancer: a randomized, open-label, phase III trial (SAKK 75/08).
        Ann Oncol. 2018; 29: 1386-1393
        • Shapiro J.
        • van Lanschot J.J.B.
        • Hulshof M.
        • van Hagen P.
        • van Berge Henegouwen M.I.
        • Wijnhoven B.P.L.
        • et al.
        Neoadjuvant chemoradiotherapy plus surgery versus surgery alone for oesophageal or junctional cancer (CROSS): long-term results of a randomised controlled trial.
        Lancet Oncol. 2015; 16: 1090-1098
        • Stahl M.
        • Walz M.K.
        • Riera-Knorrenschild J.
        • Stuschke M.
        • Sandermann A.
        • Bitzer M.
        • et al.
        Preoperative chemotherapy versus chemoradiotherapy in locally advanced adenocarcinomas of the oesophagogastric junction (POET): long-term results of a controlled randomised trial.
        Eur J Cancer. 2017; 81: 183-190
        • Tepper J.
        • Krasna M.J.
        • Niedzwiecki D.
        • Hollis D.
        • Reed C.E.
        • Goldberg R.
        • et al.
        Phase III trial of trimodality therapy with cisplatin, fluorouracil, radiotherapy, and surgery compared with surgery alone for esophageal cancer: CALGB 9781.
        J Clin Oncol. 2008; 26: 1086-1092
        • Thomas A.
        • Virdee P.S.
        • Eatock M.
        • Lord S.R.
        • Falk S.
        • Anthoney D.A.
        • et al.
        Dual Erb B Inhibition in Oesophago-gastric Cancer (DEBIOC): A phase I dose escalating safety study and randomised dose expansion of AZD8931 in combination with oxaliplatin and capecitabine chemotherapy in patients with oesophagogastric adenocarcinoma.
        Eur J Cancer. 2020; 124: 131-141
        • Urba S.G.
        • Orringer M.B.
        • Turrisi A.
        • Iannettoni M.
        • Forastiere A.
        • Strawderman M.
        Randomized trial of preoperative chemoradiation versus surgery alone in patients with locoregional esophageal carcinoma.
        J Clin Oncol. 2001; 19: 305-313
        • von Döbeln G.A.
        • Klevebro F.
        • Jacobsen A.B.
        • Johannessen H.O.
        • Nielsen N.H.
        • Johnsen G.
        • et al.
        Neoadjuvant chemotherapy versus neoadjuvant chemoradiotherapy for cancer of the esophagus or gastroesophageal junction: long-term results of a randomized clinical trial.
        Dis Esophagus. 2019; 32: doy078
        • Yamasaki M.
        • Yasuda T.
        • Yano M.
        • Hirao M.
        • Kobayashi K.
        • Fujitani K.
        • et al.
        Multicenter randomized phase II study of cisplatin and fluorouracil plus docetaxel (DCF) compared with cisplatin and fluorouracil plus Adriamycin (ACF) as preoperative chemotherapy for resectable esophageal squamous cell carcinoma (OGSG1003).
        Ann Oncol. 2017; 28: 116-120
        • Zieren H.U.
        • Müller J.M.
        • Jacobi C.A.
        • Pichlmaier H.
        • Müller R.P.
        • Staar S.
        Adjuvant postoperative radiation therapy after curative resection of squamous cell carcinoma of the thoracic esophagus: a prospective randomized study.
        World J Surg. 1995; 19: 444-449
        • Baba M.
        • Natsugoe S.
        • Shimada M.
        • Nakano S.
        • Kusano C.
        • Fukumoto T.
        • et al.
        Prospective evaluation of preoperative chemotherapy in resectable squamous cell carcinoma of the thoracic esophagus.
        Dis Esophagus. 2000; 13: 136-141
        • Boonstra J.J.
        • Kok T.C.
        • Wijnhoven B.P.
        • van Heijl M.
        • van Berge Henegouwen M.I.
        • Ten Kate F.J.
        • et al.
        Chemotherapy followed by surgery versus surgery alone in patients with resectable oesophageal squamous cell carcinoma: long-term results of a randomized controlled trial.
        BMC Cancer. 2011; 11: 181
        • Bosset J.F.
        • Gignoux M.
        • Triboulet J.P.
        • Tiret E.
        • Mantion G.
        • Elias D.
        • et al.
        Chemoradiotherapy followed by surgery compared with surgery alone in squamous-cell cancer of the esophagus.
        N Engl J Med. 1997; 337: 161-167
        • Cai X.W.
        • Zeng Y.
        • Feng W.
        • Liu M.N.
        • Yu W.
        • Zhang Q.
        • et al.
        Randomized phase II trial comparing tumor bed alone with tumor bed and elective nodal postoperative radiotherapy in patients with locoregionally advanced thoracic esophageal squamous cell carcinoma.
        Dis Esophagus. 2019; 32: doz013
        • Kelsen D.P.
        • Winter K.A.
        • Gunderson L.L.
        • Mortimer J.
        • Estes N.C.
        • Haller D.G.
        • et al.
        Long-term results of RTOG trial 8911 (USA Intergroup 113): a random assignment trial comparison of chemotherapy followed by surgery compared with surgery alone for esophageal cancer.
        J Clin Oncol. 2007; 25: 3719-3725
        • Ténière P.
        • Hay J.M.
        • Fingerhut A.
        • Fagniez P.L.
        Postoperative radiation therapy does not increase survival after curative resection for squamous cell carcinoma of the middle and lower esophagus as shown by a multicenter controlled trial. French University Association for Surgical Research.
        Surg Gynecol Obstet. 1991; 173: 123-130
        • Yang H.
        • Liu H.
        • Chen Y.
        • Zhu C.
        • Fang W.
        • Yu Z.
        • et al.
        Neoadjuvant chemoradiotherapy followed by surgery versus surgery alone for locally advanced squamous cell carcinoma of the esophagus (NEOCRTEC5010): a phase III multicenter, randomized, open-label clinical trial.
        J Clin Oncol. 2018; 36: 2796-2803
        • Park S.H.
        • Sohn T.S.
        • Lee J.
        • Lim D.H.
        • Hong M.E.
        • Kim K.M.
        • et al.
        Phase III trial to compare adjuvant chemotherapy with capecitabine and cisplatin versus concurrent chemoradiotherapy in gastric cancer: final report of the adjuvant chemoradiotherapy in stomach tumors trial, including survival and subset analyses.
        J Clin Oncol. 2015; 33: 3130-3136
        • Bang Y.J.
        • Kim Y.W.
        • Yang H.K.
        • Chung H.C.
        • Park Y.K.
        • Lee K.H.
        • et al.
        Adjuvant capecitabine and oxaliplatin for gastric cancer after D2 gastrectomy (CLASSIC): a phase 3 open-label, randomised controlled trial.
        Lancet. 2012; 379: 315-321
        • Sakuramoto S.
        • Sasako M.
        • Yamaguchi T.
        • Kinoshita T.
        • Fujii M.
        • Nashimoto A.
        • et al.
        Adjuvant chemotherapy for gastric cancer with S-1, an oral fluoropyrimidine.
        N Engl J Med. 2007; 357: 1810-1820
      2. Validity of surrogate endpoints in oncology: executive summary of rapid report A10-05, version 1.1. Institute for quality and efficiency in Health care: executive summaries. Institute for Quality and Efficiency in Health Care (IQWiG), Cologne, Germany2005