ADVERTISEMENT
Disease-free survival as surrogate for overall survival in neoadjuvant chemo(radio)therapy treatment of esophageal or gastro-esophageal junction carcinoma: An analysis of 4518 individual patients and 22 trials
Background
The use of disease-free survival (DFS) as a surrogate endpoint for overall survival may speed up clinical research in cancer trials. Validation of DFS as surrogate for OS in neoadjuvant treatment of oesophageal and gastro-oesophageal junction cancer is still uncertain.
Methods
We used Individual patient data from a network meta-analysis of neoadjuvant chemotherapy or chemoradiotherapy in esophageal or gastro-esophageal junction carcinoma (MANATEC-02). Eligible trials were randomized trials closed to accrual before 2015, comparing chemotherapy followed by surgery (CTS) versus surgery alone (S), chemoradiotherapy followed by surgery (CRTS) versus S or CTS versus CRTS, in first-line treatment of an oesophageal or gastro-oesophageal junction of locally advanced cancer with, adenocarcinoma or squamous cell carcinoma. The surrogate endpoint approach of individual and trial level correlation was used: at the patient level, correlation was estimated by Kendall’s tau and at trial level by the R2 determined with a two-stage copula model. Leave-one-trial out crossvalidation was applied. A multi-state Markov model was used to evaluate the time between recurrence and death in each trial.
Results
Both DFS and OS data were available for 4518 patients (22 of 30 trials of MANATEC-02 study), 2222 patients in CTS versus S comparison, 1908 patients in CRTS versus S and 388 in CTS versus CRTS. 3440 patients (76.1%) had a DFS event and 3303 (73.1%) died. Among the 3440 DFS events, 1235 corresponded to deaths. For CTS vs S, median OS was 16.6months [95% CI 15.8 – 17.4] and median DFS was 12.0 [11.4 - 12.9]. For CRTS vs S median OS was 30.1months [27.8 - 33.2] and median DFS was 21.9months [20 - 24.7]. For CTC vs RCTC, median OS was 28.6months [25.5 - 33.6] and median DFS was 21.9months [17.3 - 27.4]. Using an unadjusted Clayton copula model, Kendall’s tau was 0.73 [95% CI 0.73 - 0.74] and R2 correlation was 0.97 [0.97 – 0.97] for CTC vs S, Kendall’s tau was 0.75 [0.75 – 0.76] and R2 was 0.98 [0.98 – 0.98] for RCTC vs S, Kendall’s tau was 0.91 [0.91 – 0.91] and R2 was 0.96 [0.96 – 0.96] for CTC vs RCTC. The median sojourn time in recurrence state was shorter in older trials (started recruitment before 1995) vs more recent trials: mean sojourn time of 10.7 months [95% CI 10.1 – 11.3] vs 16.7 months [95% CI 15.4 - 17.6].
Conclusions
DFS is a validated surrogate endpoint for OS in trials studying neoadjuvant chemotherapy or chemoradiotherapy in oesophageal and gastro-oesophageal junction adenocarcinoma or squamous cell carcinoma. DFS may be more useful as trial endpoint when delays between recurrences and death become larger.
Legal entity responsible for the study
The author.
Funding
French government “Programme Hospitalier de Recherche Clinique”.
Disclosures
D. Cunningham: Advisory / Consultancy: OVIBIO on Scientific Advisory Board; Research grant / Funding (institution): AstraZeneca / MedImmune, Celgene, Bayer, 4SC, Eli Lilly, Clovis, Natera, Roche, Leap. G. Piessen: Advisory / Consultancy: BMS, Astellas, Nestlé; Travel / Accommodation / Expenses: medtronic. All other authors have declared no conflicts of interest.
