ADVERTISEMENT
Individual participant data network meta-analysis (IPD-NMA) of neoadjuvant chemotherapy or chemoradiotherapy in esophageal or gastro-esophageal junction carcinoma
Background
The optimal neoadjuvant treatment for resectable locally advanced carcinoma of the thoracic esophagus (TE) or gastroesophageal junction (GEJ) remains a matter of debate. Moreover, available randomized controlled trials (RCT) were not powered to study the effect according to histology (squamous cell carcinoma (SCC) or adenocarcinoma (AC)) or tumor location. The aim of this IPD-NMA was to identify the best neoadjuvant treatment with a focus on prespecified subgroups.
Methods
All, published or unpublished, RCT, closed to accrual before 2015 and having compared at least two of the following strategy were eligible: upfront surgery (S), chemotherapy followed by surgery (CS), chemoradiotherapy followed by surgery (CRS). RCT were identified by electronic database, conference proceedings and trial registries. All analyses were conducted on IPD obtained from investigators. The primary endpoint was overall survival (OS). Secondary endpoints included disease-free survival (DFS) and pattern of failure (local recurrence, distant recurrence and death without relapse). The IPD-NMA was analyzed by a one-step mixed effect Cox model. The NMA was registered in PROSPERO (CRD42018107158).
Results
Out of 33 trials identified (5807 patients), IPD were obtained for 26 RCTs (4985 pts, 3723 events): 12 comparisons (2478 pts, 2010 events) for CS vs S, 12 (2222 pts, 1545 events) for CRS vs S, and 4 (497 pts, 363 events) for CRS vs CS. The models were adjusted for age, gender, tumor location and histology subtypes. CS and CRS led to increased OS when compared to S with HR=0.86[0.75-0.99], p=0.03 and HR=0.77[0.68-0.87], p < 0.001 respectively. The NMA comparison of CRS vs CS for OS gave a HR of 0.90[0.74-1.09], p=0.27, fairly consistent in direct and indirect comparisons (p for inconsistency = 0.19) with some evidence of heterogeneity (p=0.0045). For CS vs S, a more pronounced effect on OS was observed for tumor located at the GEJ (HR= 0.68[0.54-0.86]) vs the TE (0.89[0.77-1.03]) (p=0.029). There were less evidences for an interaction with histology (p=0.21) or gender (p=0.62). For the CRS vs S, a larger effect was observed for women (HR=0.56[0.44-0.71]) than man (HR=0.83[0.74-0.93]) (p=0.003); there were less evidence for an interaction with histology (p=0.48) or location (p=0.14). For the CRS vs CS, a larger effect on OS was observed for women (HR=0.62[0.44-0.88]) than man (HR=0.99[0.81-1.20]) (p=0.011), with less evidence for an interaction with histology (p=0.67) or location (p=0.7). CS and CRS led to better DFS than S with HR=0.79[0.69-0.91], p=0.001 and HR=0.78[0.68-0.88], p < 0.001 respectively. The NMA comparison of CRS vs CS provided a HR=0.98[0.80-1.21], p=0.85. Both CS and CRS lead to lower rates of local and distant recurrences when compared to S. When compared to CS, CRS did not significantly lower the risk of local (subHR=0.88 [0.66-1.18], p=0.41) or distant (subHR=0.90 [0.73-1·.11], p=0.33) relapse.
Conclusions
In this IPD-NMA, neoadjuvant chemotherapy and chemoradiotherapy were consistently better than surgery alone across histology subtypes, but with some variation in the magnitude of treatment effect across gender and tumor location. A strong OS or DFS difference between the two combined modalities (CS vs CRS) was not identified.
Legal entity responsible for the study
The author.
Funding
French Governement via: “Programme Hospitalier de Recherche Clinique”.
Disclosures
G. Piessen: Advisory/Consultancy: BMS, Astellas, Nestlé; Travel/Accommodation / Expenses: medtronic. All other authors have declared no conflicts of interest.
