1587P - Analysis of survival outcomes according to start timing of adjuvant chemotherapy in patients with gastric cancer: A retrospective nationwide cohort study

Background

Gastrectomy with D2 lymph node dissection followed by adjuvant chemotherapy (AC) has been the standard treatment for patients with stage II and III gastric cancer (GC). Especially, 1-year S-1 and 6-month capecitabine/oxaliplatin (CAPOX) are the standard chemotherapeutic regimens for AC. However, there is no established guideline for start timing of AC for GC. We aimed to investigate treatment outcomes according to start timing of AC using big-data analysis from Korean Health Insurance and Review Assessment (HIRA).

Methods

A total of 19,140 patients with GC who underwent AC with S-1 (n=10,442) or CAPOX (n=8,698) after gastrectomy were identified from January 2014 to December 2018 in the database of HIRA. Disease-free survival (DFS) and overall survival (OS) were analyzed according to the start timing of AC from surgery classified into within 6 weeks (group A, n=12,843), from 6 to 8 weeks (group B, n=5,386), and more than 8 to 16 weeks (group C, n=911).

Results

The median follow-up duration of all patients was 3.6 years. Five-year DFS and OS were significantly inferior in group B (57%, 66%) and group C (47%, 53%) compared with group A (63%, 72%) (all p<0.0001), which were consistently observed in either S-1 or CAPOX group (all p<0.005). In multivariate analysis, group B and C were independently associated with poor DFS and OS (all p<0.0001). Moreover, DFS and OS were inferior in group B (n=5386) compared with group A (n=5386) even after propensity score matching (both p<0.0001).

Conclusions

AC for GC with S-1 or CAPOX showed the best outcomes when initiated within 6 weeks following surgery. Therefore, AC for GC should be started within 6 weeks after gastrectomy if patients have fully recovered.

Clinical trial identification

Editorial acknowledgement

Legal entity responsible for the study

The authors.

Funding

Has not received any funding.

Disclosure

All authors have declared no conflicts of interest.