gms | German Medical Science

Background: The prognostic relevance of regional lymph node (LN) involvement in EGA is widely accepted. Various studies have examined the prognostic impact of the number (no.) of metastatic LNs, the no. of LNs removed, and LN ratio (LNR), however, - despite of increasing importance of multimodal treatment - almost exclusively on patient collectives after primary surgery (PS). Aim of this study was to explore the impact of neoadjuvant treatment (NT) on lymphatic involvement and prognosis in EGA. We thus compared patients (pts.) after PS and after NT with regard to (y)pN-category, LNR, no. of positive LNs, LNs removed, and outcome.

Materials and methods: This is a retrospective analysis of 460 pts. with cT3/4, cN any, M0 EGA (120 AEGI, 136 AEGII, 41 AEGIII, 163 stomach), who underwent curative surgery (PS: 183, NT with chemotherapy (CTx): 277, R0-resection: 373) at our institution from 01/2001-06/2015. Qualitative parameters were compared by χ2-, quantitative by Mann-Whitney-U-, and survival data by log-rank-test according to Kaplan-Meier.

Results: Pathological N-category (p<0.001) and LNR (p<0.001) significantly differed between pts. with PS and NT, lymphatic spread being more pronounced after PS (PS: pN0 23%, pN1 15%, pN2 23%, pN3 39%; LNR 0.0 24%, <0.2 25%, ≥0.2 52%; NT: ypN0 42%, ypN1 22%, ypN2 15%, ypN3 21%; LNR 0.0 42%), <0.2 31%, ≥0.2 27%).

The median no. of metastatic LNs was 2±SD6.3, pts. after PS showed more positive LNs than after NT (5±SD7.0 vs. 1±SD5.6 median, p<0.001). The no. of LNs removed was 23±SD10.7 median, and did not differ between the PS and NT group (22±SD11.2 vs. 24±SD10.4 median, p=0.181).

Histopathological non-responders to NT (NR, residual tumor >10%, n=206) showed more positive LNs (2±SD5.3 vs. 0±SD2.5 median, p<0.001), higher ypN-categories (ypN0 31% vs. 79%, ypN1 26% vs. 12%, ypN2 17% vs. 7%, ypN3 37% vs. 3%, p<0.001), and poorer LNR (LNR 0.0 31% vs. 79%, <0.2 35% vs. 18%, ≥2.0 34% vs. 3%, p<0.001) than responders (R, residual tumor <10%, n=61), but there was no difference regarding the no. of LNs removed (NR 23.5±SD10.2, R 24±SD11.0 median, p=0.889).

A lower (y)pN-category and LNR were associated with improved overall survival (OS) after PS (median OS for R0-resected pts.: pN0 106.1, pN1 37.9, pN2 35.9, pN3 18.7 months, p=0.004; LNR 0.0 106.1, <0.2 35.9, ≥0.2 24.2 months, p=0.004) and NT (median OS for R0-resected pts.: ypN0 not reached (n.r.), ypN1 39.3, ypN2 39.7, ypN3 28.0 months, p<0.001; LNR 0.0 n.r., <0.2 44.2, ≥0.2 28.0 months, p<0.001).

Multivariate analysis revealed Lauren classification and LNR as independent prognostic factors in the PS group, and cT-category, surgical complications, pT-category, and LNR as independent prognosticators in the NT group. When removing LNR from the analyses, (y)pN-category became an independent prognosticator instead in both patient collectives.

Conclusion: In our collective pts. show less LN metastases after neoadjuvant CTx, especially in case of primary tumor regression, while the no. of LNs removed - or found in the resection specimen - seems uninfluenced, even in case of histopathological response. This might hint at a response of LN metastases to neoadjuvant CTx with consecutive downstaging of the ypN-category.

Pathological N-category and LNR seem to be equally important prognostic factors for pts. with NT, same as for pts. with PS. The prognostic value of the extent of lymphadenectomy shows indirectly by the LNR being a stronger independent prognosticator than the no. of metastatic LNs (pathological N-category) alone.