Background: Surgical treatment of gastric cancer is challenging and the value of extensive lymph node dissection remains unclear. The American Joint Commision on Cancer (AJCC) guidelines were revised at 1997, requiring examination of more than 15 nodes for adequate staging. The aim of this study was to determine the impact of adequate lymph nodes assessment (ALNA) on the prognosis of the patients who underwent R0 resection.
Methods: The medical records of consecutive patients with gastric cancer treated at our institution from 1/1991 through 8/2006 were reviewed. Three hundred and forty-five patients underwent curative resection without microscopic residual tumor (R0). Survival was analyzed using the Kaplan-Meier method (mean follow-up: 35 months, 0-182 months). Prognostic factors were analyzed using univariate Log-Rank test and multivariate Cox regression test.
Results: Median patient age was 66 years. Two hundred and forty-one patients (70%) were male. The tumor locations were as follows; GE junction/proximal 190 (55%), body 66 (19%), distal 86 (25%) and diffuse 3 (1%). After AJCC guideline revised (1998-2006), the number of lymph nodes examined was significantly increased compared with the number during 1991-1997 (14 vs 8, p=0.002). The percentage of ALNA (≥15 nodes examined) was significantly improved (46% vs 15%, p<0.001). The patients who had ALNA had more positive nodes than those who did not (mean 5 vs 2. p<0.001). The overall 5-year survival rate of the patients in this cohort is 38%. Multivariate analysis demonstrated that ALNA was one of the significant prognostic factors including age >66, proximal location, T stage, N stage, lymphvascular invasion (LVI). For the patients with N1 disease (<7 LNs positive), the median survival was similar between the patients with ALNA and those without ALNA (36 months vs 27 months, p=0.28).
Conclusions: These results suggest that adequate lymph node dissection has significant impact on the patient survival, probably through adequate staging. However the therapeutic value for extensive lymph node dissection is still unclear from this study.

Prognostic factors	Univariate	Multivariate
	p	Hazard ratio	p
Age >66	0.002	1.6	0.002
Proximal lesion	0.005	1.9	<0.001
Tumor size >3.5cm	0.044		0.752
T1 (ref) 2 3 4	<0.001 <0.001 0.033	1.838 1.599 2.802	0.007 0.010 0.014 0.031
N0 (ref) 1 2 3	<0.001 <0.001 0.024	1.661 1.550 4.044	<0.001 0.007 0.033 0.001
LVI	<0.001	1.437	0.023
ALNA	0.046*	0.577	0.001

* stratified with N status