Abstract
The main change in the 5th edition (1997) of the TNM classification for gastric cancer compared to the 4th edition (1987) is the use of the number of involved nodes instead of the location of positive nodes. As a result stage grouping is also altered. A second change is the requirement for the examination of at least 15 nodes to justify the N0 status. Patients with fewer examined negative nodes are unclassifiable (Nx). Data were retrieved from a randomized trial database comparing D1 and D2 dissection and 633 curatively operated patients were included. According to the criteria of the 5th edition, 39% of the node-positive patients had another N stage compared to the 4th: 21% had a lower and 18% had a higher stage. 5-year survival rates according to the 4th edition N0, N1 and N2 groups were respectively 72%, 34% and 27%. According to the 5th edition these percentages were for the N0, N1, N2, N3 and Nx groups respectively 75%, 38%, 19%, 8% and 65%. The former 1987 N1 and N2 group were significantly split into three new N 1997 groups (P = 0.006, respectively P< 0.0005). The Cox's regression analysis showed the N 1997 classification to be the most important prognostic variable, with a higher prognostic value than N 1987. In addition, the new TNM stage was also a better prognosticator. The requirement for examining at least 15 nodes, however, could not be fulfilled in 38% of all node-negative patients and we found that a minimum of 5 consecutive negative lymph nodes is a reliable number for staging purposes. We conclude that the 5th edition of the TNM classification provides a better estimation of prognosis, however, examination of at least 15 negative regional lymph nodes is too high a threshold and 5 gives similar prognostic value.
