Validity and Necessity of Sub-classification of N3 in the 7th UICC TNM Stage of Gastric Cancer

Although the 6th edition of the American Joint Committee on Cancer (AJCC) TNM classification is an excellent classification system and has been extensively used for gastric cancer staging worldwide (Greene, 2002), different lymph node staging systems between the Japanese Gastric Cancer Association (JGCA) classification and the AJCC classification had made it difficult to compare treatment outcomes in an international setting (Japanese, 1998; Ikeguchi et al., 2004; Aurello et al., 2007). In Aug, 19th, 2008, the member of AJCC, UICC and JGCA held the World Cancer Congress in Buffalo, New York, which aimed to revise and unify stage system of gastric cancer. In this conference, Japanese experts participated in revision of TNM stage for the first time. After then, in Japan, the 3th edition JGCA gastric cancer treatment guidelines were published at the beginning of 2010. The N stage of this guideline first coincided with the AJCC TNM stage. Since, the stage system of gastric cancer achieved unification among AJCC, UICC and JGCA in the world wide. The 7th edition UICC TNM classification for gastric cancer has several changes from the previous edition. In particular, the classification of metastastic lymph nodes is recognized. According to this new edition TNM stage system, N stage was categorized to N0 (no regional lymph


Introduction
Although the 6 th edition of the American Joint Committee on Cancer (AJCC) TNM classification is an excellent classification system and has been extensively used for gastric cancer staging worldwide (Greene, 2002), different lymph node staging systems between the Japanese Gastric Cancer Association (JGCA) classification and the AJCC classification had made it difficult to compare treatment outcomes in an international setting (Japanese, 1998;Ikeguchi et al., 2004;Aurello et al., 2007). In Aug, 19 th , 2008, the member of AJCC, UICC and JGCA held the World Cancer Congress in Buffalo, New York, which aimed to revise and unify stage system of gastric cancer. In this conference, Japanese experts participated in revision of TNM stage for the first time. After then, in Japan, the 3th edition JGCA gastric cancer treatment guidelines were published at the beginning of 2010. The N stage of this guideline first coincided with the AJCC TNM stage. Since, the stage system of gastric cancer achieved unification among AJCC, UICC and JGCA in the world wide.
The 7 th edition UICC TNM classification for gastric cancer has several changes from the previous edition. In particular, the classification of metastastic lymph nodes is recognized. According to this new edition TNM stage system, N stage was categorized to N0 (no regional lymph

Validity and Necessity of Sub-classification of N3 in the 7 th UICC TNM Stage of Gastric Cancer
Fang-Xuan Li 1 , Ru-Peng Zhang 2 *, Han Liang 2 , Ji-Chuan Quan 2 , Hui Liu 2 , Hui Zhang 2 node metastasis), N1 (metastasis in 1-2 regional lymph nodes), N2 (metastasis in 3-6 regional lymph nodes), and N3 (metastasis in 7 or more regional lymph nodes). N3 stage included two sub-categorization N3a (7-15 regional lymph nodes metastasis) and N3b (>15 regional lymph nodes metastasis). However, in 7th edition UICC TNM stage, N3a and N3b had not applied in TNM stage group. T1N3M0 was classified in the new IIB category, T2N3M0 was classified in IIIA stage, T3N3M0 was classified in IIIB stage, and T4a-T4bN3M0 was classified in IIIC stage. The absence of N3a and N3b in these TNM stage group is tempting to wonder whether the sub-categorization of N3 is necessary. If it is necessary, we have to suspect if the new TNM stage can predict patients' prognosis accurately. The objective of the current study was to evaluate the necessity and clinical validity of the N3 sub-classification in the 7th TNM stage system based on cumulated single institution data from a China center.

Data collection
Eligibility criteria were: (1) Patients underwent curative resection for gastric cancer between January 2001 and May 2006 at the Tianjin Medical University Cancer Institute and Hospital. (2) Patients with gastric adenocarcinoma identified with histopathologic examination, the postoperative pathological results demonstrated the margin was negative. (3) Patients without a history of prior malignancy, distant metastasis (such as liver, lung, brain, or bone-marrow metastasis), and peritoneal dissemination, (4) patients who underwent potentially curative gastrectomy plus lymphadenectomy (limited or extended), and (5) patients with no fewer than 15 dissected lymph nodes.
Follow-up plan: postoperative follow-up included clinical and laboratory examinations every 3 months for the first year, every 6 months for the second year, and annually thereafter at least until 5 years after operation or the date when the patient was dead.

Patients
Therefore, a total of 302 gastric cancer patients at pN3M0 enrolled in our study. The patients were composed of 203 males and 99 females, the radio of male to female was 2.05:1; Mean age was 59.6 years, the median age was 59.0 years. The most common site was the lower third of the stomach (127 Cases), followed by the middle third (78 cases),and 77 cases was upper third of gastric cancer and 20 cases diffused the whole stomach. Histologically, the undifferentiated type (230 cases) was more common than the differentiated type (72 cases). Concerning the depth of invasion, the number of cases with T2, T3, T4a and T4b stage was 15, 19, 159, and 108, respectively. 166 patients were at N3a stage, and 136 patients were at N3b stage.
Proximal, distal and total gastrectomy were performed in 51, 121 and 130 patient respectively. 227 Patients underwent D2 lymphadenectomy, 75 patients underwent D2+ and D3 lymphadenectomy. Of these patients, the mean and median rate of lymph node metastasis (Metastatic lymph nodes by H&E staining to the total number of dissected lymph nodes) were 66.0% and 66.6%, it ranged from 15.1%-100%. It was between 15.1% and 30% in 15 patients, between 30% and 60% in 101 patients, and 186 patients were ≥60%. The histological negative lymph node number ranged from 0 to 64. The mean and median negative lymph node number was 8.8 and 7.0. Negative lymph node number was ≤6 in 144 patients, between 7 and 15 in 104 patients, 54 patients had ≥16 negative lymph nodes. 70 patients had extranodal metastasis. 43 patients received combined devisceration. 176 patients received adjuvant chemotherapy postoperatively.

Methods
For statistical analysis, continuous variables are presented as mean (±standard deviation). Survival curves and univariate analysis were calculated according to the Kaplan-Meier method, the log-rank test was used to evaluate statistically significant differences between two groups. Cox regression analysis was used in multivariate analysis of prognostic factors. A value of P<0.050 (two-sided) was regarded as statistically significant. All analyses were performed using SPSS version 17.0. Table 1 showed the detailed classifications based on the 6 th and 7 th editions TNM stage system. The major revisions in the 7 th edition TNM classification was that the definition of N3 and T4 was altered. Newly added IIIC stage was another change in 7 th TNM stage.

Univariate analysis of prognostic factors
With the univariate analysis, 7 factors were found      (Table 2). They were: location of tumor, depth of invasion, gastric resection, combined organs resection, lymph node metastasis, rate of metastatic lymph nodes (Figure 1), negative lymph node.

Sub-classification of N3 stage
According to 7 th TNM stage, the pN3M0 gastric patients could be staged to T2N3aM0, T2N3bM0, T3N3aM0, T3N3bM0, T4aN3aM0, T4aN3bM0, T4bN3aM0, and T4bN3bM0. Because of small sample of T2 and T3 gastric cancer, T2N3M0 and T3N3M0 can not divide significantly according to sub-stage of N3 (Table 3). Patients at T2N3M0 stage had a better prognosis than other sub-classification. T3N3M0 and T4aN3aM0 patients had equal prognosis which followed the T2N3M0. T4aN3bM0 and T4bN3aM0 had lower survival rate than the formers. T4bN3bM0 had worst prognosis (Figure 2, 3).

Multivariate analysis of prognostic factors
All the aforementioned 8 variables (including our suggested Sub-classification of N3 stage) were included in a multivariate Cox proportional hazards model to adjust for the effects of covariates, ( Table 4). Rate of metastatic lymph nodes and our suggested TNM stage were independent prognostic factors for pN3M0 gastric caners (Figure 1, Figure 3).

Discussion
The rational staging system plays a crucial role in diagnosis and management of gastric cancer. It can aid the clinician in the planning of treatment, give some indication of prognosis, assist in the evaluation of the results of treatment, and facilitate the exchange of information. The TNM system is accepted as a chief standard for the staging system of gastric cancer (Sobin, 2001). The revision of the 7 th TNM stage bases on databases of different regions in the world, involving countries in Asian such as Korea and Japan. The 7 th TNM stage is the first authoritative standard for evaluation the effectiveness of treatment of gastric cancer in the world wide. It can promote clinical research of gastric cancer all over the world.
The 7 th edition UICC TNM classification for gastric cancer has several changes from the previous edition. The revision of the pN classification for the 7 th TNM classification will be more conspicuous and needs to be fully discussed Chae et al., 2011;Jung et al., 2011). An ideal lymph node staging system for gastric cancer has been controversial and has changed whenever a revision in the TNM classification has been made (Klein et al., 2001;Yamashita et al., 2008;Jung et al., 2011). Furthermore, various factors such as different surgical techniques and pathological assessments have made standardized lymph node staging difficult. To overcome these problems, some investigators have proposed the metastatic lymph node rate (Katai et al., 2004;Marchet et al., 2007;Xu, 2010;Lemmens et al., 2011). In line with previous studies, it was showed that higher rate of lymph node metastasis was related to poor prognosis. The rate of lymph node metastasis has been identified as the most intensive indicator of gastric cancer after surgery, but it cannot avoid the bias of lymph node classification originating from the varied number of dissected lymph nodes.  suggested that prognostic prediction of metastastic lymph nodes rate could be improved by associating prognosis with negative lymph nodes. By increasing the negative lymph node counts, the chance of micrometastasis was remaining decrease (Kojima et al., 2008;). It has been demonstrated that the micrometastasis were closely associated with recurrence and poor prognosis . In our studying, the less negative lymph node counts were also significant related with poor prognosis.
However, one of the drawbacks of both the N-rate and negative lymph nodes is that there are no standardized categories in literature. The role of staging according to N-rate and negative lymph nodes is therefore still not clear and should be further investigated.
Another concern for the 7 th edition pN classification is the pN3 subgroup. Although the pN3 classification was comprised of pN3a and pN3b, both subgroups were not individual determinants of the final TNM stage. It seems that there should be a compensation for this problem. There was a significant difference in the 5-year survival rate between the 7 th N3a and N3b (Ichikura et al., 1999;Chae et al., 2011). As several data indicated there was a significant difference in survivals between N2 and N3 in the 6 th N classification (Ichikura et al., 1999;). Therefore, it may be reasonable to subdivide gastric cancer TNM stage group according to the sub-stages of N3. We found that patients at T2N3M0 stage had a better prognosis than other subclassification. T3N3M0 and T4aN3aM0 patients had equal prognosis which followed the T2N3M0. T4aN3bM0 and T4bN3aM0 had lower survival rate than the formers. T4bN3bM0 had worst prognosis. Usually, advanced gastric cancer is relatively more common in china. Only 1%-4% patients with T1 and T2 gastric cancer had >15 metastastic lymph nodes (Ahn et al., 2009;Jung et al., 2011). Therefore, T1-T2N3M0 gastric cancer has a lower incidence in China, as did our specimen. Because of the narrow definition, T3 gastric cancer also has a lower incidence (Chae et al., 2011;Qiu et al., 2011;Wang et al., 2011). So a multi-center observational study is necessary in exploration of N3 sub-stage in these patients. However, in T4a and T4b gastric cancer, significant differences of prognosis can be seen between N3a and N3b sub-stage. Especially, patients at IIIC stage had significantly varied prognosis according to sub-classification of T4a, T4b and N3a, N3b. In particular, detailed staging of IIIC may be more demanded. In our opinion, the sub-stage of N3 may be useful for a more accurate prediction of patient survival and selection of therapeutic strategies. It should be applied in TNM stage group.