Validation of the seventh edition of the American Joint Committee on Cancer TNM staging system for gastric cancer

Authors

  • Hun Jung MD,

    1. Department of Surgery, Seoul St. Mary's Hospital, College of Medicine, The Catholic University of Korea, Seoul, Republic of Korea
    Search for more papers by this author
  • Han Hong Lee MD,

    1. Department of Surgery, Seoul St. Mary's Hospital, College of Medicine, The Catholic University of Korea, Seoul, Republic of Korea
    Search for more papers by this author
  • Kyo Young Song MD, PhD,

    1. Department of Surgery, Seoul St. Mary's Hospital, College of Medicine, The Catholic University of Korea, Seoul, Republic of Korea
    Search for more papers by this author
  • Hae Myung Jeon MD, PhD,

    1. Department of Surgery, Seoul St. Mary's Hospital, College of Medicine, The Catholic University of Korea, Seoul, Republic of Korea
    Search for more papers by this author
  • Cho Hyun Park MD, PhD

    Corresponding author
    1. Department of Surgery, Seoul St. Mary's Hospital, College of Medicine, The Catholic University of Korea, Seoul, Republic of Korea
    • Department of Surgery, Seoul St. Mary's Hospital, The Catholic University of Korea, 505 Banpo-dong, Seocho-gu, 137-701, Seoul, Republic of Korea
    Search for more papers by this author
    • Fax: (011) 82-2-590-1406


Abstract

BACKGROUND:

The seventh edition of the American Joint Committee on Cancer (AJCC) TNM classification for gastric cancer was published in 2010 and included major revisions. The aim of the current study was to evaluate the validity of the seventh edition TNM classification for gastric cancer based on an Asian population.

METHODS:

A total of 2916 gastric cancer patients who underwent R0 surgical resection from 1989 through 2008 in a single institute were included, and were analyzed according to the seventh edition of the TNM classification for validation.

RESULTS:

When adjusted using the seventh edition of the TNM classification, upstaging was observed in 771 patients (26.4%) and downstaging was observed in 178 patients (6.1%) compared with the sixth edition of the TNM classification. The relative risk (RR) of seventh edition pT classification was found to be increased with regular intensity compared with the sixth edition pT classification. The RR of seventh edition pN classification was found to be increased with irregular intensity compared with the sixth edition pN classification. In survival analysis, there were significant differences noted for each stage of disease, but only a marginal difference was demonstrated between stage IA and stage IB (P = .049). In the hybrid TNM classification, which combines the seventh edition pT classification and the sixth edition pN classification, both pT and pN classifications demonstrated a more ideal distribution of the RR, and 5-year survival rates also showed a significant difference for each stage (P<.01).

CONCLUSIONS:

The seventh edition of the TNM classification was considered valid based on the results of the current study. However, the hybrid TNM classification, comprised of a combination of the seventh edition pT classification and sixth edition pN classification, should be considered for the next edition. Cancer 2011. © 2011 American Cancer Society.

The TNM classification of gastric cancer is the most important independent prognostic factor and is considered to play a fundamental role in treatment. Although the sixth edition of the American Joint Committee on Cancer (AJCC) TNM classification1 is an excellent classification system and has been extensively used for gastric cancer staging worldwide, several problems have been associated with its use. First, different lymph node staging systems between the Japanese Gastric Cancer Association (JGCA) classification2 and the AJCC classification have made it difficult to compare treatment outcomes in an international setting. Anatomical lymph node staging as proposed by the JGCA has been widely used in Asian centers, but recent reports have shown that the numeric lymph node staging of the AJCC TNM classification demonstrated a greater prognostic prediction as well as higher reproducibility compared with the JGCA classification.3, 4 However, limitations such as lack of information regarding the extent of lymph node dissection and difficulty in its application to surgical treatment planning have been demonstrated as disadvantages of a numeric classification system.5-7 Moreover, to the best of our knowledge, there has been no consensus regarding the adequate grouping of lymph node staging. Another current issue in gastric cancer staging concerns the classification of depth of invasion, especially for the pT2 classification. Several investigators have proposed that the sixth edition pT2 classification should be subclassified as muscularis propria and subserosa cancer for better prediction of prognosis.8-10 In this context, the seventh edition of the AJCC TNM classification for gastric cancer was published in 2010 and included major revisions of the pT and pN classifications.11

There are additionally several differences with regard to pathophysiology and treatment pattern among Asian and Western gastric cancer patients, and these differences are associated with inconvenient application of staging systems.

The objective of the current study was to evaluate the validity and clinical utility of the seventh edition of the AJCC TNM classification based on cumulated single-institution data from an Asian center.

MATERIALS AND METHODS

Patients

This study was designed as a retrospective analysis using prospectively collected data. A total of 4793 patients underwent surgery for gastric cancer at a single institution between 1989 and 2008. All patients provided written informed consent. Data were collected from the Gastric Cancer Patient Registry of Seoul St. Mary's Hospital. The exclusion criteria for analysis were 1) synchronous malignancies, 2) palliative surgery, 3) a noncurative resection, 4) <15 retrieved lymph nodes, 5) a definitive M1 classification based on the sixth edition of the TNM classification, and 6) inaccurate medical records.

A total of 2916 patients were included in the current study and patient data were re-evaluated using the seventh edition of the AJCC TNM classification. The number of harvested lymph nodes ranged from 15 to 81, with a median value of 38 lymph nodes. Adjuvant chemotherapy was performed in 1258 patients who were pathologically confirmed as having serosal penetration by cancer cells or metastatic finding in lymph nodes. Adjuvant chemotherapy was usually performed with 5-fluorouracil-based or cisplatin-based systemic therapy. No patients were receiving neoadjuvant chemotherapy. A total of 2224 patients who underwent surgery from 1989 through 2005 were included for long-term survival analysis, and the median follow-up time was 72 months.

This study was approved by the Institutional Review Board of the Ethical Committee of the College of Medicine at the Catholic University of Korea (KC10RISE0255).

Definition of the Seventh Edition TNM Classification

For the seventh edition pT classification, the definition of T classification was not changed, only the final staging assignment of the pT classification was changed. For the seventh edition pN classification, there were considerable changes compared with the sixth edition. The sixth edition pN1 classification was separated into pN1 and pN2 classifications in the seventh edition, and the sixth edition pN2 and pN3 classifications were reclassified as pN3a and pN3b, respectively, in the seventh edition (Table 1).

Table 1. Definitions and Patient Distributions in the Sixth and Seventh Editions of the AJCC TNM Staging Systems
Sixth Edition Primary Tumor (T)Seventh Edition Primary Tumor (T)
T ClassificationDefinitionNo. of Patients (%)T ClassificationDefinitionNo. of Patients (%)
  1. AJCC indicates American Joint Committee on Cancer.

T1Mucosa or submucosa1468 (50.3)T1aMucosa901 (30.9)
T2aMuscularis propria311 (10.7)T1bSubmucosa567 (19.4)
T2bSubserosa608 (20.9)T2Muscularis propria311 (10.7)
T3Penetrates serosa517 (17.7)T3Subserosa608 (20.9)
T4Invasion of adjacent structure12 (0.4)T4aPenetrates serosa517 (17.7)
   T4bInvasion of adjacent structure12 (0.4)
Sixth Edition Regional Lymph Nodes (N)Seventh Edition Regional Lymph Nodes (N)
N ClassificationDefinitionNo. of Patients (%)N ClassificationDefinitionNo. of Patients (%)
001733 (59.4)001733 (59.4)
11-6710 (24.3)11-2387 (13.3)
27-15302 (10.4)23-6323 (11.1)
3>15171 (5.9)3a7-15302 (10.4)
   3b>15171 (5.9)

Table 2 shows the detailed classifications based on the sixth and seventh editions of the TNM classification. Among the major revisions in the seventh edition TNM classification was that the definition of stage IV disease was changed, and was only defined for patients with a definitive distant metastasis (M1 classification).

Table 2. Comparison of Patient Distribution in the Sixth and Seventh Editions and the Proposal Hybrid TNM Staging Systems
inline image

Statistical Analysis

Survival analysis was performed using the Kaplan-Meier method with the log-rank test for univariate analysis, and multivariate analysis for survival was performed using the Cox proportional hazards model with the “forward: Wald” method. Statistical analysis was performed using SPSS software (version 13.0; SPSS Inc, Chicago, Ill) and a P value <.05 was considered to indicate a statistically significant difference.

RESULTS

Change of Stage Distribution

Table 1 shows a comparison of the classification and distribution of patients using the sixth and seventh edition pT and pN classifications. For the seventh edition pT classification, the pT1a group (n = 901; 30.9%) presented the largest number of patients. The number of patients classified as pN3 according to the seventh edition (n = 473) was markedly increased compared with patients classified according to the sixth edition (16.3% vs 5.9%) and patients were reclassified as having pN3a (n = 302; 10.4%) and pN3b (n = 171; 5.9%) disease, respectively. When adjusted with the seventh edition TNM classification, upstaging was observed in 771 patients (26.4%), and downstaging was observed in 178 patients (6.1%) (Table 3).

Table 3. Change of Stage According to Each TNM Staging System
The Seventh Edition TNM Staging System Compared With the Sixth Edition TNM Staging System
DownstagingNo ChangeUpstaging
Change of StageNo. of PatientsChange of StageNo. of PatientsChange of StageNo. of Patients
IV→IIB7IA→IA1216IB→IIA328
IV→IIIA5IB→IB310II→IIIA101
IV→IIIB51II→IIA62IIIA→IIIB190
IV→IIIC115II→IIB283IIIB→IIIC152
  IIIA→IIIA96  
Total178 (6.1%) 1967 (67.5%) 771 (26.4%)
The Proposal Hybrid Staging System Compared With the Seventh Edition TNM Staging System
DownstagingNo ChangeUpstaging
Change of StageNo. of PatientsChange of StageNo. of PatientsChange of StageNo. of Patients
IIA→IB69IA→IA1216  
IIB→IIA81IB→IB310  
IIIA→IIB128IIA→IIA321  
IIIB→IIIA191IIB→IIB209  
IIIC→IIIB155IIIA→IIIA73  
  IIIB→IIIB51  
  IIIC→IIIC112  
Total624 (21.4%) 2292 (78.6%) 0 (0.0%)

Univariate Analysis of Survival Rate

The 5-year survival rates based on the seventh edition of the TNM classification were as follows (Fig. 1): IA: 98.9%; IB: 96.5%; IIA: 87.6%; IIB: 71.3%; IIIA: 65.4%; IIIB: 49.5%; and IIIC: 32.2%. There were significant differences noted among the newly established disease stages, but a marginal difference was observed between stages IA and IB (P = .049). The homogeneity of the survival rate in the same stage of the seventh edition of the TNM classification was statistically significant (data not shown). The 5-year survival rates of each pT classification based on the seventh edition were as follows: pT1a: 99.5%; pT1b: 93.6%; pT2: 89.0%; pT3: 63.9%; pT4a: 45.6%; and pT4b: 36.0%. There was a significant difference noted between pT1a and pT1b (P< .001), pT1b and pT2 (P = .002), pT2 and pT3 (P< .001), and pT3 and pT4a (P< .001). The survival difference between the pT4a and pT4b classifications was not observed because of the small number of patients in the pT4b group. The 5-year survival rates based on the seventh edition pN classification were as follows: pN0: 95.9%; pN1: 82.1%; pN2: 68.4%; pN3a: 53.4%; and pN3b: 24.2%. There were also significant differences noted among the subgroups of pN classifications (P<.001) (Table 4).

Figure 1.

(a) The 5-year survival rate according to the sixth edition of the TNM classification is shown. (b) The 5-year survival rate according to the seventh edition of the TNM classification is shown. (c) The 5-year survival rate according to the proposal hybrid TNM classification is shown.

Table 4. Results of Univariate Survival Analysis According to the Sixth and Seventh Editions and the Proposal Hybrid TNM Staging Systems
inline image

Multivariate Analysis of Survival Rate

Table 5 shows the results of multivariate survival analysis using the sixth and seventh editions of the TNM classification. The relative risk (RR) of the seventh edition pT classification increased with regular intensity along with the increment of pT classification compared with the sixth edition pT classification. The RR of each seventh edition pT classification was found to be 2.89 in pT2, 4.88 in pT3, 7.58 in pT4a and 9.45 in pT4b.

Table 5. Results of Multivariate Analysisa for Prognostic Factors According to Respective TNM Staging Systems
 PRelative Risk95% CI
  • 95% CI indicates 95% confidence interval.

  • a

    Multivariate analysis was performed using the Cox proportional hazards model using the “forward:Wald” method.

6th edition TNM staging systemT classificationT11
T2(a/b)<.0014.1582.932–4.208
T3<.0017.3765.182–10.498
T4<.0018.3853.719–18.904
N classificationN01
N1<.0013.0642.232–4.208
N2<.0015.3053.784–7.439
N3<.00110.3067.273–14.604
7th edition TNM staging systemT classificationT1(a/b)1
T2<.0012.8861.819–4.578
T3<.0014.8783.381–7.038
T4a<.0017.5765.309–10.813
T4b<.0019.4544.196–21.300
N classificationN01
N1<.0012.3321.602–3.394
N2<.0013.4382.415–4.893
N3(a/b)<.0016.4084.645–8.840
Proposal hybrid TNM staging systemT classificationT1(a/b)1
T2<.0012.9091.834–4.615
T3<.0014.8673.374–7.022
T4a<.0017.6805.384–10.954
T4b<.0018.7433.874–19.733
N classificationN01
N1<.0012.9052.108–4.002
N2<.0014.9423.509–6.960
N3<.0019.4296.618–13.435

The power of discrimination of the seventh edition pN classification was relatively weakened compared with that of the sixth edition. The RR of each seventh edition pN classification was 2.33 for pN1, 3.44 for pN2, and 6.41 for pN3(a/b) compared with pN0. In contrast, the RR of each sixth edition pN classification was found to increase with regular intensity along with the increment of pN classification. The RR of each sixth edition pN classification was 3.06 for pN1, 5.31 for pN2, and 10.31 for pN3 compared with pN0.

Proposal Hybrid TNM Staging System

We developed the proposal hybrid staging system consisted in the combination of the seventh edition pT and the sixth edition pN classifications (Table 2). Compared with the seventh edition of the TNM classification, downstaging was observed in 624 patients (21.4%) and there was no upstaging noted (Table 3). The 5-year survival rates of the proposal hybrid TNM classification were as follows: stage IA: 98.9%; stage IB: 94.7%; stage IIA: 84.9%; stage IIB: 68.3%; stage IIIA: 56.2%; stage IIIB: 41.1% and stage IIIC: 17.0%. There were significant differences noted among the survival rate of each stage (P<.05) (Table 4) (Fig. 1). The homogeneity of the survival rate in the same stage categories was also statistically significant (data not shown). Both the pT and pN classifications demonstrated similar intensity of RR with more even distribution. The RR with respect to pT1 was 2.91 for pT2, 4.87 for pT3, 7.68 for pT4a and 8.74 for pT4b. The RR with respect to pN0 was 2.91 for pN1, 4.94 for pN2 and 9.43 for pN3 (Table 5).

DISCUSSION

One of the interesting changes in the seventh edition of the AJCC TNM staging system in gastric cancer is the revision of early stage cancer. With the increase in the number of patients diagnosed with early gastric cancer in Korea, especially, the possibility of limited surgery and personalized follow-up programs for these patients have been clinically important concerns. These well-classified TNM stages would be very useful for the prediction of prognosis, but universal application could be difficult in certain circumstances. The convenient and universal TNM staging system would be clinically necessary because there are several pathophysiologic differences between Western and Asian patients with gastric cancer. That is the reason why this validation study was conducted, and the buffered hybrid TNM system is proposed based on single-institution data.

Indeed, a definite validation is difficult based on the data of single institution in Korea because there are many different aspect between Asian and Western patients with gastric cancer. It has been reported that Korean patients had earlier stage cancer with a higher number of positive lymph nodes identified, and the probability of death from gastric cancer at 5 years by stage of disease was significantly lower in all stages compared with US patients. With regard to treatment pattern, adjuvant treatment was offered to 45% of Korean patients compared with 9% of US patients, but 30.7% of US patients received neoadjuvant chemotherapy compared with Korean patients.12, 13 Neoadjuvant chemotherapy is not generally recommended for Korean patients with gastric cancer.14, 15

One of the major revisions in the seventh edition of the AJCC TNM classification is alteration of the pT2 classification. Although the sixth edition pT2 classification was subclassified as pT2a and pT2b, each subgroup is not an individual determinant of final TNM staging. However, modifying the pT2 classification has been proposed because the prognosis of patients with pT2a and pT2b disease was found to differ with statistical significance.8-10 In the current study, the 5-year survival rate of patients with the sixth edition pT2a classification was also significantly better compared with that of patients with pT2b disease (89.0% vs 63.9%; P<.05). Therefore, it appears reasonable to revise the six edition pT2(a/b) to different pT classification. Another revision of the seventh edition pT classification is that pT1 tumors were subclassified into T1a and T1b, and it is pertinent to consider consensus for the treatment modalities such as endoscopic resection in early gastric cancer patients. The seventh edition pT classification appears superior when compared with the sixth edition as a prognostic factor because the RR of the seventh edition pT classification increased with regular intensity according to the increment of pT classification compared with that of the sixth edition.

The revision of the pN classification for the seventh edition of the TNM classification will be more conspicuous and needs to be fully discussed. 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.4, 16-18 Furthermore, various factors such as the different lymph node classification systems used in Asian and Western centers, 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 ratio.19-21 The seventh edition of the TNM classification also adopted numeric classification for pN classification, and it was divided into 5 subgroups. In this study, there were significant differences noted in the 5-year survival rates for each pN classification of the seventh edition, but the RR interval between pN1 and pN2 was inadequately narrow (the RR with respect to pN0 was 2.33 for pN1 and 3.44 for pN2). This result most likely is associated with a division of pN1 and pN2 classifications with a narrow range: pN1 for 1 to 2 positive lymph nodes and pN2 for 3 to 6 positive lymph nodes. Consequently, the unwarranted stage migration would be provoked by a low power of discrimination between pN1 and pN2 disease. Another concern for the seventh 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. As a result, the RR of the seventh edition pN3 classification with respect to pN0 classification was as low as 6.41. However, there has been enough proof that the prognosis of patients with >15 positive lymph nodes is significantly worse than that of patients with 7 to 15 positive lymph nodes.17, 22, 23 In contrast, the RR of the sixth edition pN classification increased with regular intensity according to the increment of the pN classification, and was even comparable to the degree of the RR of the seventh edition pT classification.

The seventh edition of the TNM classification demonstrated significant differences in the survival rate for each stage of disease despite a more complicated structure in comparison with the sixth edition staging system. There were a few considerable arguments. First, the 5-year survival rate of most patients with stage IA to stage IIIA disease is >50%, and stage IIIB patients even demonstrate a 5-year survival rate of 49.5%. Second, the marginal difference between the seventh edition stage IA and stage IB was demonstrated, largely because of the favorable prognosis of patients with pT2N0 disease who demonstrated a similar 5-year survival rate compared with those of pT1N0 disease (data not shown). Another reason is that pT1N1 group of stage IB is defined only 1 to 2 metastatic lymph nodes patient presented more favorable survival compare to the sixth edition pN1 group. Third, the RR of the seventh edition pN classification had relatively increased in an irregular linear pattern according to the increase in pN disease, indicating that the power of discrimination was relatively weakened compared with the sixth edition pN classification.

To overcome these drawbacks, modification of the seventh edition of the TNM classification are suggested based on the findings of the current study. It is interesting to note that with a proposal hybrid TNM staging system using the pT classification of the seventh edition and the pN classification of the sixth edition, an even RR interval was demonstrated, with a more significant stratification noted. According to the proposal hybrid system, the 5-year survival rate for each stage of disease is classified with an equally significant difference with homogeneity in the same stage. The RR of the pT classification was found to have significantly increased linearly, and the RR of the pN classification also demonstrated a significantly reliable difference for each stage of disease with regular discrimination compared with the seventh edition pN classification.

Because this study was only progressed with Korean population, the multinational validation study of a proposal hybrid TNM staging systems should be suggested including gastric cancer patients who were performed neoadjuvant chemotherapy. The response rates of patients receiving neoadjuvant chemotherapy have been reported to be in the range of 24% to 80%, and the percentage of patients who were able to undergo R0 surgical resection ranged from 61% to 84%. In addition, systematic staging approaches including staging laparoscopy, histologic staging or metabolic staging with positron emission tomography/computed tomography should be tailored toward patients receiving neoadjuvant chemotherapy.24, 25

In conclusion, the application of the seventh edition of the TNM classification for gastric cancer was valid in predicting prognosis based on the findings of the current study. However, a low power of discrimination between the pN1 and pN2 classifications may result in unwarranted stage migration and the pN3 classification, which designates patients with >6 positive lymph nodes as a same group, definitely lowers the RR in patients classified as having advanced pN disease. Therefore, a new hybrid TNM classification that is comprised of a combination of the seventh edition pT classification and the sixth edition pN classification should be proposed for the future consideration.

CONFLICT OF INTEREST DISCLOSURES

The authors made no disclosures.

Ancillary