Is the prediction of prognosis not improved by the seventh edition of the TNM classification for colorectal cancer? Analysis of the surveilla006Ece, epidemiology, and end results (SEER) database
- Peng Gao†1,
- Yong-xi Song†1,
- Zhen-ning Wang1Email author,
- Ying-ying Xu1,
- Lin-lin Tong1,
- Jing-xu Sun1,
- Miao Yu1 and
- Hui-mian Xu1
© Gao et al.; licensee BioMed Central Ltd. 2013
Received: 4 June 2012
Accepted: 11 March 2013
Published: 17 March 2013
Whether the 7th edition of American Joint Committee on Cancer (AJCC) TNM staging system (AJCC-7) is a successful revision remains debatable. We aimed to compare the predictive capacity of the AJCC-7 for colorectal cancer with the 6th edition of the AJCC TNM staging system (AJCC-6).
The National Cancer Institute’s Surveillance, Epidemiology, and End Results (SEER) dataset consisting of 158,483 records was used in this study. We evaluated the predictive capacity of the two editions of the staging system using Harrell’s C index and Bayesian Information Criterion (BIC).
There was a significant prognostic difference between patients at stage IIB and IIC (P < 0.001). Stage III patients with similar prognoses were adequately sub-grouped in the same stage according to AJCC-7. The Harrell’s C index revealed a value of 0.7692 for AJCC-7, which was significantly better than 0.7663 for AJCC-6 (P < 0.001). BIC analysis provided consistent results (P < 0.001).
This study demonstrates that AJCC-7 is superior to the AJCC-6 staging system in predictive capacity.
KeywordsTNM system Colorectal cancer Prognostic prediction
T4 lesions are subdivided as T4a (Tumor penetrates the surface of the visceral peritoneum) or T4b. (Tumor directly invades or is histologically adherent to other organs or structures);
T1-2 lesions that lack regional lymph node metastasis but exhibit tumor deposit(s) are classified in addition as N1c;
N1 tumors are subdivided as N1a (metastasis in 1 regional node) and N1b (metastasis in 2–3 nodes), and N2 tumors are subdivided as N2a (metastasis in 4–6 nodes) and N2b (metastasis in 7 or more nodes);
Stage Group II is subdivided into IIA (T3N0), IIB (T4aN0), and IIC (T4bN0);
T4bN1 is reclassified from IIIB to IIIC;
T1N2a is reclassified as IIIA, and T1N2b, T2N2a-b, and T3N2a are all reclassified as IIIB;
M1 tumors are subdivided into M1a (single metastatic site) and M1b (multiple metastatic sites).
After the publication of AJCC-7, its prognostic validity was discussed worldwide. Whether AJCC-7 is a successful revision remains debatable. Some previous studies have validated supported the revisions introduced [6–9]. However, reports by Mori  and Nagtegaal [11, 12] have raised objections to the revisions; in particular, Nitsche  suggested that AJCC-7 did not provide greater accuracy in predicting colorectal cancer patients’ prognosis compared with AJCC-6.
This study aimed to analyze the prognostic validity of AJCC-7 for colorectal cancer in comparison to AJCC-6; we applied the National Cancer Institute’s Surveillance, Epidemiology, and End Results (SEER) dataset for statistical analysis.
Data from the SEER program
The dataset we used is the National Cancer Institute’s SEER dataset, 1973–2008. SEER collects data on cancer cases from various locations and sources throughout the United States and the program is regarded as a model population-based tumor registry. This national program includes 17 regional registries that cover approximately 28% of the US population. The number of records included in the SEER dataset reaches 6,551,087, including 5,937,405 malignant cases. Among these patients, more than 670,000 patients suffered from colorectal cancer. Patients diagnosed from 1991 through 2003 were selected for analysis. The primary study endpoint was cancer-specific survival.
We selected tumors according to the primary site as follows: cecum, ascending colon, hepatic flexure, transverse colon, splenic flexure, descending colon, sigmoid colon, overlapping lesion of colon, colon NOS (not otherwise specified), rectosigmoid junction, and rectum NOS. We further restricted the tumors included by specific histologic type, as defined by the following individual International Classification of Diseases for Oncology, third edition (ICD-O-3), codes: 8000–8152, 8154–8231, 8243–8245, 8250–8576, 8940–8950, 8980–8981 in accordance with AJCC-7 . Patients were excluded from this study if they exhibited: 1) prior non-colorectal cancer; 2) in situ tumors; 3) incomplete pathological data entries; or 4) died during the immediate postoperative period (within 30 days).
Clinicopathologic features of the colorectal cancer patients
68.45 ± 13.07
Number of lymph nodes retrieved*
12.10 ± 8.87
This study was in compliance with the Helsinki Declaration. We have got permission to access the research data file in SEER program and the reference number was 10188-Nov2011. The study was also approved by the Research Ethics Committee of China Medical University, China and the reference number was 96.
Continuous data are presented as mean ± standard deviation (SD). Cancer-specific survival was analyzed using Kaplan-Meier survival curves, and comparisons were made using the log-rank test.
We evaluated the predictive capacity of the different categories by investigating measures of discrimination. Discrimination refers to the ability to distinguish between high-risk and low-risk patients, and we quantified discrimination using the Harrell’s C index [14, 15] and the Bayesian Information Criterion (BIC) . A model with perfect predictive capacity (sensitivity and specificity of 100%) would have a Harrell’s c-index of 1.00; a category that exhibited a higher Harrell’s C index was considered to exhibit a more accurate predictive capacity. BIC was used to assess the overall prognostic performance of different classification systems via bootstrap-resampling analysis. A smaller BIC value indicates a more desirable model for predicting outcome.
All statistical analyses and graphics were performed using the PASW Statistics 18.0 software (SPSS, Inc., Somers, NY, USA) and STATA MP ver. 10 (StataCorp LP, College Station, TX) statistical software. For all analyses, P < 0.05 was considered to indicate a significant result.
Survival outcomes by two editions of staging system
Comparison of 5-year survival rate based on the 6th edition of TNM system according to the 7th edition of TNM system
The 7th edition (n, 5-YSRb)
The 6th edition
Comparison of the prognostic capacity between AJCC-7 and AJCC-6
Compare Harrell’s C index and Bayesian information criterion between the 6th and 7th edition TNM staging system
Harrell’s C index
The 6th edition TNM stage
The 7th edition TNM stage
Accurate staging is necessary to evaluate the prognosis of patients and is a critical element in determining appropriate treatment based on the experiences and outcomes of groups of prior patients with a similar stage. The AJCC TNM system is currently the most clinically useful staging system. However, unresolved issues in the latest edition remain and further investigation is warranted. Whether AJCC-7 exhibits a greater predictive capacity compared to AJCC-6 remains unclear [10, 11, 13]. For the first time, we compared the predictive capacity between AJCC-7 and AJCC-6 using two independent datasets.
Compared with AJCC-6, stage IIB is subdivided into IIB (T4aN0) and IIC (T4bN0) in AJCC-7. Lan  and Nitsche  proposed that there was not a marked difference in outcome between the two sub-stages. Kim  also found no significant difference in survival rate between patients with stage IIB and stage IIC cancer, but proposed that there was a trend toward a better survival rate for patients with stage IIB cancer. Based on the SEER dataset, a report by Gunderson indicated that there were large differences in the 5-year survival rate between patients with stage IIB and stage IIC cancer . In the present study, we found a significant difference in prognosis between AJCC-7 stage IIB and IIC patients (P < 0.001, Figure 3). It is valuable to subdivide patients into IIB and IIC based on whether the tumor directly invades or is histologically adherent to other organs or structures.
For stage III patients, those who exhibited a similar prognosis were adequately sub-grouped into the same stage according to AJCC-7 (Figure 4a). However, based on AJCC-6, some patients were not sub-grouped into a reasonable stage (Figure 4b). Therefore, the results were more in accordance with AJCC-7.
Based on the SEER dataset, we found that there were significant differences in survival for AJCC-6 stage IIB, IIIB, and IIIC patients when stratified by AJCC-7. However, no significant differences in prognosis were observed for the AJCC-7 stage IIIA and IIIC patients when stratified by AJCC-6 (Table 2). Although there was a significant difference in survival for AJCC-7 stage IIIB when stratified by AJCC-6, the staging was reasonable because the difference in survival between the migrated patients and patients both in stage IIIC of AJCC-6 and AJCC-7 was even more significant (5-year survival rate: 54.8% vs. 34.0%, p < 0.001). These results indicated that using AJCC-7 could potentially provide more detailed classification and greater power to subgroup patients with a more homogenous prognosis compared to AJCC-6.
Two comprehensive statistical methods, Harrell’s C index and BIC, were used to compare the predictive capacity of AJCC-7 and AJCC-6. Both of the two tests revealed that the prognostic performance of AJCC-7 was significantly better compared to AJCC-6 (Table 3). Nitsche  proposed that AJCC-7 did not provide greater accuracy in predicting colorectal cancer patient prognosis compared with AJCC-6 based on a European single-center collective. However, in their published data we found that AJCC-7 exhibited a better predictive capacity according to the results of c-index and BIC, although the differences were not significant. It is possible that the differences between the results of the two tests were caused by different sample sizes. Stage migration occurred in only 8.48% (13442/158483) of patients from the SEER dataset used in this study and 13.5% (302/2229) of patients from the study by Nitsche. The total number of reclassified patients was small. Therefore, only when the comparison was based on a large dataset such as SEER the difference between AJCC-6 and AJCC-7 was significantly validated. When the comparisons were based on a relatively small dataset, it was difficult to obtain significant results, although a tendency that AJCC-7 exhibited a greater predictive capacity compared to AJCC-6 could be observed.
The classification of tumor deposits varied enormously in AJCC-7. These changes have been criticized as confusing and as exhibiting low reproducibility in the study by Nagtegaal ; however, they also proposed that AJCC-7 exhibited greater prognostic value than other staging systems when all tumor deposits irrespective of size or contour were included as lymph nodes. In our previous study , we found that the classification of tumor deposits in AJCC-7 satisfactorily predicted patients’ outcome for those without lymph node metastases. However, we also found that patients who were categorized as T3-4N2b with tumor deposits should be reclassified as stage IV. In the present study, it was not possible to compare the classification of tumor deposits in AJCC-7 with that in AJCC-6 because the SEER dataset did not identify the status of tumor deposits. Furthermore, the number of metastatic organ/site was unknown in the SEER program. Considering these two points, this study answers only well defined parts of the question about the comparison of AJCC-7 with AJCC-6.
Our study does have some limitations. It is a retrospective exploratory study. Clinical and pathologic patient information can be heterogeneous because SEER collects information from 12 population-based cancer registries. On the other hand, in the SEER dataset, data on adjuvant therapy is limited to information on radiation therapy only and the information of chemotherapy is unavailable.
This study demonstrates that AJCC-7 is superior to the AJCC-6 staging system in predictive capacity. We recommend that this staging system be considered for clinical application.
Written informed consent was obtained from the patient for publication of this report and any accompanying images.
American Joint Committee on Cancer
National Cancer Institute’s Surveillance, Epidemiology, and End Results
Bayesian Information Criterion
International Classification of Diseases for Oncology, third edition
5-year survival rate.
This work was supported by National Science Foundation of China (No. 30972879 and No. 81172370), the Program of Scientific and Technological Department of Liaoning Province (No. 2010225032) and the Program of Education Department of Liaoning Province (L2011137).
- Siegel R, Ward E, Brawley O, Jemal A: Cancer statistics, 2011: the impact of eliminating socioeconomic and racial disparities on premature cancer deaths. CA Cancer J Clin. 2011, 61 (4): 212-236. 10.3322/caac.20121.View ArticlePubMedGoogle Scholar
- O’Connell JB, Maggard MA, Ko CY: Colon cancer survival rates with the New American joint committee on cancer sixth edition staging. J Natl Cancer Inst. 2004, 96 (19): 1420-1425. 10.1093/jnci/djh275.View ArticlePubMedGoogle Scholar
- Greene FL: Current TNM staging of colorectal cancer. Lancet Oncol. 2007, 8 (7): 572-573. 10.1016/S1470-2045(07)70185-7.View ArticlePubMedGoogle Scholar
- Edge SB, American Joint Committee on Cancer: AJCC cancer staging manual. 2010, New York: Springer, 7Google Scholar
- Greene FL, American Joint Committee on Cancer., American Cancer Society: AJCC cancer staging manual. 2002, New York: Springer-Verlag, 6View ArticleGoogle Scholar
- Gunderson LL, Jessup JM, Sargent DJ, Greene FL, Stewart AK: Revised TN categorization for colon cancer based on national survival outcomes data. J Clin Oncol. 2010, 28 (2): 264-271. 10.1200/JCO.2009.24.0952.View ArticlePubMedGoogle Scholar
- Gunderson LL, Jessup JM, Sargent DJ, Greene FL, Stewart A: Revised tumor and node categorization for rectal cancer based on surveillance, epidemiology, and end results and rectal pooled analysis outcomes. J Clin Oncol. 2010, 28 (2): 256-263. 10.1200/JCO.2009.23.9194.View ArticlePubMedGoogle Scholar
- Kim KH, Yang SS, Yoon YS, Lim SB, Yu CS, Kim JC: Validation of the seventh edition of the American Joint Committee on Cancer tumor-node-metastasis (AJCC TNM) staging in patients with stage II and stage III colorectal carcinoma: analysis of 2511 cases from a medical centre in Korea. Colorectal Dis. 2011, 13 (8): e220-e226. 10.1111/j.1463-1318.2011.02625.x.View ArticlePubMedGoogle Scholar
- Lan YT, Yang SH, Chang SC, Liang WY, Li AF, Wang HS, Jiang JK, Chen WS, Lin TC, Lin JK: Analysis of the seventh edition of American Joint Committee on colon cancer staging. Int J Colorectal Dis. 2012, 27 (5): 657-663. 10.1007/s00384-011-1366-6.View ArticlePubMedGoogle Scholar
- Mori T: A Comparison of the New (Planned) TNM Classification and Japanese General Rule for Staging Colorectal Cancer. Cancer Invest. 2010, 28 (4): 387-392.View ArticlePubMedGoogle Scholar
- Nagtegaal ID, Tot T, Jayne DG, McShane P, Nihlberg A, Marshall HC, Pahlman L, Brown JM, Guillou PJ, Quirke P: Lymph nodes, tumor deposits, and TNM: Are We getting better?. J Clin Oncol. 2011, 29 (18): 2487-2492. 10.1200/JCO.2011.34.6429.View ArticlePubMedGoogle Scholar
- Nagtegaal ID, Quirke P: Revised staging: is it really better, or Do We Not know?. J Clin Oncol. 2010, 28 (23): e397-e398. 10.1200/JCO.2010.28.5726.View ArticlePubMedGoogle Scholar
- Nitsche U, Maak M, Schuster T, Künzli B, Langer R, Slotta-Huspenina J, Janssen K-P, Friess H, Rosenberg R: Prediction of prognosis is Not improved by the seventh and latest edition of the TNM classification for colorectal cancer in a single-center collective. Ann Surg. 2011, 254 (5): 793-801. 10.1097/SLA.0b013e3182369101.View ArticlePubMedGoogle Scholar
- Harrell FE, Lee KL, Mark DB: Multivariable prognostic models: issues in developing models, evaluating assumptions and adequacy, and measuring and reducing errors. Stat Med. 1996, 15 (4): 361-387. 10.1002/(SICI)1097-0258(19960229)15:4<361::AID-SIM168>3.0.CO;2-4.View ArticlePubMedGoogle Scholar
- Newson RB: Comparing the predictive powers of survival models using Harrell’s C or Somers’ D. Stata J. 2010, 10: 339-358.Google Scholar
- Volinsky CT, Raftery AE: Bayesian information criterion for censored survival models. Biometrics. 2000, 56 (1): 256-262. 10.1111/j.0006-341X.2000.00256.x.View ArticlePubMedGoogle Scholar
- Tong LL, Gao P, Wang ZN, Song YX, Xu YY, Sun Z, Xing CZ, Xu HM: Is the seventh edition of the UICC/AJCC TNM staging system reasonable for patients with tumor deposits in colorectal cancer?. Ann Surg. 2012, 255 (2): 208-213. 10.1097/SLA.0b013e31821ad8a2.View ArticlePubMedGoogle Scholar
- The pre-publication history for this paper can be accessed here:http://www.biomedcentral.com/1471-2407/13/123/prepub
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.