|Home | About | Journals | Submit | Contact Us | Français|
The sixth edition of the American Joint Committee on Cancer (AJCC) rectal cancer staging subdivided stage II into IIA (T3N0) and IIB (T4N0) and stage III into IIIA (T1-2N1M0), IIIB (T3-4N1M0), and IIIC (anyTN2M0). Subsequent analyses supported revised substaging of stage III as a result of improved survival with T1-2N2 versus T3-4N2 and survival of T4N1 more similar to T3-4N2 than T3N1. The AJCC Hindgut Taskforce sought population-based validation that depth of invasion interacts with nodal status to affect survival.
Surveillance, Epidemiology, and End Results (SEER) population-based data from January 1992 to December 2004 for 35,829 patients with rectal cancer were compared with rectal pooled analysis data (3,791 patients). T4N0 cancers were stratified by tumors that perforate visceral peritoneum (T4a) versus tumors that invade or are adherent to adjacent organs or structures (T4b). N1 and N2 were stratified by number of positive nodes as follows: N1a/N1b (one v two to three nodes) and N2a/N2b (four to six v ≥ seven nodes). Five-year observed and relative survival rates were obtained for each TN category.
SEER rectal cancer analyses confirm that T1-2N2 cancers have better prognosis than T3-4N2, T4bN1 have similar prognosis to T4N2, T1-2N1 have similar prognosis to T2N0/T3N0, and T1-2N2a have similar prognosis to T2N0/T3N0 (T1N2a) or T4aN0 (T2N2a). Prognosis for T4a lesions is better than T4b by N category. The number of positive nodes affects prognosis.
This SEER population-based rectal cancer analysis validates the rectal pooled analyses and supports the shift of T1-2N2 lesions from IIIC to IIIA or IIIB and T4bN1 from IIIB to IIIC. SEER outcomes support subdividing T4, N1, and N2 and revised substaging of stages II and III. Survival by TN category suggests a complex biologic interaction between depth of invasion and nodal status.
Survival and disease relapse after surgery alone or combined with adjuvant treatment for rectal cancer are a function of both degree of bowel wall penetration of the primary lesion (T classification) and nodal status (N classification), as suggested for 40 to 50 years.1–3 Nodal involvement alone is inadequate as the sole pathologic factor to predict survival and relapse rates.4–13 However, through the fifth edition of the American Joint Committee on Cancer (AJCC) staging manual, marked differences in survival by TN category of disease within stages II and III were not taken into account by appropriate substaging.
In the sixth edition of AJCC staging, stage II was subdivided into IIA (T3N0) and IIB (T4N0), and stage III was subdivided into IIIA (T1-2N1M0), IIIB (T3-4N1M0), and IIIC (anyTN2M0).14 The addition of substaging for stages II and III was based on existing outcomes for IIA versus IIB and for IIIA and IIIB. The placement of all TN2 patients into IIIC was based on data that patients with N2 cancers (four or more positive nodes) had poorer outcomes than patients with N1 cancers (one to three positive nodes).
Subsequent rectal cancer pooled analyses demonstrated the independent prognostic significance of each TN and NT category of resected rectal cancer (N subcategory within T category and T subcategory within N category).15,16 The outcomes in the rectal pooled analyses supported revised substaging of stage III as a result of improved survival for patients with T1-2N2 cancers versus T3-4N2 and survival rates with T4N1 lesions that are more similar to rates seen with T3-4N2 than T3N1.
Before making such changes in the seventh edition of AJCC staging, the AJCC Hindgut Taskforce (HTF) sought validation in a population-based data set that depth of invasion interacts with nodal status to impact survival. Data were obtained for patients with both rectal and colon cancers; the colon cancer data in 109,953 evaluable patients with invasive cancer are reported in a separate article.16a
Surveillance, Epidemiology, and End Results (SEER) population-based data were obtained from January 1, 1992 to December 2004 for 55,011 patients with rectal and rectosigmoid cancer (C19.9 and 20.9); 35,829 patients had invasive rectal cancer and evaluable TN category of disease (T1-4N0-2), 17,302 were categorized as NX, and 1,880 were categorized as Tis or T1 polyp. Data of the 35,829 patients with invasive cancers and evaluable TN category of disease were compared with cooperative group rectal pooled analysis data on 2,551 and 3,791 patients (pooled analyses 1 and 2, respectively).15,16 Patients who died within 30 days of surgical resection were not included in the current analysis. The effects of different treatment approaches (surgical technique [total mesorectal resection], adjuvant chemoradiotherapy [pre- or postoperative], adjuvant chemotherapy, or other) were not felt to be pertinent to the current analysis; such data were not analyzed in depth. Of the total group of 55,011 patients, only 4,821 (8.8%) were recorded as having received preoperative irradiation as a component of treatment (T1-4N0-2, 3,353 patients [9.4% of 35,829 patients with evaluable TN category of disease]; NX category, 1,400 patients; Tis/T1 polyp N0-2, 68 patients).
Tumors were stratified by SEER's extent of disease and number of positive nodes coding schemes. T4N0 cancers were stratified by tumors that perforate visceral peritoneum (T4a) versus tumors that invade or are adherent to adjacent organs or structures (T4b). N1 (metastasis in one to three regional nodes) and N2 (metastasis in ≥ four regional nodes) were stratified by number of positive lymph nodes, as follows: N1a (one positive node), N1b (two to three positive nodes), N2a (four to six positive nodes), and N2b (≥ seven positive nodes). T1 to T3 categories were defined as per prior AJCC staging (T1 = tumor invades submucosa; T2 = tumor invades muscularis propria; T3 = tumor invades pericolorectal tissues).
Both observed and relative survival data were obtained for each TN category of disease. Observed survival ( overall survival [OS]) is the proportion of cancer patients surviving for a specified time interval after diagnosis. Relative survival (survival corrected by age-related morbidity; disease-specific survival) is a net survival measure representing cancer survival in the absence of other causes of death.
Observed survival outcomes at 5 years in the current SEER rectal cancer analysis were compared with 5-year OS outcomes in the two rectal cancer pooled analyses for patients with invasive rectal cancer and evaluable TN category of disease (SEER database, 35,829 patients; rectal pooled analysis 1, 2,551 patients; rectal pooled analysis 2, 3,791 patients). As shown in Table 1, SEER analyses confirmed the findings of the two pooled analyses with regard to differential prognosis by NT/TN category of disease for patients with stage II and III cancers. For most TN categories, survival rates in the SEER analysis were 7% to 10% lower than in the rectal pooled analyses.
For patients with stage II cancers, the differential survival outcomes for T3N0 and T4N0 lesions are listed in Table 1. In both pooled analyses, patients with T3N0 lesions had improved 5-year OS relative to patients with T4N0 (P = .046 and P = .02). In the SEER rectal analysis, large differences in 5-year observed survival were found between T3 versus T4N0 cancers, and SEs were small based on large patient numbers (T3N0, 64.0% ± 0.5%, n = 10,615; T4N0, 50.5% ± 1.4%, n = 1,587).
For stage III cancers, patients with T1-2 lesions (confined to rectal wall) had much better prognoses than patients with T3-4 lesions for both N1 and N2 category of disease. These differences were statistically significant in both the SEER and rectal pooled analyses. In addition, patients with T4N1 lesions had prognoses more akin to those of patients with T3N2 or T4N2 cancers in both the SEER and rectal pooled analyses.
The 5-year survival of patients with T1-2 lesions was better than expected for both N1 and N2 category of disease in both the SEER and rectal pooled analyses. For patients with T1-2N1 lesions, 5-year survival was similar to that of patients with T3N0 lesions. For patients with T1-2N2 lesions, 5-year survival was more akin to that of patients with T3-4N0 or T3N1 lesions.
The large patient numbers in the SEER rectal cancer database allowed the evaluation of both observed and relative survival at 5 years for each TN category of disease, including patients with Tis, T1 polyp, and NX lesions (Tables 2 and and3).3). Analyses of highest pertinence to this article included the 35,829 patients with invasive T1-4N0-2 cancers.
A clear survival difference for patients with stage I versus IIA cancers is seen in the SEER analysis (Tables 1 to to3).3). For patients with T1-2N0 lesions, 5-year observed survival was 77.6% ± 0.5% v 64.0% ± 0.5% for patients with T3N0 lesions.
Subcategories of patients with T1-2N2 cancers had survival rates similar to those in patients with N0 disease. Patients with T1N2a lesions had similar 5-year observed survival to patients with T2N0 or T3N0 cancers (T2N0, 75.7% ± 0.6%; T1N2a, 73.8% ± 6.2%; T3N0, 64.0% ± 0.5%). Patients with T2N2a lesions had similar 5-year observed survival to patients with T4aN0 cancers (T2N2a, 58.2% ± 3.4%; T4aN0, 55.7% ± 1.9%).
Prognosis for patients with T4a lesions (tumor penetrates to the surface of visceral peritoneum; revised definition, AJCC seventh edition) is better than the prognosis for patients with T4b lesions (tumor directly invades or is adherent to other organs or structures) for each N category. Relative and observed 5-year survival rates for T4aN0 versus T4bN0 lesions were 69.2% ± 2.4% v 53.6% ± 2.5% (relative) and 55.7% ± 1.9% v 44.7% ± 2.1% (observed), respectively. For N1 and N2 categories, the survival differences for T4a versus T4b were even more striking, as shown in Tables 2 and and33 (N1: relative, 58.7% ± 3.1% v 28.5% ± 2.9%; observed, 48.2% ± 2.5% v 24.3% ± 2.5%; N2: relative, 40.6% ± 3.2% v 18.4% ± 3.0%; observed, 34.3% ± 2.7% v 15.6% ± 2.5%, respectively).
The number of positive nodes affects prognosis for most TN categories of disease (Tables 2 and and3).3). Patients with only one metastatic regional node (N1a) have a 3% to 10% better 5-year relative and observed survival than patients with two to three positive nodes (N1b) for most TN categories of disease. Patients with four to six involved nodes (N2a) have a 5% to 20% better 5-year survival than patients with ≥ seven positive nodes (N2b) by TN category (Tables 2 and and33).
The relative survival impact of both the number of positive nodes and number of nodes evaluated by the pathologist is presented in Figure 1. Relative survival improves for some TN categories as the number of nodes examined increases, most obvious in the larger T3 category of disease (Fig 1C). Prognosis as a function of percentage of nodes involved was not evaluated in the current analysis.
Survival and disease relapse after surgery alone1–9,13 or combined with adjuvant treatment10–12,14-16,18–44 for rectal cancer patients are a function of both degree of bowel wall penetration of the primary lesion and nodal status. However, nodal involvement alone does not determine survival and relapse rates. Invasion through the bowel wall and number of involved lymph nodes are independent high-risk factors for both relapse and survival.
For patients with a single high-risk factor of either direct tumor extension beyond the wall, nodes negative (T3N0), or positive nodes but primary tumor confined to the wall (T1-2N1-2), local relapse rates published in older surgical series have ranged from 20% to 40%.2–7 For patients with both positive nodes and extension beyond the wall (T3-4N1-2), the risk of pelvic relapse was nearly additive (40% to 65% in clinical series and 70% in a reoperative series).2–7
The rate of systemic metastases is significantly higher for patients with both high-risk pathologic factors (extension beyond rectal wall and positive nodes; T3-4N1-2), as opposed to patients with only a single risk factor (T3-4N0, T1-2N1). In published data from adjuvant rectal cancer patients irradiated at either Massachusetts General Hospital10,11 or Mayo Clinic,12 the incidence of subsequent systemic relapse was approximately 20% for patients with T3-4N0 and T1-2N+ lesions versus 40% to 60% for patients with T3-4N+ lesions.
Single-institution analyses from Massachusetts General Hospital10 and Mayo Clinic12 had previously suggested that patients with T1-2N1-2 lesions who were treated with postoperative irradiation alone or combined with chemotherapy had outcomes similar to T3N0 and T4N0 patients, but patient numbers in each stage subset were small. In the rectal pooled analyses with larger numbers of patients, the 5-year OS rate observed for the T1-2N1 patients was similar to that for T3N0 patients, and the 5-year survival for T1-2N2 patients was similar to that for patients with T4N0 or T3N1 lesions (Table 1).15,16 Results by N2 category were rarely available before the first pooled analysis,15 and results by T subcategory for patients with N2 disease (ie, T1-2N2, T3N2, T4N2) were nonexistent.
Data from the rectal cancer pooled analyses (Table 1)15,16 strongly supported substaging of TNM stages II and III, as accomplished in the sixth edition (2002) of TNM staging.14,18 As shown in Table 1, for TNM stage III patients, three separate prognostic subgroups of lesions exist (intermediate risk, T1-2N1; moderately high risk, T1-2N2 and T3N1; and high risk, T3N2 and T4N1-2). To combine or merge all of these patients into TNM stage III (Dukes C) does not provide full prognostic information for patients or physicians. However, patients with T1-2N1-2 disease had a more favorable prognosis than previously thought, and patients with T4N1 lesions (stage IIIB, AJCC sixth edition, along with T3N1 lesions) had prognoses more akin to those of patients with T3-4N2 lesions (stage IIIC, sixth edition).
For patients with N2 disease, data from the rectal pooled analyses demonstrated that N2 disease does not by itself confer poor prognosis.15,16 Substaging by T category influenced both 5-year OS (N2T1-2, 67%; N2T3, 44%; and N2T4, 37%; P < .001; Table 1) and 5-year disease-free survival (N2T1-2, 58%; N2T3, 36%; and N2T4, 30%; P < .001). Placement of all N2 patients in AJCC IIIC substage in the sixth edition16a did not reflect the markedly different prognosis of N2 patients observed in the rectal pooled analyses.
As shown in Table 1, data in the current large SEER population-based rectal cancer analysis validates the rectal pooled analyses with regard to the more favorable prognosis of patients with T1-2N1-2 lesions (stage IIIC, AJCC sixth edition) and less favorable prognosis of patients with T4N1 cancers (stage IIIB, sixth edition). Both SEER and rectal pooled analyses data support the shift of T1-2N2 lesions from stage IIIC to an earlier stage of disease (IIIA/IIIB) and T4N1 lesions from stage IIIB to IIIC (Tables 4 and and55).
Expanded SEER rectal cancer outcomes data (Tables 2 to to5)5) also support subdividing T4, N1, and N2 categories of disease. Patients with T4a lesions (penetrates to the surface of visceral peritoneum [revised definition, AJCC seventh edition]) have a better prognosis than patients with T4b lesions (directly invades or is adherent to other organs or structures) for each N category of disease (N0, N1, and N2). For patients with N0T4a versus N0T4b lesions, there is an approximately 10% improvement in absolute 5-year relative survival and OS, and for patients with N1T4a versus N1T4b and N2T4a versus N2T4b disease, there is a nearly 20% improvement in 5-year survival. Patients with one positive node (N1a) have a better prognosis than patients with two to three positive nodes (N1b), and patients with four to five positive nodes (N2a) have a better prognosis than patients with ≥ seven positive nodes (N2b) by T category.
Previous analyses with much smaller data sets had suggested that patients with perforated T4 lesions may have a worse prognosis than patients with invasion of or adherence to other organs or structures.45,46 However, as shown in the current SEER analysis with large data sets for each TN category of disease, patients with T4 lesions that penetrate to the surface of visceral peritoneum (T4a in AJCC seventh edition) have a more favorable prognosis than patients with invasion of or adherence to other organs or structures (T4b in AJCC seventh edition).
Data in the current SEER analyses combined with rectal pooled analyses data support revised substaging of stages II and III (Tables 4 and and5).5). The AJCC seventh edition HTF recommended the following changes (Table 5): subdivide IIB into IIB (T4aN0) and IIC (T4bN0); shift more favorable TN2 categories to either IIIA (T1N2a) or IIIB (T2N2a, T1-2N2b, T3N2a); and shift less favorable T4N1 lesions from IIIB to IIIC (T4bN1).
Survival outcomes by TN/NT category in the rectal pooled analyses15,16 and the current SEER rectal cancer analysis suggest a complex biologic interaction between depth of invasion and nodal status. As shown in Tables 1 to to5,5, some TN categories of patients with positive nodes (T1-2N1 and T1-2N2) have a similar or better prognosis than patients with negative nodes with regard to both relapse (rectal pooled) and survival (SEER, rectal pooled). Patients with T1-2N1 lesions have better 5-year OS (rectal pooled), relative survival, and observed survival (SEER) than patients with T3N0 or T4N0 cancers (Tables 1 to to5),5), with outcomes more akin to patients with T2N0 lesions (Tables 4 to to5).5). Accordingly, the indications for adjuvant chemoradiotherapy or chemotherapy in patients with T1-2N1 disease should continue to be evaluated. Patients with N2a category (four to six involved nodes) but limited invasion have 5-year survival outcomes similar to those of patients with T2N0 and T3N0 cancers (T1N2a) or T4aN0 cancers (T2N2a) in the current analysis.
Survival outcomes by TN category of disease in the SEER rectal cancer analysis are more similar to SEER colon cancer outcomes than expected.16a Because of the similarities, the AJCC seventh edition HTF recommended continuance of a common staging system for patients with rectal and colon cancers. These similarities may be the result of common tumor biology, the impact of adjuvant chemoradiotherapy (preoperative or postoperative), and/or adjuvant chemotherapy or other factors.
This revision of TNM classification for rectal cancer demonstrates the critical role of formulating postulates and then assessing them in data sets that are larger than single-institution series. As the AJCC proceeds to the next edition, it will be important to collect data on other points of consideration that include, but are not limited to, the number of peritumoral deposits, the number of positive and total nodes examined, and the magnitude of the circumferential radial margin. Only through prospective data collection will usable data exist that can guide decisions relative to the next edition of the staging manual, when hopefully several molecular markers can be incorporated as an adjunct or modifier to the TNM categories of disease.
We acknowledge Lynn A.G. Ries, MS Health Statistician, Cancer Statistics Branch, Cancer Surveillance Research Program, Division of Cancer Control and Population Sciences, National Cancer Institute, Bethesda, MD, for her invaluable assistance in helping navigate the SEER*Stat to retrieve the data used in the current Surveillance, Epidemiology, and End Results (SEER) rectal cancer analysis.
American Joint Committee on Cancer (AJCC) Hindgut Taskforce, 7th Edition: John Milburn Jessup, MD, FACS, Chair; Leonard L. Gunderson, MD, MS, FASTRO, Vice-Chair; Jaffer Ajani, MD; Robert W Beart, Jr, MD, FACS; Al B. Benson III, MD; James Brierly, MD, FRCR, FRCPC (International Union Against Cancer [UICC] Representative); John M. Carethers, MD; Paul Catalano, SCD (Statistical Task Force Liaison); George J. Chang, MD; Carolyn A. Compton, MD, PhD (Editorial Board Representative); Julio Garcia-Aguilar, MD, PhD, FACS; Richard M. Goldberg, MD (Cooperative Group Representative; Cancer and Leukemia Group B); Frederick L. Greene, MD, FACS (Editorial Board Representative); Daniel G. Haller, MD, FACP; Stanley R. Hamilton, MD; Donald E. Hanson, MD; Vencine Kelly, CTR; Bruce D. Minsky, MD; Heidi Nelson, MD; Stephen Rubesin, MD; Daniel J. Sargent, PhD (Statistical Task Force Liaison); Leslie H. Sobin, MD (UICC Representative); Mary Kay Washington, MD (College of American Pathologists Representative); Martin Weiser, MD (AJCC grant recipient); and Mark Lane Welton, MD, FACS, FASCRS.
See accompanying article on page 264
Written on behalf of the American Joint Committee on Cancer Hindgut Taskforce.
Authors' disclosures of potential conflicts of interest and author contributions are found at the end of this article.
The author(s) indicated no potential conflicts of interest.
Conception and design: Leonard L. Gunderson, John Milburn Jessup, Daniel J. Sargent, Frederick L. Greene
Administrative support: Leonard L. Gunderson, John Milburn Jessup
Provision of study materials or patients: John Milburn Jessup
Collection and assembly of data: Leonard L. Gunderson, John Milburn Jessup
Data analysis and interpretation: Leonard L. Gunderson, John Milburn Jessup, Daniel J. Sargent, Frederick L. Greene, Andrew Stewart
Manuscript writing: Leonard L. Gunderson, John Milburn Jessup, Daniel J. Sargent, Frederick L. Greene, Andrew Stewart
Final approval of manuscript: Leonard L. Gunderson, John Milburn Jessup, Daniel J. Sargent, Frederick L. Greene, Andrew Stewart