Originally published as JCO Early Release 10.1200/JCO.2004.02.946 on April 5 2004

This Article Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Services Email this article to a colleague Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Save to my personal folders Download to citation manager Rights & Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Benson, A. B. Articles by Catalano, P. J. Search for Related Content PubMed PubMed Citation Articles by Benson, A. B., III Articles by Catalano, P. J. Social Bookmarking                            What's this?

A Rose, Is a Rose, Is a Rose...or Not

¹ Division of Hematology/Oncology, Robert H. Lurie Comprehensive Cancer Center, Northwestern University, Chicago, IL
² Department of Biostatistical Science, Dana-Farber Cancer Institute, and the Department of Biostatistics, Harvard School of Public Health, Boston, MA

Refinements in colorectal cancer staging systems have evolved from the Dukes' classification to the most recent subgroup stratification of patients as reflected in the sixth edition of the tumor-node-metastasis system Cancer Staging Manual.¹ Modifications of the staging system have proceeded in recognition that neither all stage II nor all stage III colon or rectal cancer patients can be aggregated for prognostic purposes within a single stage. Patients are truly not all alike, and most certainly are not in terms of the likelihood of surviving their cancer. In the future, it is possible that subset characteristics within a stage may also serve to predict likelihood of treatment benefit.

This edition of the Journal of Clinical Oncology includes three large retrospective analyses of colon cancer and rectal cancer patients, that evaluate stage and additional potential prognostic factors to assess risk of relapse, survival, and benefit of therapy.^2–4 The articles by Gill et al, Gunderson et al, and Greene et al provide valuable data describing nuances of staging for colorectal cancers. These authors have demonstrated that both T and N stage classifications are critical in defining risk, both in terms of recurrence and in overall survival. They build on recently published intergroup subset analyses confirming that more accurate staging and improved survival is achieved for colon and rectal cancer patients based on the number of lymph nodes that are analyzed by the pathologist.^5,6 Whether one focuses on the more stratified and refined American Joint Committee on Cancer (AJCC), Sixth Edition, staging, for example, in the substrata of stage III, or whether one adopts the risk classifications of Gunderson et al for rectal cancer—in particular, distinguishing between the moderately high- and high-risk categories—it is clear that important prognostic information is already available to help guide clinicians and patients as they consider current therapy options. What the risk classifications imply about the underlying biology of the disease and its associated implications for directing patients to individualized therapy is a separate, but related, issue. The integration of stage and other pathologic and molecular determinants is essential to the design of the next generation of colorectal adjuvant studies.

We have learned that the three stage III subclassifications provide statistically important prognostic information. What we have not been able to understand is whether these subgroups are predictive in general. In the context of the present three studies, the goal of defining robust predictive characteristics remains elusive, for a variety of particular reasons clearly described by the authors. For example, Gill et al state that their prognostic and predictive model is limited, since many studies might have adjusted prognostic factors for stage, but other factors, such as extent of lymph node involvement and grade, are not uniformly evaluated. Therefore, the Gill model is confined to just a few prognostic factors. The Gunderson article, as a second example, is limited by the use of descriptive data to evaluate treatment outcome relative to treatment method since the treatment varied from trial to trial, making it difficult to compare the same regimen in different trials. In addition, the studies included in the rectal cancer pooled analysis accrued patients over a 13-year period, introducing a variety of surgical techniques and treatment methods. If the current staging system is to have as much value in guiding clinical research as it does in defining risk, it is critical that robust predictive factors must also be identified and integrated into prospective clinical trial design.

It seems that an outstanding fundamental question remains: Do the refinements of categories within stage via T and N (and T within N, and vice versa) reflect different underlying biology or do they more simply our ability to statistically characterize the natural history of the disease at the time of presentation? The latter would imply that it is possible to direct therapy to those subgroups numerically most likely to benefit from current adjuvant treatments; however, attempting to target therapy within stage groups or risk classifications could be a fruitless exercise. We need to learn whether, for example in the moderately high-risk group of Gunderson et al, T1-2/N2 lesions are biologically different than T4/N0 cancers and, if so, how to best direct current therapies and to study new ones.

In the absence of more information, can we use current staging information to assist in developing the next generation of adjuvant colorectal trials? Simple calculation shows that to rigorously study the effectiveness of new agents within substrata in stage III (ie, as the main goal of a traditional factorial arrangement of treatment-by-stage subclasses) one trial would require many thousands of patients. A trial that just compares two treatments would require this large number of patients because the contrast of treatment effects within substage would be of primary interest. Essentially, the entire US Intergroup would have to commit its resources for many years, and, by implication, only a single study could be done.

How do we move past these barriers? Perhaps molecular studies can help elucidate underlying biologic characteristics to assist us in directing new therapies, especially targeted therapies, to those most likely to benefit within stage subclasses (eg, to distinguish more clearly between the A and B categories within stages II and III). Without a clearer biologic directive on how to proceed in trial design, brute force factorial design will have little chance of producing results on best therapies within substages since that approach is hampered scientifically (no clear direction on exactly where to go) and practically (paucity of patients). Stage II trials with the same goals are even more problematic, given the low failure rates and associated small numbers of events typically observed in this stage. It would require a massive study to establish the required synergistic effects in a stage-by-treatment evaluation, unless there was a substantial difference in treatment effect. A trial of this magnitude is unlikely to develop given the incremental steps that have been achieved to bring us to the current state of the art in this disease.

An immediate recommendation must be made to include detailed review of each colorectal cancer patient's pathology with agreement as to the precise stage by the surgeon, pathologist, and oncologist. In particular, attention to the number of lymph nodes evaluated by the pathologist is of paramount importance, particularly for patients with stage II colon cancer. It is clear that the number of lymph nodes evaluated and reported is a critical determinant of stage and prognosis and may affect treatment recommendations. The routine use of staging criteria is outlined in the sixth edition of the AJCC staging manual and must be emphasized. Both the AJCC and the College of American Pathologists urge examination of at least 12 lymph nodes.^{1, 7} Patients who do not have an evaluation of a sufficient number of lymph nodes must be considered inadequately staged. When the written pathology report suggests limited lymph node evaluation, it is appropriate to request additional review of lymph nodes from the surgical specimen, whenever possible. Of note, the pathologic lymph node assessment is often affected (ie, downstaged) for those patients with clinical stage II and III rectal cancer who have received preoperative chemoradiation.

Both physician discussions with patients and current and future colorectal cancer clinical trials should address risks associated with sub-set staging descriptions (stage IIA v Stage IIB; stage IIIA v Stage IIIB v Stage IIIC.) For example, overall survival and disease-free survival can be discussed with a rectal cancer patient based on TN Stage (N substage within T stage) and NT Stage (T substage within N stage) as described by Gunderson et al. Discussion of intermediate, moderately-high, and high risk with an associated overall survival range with chemotherapy and chemo-radiation from the pooled analyses of cooperative group trials can provide the rectal cancer patient with helpful information including an estimate of risk and benefit. Furthermore, the cooperative group analyses suggest that one treatment approach for all rectal adjuvant patients may not be warranted. As discussed by Tepper et al in the final report of a recent intergroup rectal cancer trial, there may be a sub-set of patients who do not require pelvic radiation therapy.⁸ Patients who have had a confirmed total mesorectal excision with T1-2/N+ or T3/NO tumors that are more proximally located in the rectum, with adequate lymph node evaluation and negative surgical margins, carry a low risk of local recurrence and may be considered for chemotherapy alone.

Similarly, colon cancer patients may find useful information to assist in making an adjuvant therapy decision by using the prognostic model as described by Gill et al, which includes the prognostic factors of nodal status, histological grade, depth of tumor invasion into the bowel wall, and age (http://www.mayoclinic.com/calcs). There are a number of other reported prognostic markers, however, that might also enter into discussions of risk including preoperative carcinoembryonic antigen value, lymphatic or vascular invasion, perforation, microsatellite instability, and loss of heterozygosity of 18q.⁷ However. these additional prognostic factors cannot at this time be used to predict outcome after adjuvant therapy, as there are insufficient data to select therapy based on a particular pathologic or molecular characteristic.

Individualized patient strategies for treatment based on prognostic and predictive molecular markers are an important future goal. Current and planned adjuvant colon and rectal clinical trials now include tissue acquisition with detailed stage and pathology analysis which have the potential to provide the necessary prospective data to reach this most critical goal.

Authors' Disclosures of Potential Conflicts of Interest

The authors indicated no potential conflicts of interest.

REFERENCES

1. Greene FL, Page DL, Fleming ID, et al: AJCC Cancer Staging Manual, Sixth Edition, New York, NY, Springer-Verlag, 2002

2. Greene FL, Stewart AK, Norton HJ: New tumor-node-metastasis staging strategy for node-positive (stage III) rectal cancer. J Clin Oncol 22:1778–1784, 2004[Abstract/Free Full Text]

3. Gill S, Loprinzi CL, Sargent DJ, et al: Pooled analysis of fluorouracil-based adjuvant therapy for stage II and III colon cancer: Who benefits and by how much? J Clin Oncol 22:1797–1806, 2004[Abstract/Free Full Text]

4. Gunderson LL, Sargent DJ, Tepper JE, et al: Impact of T and N stage and treatment on survival and relapse in adjuvant rectal cancer: A pooled analysis. J Clin Oncol 22:1785–1796

5. Le Voyer TE, Sigurdson ER, Hanlon AL, et al: colon cancer survival is associated with increasing number of lymph nodes analyzed: A secondary survey of Intergroup Trial INT-0089. J Clin Oncol 21: 2912–2919, 2003[Abstract/Free Full Text]

6. Tepper JE, O'Connell MJ, Niedzweicki D, et al: Impact of number of nodes retrieved on outcome in patients with rectal cancer. J Clin Oncol 19:157–163, 2001[Abstract/Free Full Text]

7. Compton CC, Fielding LP, Burgart LJ, et al: Prognostic factors in colorectal cancer: College of American Pathologists consensus statement. Arch Path Lab Med 124:979–994, 2000

8. Tepper JE, O'Connell M, Niedzweicki D, et al: Adjuvant therapy in rectal cancer: Analysis of stage, sex, and local control–final report of Intergroup 0114. J Clin Oncol 20:1744–1750, 2002[Abstract/Free Full Text]

CiteULike    Complore    Connotea    Del.icio.us    Digg    Facebook    Reddit    Technorati    Twitter    What's this?

This article has been cited by other articles:

				J. E. Tepper Patient Selection, Realism, and Randomized Clinical Trials J. Clin. Oncol., August 15, 2004; 22(16): 3215 - 3217. [Full Text] [PDF]

This Article Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Services Email this article to a colleague Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Save to my personal folders Download to citation manager Rights & Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Benson, A. B. Articles by Catalano, P. J. Search for Related Content PubMed PubMed Citation Articles by Benson, A. B., III Articles by Catalano, P. J. Social Bookmarking                            What's this?