|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
© 2003 American Society for Clinical Oncology
The Surveillance, Epidemiology, and End-Results Program Database as a Resource for Conducting Descriptive Epidemiologic and Clinical StudiesNational Cancer Institute, Bethesda, MD IN THIS issue of the Journal of Clinical Oncology, Anderson et al1 analyze data from the Surveillance, Epidemiology, and End-Results (SEER) Program to determine whether inflammatory breast cancer (IBC) and locally advanced breast cancer (LABC) are distinctly different clinicopathologic entities or variations of the same entity. This distinction would be particularly important if it assists in the identification of risk factors for each and discriminates among therapeutic approaches. The findings are used to illustrate the utility of the SEER database for addressing selected clinical and epidemiologic issues. The authors conclude that the two cancers are distinct biologic entities. This conclusion is based on four major differences between IBC and LABC. The first is the age-specific incidence rates. The mean age for women diagnosed with IBC is only slightly more than 2 years younger than the mean age of women diagnosed with LABC. However, the age distributions are strikingly different. The peak incidence of IBC occurs in the women in their 50s, and then plateaus. This contrasts with the progressive increase of LABC throughout life. Second, the difference in estrogen receptor (ER) status, with a larger percentage of women diagnosed with IBC having ER-negative tumors, also supports a biologic distinction between IBC and LABC. Women diagnosed with IBC also more often have tumors with poorer histologic grade than women diagnosed with LABC. Finally, women diagnosed with IBC have considerably worse survival, although ER status is an independent predictor of cause-specific survival in both IBC and LABC. The median survival for women diagnosed with ER-negative IBC is approximately 30 months, whereas the median survival time for women with ER-positive IBC is about twice as long. Survival is better for women diagnosed with LABC, in which the median survival time of women with ER-negative LABC is only slightly less than for women with ER-positive IBC, and the median survival time of women with ER-positive LABC is greater than the follow-up time of this study. These distinct clinical and epidemiologic features of IBC and LABC support the thesis of the authors that these two types of breast cancer are different entities that develop differently and have different prognosis, but that receptor status influences the outcome of both. The ability to conduct analyses on epidemiologic and clinical differences of cancers is facilitated by the availability of a large comprehensive database such as SEER.2 SEER contains a wealth of data that can be used to categorize cancers in a number of ways. It is important for researchers to understand how these variables are defined because these definitions affect the analysis and interpretation of findings. For example, IBC can be categorized using multiple variables within the SEER data set. The authors compare alternate schemes for classifying IBC incidence by age. The curves of the age-specific incidence are similar with the alternative approach, adding strength to their argument. Anderson et al1 defined LABC as stage III breast cancer cases remaining after cases of IBC are removed. The authors report that women with IBC more often had negative nodes than did women with LABC. However, this is the result of the definition of stage III breast cancer; nodal status is one variable used in the assignment to stage III and LABC, but not to IBC. By definition, the majority of the women diagnosed with LABC would have positive lymph nodes. Only tumors that extend to the chest wall, skin of the breast, or both may have negative nodes and still be classified as LABC. All other tumors must have positive lymph nodes to be classified as LABC. In addition, the authors noted that women with IBC have larger tumors than women with LABC when the size of tumor is measured. Nearly 72% of all women classified as having IBC have tumors that are recorded in a code defined as "diffuse, widespread: 3/4 or more of breast, inflammatory breast cancer" compared with only 3% of the cases in LABC that are coded in that category. Tumor size is "unknown" for approximately 7% of both IBC and LABC cases. The SEER database has made important contributions to our clinical understanding of cancers.3–6 Issues of incidence, treatment outcomes, complications, costs, and survival have been addressed by both the basic data collected on all cancers and special studies in which additional data of interest are collected on samples of patient cases. SEER will celebrate its 30th anniversary this year. There are currently about 3 million incident cancer patient cases included on the SEER public use file. Data since 1992 provide coverage of about 14% of the United States population; beginning in 2000 and beyond, the SEER Program will cover 26% of the total United States population. The populations in SEER are ethnically, geographically, and socioeconomically diverse, providing an unparalleled opportunity to assess these factors. The SEER data set is sufficiently large, with an adequate age range, recording of extensive tumor characteristics, and active follow-up for vital status, to provide a comparison of IBC (an uncommon tumor) and LABC. This database is publicly available to researchers, and to facilitate its use, the National Cancer Institute has provided a data analysis package.7 The population-based nature of the SEER registries requires that all cancers occurring in a defined geographic region be collected. This minimizes potential biases that might be found in facility-based studies where referral patterns may differ with more severe, perhaps more diffuse, disease being referred to highly specialized facilities. Both the comprehensive nature of the database and its limitations should be appreciated. The registries collect detailed information on tumor characteristics, demographic data, surgical intervention, and whether radiation therapy was given. SEER includes active follow-up for vital status and cause of death for all cancer patient cases. Data are collected primarily from the pathology report, hospital, or outpatient surgical center record, and radiation facilities, but physicians’ offices are not routinely visited. The SEER Program standard for patient case ascertainment is 98%.8 Comparison studies have found that major surgery and radiation are accurately recorded in the SEER data.9,10 Chemotherapy and hormonal therapy are substantially underreported and are not available on the SEER public use files. A quality control program is conducted by the National Cancer Institute each year to evaluate the quality and completeness of the data. Patient cases are reabstracted to determine the accuracy of the data, and information is gathered to identify and correct problems. Training workshops are conducted by the National Cancer Institute as well as local training and in-service programs for the registry abstractors. However, there is no standardization of the medical records, laboratory tests, or pathology reports from which data are collected. The lack of standardization may contribute to considerable variability. The variability of the data is likely to attenuate or bias against finding associations in clinical data. The article by Anderson et al1 is an excellent demonstration of a creative and thoughtful use of this unique resource. These investigators have used SEER to define the characteristics of IBC that set it apart from LABC. The SEER resource allows examination of large numbers of patient cases that are drawn from an unselected population. The public availability of the SEER data and associated programming allows future innovative research on similar topics that address important questions in population health. REFERENCES
1. Anderson WF, Chu KC, Shine C: Inflammatory breast carcinoma and non-inflammatory locally advanced breast carcinoma: Distinct clinicopathologic entities? J Clin Oncol 21:2254–2259, 2003
2. Hankey BF, Ries LA, Edwards BK: The surveillance, epidemiology, and end results program: A national resource. Cancer Epidemiol Biomarkers Prev 8:1117–1121, 1999
3. Groome PA, O’Sullivan B, Irish JC, et al: Management and outcome differences in supraglottic cancer between Ontario, Canada, and the Surveillance, Epidemiology, and End Results Areas of the United States. J Clin Oncol 21:496–505, 2003
4. Du XL, Osborne C, Goodwin JS: Population-based assessment of hospitalizations for toxicity from chemotherapy in older women with breast cancer. J Clin Oncol 20:4636–4642, 2002
5. Warren JL, Brown ML, Fay MP, et al: Costs of treatment for elderly women with early-stage breast cancer in fee-for-service settings. J Clin Oncol 20:307–316, 2002
6. Gaffney DK, Alexander T, Wiggins CL: Diminished survival in patients with inner versus outer quadrant breast cancers. J Clin Oncol 21:467–472, 2003 7. National Cancer Institute: SEER 1973–1999 public-use data. http://seer.cancer.gov/publicdata/ 8. National Cancer Institute: Data quality. http://seer.cancer.gov/about/quality.html 9. Cooper GS, Virnig B, Klabunde CN, et al: Use of SEER-Medicare data for measuring cancer surgery. Med Care 40:IV43–IV48, 2002 (8 suppl) 10. Virnig BA, Warren JL, Cooper GS, et al: Studying radiation therapy using SEER-Medicare-linked data. Med Care 40:IV49–IV54, 2002 (8 suppl)
Related Correspondence
Related Reply
This article has been cited by other articles:
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||
|
|||||||||||
|
|
Copyright © 2003 by the American Society of Clinical Oncology, Online ISSN: 1527-7755. Print ISSN: 0732-183X
|
