Background: Various parameters assist in the definition of prognosis and in the choice of therapy for breast cancer. This study evaluates the effects of prognostic factors on disease outcome in elderly women.
Methods: A retrospective cohort analysis was performed on 1,267 consecutive patients with locoregional breast cancer, including 374 patients 65 years of age or older, who were referred to a university cancer center over an eight-year period. Information on prognostic factors, disease outcome, and survival was analyzed.
Results: Women 65 years of age or older were more likely to have early-stage cancers, lower histologic grade, higher hormone receptor levels, and lower S-phase fractions. They experienced a longer time to disease recurrence and overall longer survival than their younger counterparts.
Conclusions: Advancing age should not be considered an unfavorable risk factor for breast cancer. Prognosis and treatment should be based on disease stage and histologic and biologic parameters rather than patient age.

Introduction

It is estimated that more than 185,000 women will be diagnosed with breast cancer in the United States in 1996, representing 31% of all newly diagnosed invasive malignancies in women in this country.[1] Nearly 45,000 women will die of the disease in the same year.[1] Breast cancer represents a leading cause of morbidity and mortality in the elderly. For those over 65 years of age, the annual age-adjusted incidence and mortality rates are 440 and 125 per 100,000, respectively.[2] The median age of women with breast cancer reported to the Surveillance, Epidemiology, and End Results (SEER) program is 63 years.[2] While the overall age-adjusted mortality from breast cancer has decreased little this century, the relative five-year survival has increased to more than 80%.[2]

Management of breast cancer in the elderly remains controversial.[3] Several studies have suggested prognostic factors such as disease stage differ with regard to age.[4,5] Treatment offered to breast cancer patients also has been shown to differ with regard to age.[6-9]The elderly tend to be offered conventional therapy less often than younger patients. Nevertheless, data reported to the SEER Program suggest that mortality rates and relative survival are as high in the elderly as in younger subjects.[2]

Our study was undertaken to further elucidate the effects of age on the demographic, diagnostic, and prognostic factors commonly evaluated in breast cancer patients and to assess the outcome of treatment in elderly women with locoregional breast cancer at diagnosis.

Materials and Methods

All newly diagnosed breast cancer patients with locoregional disease at presentation who were referred to the University of South Florida (USF) and the H. Lee Moffitt Cancer Center & Research Institute at USF over an eight-year period were studied through a retrospective chart review. Systematic data collection for each patient included background demography, medical, family, and social history, diagnostic and staging information, and initial treatment. Background information included age, race, sex, family history of breast cancer, history of smoking, number of pack years, history of thyroid disease, psychiatric illness, menstrual history, menopausal status, number of pregnancies, history of oophorectomy, and history of exogenous hormone use. Diagnostic and staging information included stage, number of nodes sampled, number of positive nodes, size of primary tumor, presence of bilateral disease, presence of positive margins, histologic type, tumor grade, DNA index, ploidy, S-phase fraction, HER-2/neu oncogene expression, epidermal growth factor receptor, cathepsin D, estrogen and progesterone receptor levels, and preoperative carcinoembryonic antigen level. All diagnoses were confirmed by institutional review of the original biopsy material including axillary contents. Treatment information included dates, type and place of surgery, and use of adjuvant radiation therapy, adjuvant chemotherapy, and adjuvant tamoxifen. Follow-up information included dates of diagnosis, site and date of recurrence, and cause and date of death. Data confirmation was obtained through a medical record audit of each patient by a trained auditor.

Summary measures were calculated for each variable consisting of the mean or median for continuous measures and proportions for categorical measures. For all variables with distributions approximating normality, means were compared between groups using the t test for the comparison of two independent means based on a pooled variance. For measures with distributions far from normality, nonparametric methods of inference were used. The Mann-Whitney U Test (Wilcoxon Rank Sum Test) for each variable was used to test the equality of distributions between two groups.[10] The distribution of categorical variables was compared using the chi-square method for testing the null hypothesis of no association. Survival time and time to disease recurrence were measured from the date of diagnosis. Survival functions were estimated by the product-limit method of Kaplan and Meier for censored data.[11] Survival functions were compared using the log-rank (Mantel-Haenszel) test of the null hypothesis for equality of the survival functions.[12,13] The cumulative probability of survival and the cumulative risk of recurrence are presented for each group as well as the standard error as a percentage. Multivariate analysis was performed using the proportional hazards method of Cox.[14] Coefficients were estimated by the method of maximum likelihood, and the likelihood ratio test was used to test the null hypothesis that the model coefficients were zero. Coefficient estimates were obtained in a forward stepwise fashion after satisfying the proportional hazards assumptions of the model.[15]

Results

Results are available on 1,267 women with breast cancer - including 374 patients who were 65 years of age or older - who were referred to our institution between 1985 and 1993. Fig 1 displays the age distribution of all patients with invasive (n=1,136) or noninvasive (n=131) carcinoma. This age distribution is somewhat younger than population-based registries, but it is probably representative of experience at other tertiary referral cancer centers. Comparison of patients less than 65 years of age and those age 65 years and over revealed no significant difference in race, family history of breast cancer, number of pregnancies, or prior oophorectomy. Older patients more frequently reported a history of thyroid disorder (14.2%) than younger patients (8%) (P=.0042). Younger patients more commonly reported a history of smoking (41.9%) compared to older patients (30.5%) (P=.0002). While older patients were uniformly postmenopausal, 408 (45.7%) women under age 65 years were premenopausal (P<.0001).

Diagnostic and staging results showed no significant difference in histologic classification between older and younger age groups. All but one of the 19 inflammatory carcinomas were diagnosed in women less than 65 years of age. However, women age 65 years and older more often had a history of synchronous or metachronous bilateral disease (9.4%) compared to younger women (4.8%) (P=.0031). Table 1 compares diagnostic and staging variables between with older and younger patients. Patients age 65 years and over more often presented with earlier-stage disease, (stage I, tumor size less than 2 cm, and negative nodes). Alternatively, patients under 65 years of age more often presented with more advanced-stage disease (stage III, tumor size more than 5 cm, or more than four lymph nodes involved). Older patients more often presented with lower tumor grade (grade I or II), and a greater proportion were estrogen or progesterone receptor-positive than younger patients.

Primary treatment in the form of surgery, radiation therapy, chemotherapy, and hormonal therapy was compared between the two age groups. Among women age 65 years and over, 191 (51%) underwent a modified radical mastectomy, 148 (40%) underwent a lumpectomy with axillary lymph node dissection, and the remainder underwent a limited breast resection without axillary node dissection. There was no significant difference in the distribution of surgical procedures among older subjects in comparison to those offered to younger patients. Radiation therapy was provided to 162 (45%) older patients compared to 418 (49%) younger patients (P=.225). There also was no significant difference in the use of radiation therapy after adjusting for the type of surgical resection; all patients undergoing breast-conserving procedures received radiation therapy. Adjuvant chemotherapy was offered to 212 (58%) older patients compared to 545 (63%) younger patients (P=.110). There was no significant difference in the use of adjuvant chemotherapy after adjusting for the stage of disease. More patients age 65 years and over were treated with adjunctive tamoxifen alone (49%) compared to younger patients (21%) (P=.0007). However, there was no significant difference in the use of tamoxifen after adjusting for hormone receptor status. The median time of follow-up in this study was 31 months.

Several variables were found to have significant prognostic importance for recurrence in women with invasive breast cancer. These variables include stage, tumor size, number of positive nodes, histologic type, tumor grade, estrogen receptor-positivity, S-phase fraction, ploidy, DNA content, menopausal status, and age. Table 2 compares several biologic variables used to estimate prognosis between older and younger patients with invasive breast cancer. The distributions of these measurements were generally skewed. Nonparametric methods were used to test the null hypothesis of equal distributions. Tumors in patients age 65 years and over had higher levels of measured estrogen and progesterone receptors, while tumors in younger subjects had higher S-phase fractions and epidermal growth factor receptor levels. There was no significant difference in tumor DNA content, HER-2/neu oncogene expression, or the level of cathepsin D between the two age groups.

Fig 2A displays the overall survival for patients with invasive breast cancer. Women 65 years of age and over survived longer than younger patients with five-year overall survivals of 93% ± 2.7% (proportion ± standard error) and 86% ± 1.9%, respectively (P=.0071). There was no significant difference in the cause of death, with breast cancer accounting for death in 67% of women 65 years of age and older and in 87% of those under 65 years of age (P=.3184). Fig 2B displays the time to recurrence for all women with invasive breast cancer. Older patients with breast cancer experienced a longer time to disease recurrence than younger patients with five-year cumulative risk of recurrence of 15% ± 3.8% and 22% ± 2.2%, respectively (P=.0034).

Table 3 compares the cumulative risk of recurrence at five years between older and younger patients with invasive breast cancer for several important prognostic factors. There was no significant difference in time to recurrence between older and younger individuals when considering only postmenopausal patients. Likewise, there was no significant difference in the time to recurrence between the two age groups for patients with negative lymph nodes, small or large tumor size, and negative estrogen receptor status. There also was no significant difference in outcome when stratified by stage or tumor grade. Older patients have a significantly longer time to recurrence when considering patients with positive nodes, intermediate tumor size, or positive receptors.

Interactions exist between age and menopausal status, nodal status and, estrogen receptor status. Older patients and younger postmenopausal patients have nearly identical recurrence risk. Premenopausal patients under 65 years of age experience disease recurrence significantly earlier than postmenopausal patients both over and under 65 years of age. Among patients with positive nodes, those under 65 years of age recur significantly earlier than older patients. Among patients age 65 and over, those with positive nodes experience recurrence earlier than patients with negative nodes (P=.043). As shown in Fig 3, the recurrence function for older patients with positive nodes closely follows the negative node recurrence function for the first four to five years, but late relapses occur such that the cumulative risk of recurrence at seven to eight years is virtually the same as younger patients with positive nodes. Age has little influence on relapse risk in patients with negative estrogen receptors. Positive estrogen receptors were associated with a longer time to recurrence in both those under age 65 (P=.0447) and those 65 years of age and over (P=.0398). As shown in Fig 4, patients with positive estrogen receptors under age 65 experience earlier recurrence than such patients age 65 and over (P=.0281).

The hazard rate for recurrence and mortality was modeled in patients with invasive breast cancer using the multivariate regression method of Cox.[9] Interaction terms between the main covariates failed to achieve statistical significance and therefore are not included in the models. Table 4 provides the results of the Cox regression models for disease recurrence with age included as a model covariate. With time to recurrence as the outcome of interest, the effect of age was significant after adjusting for nodal status, tumor size, and histologic subtype. No other variables, including receptor levels and other biologic parameters, contributed significantly to predicting the risk of recurrence. Specifically, with age at diagnosis in the model, menopausal status was no longer a significant predictor of outcome. If analysis is limited to patients age 65 years and over, the only significant covariate is that of tumor size (P=.0001).

Discussion

Breast cancer is the most common malignancy in women in the United States.[1] Breast cancer incidence and mortality increase with advancing age up to the eighth decade of life. Nevertheless, the five-year relative survival in women with breast cancer 65 years of age and older reported to the SEER program is virtually identical to that in younger women.[2]

We analyzed the data from 1,267 consecutive women with breast cancer treated over the past eight years at our university cancer center. While the age distribution of this series was younger than that of population-based registries, the age-specific distribution of stage and other prognostic factors is similar. Patients 65 years of age and over were found to present with earlier-stage disease (smaller tumor size and fewer lymph nodes involved), lower tumor grade, and a greater proportion of hormone receptor-positive tumors. Prognostic factors significantly associated with disease recurrence in our population were stage, nodal status, tumor size, tumor grade, estrogen and progesterone receptor status, menopausal status, S-phase fraction, epidermal growth factor receptor, HER-2/neu oncogene expression, and age. In our experience, women age 65 years and over had a significantly lower rate of disease recurrence and mortality than younger patients. These contradict Swedish data suggesting that prognosis worsens with increasing age after 50 years of age.[16] In our study, age was significantly related to several prognostic factors including stage, nodal involvement, tumor size, and estrogen and progesterone receptor activity in the present study. While the sample size of our data is limited, older patients with invasive breast cancer had higher quantitative levels of estrogen and progesterone receptors, higher epidermal growth factor receptor levels, and lower S-phase fractions. In multivariate analysis, increasing age was a significant independent predictor of delayed disease recurrence. These findings contradict those reported to SEER in which elderly women with breast cancer in metropolitan Detroit had a less favorable prognosis.[17] However, our data are not population-based and are from a single tertiary referral center.

Although older women were more likely to be offered tamoxifen at our institution, they were just as likely as younger patients to be offered lumpectomy, radiation therapy, and adjuvant chemotherapy. This is in contrast to other data suggesting that definitive treatment is less likely to be offered to older patients, even those with early-stage disease.[6,7] Based on the results presented here, treatment decisions should be based on standard prognostic factors rather than age.

Several prognostic factors for breast cancer confound or are confounded by patient age. Older patients with breast cancer presented with earlier-stage disease, favorable histologic characteristics, and positive receptor status. Age should therefore be considered a potential confounding factor in prognostic factor evaluation in patients with breast cancer. Failure to consider confounding factors in prognostic models may result in invalid conclusions relative to the importance of the variables under study. Age is also a potential effect modifier by virtue of interaction with other prognostic factors. Patient age should be considered in all studies of breast cancer prognostic factor analysis.

Appreciation is expressed to Nicole Beale, Kathy McKinley, and Dorothy Allen for their technical assistance in the preparation of this manuscript.

References

1. Parker SL, Tong T, Bolden S, et al. Cancer statistics, 1996. CA Cancer J Clin. 1996;46:5-27.
2. Miller BA, Ries LA, Hankey BF, et al, eds. SEER Cancer Statistics Review: 1973-90. National Cancer Institute. NIH Pub No 93-2789, 1993.
3. Balducci L, Schapira DV, Cox CE, et al. Breast cancer of the older woman: an annotated review. J Am Geriatr Soc. 1991;39:1113-1123.
4. Goodwin JS, Samet JM, Key CR, et al. Stage at diagnosis of cancer varies with age of the patient. J Am Geriatr Soc. 1986;34:20-26.
5. Peer PG, van Dijck JA, Hendricks JH, et al. Age-dependent growth rate of primary breast cancer. Cancer. 1993;71:3547-3551.
6. Samet J, Hunt WC, Key C, et al. Choice of cancer therapy varies with age of patient. JAMA. 1986;255:3385-3390.
7. Greenfield S, Blanco DM, Elashoff RM. Patterns of care related to age of breast cancer patients. JAMA. 1987;257:2766-2770.
8. Yancik R, Ries LG, Yates JW. Breast cancer in aging women: a population-based study of contrasts in stage, surgery, and survival. Cancer. 1989:63:976-981.
9. Balducci L, Lyman GH. Adjunctive treatment of cancer in older persons. In: Balducci L, Lyman GH, Ershler WB, eds. Geriatric Oncology. Philadelphia, Pa: JB Lippincott Co; 1992.
10. Wilcoxon F. Individual comparisons by ranking methods. Biometr Bull. 1945;1:80-83.
11. Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc. 1958;52:457-481.
12. Mantel N. Evaluation of survival data and two new rank order statistics arising in its consideration. Cancer Chemother Rep. 1966;50:163-170.
13. Peto R, Pike MC, Armitage P, et al. Design and analysis of randomized clinical trials requiring prolonged observation of each patient. II. Analysis and examples. Br J Cancer. 1977;35:1-39.
14. Cox DR. Regression models and life tables. J R Stat Soc. 1972;34:187-220.
15. Bartolucci AA, Fraser MD. Comparative step-up and composite tests for selecting prognostic indicators associated with survival. Biometr J. 1977;19:437-448.
16. Adami HO, Malker B, Holmberg L, et al. The relation between survival and age at diagnosis in breast cancer. N Engl J Med. 1986;315:559-563.
17. Satariano WA, Belle SH, Swanson GM. The severity of breast cancer at diagnosis: a comparison of age and extent of disease in black and white women. Am J Public Health. 1986;76:779-782.

From the departments of Internal Medicine (GHL, SL, LB, JH), Surgery (NK, DR, CC), Pathology (PB), Radiology (HG) and Epidemiology and Biostatistics (GHL), H. Lee Moffitt Cancer Center & Research Institute at the University of South Florida, Tampa, Fla, and the Albert-Ludwigs Universität (NK), Freiberg, Germany.

Back to Cancer Control Journal Volume 3 Number 5