- Email: [email protected]
The influence of older age on breast cancer treatment decisions and outcome
Int. J. Radiation
Oncology
Bid
PhyT., Vol. 34. No. 3. pp. X15-570. IYYh Copyright I( 1 I YY6 Elsevicr Science Inc. Printed in the USA. All rights reserved 036030 I6/9h s IS.00 + .oo
0360-3016(95)02167-l
ELSEVIER
0 Clinical
Original
THE INFLUENCE
Contribution OF OLDER AGE ON BREAST CANCER DECISIONS AND OUTCOME
TREATMENT
THOMAS E. MERCHANT, D.O., PH.D., * BERYL MCCORMICK, M.D., * JOACHIM YAHALOM, M.D.* AND PATRICK BORGEN, M.D.+ Departments
of *Radiation
Oncology and ’ Surgery, Memorial Sloan-Kettering 1275 York Avenue, New York, NY 10021
Cancer Center,
Purpose: Information concerning the differences between older and younger women with breast cancer, treated with standard therapy, is lacking from many prospective series.The purpose of thii study is to identify factors that htfluence treatment decisionsand determine if women age 65 and older are treated differently than younger women.The outcomesof older women would then be compared to younger women to determine if treatment differences influence outcome. Methods and Materials: The records of 558 women with early invasive breast cancer who were treated with breast conserving surgery and radiation therapy were retrospectively reviewed. Four hundred thirtytwo women under the age of 65 (range: 24-64) and 126 women age 65 and older (range: 65-85) were assessed for treatment differencesincluding breast reexcision, extent of axillary dissection,extent of breast and nodal hradiation, and the useof chemotherapy or hormonal therapy. Differences in the treatment of the two groups were determined and the end points of local control, disease-freesurvival, and overall survival were compared. Median follow-up was 5.5 years. Results: The two treatment groups had identical pathologic TNM staging with the exception that 21% of the older age group and 5% of the younger age group did not undergo axillaty dissection.Women age 65 and older were lesslikely to have a reexcision, extensiveaxillary dissection,chemotherapy, or nodal irradiation. They were more likely to receive hormonal therapy. Reexcisionin older women was positively influenced by a family history of breast cancer and negatively influenced by a history of previous malignancy. Noneof the patients who were treated without an axillary dissectionsuffered a regional recurrence. Although local control was better in older patients, there were no differences in disease-freeor overall survival for the two groups. Discussion:The findings of this study reveal that older patients have sign&ant treatment differences as compared to younger patients; however, despite thesedifferences, similar local control and survival were a&i&d at 5 and 10 years. With the expectedsurvival of older womenincreasing,the prospective evaluation of treatment options for older womenshould be considered. Breast cancer, Breast-conserving therapy, Older-aged women.
( 10). Both Kantrowitz et al. (9) and Goodwin and Samet (8) demonstrated that less aggressive treatment for women in this age group resulted in inferior outcome, but others have not seen this outcome difference. Large group studies are often a source for benchmark numbers. However, information concerning differences between older and younger women, treated with standard therapy, is lacking from many large prospective series because entry criteria frequently excluded patients over the age of 70 years. In this study, older women with breast cancer (age 65 and above) were retrospectively compared to younger women (age less than 65 years) to identify
INTRODUCTION
There is convincing evidence that the treatment of breast cancer in older women is more often influenced by nonmedical and noncancer factors than in younger women. Considering the current trend towards increasing longevity in women older than 65 years (3 ) , it is important to compile information on older women to identify factors that influence treatment decisions and eventual treatment outcomes. Older women are at greatest risk for the development of breast cancer and they experience the greatest decline in survival as a result of breast cancer mortality
This manuscript was presented in part at the American Society for Therapeutic Radiology and Oncology, San Francisco, CA, October 1994.
Reprint requests to: Thomas E. Merchant, D.O., Ph.D. Accepted for publication 2 October 1995.
565
566
I. J. Radiation
Oncology
0 Biology
0 Physics
Table 1. Disease-free survival and overall survival by age for 558 women treated with breast conserving surgery and radiation therapy
Age Disease-free survival Overall survival
n n n n
= = = =
432 < 126 2 432 < 126 2
65 years 65 years 65 years 65 Years
5-Year
lo-Year
89%
75%
90% 93%
75% 81% 83%
90%
Median follow-up 5.5 years.
factors that influenced treatment differences and to determine if these differences influence outcome. METHODS
AND MATERIALS
The records of 558 women with early invasive breast cancer [American Joint Committee on Cancer (AJC) Stage I or II] who were treated with breast-conserving surgery and radiotherapy were retrospectively analyzed to determine if age and clinical characteristics were predictive of treatment differences. All patients were treated from 1980 to 1990 in the department of radiation oncology: 80% also had their surgery at Memorial Sloan-Kettering Cancer Center, and most, if any systemic treatment. The patients were separated into two age groups. One group included 432 women who were under the age of 65 (range: 24-64 years) and the other included the remaining 126 women who were age 65 and older (range 65-85 years). TNM staging in both age groups was identical with the exception that 20.6% (26 out of 126) of the older age group did not undergo nodal dissection and were thus Nx, whereas 5.3% (23 out of 432) of the younger age group were Nx. The final margin status was also similar in both age groups (Fig. la). Clinical and treatment variables The clinical variables of age, family history of breast cancer, prior history of cancer detection by screening mammography, tumor size (T 1 vs. > Tl ) and pathologic lymph node status (N+ vs. N- ) were analyzed to determine their influence on local and systemic treatment variables. The treatment variables included additional surgery beyond initial excisional biopsy, extent of axillary dissection (Level I or no axillary dissection vs. Level II or III dissection), extent of radiation therapy portals (breastonly treatment vs. breast and nodal irradiation), hormonal therapy (tamoxifen vs. no tamoxifen) , and systemic treatment (any hormonal therapy and/or combination chemotherapy vs. no systemic therapy). The clinical and treatment variables that characterize the two patient groups are presented in Figs. la and b. Using logistic regression, the clinical variables were entered in a forward, stepwise manner using p < 0.05 as criterion for inclusion in the model. For evaluation of end
Volume
34, Number
3, 1996
points of local control and survival for the two age groups, Kaplan-Meier methodology was used with WilcoxonBreslow test statistics. The median follow-up was 5.5 years. The date of final breast/axillary surgical procedure prior to radiation therapy was used as the date of diagnosis. All failures, not exclusively first failures, were included in the analysis. RESULTS Table 1 displays the overall survival and disease-free survival of these two patient groups. No differences were observed regarding either end point, establishing similar breast cancer survival outcomes for both age groups, but more importantly, similar overall survival. In other words, in this selected group of patients of 65 years of age and older, competing causes of mortality, which might impact on overall survival and decrease the validity of diseasefree survival, were not observed (Fig. 2b and c). Local control in the treated breast is illustrated in Fig. 2a. At both 5 and 10 years, statistically significant improvement in local control was observed in the older patient cohort. Despite these similarities in outcome, differences in both the clinical and treatment characteristics of the two patient groups were noted. As Fig. la illustrates, the older patient group had a trend towards more positive family histories for breast cancer, more prior cancers, and were more likely to have their lesion detected mammographically than their younger counterparts. However, they were far less likely to have any (in addition to Level I) axillary surgery. With regard to tumor characteristics, both groups were well balanced in terms of Tl vs. T2 lesions, the proportion of patients with involved axillary lymph nodes, and the proportion of patients with known or clear margins. Treatment characteristics also varied by age, as displayed in Fig. lb. Older patients were less likely to have additional surgery, defined as a reexcision after an initial lumpectomy or excisional biopsy. They were also less likely to have nodes dissected beyond Level I. Looking at any systemic therapy, the older patient group was less likely to receive a prescription for this treatment, but as seen in the last two columns in Fig. lb, although this was highly significantly true for systemic chemotherapy, the reverse was true with regards to hormonal therapy, i.e., tamoxifen. Figure lc displays the results of the Cox regression analysis with regards to the older cohort. All of the clinical variables included in the analysis were identified as being independently significant in predicting at least one of the treatment variables. The treatment of older women differed from younger women in that both age alone and age in combination with clinical covariants were predictive of several treatment variables as noted. Of particular note are the impact of age on the recommendation for
Age
influences
breast
cancer
treatment
and
decisions
0 T. E. MERCHANT
Clinical Characteristics
Mammo-
Family History
wphy
Tumor Size > Tl
Involved Nodes
567
er al.
by Age
Clear Margins
No Axillary Surgery
a
Treatment
Characteristics
bv Aae
60%*
additional surgery
dissection
extended RT fields
systemic treatment
chemotherapy
hormonal therapy
b
IAge
Family History
Previous Cancer
Not Detected by
Tumor > T 1
Nodal Involvement
Mammography (‘)Jl-< 0.0001
(-)J < 0.03
<+,pc 0.0001 Chemotherapy
(-) p < 0.0001
Hormonal Therapy
(+)p < 0.0001
-A--=-i I 1
C
Fig. 1. Clinical characteristics (a) and treatment characteristics (b) by group for the 432 women age less than 65 years and 126 women ages 65 years and older evaluated for this study. The results of a Cox regression analysis (c) show the influence of age and other clinical characteristics on treatment selection for older women (age 2 65 years).
568
I. J. Radiation Oncology l Biology l Physics
additional breast surgery, extended axillary node dissection, systemic chemotherapy, and in a positive way, on the use of tamoxifen. A similar Cox regression analysis for the younger age group was significant only in that young age and tumor size greater than 2 cm was highly statistically significant in the prediction for systemic use of chemotherapy (p > 0.0001). No other clinical variables were predicted for the treatments listed for this younger cohort. Data were also analyzed to evaluate the impact of any axillary node surgery in the older cohort with regard to local failure and disease-free survival. Twenty-six patients, with 27 breast cancers, did not undergo any type of surgical node dissection. None of these patients developed regional recurrences. Also, the disease-free and overall survival curves (Figs. 2b and c) were identical for the older cohort with an axillary node dissection at any level, compared to those with no nodal dissection. At 10 years, the survival, with no evidence of disease (NED), for the older cohort without an axillary node dissection was 78%, compared to 72% for the older cohort with an axillary node dissection. DISCUSSION The women in this study may not be representative of all breast patients in the United States because they opted to be treated with breast-conserving surgery at a major cancer referral center. Despite this potentially homogenizing effect, older age alone did appear to influence selection of treatments by our definition. That older age influenced additional surgery, extent of axillary node dissection, and chemotherapy and hormone therapy corresponds intuitively to what we think we know about older women and breast cancer. In other studies, it is well demonstrated that the older patient is less likely to have regular clinical examinations and mammograms, and more likely to be widowed, poor, and multiply dependant. She will often have comorbid medical conditions and, consequently, may be viewed as someone who may not be able to tolerate therapy as well as a younger woman. Differences in clinical presentation may account for the fact that breast self examination, and not screening mammography, is the principle mode of breast cancer detection in older women. Thus, this group often presents with locally advanced disease (4). Biologically, older women with breast cancer have an increased prevalence of hormone receptor positivity in their tumors, and clinically are observed to have a more indolent course than their younger counterparts (5, 6). Screening is used to reduce cancer-related mortality. However, there is little data regarding mass screening programs for women older than the age of 70 years. An age-related decline in screening compliance has been documented (2). In contrast, 40% of the older cohort in our group had their primary lesions detected by mammogra-
Volume 34, Number 3, 1996
phy. This high proportion again supports the hypothesis that our older patient cohort may not be representative of all older women in the country. Regional variations in the treatment of breast cancer have also emerged. Women living in rural settings are likely to be treated differently than those living in urban areas. Regional differences and practice patterns must be assumed to influence the treatment of younger as well older women. One of the best examples of this regional variation is the rate of breast-conserving surgery, which is significantly higher for women who live in New England, in the Mid-Atlantic states, or near large cities in California, when compared to other areas of the country ( 11). There are data that suggest that withholding treatment from older women can result in poor therapeutic outcome; however, geographic variation and treatment and the heterogeneity of the older age groups make it very difficult to define a minimum or standard treatment for these patients. Social characteristics are also predictive for the receipt of definitive treatment (8). Limited to the treatment variables assessed in our study, our data demonstrated that our older women did receive treatment that was demonstrably different from their younger cohorts, but that these differences in these treatments did not appear to influence outcome as we defined it here. A recent report from the University of Michigan presented findings consistent with those of our study ( 1). Seventy-seven patients in that study age 65 years and older were evaluated with regard to 51 younger patients. Again, two treatment groups were matched for stage. Patients were classified as having standard therapy that consisted of local treatment, breast-conserving surgery with axillary dissection and radiation or a modified mastectomy, and systemic treatment, which could consist of either chemotherapy or hormone therapy with Stage II disease. Older patients in that study were less likely to receive this standard treatment, although the analysis of outcome was not presented in this study. Investigators from the M. D. Anderson Cancer Center retrospectively evaluated the clinical behavior and outcome of 184 breast cancer patients older than the age of 69 years ( 14). Only 10% of this patient cohort underwent breast-conserving surgery, and only one-third of the patients had Stage I breast cancer. Of the 160 patients in the group that underwent an axillary node dissection, twothirds of them were found to have positive lymph nodes. However, in contrast to what one would anticipate with this finding, only 30% of all patients received any postoperative radiation therapy, only 8% received hormonal therapy, and 4% received systemic chemotherapy. With a median follow-up of 80 months, however, the cancerspecific survival rate at 7 years was 79% for this older age group. The University of Rochester group retrospectively ana-
Age influences
breast cancer treatment
Local Control
60
i 2”
Median
F/U 67.7 months,
p=O.Ol
>
0
120
QO
Months
a Disease-Free Survival
.9u 90 2 2 2 80 .-s z z 70
--a-
Age
i? a
---+--
Age < 69 years, n=432
@
2 65 pears,
n=126
h
60-
00 Median
50 I 0
FAJ 68.8 months,
QOWOQ
p=n.s.
60
Ii0
Months
b
Overall Survival
90 2 T
3 80 = e g 70
’
,j
-G--
Age 2 6i years, ~126
-
Age < 65 years, ~432
Median 0
F”;’ 68.7 months, 60
p=n.s.
, ‘7 120
Months C
Fig. 2. Kaplan-Meier curves of the duration of local control (a), disease-freesurvival (b), and overall survival (c) for womenages< 65 years vs. womenages2 65 years treated with breastconservingsurgeryand postoperativeexternalbeam radiation therapy.
and decisions
0 T. E. MERCHANI
et al.
569
lyzed outcome in patients greater than 60 years of age, treated with segmentalmastectomy, with or without postoperative radiation (9). Of note, their results demonstrated that there were marked differences in outcome depending on whether or not radiation was applied, with the advantage favoring radiation therapy. Two prospective randomized trials have compared mastectomy or wide local excision to tamoxifen in elderly women ages 70 years and older. Tamoxifen was selected as a treatment based upon the reported experiences, indicating that tamoxifen could be preferentially used for infirm patients with good response rates. In the first of the two trials (7 ) , 166 women with medically operable, surgically resectable breast cancer were randomized to surgery (wide local excision or mastectomy) or tamoxifen. Of the 56 patients in the surgical arm, 68% had Tl or T2 tumors and 82% were treated with wide local excision. The remaining patients were treated with mastectomy. With a median follow-up of 3 years, local progression was identified in 25% of the tamoxifen group and 38% in the surgical group. There was no difference in disease-free survival. Recurrences were generally managed by cross-over to the opposite treatment arm. In the second trial, Robertson ef al. ( 12) reported the results of a similar randomized trial where 135 consecutive patients were treated with wedge resection or tamoxifen. Criteria for inclusion included women with tumors less than 5 cm. Recurrences were managed with further local treatment and tamoxifen according to individual circumstances. With a median follow-up of 2 years, local control rates of 70% in the surgical arm and 47% in the tamoxifen arm were achieved. In an update ( 13) , the failure of local regional control was significantly worse for the tamoxifen group at a median follow-up of 65 months. In an effort to further define an acceptable minimum treatment, two randomized trials have been performed to determine the outcomes of “low-risk,” node-negative patients with breast cancer treated with breast-conserving surgery with or without adjuvant radiation therapy. These trials included elderly patients who were analyzed in a subset of patients with ages greater than 50 or 55 years. In the study reported by Clark and others (4 ), breast irradiation reduced relapse from 25.7 to 5.5%. A subset of women at low risk for relapse could not be defined based upon age, tumor size, hormone receptor status, differentiation, and nuclear grade. Veronesi ef nl. ( 15) randomized node-negative women with tumors lessthan 2.5 cm to quadrantectomy with or without adjuvant radiation therapy. With a median follow-up of 39 months, the incidence of local recurrence was reduced from 8.8 to 0.3% with the addition of radiation therapy. However, for women age greater than 55 years, the local recurrence rate was only 3.8% in the women who did not receive radiation therapy.
570
I. J. Radiation Oncology 0 Biology 0 Physics
Volume 34, Number 3, 1996
CONCLUSION
While all of the patients in our study received breast radiation, our small numbers suggestthat in selectedolder patients with clinically negative axilla, radiation can be substituted for an axillary node dissection, providing excellent local control. Although these patients lack the hypothetical benefit of the prognostic knowledge of the status of the nodes, this did not impact on local control, or diseasefree survival in our patient cohort. With the expected survival of the American woman increasing, more studies should be directed to prospectively evaluating treatment options for the older woman with breast cancer.
The findings in this study reveal that older patients did have significant treatment differences when compared to their younger counterparts. Although all local treatment to the breast, i.e., limited resection and radiation, was identical, older patients were less likely to have nodal radiation and axillary node dissections.As well, they were more likely to be treated by tamoxifen as a systemic therapy, and lesslikely to receive chemotherapy. Despite these differences, similar local control and survival was achieved at 5 and 10 years in both groups.
REFERENCES 1. August, D. A.; Rea,T.; Sondak,V. K. Age-relateddifferencesin breastcancertreatment.Ann. Surg. Oncol. 1:4552; 1994. 2. Beghe,C.; Balducci, L.; Cohen,H. Secondaryprevention of breast cancer in tbe older woman: Issuesrelated to screening.CancerControl 1:320-326; 1994. 3. Brody, J.; Cassell,C. K. Normal humanaging theory, demography.In: Cassel,C. K.; Riesenberg,D. E.; Sorensen, L., eds.Geriatric medicine:Principlesandpractice,2nd ed. New York: SpringerVerlag; 1990:16-27. 4. Clark, R. M.; McCulloch, P. B.; Levine, M. N.; Lipa, M.; Wilkinson, R. H.; Mahoney, L. J.; Basrur, V. R. Nair, B. D. Randomizedclinical trials to assess the effectiveness of breast irradiation following lumpectomy and axillary dissectionfor node-negativebreastcancer.J. Natl. Cancer Inst. 84:683-689; 1992. 5. Foster,R. S. Managementof primary breastcancerin older patients: Treatment options. Cancer Control 1:339-343; 1994. 6. Foster,R. S.; Lang, S. P.; Costanza,M. C.; Worden,J. K; Haines,C. R.; Yates, J. W. Breastself-examinationpractices and breastcancer stage.N. Engl. J. Med. 299:265270; 1978. 7. Gazet, J. C.; Markopoulos, Ch.; Ford, H. T.; Coombes, R. C.; Bland, J. M.; Dixon, R. C. Prospectiverandomised trial of tamoxifen vs. surgeryin elderly patientswith breast cancer.Lancet i:679-681; 1988. 8. Goodwin,J. S.; Samet,J. M. Carereceivedby olderwomen
diagnosedwith breastcancer.CancerControl 1:313-319; 1994. 9. Kantorowitz, D. A.; Poulter, C. A.; Sischy, B.; Paterson, E.; Sobel, S. H.; Rubin, P.; Dvoretsky, P. A.; Mishalak, W.; Doane,K. L. Treatmentof breastcanceramongelderly womenwith segmentalmastectomyor segmentalmastectomy plus postoperativeradiotherapy.Int. J. Radit. Oncol. Biol. Phys. 15:263-270; 1988. 10. Manton, K. G.; Wrigley, J. M.; Cohen, H. J.; Woodbury, M. A. Cancermortality, aging and patternsof comorbidity in the United States:1968- 1986.J. Gerontol.46:225-234; 1991. 11. Nattinger, A. B.; Goodwin, J. S. Geographicand hospital variation in tbe managementof older womenwith breast cancer.CancerControl 1:334-338; 1994. 12. Robertson,J. F. R.; Ellis, I. 0.; Elston, C. W.; Blarney, R. W. Mastectomyor tamoxifen asinitial therapyfor operable breastcancerin elderly patients:5-year follow-up. Eur. J. Cancer28A:908-910; 1992. 13. Robertson,J. F. R.; Todd, J. H.; Ellis, I. 0.; Elston,C. W.; Blarney, R. W. Comparisonof mastectomywith tamoxifen for treatingelderly patientswith operablebreastcancer.Br. Med. J. 297:511-514; 1988. 14. Singletary, S. E.; Shallenberger,R.; Guinee, V. F. Breast cancerin tbe elderly. Ann. Surg. 218:667-671; 1993. 15. Veronesi,U.; Luini, A.; Del Vecchio, M.; Greco, M.; Galimberti, V.; Merson,M.; Rilke, F.; Sacchini,V.; Saccozzi,R.; Savio, T. Radiotherapyafter breast-preservingsurgery in womenwith localizedcancerof the breast.N. Engl. J. Med. 328:1587- 1591; 1993.
Oncology
Bid
PhyT., Vol. 34. No. 3. pp. X15-570. IYYh Copyright I( 1 I YY6 Elsevicr Science Inc. Printed in the USA. All rights reserved 036030 I6/9h s IS.00 + .oo
0360-3016(95)02167-l
ELSEVIER
0 Clinical
Original
THE INFLUENCE
Contribution OF OLDER AGE ON BREAST CANCER DECISIONS AND OUTCOME
TREATMENT
THOMAS E. MERCHANT, D.O., PH.D., * BERYL MCCORMICK, M.D., * JOACHIM YAHALOM, M.D.* AND PATRICK BORGEN, M.D.+ Departments
of *Radiation
Oncology and ’ Surgery, Memorial Sloan-Kettering 1275 York Avenue, New York, NY 10021
Cancer Center,
Purpose: Information concerning the differences between older and younger women with breast cancer, treated with standard therapy, is lacking from many prospective series.The purpose of thii study is to identify factors that htfluence treatment decisionsand determine if women age 65 and older are treated differently than younger women.The outcomesof older women would then be compared to younger women to determine if treatment differences influence outcome. Methods and Materials: The records of 558 women with early invasive breast cancer who were treated with breast conserving surgery and radiation therapy were retrospectively reviewed. Four hundred thirtytwo women under the age of 65 (range: 24-64) and 126 women age 65 and older (range: 65-85) were assessed for treatment differencesincluding breast reexcision, extent of axillary dissection,extent of breast and nodal hradiation, and the useof chemotherapy or hormonal therapy. Differences in the treatment of the two groups were determined and the end points of local control, disease-freesurvival, and overall survival were compared. Median follow-up was 5.5 years. Results: The two treatment groups had identical pathologic TNM staging with the exception that 21% of the older age group and 5% of the younger age group did not undergo axillaty dissection.Women age 65 and older were lesslikely to have a reexcision, extensiveaxillary dissection,chemotherapy, or nodal irradiation. They were more likely to receive hormonal therapy. Reexcisionin older women was positively influenced by a family history of breast cancer and negatively influenced by a history of previous malignancy. Noneof the patients who were treated without an axillary dissectionsuffered a regional recurrence. Although local control was better in older patients, there were no differences in disease-freeor overall survival for the two groups. Discussion:The findings of this study reveal that older patients have sign&ant treatment differences as compared to younger patients; however, despite thesedifferences, similar local control and survival were a&i&d at 5 and 10 years. With the expectedsurvival of older womenincreasing,the prospective evaluation of treatment options for older womenshould be considered. Breast cancer, Breast-conserving therapy, Older-aged women.
( 10). Both Kantrowitz et al. (9) and Goodwin and Samet (8) demonstrated that less aggressive treatment for women in this age group resulted in inferior outcome, but others have not seen this outcome difference. Large group studies are often a source for benchmark numbers. However, information concerning differences between older and younger women, treated with standard therapy, is lacking from many large prospective series because entry criteria frequently excluded patients over the age of 70 years. In this study, older women with breast cancer (age 65 and above) were retrospectively compared to younger women (age less than 65 years) to identify
INTRODUCTION
There is convincing evidence that the treatment of breast cancer in older women is more often influenced by nonmedical and noncancer factors than in younger women. Considering the current trend towards increasing longevity in women older than 65 years (3 ) , it is important to compile information on older women to identify factors that influence treatment decisions and eventual treatment outcomes. Older women are at greatest risk for the development of breast cancer and they experience the greatest decline in survival as a result of breast cancer mortality
This manuscript was presented in part at the American Society for Therapeutic Radiology and Oncology, San Francisco, CA, October 1994.
Reprint requests to: Thomas E. Merchant, D.O., Ph.D. Accepted for publication 2 October 1995.
565
566
I. J. Radiation
Oncology
0 Biology
0 Physics
Table 1. Disease-free survival and overall survival by age for 558 women treated with breast conserving surgery and radiation therapy
Age Disease-free survival Overall survival
n n n n
= = = =
432 < 126 2 432 < 126 2
65 years 65 years 65 years 65 Years
5-Year
lo-Year
89%
75%
90% 93%
75% 81% 83%
90%
Median follow-up 5.5 years.
factors that influenced treatment differences and to determine if these differences influence outcome. METHODS
AND MATERIALS
The records of 558 women with early invasive breast cancer [American Joint Committee on Cancer (AJC) Stage I or II] who were treated with breast-conserving surgery and radiotherapy were retrospectively analyzed to determine if age and clinical characteristics were predictive of treatment differences. All patients were treated from 1980 to 1990 in the department of radiation oncology: 80% also had their surgery at Memorial Sloan-Kettering Cancer Center, and most, if any systemic treatment. The patients were separated into two age groups. One group included 432 women who were under the age of 65 (range: 24-64 years) and the other included the remaining 126 women who were age 65 and older (range 65-85 years). TNM staging in both age groups was identical with the exception that 20.6% (26 out of 126) of the older age group did not undergo nodal dissection and were thus Nx, whereas 5.3% (23 out of 432) of the younger age group were Nx. The final margin status was also similar in both age groups (Fig. la). Clinical and treatment variables The clinical variables of age, family history of breast cancer, prior history of cancer detection by screening mammography, tumor size (T 1 vs. > Tl ) and pathologic lymph node status (N+ vs. N- ) were analyzed to determine their influence on local and systemic treatment variables. The treatment variables included additional surgery beyond initial excisional biopsy, extent of axillary dissection (Level I or no axillary dissection vs. Level II or III dissection), extent of radiation therapy portals (breastonly treatment vs. breast and nodal irradiation), hormonal therapy (tamoxifen vs. no tamoxifen) , and systemic treatment (any hormonal therapy and/or combination chemotherapy vs. no systemic therapy). The clinical and treatment variables that characterize the two patient groups are presented in Figs. la and b. Using logistic regression, the clinical variables were entered in a forward, stepwise manner using p < 0.05 as criterion for inclusion in the model. For evaluation of end
Volume
34, Number
3, 1996
points of local control and survival for the two age groups, Kaplan-Meier methodology was used with WilcoxonBreslow test statistics. The median follow-up was 5.5 years. The date of final breast/axillary surgical procedure prior to radiation therapy was used as the date of diagnosis. All failures, not exclusively first failures, were included in the analysis. RESULTS Table 1 displays the overall survival and disease-free survival of these two patient groups. No differences were observed regarding either end point, establishing similar breast cancer survival outcomes for both age groups, but more importantly, similar overall survival. In other words, in this selected group of patients of 65 years of age and older, competing causes of mortality, which might impact on overall survival and decrease the validity of diseasefree survival, were not observed (Fig. 2b and c). Local control in the treated breast is illustrated in Fig. 2a. At both 5 and 10 years, statistically significant improvement in local control was observed in the older patient cohort. Despite these similarities in outcome, differences in both the clinical and treatment characteristics of the two patient groups were noted. As Fig. la illustrates, the older patient group had a trend towards more positive family histories for breast cancer, more prior cancers, and were more likely to have their lesion detected mammographically than their younger counterparts. However, they were far less likely to have any (in addition to Level I) axillary surgery. With regard to tumor characteristics, both groups were well balanced in terms of Tl vs. T2 lesions, the proportion of patients with involved axillary lymph nodes, and the proportion of patients with known or clear margins. Treatment characteristics also varied by age, as displayed in Fig. lb. Older patients were less likely to have additional surgery, defined as a reexcision after an initial lumpectomy or excisional biopsy. They were also less likely to have nodes dissected beyond Level I. Looking at any systemic therapy, the older patient group was less likely to receive a prescription for this treatment, but as seen in the last two columns in Fig. lb, although this was highly significantly true for systemic chemotherapy, the reverse was true with regards to hormonal therapy, i.e., tamoxifen. Figure lc displays the results of the Cox regression analysis with regards to the older cohort. All of the clinical variables included in the analysis were identified as being independently significant in predicting at least one of the treatment variables. The treatment of older women differed from younger women in that both age alone and age in combination with clinical covariants were predictive of several treatment variables as noted. Of particular note are the impact of age on the recommendation for
Age
influences
breast
cancer
treatment
and
decisions
0 T. E. MERCHANT
Clinical Characteristics
Mammo-
Family History
wphy
Tumor Size > Tl
Involved Nodes
567
er al.
by Age
Clear Margins
No Axillary Surgery
a
Treatment
Characteristics
bv Aae
60%*
additional surgery
dissection
extended RT fields
systemic treatment
chemotherapy
hormonal therapy
b
IAge
Family History
Previous Cancer
Not Detected by
Tumor > T 1
Nodal Involvement
Mammography (‘)Jl-< 0.0001
(-)J < 0.03
<+,pc 0.0001 Chemotherapy
(-) p < 0.0001
Hormonal Therapy
(+)p < 0.0001
-A--=-i I 1
C
Fig. 1. Clinical characteristics (a) and treatment characteristics (b) by group for the 432 women age less than 65 years and 126 women ages 65 years and older evaluated for this study. The results of a Cox regression analysis (c) show the influence of age and other clinical characteristics on treatment selection for older women (age 2 65 years).
568
I. J. Radiation Oncology l Biology l Physics
additional breast surgery, extended axillary node dissection, systemic chemotherapy, and in a positive way, on the use of tamoxifen. A similar Cox regression analysis for the younger age group was significant only in that young age and tumor size greater than 2 cm was highly statistically significant in the prediction for systemic use of chemotherapy (p > 0.0001). No other clinical variables were predicted for the treatments listed for this younger cohort. Data were also analyzed to evaluate the impact of any axillary node surgery in the older cohort with regard to local failure and disease-free survival. Twenty-six patients, with 27 breast cancers, did not undergo any type of surgical node dissection. None of these patients developed regional recurrences. Also, the disease-free and overall survival curves (Figs. 2b and c) were identical for the older cohort with an axillary node dissection at any level, compared to those with no nodal dissection. At 10 years, the survival, with no evidence of disease (NED), for the older cohort without an axillary node dissection was 78%, compared to 72% for the older cohort with an axillary node dissection. DISCUSSION The women in this study may not be representative of all breast patients in the United States because they opted to be treated with breast-conserving surgery at a major cancer referral center. Despite this potentially homogenizing effect, older age alone did appear to influence selection of treatments by our definition. That older age influenced additional surgery, extent of axillary node dissection, and chemotherapy and hormone therapy corresponds intuitively to what we think we know about older women and breast cancer. In other studies, it is well demonstrated that the older patient is less likely to have regular clinical examinations and mammograms, and more likely to be widowed, poor, and multiply dependant. She will often have comorbid medical conditions and, consequently, may be viewed as someone who may not be able to tolerate therapy as well as a younger woman. Differences in clinical presentation may account for the fact that breast self examination, and not screening mammography, is the principle mode of breast cancer detection in older women. Thus, this group often presents with locally advanced disease (4). Biologically, older women with breast cancer have an increased prevalence of hormone receptor positivity in their tumors, and clinically are observed to have a more indolent course than their younger counterparts (5, 6). Screening is used to reduce cancer-related mortality. However, there is little data regarding mass screening programs for women older than the age of 70 years. An age-related decline in screening compliance has been documented (2). In contrast, 40% of the older cohort in our group had their primary lesions detected by mammogra-
Volume 34, Number 3, 1996
phy. This high proportion again supports the hypothesis that our older patient cohort may not be representative of all older women in the country. Regional variations in the treatment of breast cancer have also emerged. Women living in rural settings are likely to be treated differently than those living in urban areas. Regional differences and practice patterns must be assumed to influence the treatment of younger as well older women. One of the best examples of this regional variation is the rate of breast-conserving surgery, which is significantly higher for women who live in New England, in the Mid-Atlantic states, or near large cities in California, when compared to other areas of the country ( 11). There are data that suggest that withholding treatment from older women can result in poor therapeutic outcome; however, geographic variation and treatment and the heterogeneity of the older age groups make it very difficult to define a minimum or standard treatment for these patients. Social characteristics are also predictive for the receipt of definitive treatment (8). Limited to the treatment variables assessed in our study, our data demonstrated that our older women did receive treatment that was demonstrably different from their younger cohorts, but that these differences in these treatments did not appear to influence outcome as we defined it here. A recent report from the University of Michigan presented findings consistent with those of our study ( 1). Seventy-seven patients in that study age 65 years and older were evaluated with regard to 51 younger patients. Again, two treatment groups were matched for stage. Patients were classified as having standard therapy that consisted of local treatment, breast-conserving surgery with axillary dissection and radiation or a modified mastectomy, and systemic treatment, which could consist of either chemotherapy or hormone therapy with Stage II disease. Older patients in that study were less likely to receive this standard treatment, although the analysis of outcome was not presented in this study. Investigators from the M. D. Anderson Cancer Center retrospectively evaluated the clinical behavior and outcome of 184 breast cancer patients older than the age of 69 years ( 14). Only 10% of this patient cohort underwent breast-conserving surgery, and only one-third of the patients had Stage I breast cancer. Of the 160 patients in the group that underwent an axillary node dissection, twothirds of them were found to have positive lymph nodes. However, in contrast to what one would anticipate with this finding, only 30% of all patients received any postoperative radiation therapy, only 8% received hormonal therapy, and 4% received systemic chemotherapy. With a median follow-up of 80 months, however, the cancerspecific survival rate at 7 years was 79% for this older age group. The University of Rochester group retrospectively ana-
Age influences
breast cancer treatment
Local Control
60
i 2”
Median
F/U 67.7 months,
p=O.Ol
>
0
120
QO
Months
a Disease-Free Survival
.9u 90 2 2 2 80 .-s z z 70
--a-
Age
i? a
---+--
Age < 69 years, n=432
@
2 65 pears,
n=126
h
60-
00 Median
50 I 0
FAJ 68.8 months,
QOWOQ
p=n.s.
60
Ii0
Months
b
Overall Survival
90 2 T
3 80 = e g 70
’
,j
-G--
Age 2 6i years, ~126
-
Age < 65 years, ~432
Median 0
F”;’ 68.7 months, 60
p=n.s.
, ‘7 120
Months C
Fig. 2. Kaplan-Meier curves of the duration of local control (a), disease-freesurvival (b), and overall survival (c) for womenages< 65 years vs. womenages2 65 years treated with breastconservingsurgeryand postoperativeexternalbeam radiation therapy.
and decisions
0 T. E. MERCHANI
et al.
569
lyzed outcome in patients greater than 60 years of age, treated with segmentalmastectomy, with or without postoperative radiation (9). Of note, their results demonstrated that there were marked differences in outcome depending on whether or not radiation was applied, with the advantage favoring radiation therapy. Two prospective randomized trials have compared mastectomy or wide local excision to tamoxifen in elderly women ages 70 years and older. Tamoxifen was selected as a treatment based upon the reported experiences, indicating that tamoxifen could be preferentially used for infirm patients with good response rates. In the first of the two trials (7 ) , 166 women with medically operable, surgically resectable breast cancer were randomized to surgery (wide local excision or mastectomy) or tamoxifen. Of the 56 patients in the surgical arm, 68% had Tl or T2 tumors and 82% were treated with wide local excision. The remaining patients were treated with mastectomy. With a median follow-up of 3 years, local progression was identified in 25% of the tamoxifen group and 38% in the surgical group. There was no difference in disease-free survival. Recurrences were generally managed by cross-over to the opposite treatment arm. In the second trial, Robertson ef al. ( 12) reported the results of a similar randomized trial where 135 consecutive patients were treated with wedge resection or tamoxifen. Criteria for inclusion included women with tumors less than 5 cm. Recurrences were managed with further local treatment and tamoxifen according to individual circumstances. With a median follow-up of 2 years, local control rates of 70% in the surgical arm and 47% in the tamoxifen arm were achieved. In an update ( 13) , the failure of local regional control was significantly worse for the tamoxifen group at a median follow-up of 65 months. In an effort to further define an acceptable minimum treatment, two randomized trials have been performed to determine the outcomes of “low-risk,” node-negative patients with breast cancer treated with breast-conserving surgery with or without adjuvant radiation therapy. These trials included elderly patients who were analyzed in a subset of patients with ages greater than 50 or 55 years. In the study reported by Clark and others (4 ), breast irradiation reduced relapse from 25.7 to 5.5%. A subset of women at low risk for relapse could not be defined based upon age, tumor size, hormone receptor status, differentiation, and nuclear grade. Veronesi ef nl. ( 15) randomized node-negative women with tumors lessthan 2.5 cm to quadrantectomy with or without adjuvant radiation therapy. With a median follow-up of 39 months, the incidence of local recurrence was reduced from 8.8 to 0.3% with the addition of radiation therapy. However, for women age greater than 55 years, the local recurrence rate was only 3.8% in the women who did not receive radiation therapy.
570
I. J. Radiation Oncology 0 Biology 0 Physics
Volume 34, Number 3, 1996
CONCLUSION
While all of the patients in our study received breast radiation, our small numbers suggestthat in selectedolder patients with clinically negative axilla, radiation can be substituted for an axillary node dissection, providing excellent local control. Although these patients lack the hypothetical benefit of the prognostic knowledge of the status of the nodes, this did not impact on local control, or diseasefree survival in our patient cohort. With the expected survival of the American woman increasing, more studies should be directed to prospectively evaluating treatment options for the older woman with breast cancer.
The findings in this study reveal that older patients did have significant treatment differences when compared to their younger counterparts. Although all local treatment to the breast, i.e., limited resection and radiation, was identical, older patients were less likely to have nodal radiation and axillary node dissections.As well, they were more likely to be treated by tamoxifen as a systemic therapy, and lesslikely to receive chemotherapy. Despite these differences, similar local control and survival was achieved at 5 and 10 years in both groups.
REFERENCES 1. August, D. A.; Rea,T.; Sondak,V. K. Age-relateddifferencesin breastcancertreatment.Ann. Surg. Oncol. 1:4552; 1994. 2. Beghe,C.; Balducci, L.; Cohen,H. Secondaryprevention of breast cancer in tbe older woman: Issuesrelated to screening.CancerControl 1:320-326; 1994. 3. Brody, J.; Cassell,C. K. Normal humanaging theory, demography.In: Cassel,C. K.; Riesenberg,D. E.; Sorensen, L., eds.Geriatric medicine:Principlesandpractice,2nd ed. New York: SpringerVerlag; 1990:16-27. 4. Clark, R. M.; McCulloch, P. B.; Levine, M. N.; Lipa, M.; Wilkinson, R. H.; Mahoney, L. J.; Basrur, V. R. Nair, B. D. Randomizedclinical trials to assess the effectiveness of breast irradiation following lumpectomy and axillary dissectionfor node-negativebreastcancer.J. Natl. Cancer Inst. 84:683-689; 1992. 5. Foster,R. S. Managementof primary breastcancerin older patients: Treatment options. Cancer Control 1:339-343; 1994. 6. Foster,R. S.; Lang, S. P.; Costanza,M. C.; Worden,J. K; Haines,C. R.; Yates, J. W. Breastself-examinationpractices and breastcancer stage.N. Engl. J. Med. 299:265270; 1978. 7. Gazet, J. C.; Markopoulos, Ch.; Ford, H. T.; Coombes, R. C.; Bland, J. M.; Dixon, R. C. Prospectiverandomised trial of tamoxifen vs. surgeryin elderly patientswith breast cancer.Lancet i:679-681; 1988. 8. Goodwin,J. S.; Samet,J. M. Carereceivedby olderwomen
diagnosedwith breastcancer.CancerControl 1:313-319; 1994. 9. Kantorowitz, D. A.; Poulter, C. A.; Sischy, B.; Paterson, E.; Sobel, S. H.; Rubin, P.; Dvoretsky, P. A.; Mishalak, W.; Doane,K. L. Treatmentof breastcanceramongelderly womenwith segmentalmastectomyor segmentalmastectomy plus postoperativeradiotherapy.Int. J. Radit. Oncol. Biol. Phys. 15:263-270; 1988. 10. Manton, K. G.; Wrigley, J. M.; Cohen, H. J.; Woodbury, M. A. Cancermortality, aging and patternsof comorbidity in the United States:1968- 1986.J. Gerontol.46:225-234; 1991. 11. Nattinger, A. B.; Goodwin, J. S. Geographicand hospital variation in tbe managementof older womenwith breast cancer.CancerControl 1:334-338; 1994. 12. Robertson,J. F. R.; Ellis, I. 0.; Elston, C. W.; Blarney, R. W. Mastectomyor tamoxifen asinitial therapyfor operable breastcancerin elderly patients:5-year follow-up. Eur. J. Cancer28A:908-910; 1992. 13. Robertson,J. F. R.; Todd, J. H.; Ellis, I. 0.; Elston,C. W.; Blarney, R. W. Comparisonof mastectomywith tamoxifen for treatingelderly patientswith operablebreastcancer.Br. Med. J. 297:511-514; 1988. 14. Singletary, S. E.; Shallenberger,R.; Guinee, V. F. Breast cancerin tbe elderly. Ann. Surg. 218:667-671; 1993. 15. Veronesi,U.; Luini, A.; Del Vecchio, M.; Greco, M.; Galimberti, V.; Merson,M.; Rilke, F.; Sacchini,V.; Saccozzi,R.; Savio, T. Radiotherapyafter breast-preservingsurgery in womenwith localizedcancerof the breast.N. Engl. J. Med. 328:1587- 1591; 1993.