Fax: (919) 668-7345
Age-related disparities in the use of radiotherapy for treatment of localized soft tissue sarcoma†
Article first published online: 8 MAR 2011
Copyright © 2011 American Cancer Society
Volume 117, Issue 17, pages 4033–4040, 1 September 2011
How to Cite
Horton, J. K., Gleason, J. F., Klepin, H. D., Isom, S., Fried, D. B. and Geiger, A. M. (2011), Age-related disparities in the use of radiotherapy for treatment of localized soft tissue sarcoma. Cancer, 117: 4033–4040. doi: 10.1002/cncr.25996
Preliminary results from this study were presented at the 94th Scientific Assembly and Annual Meeting of the Radiological Society of North America; Chicago, Illinois; November 30 to December 5, 2008.
- Issue published online: 19 AUG 2011
- Article first published online: 8 MAR 2011
- Manuscript Accepted: 16 DEC 2010
- Manuscript Revised: 7 DEC 2010
- Manuscript Received: 28 AUG 2010
- age factors;
- healthcare disparities;
Many elderly patients with cancer experience increased cancer-related morbidity and mortality compared with younger patients. In soft tissue sarcoma, adjuvant radiotherapy is an integral part of definitive therapy for limb preservation. The authors of this report hypothesized that age-related disparities exist in the use of radiation.
Surveillance, Epidemiology, and End Results (SEER) data were used to conduct a retrospective cohort study among patients aged ≥25 years who were diagnosed from 1998 to 2004 with nonmetastatic, biopsy-proven, high-grade soft tissue sarcoma of the extremities and underwent a limb-sparing procedure. Patients were stratified according to age (ages <50 years, 50-70 years, and >70 years). Logistic regression was used to determine the association between age and the receipt of radiotherapy adjusting for histology, tumor location, tumor size, surgery, sex, race, and marital status. A Cox proportional hazards model was used to compare disease-specific and all-cause mortality.
Among 1354 eligible patients; 37.1% were aged >70 years, 44.3% were women, and 84.4% were Caucasian. Although 73.8% of the cohort received radiotherapy, receipt decreased from 78.2% among patients aged <50 years to 69.6% among patients aged >70 years (test for trend; P = .006). After adjusting for demographic and tumor factors, older patients remained less likely to receive radiotherapy (odds ratio, 0.66; 95% confidence interval, 0.47-0.92) and more likely to experience disease-specific death (hazard ratio, 2.4; 95% confidence interval, 1.4-4.1) compared with the youngest group.
Older adults appeared to be less likely to receive definitive therapy for soft tissue sarcoma of the extremities. In the absence of clinical trials and treatment guidelines tailored to this population, under treatment may disadvantage elderly patients, who have increased cancer-related morbidity and mortality. Cancer 2011;. © 2011 American Cancer Society.
The US Census Bureau indicates that the number of adults aged >65 years will increase dramatically from the current estimate of 40 million to >70 million by 2030.1 Cancer disproportionately afflicts older adults, and >50% of all new cancer diagnoses between 2003 and 2007 occurred among adults aged ≥65 years.2 The average life expectancy for a healthy adult aged 70 years in the United States is 12 to 15 years, and even those with significant comorbidities are expected to live for 5 to 7 years.3
Despite this significant predicted longevity, many elderly patients with cancer experience increased cancer-related morbidity and mortality compared with younger patients.4-7 Surveillance, Epidemiology, and End Results (SEER) statistics from the National Cancer Institute indicate that the age-adjusted mortality rate for all cancer types combined is 15 times greater for older adults.4, 7 Several studies have suggested that under-representation in clinical trials8, 9 and the administration of substandard treatment based on age contributes to this outcome disparity in older adults.10-13
Approximately 10,000 individuals will be diagnosed with cancer of the soft tissues in 2008. In those who have high-grade tumors of the extremities, radiotherapy (RT) has been established as an integral component of treatment through several randomized trials,14-16 which demonstrated a significant reduction in the rates of local recurrence when RT was combined with limb-sparing surgery. When no RT is given, the majority of local failures occur within the first 2 to 4 years. Because salvage options are limited, failure to prevent local recurrences with RT certainly will increase patient morbidity and may increase the probability of dying from sarcoma.
The randomized trials that established limb preservation in soft tissue sarcoma included very few patients aged >70 years. It is unclear whether this standard of care is applied uniformly across age groups in everyday practice. This is particularly relevant because 38% of new soft tissue cancer diagnoses are in adults aged >65 years.2 Thus, in the current study, we explored potential age-related disparities using an existing database to evaluate the hypothesis that older adults with soft tissue sarcoma of the extremity would be less likely to receive adjuvant RT compared with their younger counterparts and that this would translate into decreased survival.
MATERIALS AND METHODS
Setting and Design
We conducted a retrospective cohort study using the National Cancer Institute's SEER database to obtain data on the treatment of patients with localized soft tissue sarcoma in the United States. The institutional review board at Wake Forest University determined that the use of deidentified data in this analysis did not meet the federal definition of human subject research and, thus, was exempt from review.
Our cohort included patients with nonmetastatic, biopsy-proven, high-grade soft tissue sarcoma of the extremity who underwent a limb-sparing surgical procedure between the years 1998 and 2004. The seminal trials that established the use of adjuvant RT were published during or before 1998; thus, we expected that treatment patterns would have been relatively stable during this period. In addition, we restricted the cohort to adults aged ≥25 years. Tumors occuring in younger patients were felt to represent a distinct population with differing prognoses and treatment options. Patients with tumors that represented unique biologic entities, such as Kaposi sarcoma, were excluded for the same reason. Finally, we excluded patients who were missing data on surgery or RT, age, sex, race/ethnicity, or marital status.
SEER is a population-based tumor registry that contains demographic, tumor, treatment, and vital status information on all incident cancers within specific geographic regions. Currently, the data contained in the SEER Program cover approximately 25% of the US population. Data from the following registries were included as available: Atlanta, Connecticut, Detroit, Hawaii, Iowa, New Mexico, San Francisco-Oakland, Seattle-Puget Sound, Utah, Los Angeles, San Jose-Monterey, Rural Georgia, Alaska, Greater California, Kentucky, Louisiana, and New Jersey.
Our main interest was in the association between age, receipt of RT, and correlation with patient outcomes. Patients who satisfied the inclusion criteria were stratified into 3 groups according to age (ages <50 years, 50-70 years, and >70 years) in order to evaluate the receipt of RT and treatment-related outcomes in these 3 distinct populations. Age brackets were chosen to reflect geriatric oncology practice trends, consistent with recent publications in which age 70 years is used to define a geriatric population.17, 18 RT was included in any of the following forms: beam RT, radioactive implants, radioisotopes, a combination of the previous 3 forms, or RT not otherwise specified.
We controlled for additional variables as follows: Three racial groups were considered: white, black, and other. Marital status was divided into married or currently not married, single, separated/divorced, and widowed. Histology was categorized using the SEER Histology Validation List. Selected codes are reported in Table 1. High-grade tumors, those defined as moderately differentiated, poorly differentiated, or undifferentiated, were defined according to the Term Grade SEER code (2/3, 3/3, and 2/2). Surgery was considered limb-sparing when it was coded as a local excision or radical excision/resection of lesion with limb salvage. Tumors that measured ≤50 mm in greatest dimension were recorded as small, and tumors that measured >50 mm in greatest dimension were recorded as large. International Classification of Diseases (9th and 10th revision) codes were used to determine cause of death. The codes for death from soft tissue malignancy (164.1, 171, C47, C49, C38.0, and C45.2) were selected to determine disease-specific survival.
|Codea||Histology||No. of Patients (%)|
|8800||Sarcoma, NOS||76 (5.6)|
|8801||Spindle cell sarcoma||45 (3.3)|
|8802||Giant cell sarcoma||62 (4.6)|
|8803||Small cell sarcoma||1 (0.1)|
|8804||Epithelioid sarcoma||10 (0.7)|
|8805||Undifferentiated sarcoma||2 (0.2)|
|8810||Fibrosarcoma, NOS||17 (1.3)|
|8830||Fibrous histiocytoma, malignant||576 (42.5)|
|8832||Dermatofibrosarcoma, NOS||3 (0.2)|
|8850||Liposarcoma, NOS||20 (1.5)|
|8851||Liposarcoma, well differentiated||1 (0.1)|
|8852||Myxoid liposarcoma||43 (3.2)|
|8853||Round cell liposarcoma||27 (2)|
|8854||Pleomorphic liposarcoma||74 (5.5)|
|8855||Mixed type liposarcoma||13 (1)|
|8858||Dedifferentiated liposarcoma||33 (2.4)|
|8890||Leiomyosarcoma, NOS||141 (10.4)|
|8891||Epithelioid leiomyosarcoma||6 (0.4)|
|8896||Myxoid leiomyosarcoma||3 (0.2)|
|8900||Rhabdomyosarcoma, NOS||1 (0.1)|
|8901||Pleomorphic rhabdomyosarcoma, adult type||12 (0.9)|
|8910||Embryonal rhabdomyosarcoma||2 (0.2)|
|8920||Alveolar rhabdomyosarcoma||4 (0.3)|
|9040||Synovial sarcoma, NOS||31 (2.3)|
|9041||Synovial sarcoma, spindle cell||31 (2.3)|
|9042||Synovial sarcoma, epithelioid cell||2 (0.2)|
|9043||Synovial sarcoma, biphasic||12 (0.9)|
|9044||Clear cell sarcoma, NOS [except of kidney M-8964/3]||2 (0.2)|
|9133||Epithelioid hemangioendothelioma, malignant||1 (0.1)|
|9150||Hemangiopericytoma, malignant||2 (0.2)|
|9540||Malignant peripheral nerve sheath tumor||22 (1.6)|
|9560||Neurilemoma, malignant||2 (0.2)|
|9581||Alveolar soft part sarcoma||6 (0.4)|
We used the Cochran-Armitage test of trend to examine the association between age and the receipt of RT, and we used Pearson chi-square tests to examine associations between the receipt of RT and other variables of interest. A logistic regression model was used to examine the effect of age on RT use while adjusting for other variables that also probably influenced the receipt of RT. These other variables included demographics (sex, race, marital status) and tumor characteristics (site, histology, size) as well as the type of surgical procedure. Additional potential confounders, such as margin status, comorbidities, and performance status, are not recorded in the SEER database and, thus, are not accounted for in the current analysis. Tumor depth can be reported but is captured poorly in the SEER database; only 58% of our patients had data on tumor depth. Similarly, data on local control are not included in the SEER database; and, thus, the impact of RT on local recurrence rates could not be examined. A Cox proportional hazards model was used to assess age and the receipt of RT on all-cause and disease-specific mortality. All data were analyzed using the SAS statistical software package (version 9.2; SAS Institute, Inc., Cary, NC). The results were considered statistically significant when P < .05.
We identified 1354 patients who met our eligibility criteria (Fig. 1). Within the cohort, 354 patients (26.1%) were aged <50 years, 497 patients (36.7%) were ages 50 to 70 years, and 503 patients (37.2%) were aged >70 years (Table 2). Men (754 patients; 55.7%) and whites (1134 patients; 84.4%) made up the majority of the patient cohort. Malignant fibrous histiocytoma was the most common histologic diagnosis (576 patients; 42.5%), and “other” histologies (a heterogeneous collection of tumors) also were common (487 patients; 36%). Large tumors were more prevalent (805 patients; 59.5%) than small tumors.
|No. of Patients (%)|
|Demographic||Aged <50 Years||Ages 50-70 Years||Aged >70 Years||Total No.||P|
|354 (26.1)||497 (36.7)||503 (37.1)||1354|
|Fibrous histiocytoma, malignant||101 (28.5)||222 (44.7)||253 (50.3)||576|
|Leiomyosarcoma||29 (8.2)||50 (10.1)||62 (12.3)||141|
|Pleomorphic liposarcoma||12 (3.4)||34 (6.8)||28 (5.6)||74|
|Sarcoma, not otherwise specified||17 (4.8)||29 (5.8)||30 (6)||76|
|Other||195 (55.1)||162 (32.6)||130 (25.8)||487|
|Lower extremity||261 (73.7)||363 (73)||366 (72.8)||990|
|Upper extremity||93 (26.3)||134 (27)||137 (27.2)||364|
|Small: ≤5 cm||142 (40.1)||211 (42.5)||196 (39)||549|
|Large: >5 cm||212 (59.9)||286 (57.5)||307 (61)||805|
|Local excision||104 (29.4)||173 (34.8)||192 (38.2)||469|
|Radical resection||250 (70.6)||324 (65.2)||311 (61.8)||885|
|Men||186 (52.5)||299 (60.2)||269 (53.5)||754|
|Women||168 (47.5)||198 (39.8)||234 (46.5)||600|
|White||286 (80.8)||426 (85.7)||431 (85.7)||1143|
|Black||47 (13.3)||44 (8.9)||36 (7.2)||127|
|Other||21 (5.9)||27 (5.4)||36 (7.2)||84|
|Currently not married||139 (39.3)||150 (30.2)||206 (41)||495|
|Married||215 (60.7)||347 (69.8)||297 (59)||859|
|No radiotherapy||77 (21.8)||125 (25.2)||153 (30.4)||355|
|Radiotherapy||277 (78.2)||372 (74.8)||350 (69.6)||999|
Overall, 999 patients (73.8%) in the study population received RT. The overwhelming majority (960 patients; 96.1%) received external-beam RT. However, the receipt of RT decreased from 78.3% in patients aged <50 years to 74.9% in those ages 50 to 70 years and decreased again to 69.6% in patients aged >70 years (test for trend; P = .006). Conversely, undergoing local excision as the definitive surgical procedure increased from 29.4% in patients aged <50 years to 38.2% in patients aged >70 years (P = .029). Older patients also were more likely to be married, white men, with a malignant fibrous histiocytoma histology.
Age retained a strong association with the receipt of RT in a multivariate model (Table 3). Compared with patients aged < 50 years, patients aged ≥70 years were statistically significantly less likely to receive RT (odds ratio [OR], 0.66; 95% confidence interval [CI], 0.47-0.92). Patients ages 50 to 70 years also were less likely to receive RT, although this difference had only borderline statistical significance (OR, 0.81; 95% CI, 0.57-1.14). Pleomorphic liposarcoma, radical resection, large size, and being married all retained an independent association with the receipt of RT in the adjusted model. On further analysis of the 74 patients with pleomorphic sarcomas, 73% had tumors that measured >5 cm in greatest dimension.
|Characteristic||OR||95% CI||OR||95% CI|
|Fibrous histiocytoma, malignant||1.00||Referent|
|Sarcoma, not otherwise specified||1.14||0.66-1.98||1.28||0.73-2.26|
|Currently not married||1.00||Referent|
In an adjusted model, age and the receipt of RT together were associated with survival (Table 4). Regardless of age, not receiving RT was associated statistically with all-cause mortality and had a borderline association with disease-specific death. Age increased these effects. For example, relative to patients aged <50 years, patients aged ≥70 years who received RT appeared to have a slightly elevated chance of disease-specific death (hazard ratio [HR], 1.3; 95% CI, 0.9-2.1). However, older patients who did not receive RT were even more likely to experience disease-specific death (HR, 2.4; 95% CI, 1.4-4.1).
|All-Cause Mortality||Disease-Specific Mortality|
|Characteristic||HR||95% CI||HR||95% CI|
|Aged <50 y with RT||1.0||Referent||1.0||Referent|
|Aged <50 y without RT||1.7||1.0-2.9||1.4||0.7-2.9|
|Ages 50-70 y with RT||1.4||1.0-2.1||1.5||1.0-2.3|
|Ages 50-70 y without RT||2.1||1.3-3.4||1.8||1.0-3.3|
|Aged ≥70 y with RT||2.0||1.4-2.9||1.3||0.9-2.1|
|Aged ≥70 y without RT||4.0||2.8-6.3||2.4||1.4-4.1|
|Fibrous histiocytoma, malignant||1.00||Referent||1.0||Referent|
|Sarcoma, not otherwise specified||1.8||1.1-2.7||2.7||1.6-4.5|
|Currently not married||1.0||Referent||1.0||Referent|
In the current study, as expected, >33% of the patients in our study who were diagnosed with high-grade soft tissue sarcoma of the extremity were aged ≥70 years. Factors that are known to correlate with the receipt of RT were predictive of treatment in our series. For example, married patients who had greater family support more frequently received RT, and those with larger tumors who, thus, were at greater risk, also were treated more frequently. In addition, patients who underwent more radical surgery, which implies the possibility of both a greater risk and a more robust patient, more commonly received adjuvant RT. In contrast and despite an increasing incidence of sarcoma, the receipt of adjuvant RT declined with age. This trend remained even after adjusting the analysis for tumor histology, site, tumor size, and surgical procedure. In turn, the receipt of RT was linked closely to survival. Across all age groups, overall survival was worse in patients who did not receive RT, and a trend toward decreased disease-specific survival was observed with increasing age.
To our knowledge, no other report has focused on age differences and patient outcomes related to the receipt of RT for soft tissue sarcoma of the extremity. Our findings suggest that omitting RT for older adults with high-grade sarcoma may result in inferior outcomes. The rationale for differential cancer treatment by increasing age has been challenged in recent years, and some reports have suggested that there is an equal benefit from equal treatment in various tumor types.19-24 For example, Muss et al performed a meta-analysis of breast cancer trials focused on locally advanced disease and reported an equal benefit for adjuvant chemotherapy in patients aged >65 years compared with patients aged <65 years.20 Similar findings have been reported in the treatment of colon cancer with respect to mortality.25 In addition, a few studies have demonstrated that attenuated dose treatment regimens for older patients with non-Hodgkin lymphoma and small cell lung cancer may have a negative impact on treatment outcomes.22-24, 26 Those studies challenged the previous notion of limiting therapeutic options based on chronologic age alone.
Unfortunately, existing clinical trials provide limited data on the treatment of older adults. Older adults comprise the majority of incident and prevalent cancer cases, but they make up the minority of patients previously and currently enrolled on clinical trials. In a recent publication by Lewis et al, who evaluated participation of the elderly (aged ≥65 years) in clinical trials sponsored by the National Cancer Institute, only 32% of trial participants were elderly.9 Another report published by Hutchins et al described the representation of older adults in Southwest Oncology Group treatment trials between 1993 and 1996.8 Over 16,000 patients were enrolled consecutively during that time frame, and only 25% of those patients were aged ≥65 years. Similarly, there were very few patients aged >65 years in the 4 sentinel randomized trials that established the use of RT in soft tissue sarcomas.14-16, 27 Thus, although clinical trials provide critical data guiding patient care, these data are difficult to extrapolate to the elderly population and, thus, may contribute to treatment disparity.
A key limitation of our study was the lack of data on local recurrence. These data are not captured in the SEER database and would be useful for measuring the direct impact of RT. However, randomized data clearly support a significant reduction in local recurrences among patients with high-grade tumors who receive local RT. In those trials, a survival benefit was not reported; however, with the small number of patients, a modest improvement in survival may have been missed. Koshy et al, using the SEER database in a larger patient population similar to ours, concluded that the receipt of RT was associated with improvements in overall survival.28
Whether or not this improvement extends to an older population with generally poorer health is less clear. The significant decrease in overall survival observed in our study when RT was not received may have been related to poor overall health in the older population. This may strongly influence treatment decision-making on the part of the practitioner or the patient. For example, patients with significant comorbidities or poor performance status may not be selected for aggressive surgical therapy (ie, radical resection) or postoperative radiotherapy. The SEER database does not capture data on overall health and function, and it is possible that these missing data explain some or all of the trends observed in our study. However, the trend toward decreased disease-specific survival with increasing age suggests that some patients who may benefit from RT may not be receiving it. Even older patients with significant comorbidities are expected to live for 5 to 7 years,3 and local recurrence in high-grade soft tissue sarcomas often is concentrated in the 2 to 4 years after surgical excision.16 It may be that even older patients with comorbidities still may benefit from local therapy, particularly when accounting for limited salvage options. It is important to note, however, that specific data on comorbidities and performance status, factors that often shape treatment recommendations, also are not available in the SEER database. Because clinical trials often exclude patients with comorbidities,7 observational studies that incorporate comorbidity and performance status may prove useful for understanding the potential benefit of RT. Clinical trials addressing these issues would also be useful.
Without evidence-based treatment options, clinicians are faced with either extrapolating results obtained from younger, healthier patients or tailoring treatment without the benefit of scientific evidence. Improving outcomes and treatment decision-making for older adults with malignancy depends on a better understanding of the specific factors that contribute to response and adverse outcomes in this population. Our current results suggest that future clinical trials and observational studies need to address explicitly the appropriate treatment for older adults with and without comorbidities. In the absence of additional data, the growing population of older adults with high-grade soft tissue sarcoma of the extremities may experience increased cancer-related morbidity and mortality because of under treatment.
CONFLICT OF INTEREST DISCLOSURES
The authors made no disclosures.
- 1Population Division, US Census Bureau. Table 2. Projections of the Population by Selected Age Groups and Sex for the United States: 2010 to 2050 (NP2008-T2). Released August 14, 2008. Census Bureau web site. Available at: http://www.census.gov/.
- 2Surveillance Epidemiology and End Results. SEER Cancer Statistics Review 1975-2007, Table 1.10. Age-Distribution of Incidence Cases, Reviewed 2011. Available at: http://www.seer.cancer.gov/csr/1975_2007/browse_csr.php?section=1&page=sect_01_table.10.html
- 4Ries LAG, Harkins D, Krapcho M, et al. eds. SEER Cancer Statistics Review, 1975-2003. Bethesda, MD: National Cancer Institute; 2006. Based on the November 2005 SEER data submission, posted to the SEER web site 2006.
- 22Platinum-etoposide chemotherapy in elderly patients with small-cell lung cancer: results of a randomized multicenter phase II study assessing attenuated-dose or full-dose with lenograstim prophylaxis—a Forza Operativa Nazionale Italiana Carcinoma Polmonare and Gruppo Studio Tumori Polmonari Veneto (FONICAP-GSTPV) study. J Clin Oncol. 2005; 23: 569-575., , , et al.