Treatment patterns and survival among older adults in the United States with advanced soft-tissue sarcomas

Soft-tissue sarcoma (STS) refers to a rare and heterogeneous group of malignant tumors comprising more than 50 histologic subtypes derived from connective tissues and other cells of mesenchymal origin. Soft-tissue sarcoma accounts for approximately 1% of all incident malignancies [1] and an estimated 13,040 new cases will be diagnosed in the United States (US) in 2018, with 7370 dying from the disease [2]. In 2014, the age-adjusted incidence rate, for STS for patients 65 years of age and older was 11.3 per 100,000, compared to 2.3 per 100,000 for patients younger than 65 years of age [3].

Among all patients diagnosed with STS, the 5-year overall survival is approximately 50% [4]. Tumor histology, stage, and primary site are prognostic factors [4, 5]; 5-year overall survival is 83% in patients with localized disease and 16% in patients with distant metastases at initial diagnosis [6]. Eventually, either at initial diagnosis or after recurrence from a more limited extent of disease, 40–50% of patients with STS will have metastatic or unresectable locally-advanced disease (collectively referred to as “advanced STS”) [7].

Treatment options for patients with advanced STS have historically been limited; cytotoxic chemotherapy has been the mainstay of therapy for decades. Active drugs include doxorubicin, ifosfamide, gemcitabine, taxanes, and several others; these may be administered either as single agents or in combination regimens [4, 8]. The intent of these treatments is palliative for the majority of patients [4, 8, 9]. Until recently, clinical trials—including those demonstrating superiority of combination chemotherapy over monotherapy and those leading to the regulatory approval of newer agents such as olaratumab, trabectedin, and pazopanib—have found improvements in response rates and progression-free survival [4, 10‐14]. However, olaratumab is the only newer agent to date that has also demonstrated a statistically significant improvement in overall survival [11].

Few studies have examined utilization of systemic treatments for STS in the general population of the US, especially in older patients with advanced disease. Furthermore, there is little published information on variation in outcomes according to tumor histology, especially among older adults. Thus, the objectives of this study were to describe real-world, population-based treatment patterns and survival among older adult patients with advanced STS, overall and by groupings of specific histologic subtypes (hereafter “histologic categories”).

A total of 4274 patients with advanced STS met the inclusion criteria for this study (Table 1). Patients were on average 77.8 years of age (SD, 7.3; range 65–104 years) at the time of advanced STS diagnosis and 49.2% were male (Table 2). The mean (SD) CCI score was 2.8 (2.33), and 69.2, 18.3, and 6.5% of patients had a history of hypertension, congestive heart failure, and myocardial infarction, respectively.

At initial diagnosis, 36.0% of patients had metastatic (distant) disease; the remaining majority (64.0%) were diagnosed at earlier stages of disease and had claims based indicators of progression to advanced STS (Table 2) with a mean (SD) interval of 16.6 (23.0) months from initial diagnosis. Nodal disease (without any distant metastases) was observed only among 1.6% of patients at initial diagnosis. The most common histologic category was leiomyosarcoma (n = 922; 21.6%), followed by UPS (n = 652; 15.3%), liposarcoma (n = 554; 13.0%), vascular sarcoma (n = 357; 8.4%), fibroblastic/myofibroblastic sarcoma (n = 227; 5.3%), nerve sheath sarcoma (n = 106; 2.5%), rhabdomyosarcoma (n = 98; 2.3%), and synovial sarcoma (n = 49; 1.1%); the remaining 30.6% (n = 1309) had other or NOS histologic categories. A few variations in patient and tumor characteristics by tumor histology were evident: 66.7% of patients with leiomyosarcoma were female; patients with synovial sarcoma had a mean age of 74.1 years, and 54.1, 40.8, 48.7% of patients with rhabdomyosarcoma, synovial sarcoma, and other/NOS histology, respectively, had distant stage of disease at initial STS diagnosis (Table 2). Overall, the most common specified anatomic tumor site was lower limb (21.9%) (Appendix 2: Table 7), and the most common known tumor grade was undifferentiated (30.7%) (Table 2). The average observed follow-up time from advanced STS diagnosis was 1.7 years (SD, 2.36) (Table 2).

Cancer-directed treatment was received by 62.1% of patients. Radiation was received by 40.0% of patients and surgery was received by 10.3% of patients. The mean (SD) age of patients who received cancer-directed treatment was 76.5 (6.9) years and the mean (SD) CCI score was 2.5 (2.1), with hypertension, diabetes (without complications), chronic pulmonary disease, peripheral vascular disease, and congestive heart failure observed among 68.9, 28.4, 24.4, 18.1, and 14.6% of patients, respectively (Additional file 1: Table S1). Chemotherapy, biologic therapy, and targeted therapy were received by 27.5, 1.9, and 1.3% of patients, respectively. Among the 28.7% of all patients with advanced STS who received chemotherapy, biologic therapy, or targeted therapy (n = 1227), the mean duration of first-line therapy was 4.1 months (SD, 4.1) (Table 3). Second-line therapy was received by 11.1% (n = 476) of all patients; the mean duration of second-line treatment was 4.6 months (SD, 5.3). Among the 4.4% of all patients with advanced STS who received at least third-line therapy (n = 189), the mean duration of third-line therapy was 4 months (SD, 3.4). The proportion of patients receiving up to three lines of treatment and duration of each line of treatment by histologic category is shown in Table 3. First-line chemotherapy was most commonly received by patients with synovial sarcoma (42.9%), leiomyosarcoma (37.0%), and vascular sarcoma (36.4%).

Overall, the most common regimen during first-line therapy was docetaxel plus gemcitabine (26.5%), followed by doxorubicin monotherapy (18.8%), gemcitabine monotherapy (9.1%), paclitaxel monotherapy (5.5%), and bevacizumab monotherapy (3.3%). Docetaxel plus gemcitabine was the most common first-line therapy for patients with all histologic categories except liposarcoma (for whom it was doxorubicin monotherapy [24.3%]) and vascular sarcoma (for whom it was paclitaxel monotherapy [40.0%]) (Table 4). During first-line therapy, doxorubicin plus ifosfamide was received by 3.0% and ifosfamide monotherapy was received by 2.7% of patients. Among all patients who received doxorubicin alone or a doxorubicin-based first-line therapy (n = 371), 29.7% received liposomal doxorubicin, and 5.9% received the cardioprotective agent dexrazoxane.

Among patients who received second-line (n = 476) and third-line (n = 189) therapy, doxorubicin monotherapy (second line, 21.2%; third line, 14.8%), docetaxel plus gemcitabine (second line, 17.4%; third line, 11.1%), and gemcitabine monotherapy (second line, 8.6%; third line, 11.1%) were most common (Table 4). Among the 476 patients who received second-line therapy, the most common treatment sequence was first-line docetaxel plus gemcitabine followed by second-line doxorubicin monotherapy (n = 56, 11.8%). Among the 189 patients who received third-line therapy, the most common sequence was first-line doxorubicin followed by second-line gemcitabine and then third-line docetaxel plus gemcitabine (n ≤ 11). Among patients who received second- or third-line therapy, ≤ 11 received dexrazoxane.

Among patients who received supportive care only (n = 1618, 37.86%), the mean (SD) age was 80.0 (7.5) years (Additional file 1: Table S1). Hypertension (69.6%), diabetes without complications (28.7%), chronic pulmonary disease (28.1%), congestive heart failure (24.5%), and peripheral vascular disease (22.6%) were the most common baseline comorbidities and the mean (SD) CCI score was 3.1 (2.6).

The majority of patients (n = 3565; 83.4%) died during study follow-up, with a median survival (95% CI) of 8.9 months (8.3, 9.7) from the time of advanced STS diagnosis. Median survival (95% CI) from advanced STS diagnosis among patients who received cancer-directed therapy was 13.6 months (12.9, 14.6), and among those who received supportive care it was 2.8 months (2.6, 3.4) (Table 5 and Fig. 1). Survival estimates for each histologic category is shown in Table 5. Estimated survival from advanced STS diagnosis varied by histologic category with median survival ranging from 21.4 months (15.6, 26.8) for patients with fibroblastic/myofibroblastic sarcoma to 3.0 months (1.8, 5.7) for patients with rhabdomyosarcoma.

This study assessed recent real-world treatment patterns and estimated survival of a population-based cohort of older adults diagnosed with advanced STS in the US. More than 60% of patients were initially diagnosed at an earlier stage of disease and progressed to advanced STS with mean time to progression of 16.6 months. Consistent with previous studies [25‐27], leiomyosarcoma, UPS and liposarcoma were the most common histological categories of STS in this study population.

Among patients initiating first-line systemic therapy, the most common regimen was docetaxel plus gemcitabine combination therapy (26.5%); the next most common first-line regimen (18.8%) was doxorubicin only. Both a recent medical record review study in the US [25] and the Sarcoma Treatment and Burden of Illness in North America and Europe chart review study [26] found that anthracyclines were used most commonly among first-line therapies (either alone or in combination) in younger patient populations. In this study of older adults, docetaxel plus gemcitabine was the most commonly used first-line therapy for most histologic categories except for vascular sarcomas (for which paclitaxel was most commonly used) and liposarcoma (for which doxorubicin was most commonly used). The use of paclitaxel for vascular sarcomas is consistent with evidence suggesting that vascular sarcomas may be relatively more responsive to taxanes and as such are recommended by NCCN guidelines [24, 28‐30]. A relatively lower use of doxorubicin or related drugs was observed in this study, plausibly because the older age of the study population (mean age, 77.8 years) may increase concerns about adverse events, particularly doxorubicin-associated cardiotoxicity [27]. In the current study, 18.3% of the patients had a history of congestive heart failure and 25.8% of patients had a history of chronic pulmonary disease. Additionally, 69.2% of the patients in this study had hypertension that, along with age, is known to increase the risk of doxorubicin-associated cardiotoxicity [31, 32]. Prior research has addressed the issues of doxorubicin-based therapies and cardiovascular risk factors in the older adult population [33‐35]. It is possible that the low use of doxorubicin observed in this study cohort is related to these issues, but the study design and available data do not allow for a thorough investigation of this potential relationship. Further understanding of physician preferences with regard to administration of doxorubicin-based therapy among older adult patients is needed to adequately comprehend observed treatment patterns. Knowledge about the appropriate care of the older patient remains limited, but data from this study begin to fill this gap and suggest directions for future research related to toxicity risk, disease progression, recurrence and survival outcomes, and their relationship with patient age in the setting of advanced STS.

As observed in other studies, gemcitabine was used most commonly during second-line therapy (either alone or in combination) [25, 26]. We also noted that dexrazoxane was not at all commonly used in this population among those who were treated with doxorubicin.

Approximately 62% of the study population received cancer-directed therapy. As compared to patients receiving cancer-directed therapy, patients receiving supportive care only had a mean age of 80.0 years (cancer-directed therapy: 76.5 years) at the time of being diagnosed with advanced disease and had a baseline comorbidity burden (mean CCI score) of 3.1 (cancer-directed therapy: 2.5), and 24.5% (cancer-directed therapy: 14.6%) had a history of congestive heart failure. Therefore, it seems likely that many of these patients may have been considered to not be candidates for chemotherapy, particularly with a cardiotoxic agent. The mean non-cancer CCI scores observed in this study for the overall population, patients receiving cancer-directed treatment, and patients receiving supportive care only are in line with those reported by Davis et al. among metastatic lung cancer patients [36].

Median overall survival in the entire study population was estimated to be less than 9 months. Patients who received supportive care lived only less than 3 months. The shorter survival of the group that received only supportive care, in addition to being related to not receiving anticancer treatment per se, is likely to be confounded by selection of patients for treatment who had a more favorable prognosis or better ability to tolerate treatment.

A study by Italiano et al. [7] (median age range 53–59 years), which included patients from the time frame of this study (i.e., 2002–2006), found that patients with synchronous or metachronous metastatic STS had an overall survival of 18 months from the time of metastatic diagnosis. Patients in the study by Italiano et al. [7] had relatively longer survival than those in this study, plausibly because patients in this study were considerably older. Similar to this study, survival did vary by histologic category in the study by Italiano et al. [7]. For example, patients with leiomyosarcoma, UPS, and nerve sheath sarcoma had a median survival of 12.9, 9.6, and 6.2 months, respectively, in this study and 19.4, 11.2, and 8.6 months, respectively, in the Italiano et al. [7] study.

An array of treatments (e.g. eribulin, olaratumab, pazopanib, trabectedin) became available for this population in the last decade [11, 13, 14, 37]; however, effectiveness of these treatments specifically among older adult patients with advanced STS is yet to established. Other treatments like oral cyclophosphamide plus prednisone [34] may also be feasible for older adult patients for whom treatment with doxorubicin may not be an option. Overall survival, progression-free survival, and response rates of patients with advanced STS may improve as these treatments will become part of routine care provided to this population.

This study is subject to several limitations inherent in analyses of Medicare claims data and the use of such data in studies of advanced cancer, in particular. For patients who were not initially diagnosed at the metastatic stage of the disease, ICD-9-CM diagnosis codes were used to identify evidence of metastatic disease during the follow-up period. However, the use of ICD-9-CM codes to identify metastatic disease has been shown to have sensitivity, specificity, and positive predictive value of less than 80%, and thus the use of ICD-9-CM diagnosis codes in the Medicare claims data may have resulted in inaccurate or under-identification of an advanced STS population [38, 39]. As described in the study methods, one of the criteria used to identify progression to metastatic disease was the initiation of systemic therapy at least 6 months after surgery. This criterion may have resulted in selection bias because patients who had disease progression but did not receive systemic therapy would have been omitted from the supportive care-only group. Although, the 6-month lag period after surgery was used to avoid interpreting adjuvant therapy as treatment for advanced disease, patients who progressed and received systemic therapy for advanced disease would have been omitted from the cancer-directed therapy group. Lines of therapy are not reported in claims data; therefore, an algorithm had to be defined to estimate the lines of therapy. This may have misclassified treatments by line of therapy, as the reasons for treatment changes were not available in the data. The Medicare Part D database was available for only a subset of the cohort (32.4%); however, only a small proportion of all chemotherapy (2%) claims were identified from the Medicare Part D database, so the risk of missing important treatment data is relatively low despite this limitation. An array of treatments (i.e., eribulin, olaratumab, pazopanib, trabectedin) became available for this population in the last decade, although these agents are not fully represented in the study dataset. Future research should evaluate the use and outcomes of these novel treatments in the older adult population. Finally, this study included only patients aged 65 years or older, and although SEER-Medicare data is representative of the US population 65 years and older for age and gender, participating SEER sites may not be representative with regard to distribution of race, income, urban residence, HMO enrollment as well as cancer mortality [15] and thus the results should not be generalized to the entire population of older adult patients with advanced STS.

Despite these limitations, this study documents real-world treatment patterns that may help inform providers, researchers, and policymakers about the care of older patients with STS in the US. As real-world data including new therapeutic options become available, our results provide a basis for analyzing changes in treatment patterns and outcomes over time. This study demonstrates that the prognosis is poor for older adult patients with advanced STS, highlighting the unmet medical need in this population.

Although previous studies of younger populations reported anthracycline-based therapy predominated in first line, our study of older adults with advanced STS found that doxorubicin was not commonly used. Despite variation by histologic category, prognosis was observed to be poor for older adult patients with advanced STS in this study.