Treatment Patterns and Survival among Adult Patients with Advanced Soft Tissue Sarcoma: A Retrospective Medical Record Review in the United Kingdom, Spain, Germany, and France

Objective To describe real-world treatment patterns and outcomes for patients with advanced soft tissue sarcoma (STS) not amenable to surgery or radiotherapy in the United Kingdom, Spain, Germany, and France. Methods Physicians completed a web-based medical record abstraction for adult patients with advanced STS (other than Kaposi's sarcoma or gastrointestinal stromal tumor) who received ≥1 line of systemic therapy. Clinical characteristics, treatments, tumor responses, and mortality data were recorded. Results A total of 130 physicians provided data for 807 patients. Patients' mean age at advanced STS diagnosis was 57.1 (±12.3) years; 59% were male. The most commonly identified histologic categories were leiomyosarcoma (28%), liposarcoma (13%), and rhabdomyosarcoma (11%). Overall, 57% of patients received only 1 line of therapy, 32% received 2 lines of therapy, and 11% received ≥3 lines of therapy. The most common first-line regimens were doxorubicin alone (41%), doxorubicin plus ifosfamide (19%), docetaxel plus gemcitabine (9%), paclitaxel alone (4%), and ifosfamide (4%). Median overall survival from start of treatment was estimated to be 17.6 months (95% confidence interval, 15.6–19.0 months). Conclusions In real-world clinical practice, advanced STS is most commonly treated with older therapies in the United Kingdom, Spain, Germany, and France. New therapies that improve overall survival in advanced STS are needed.


Introduction
Soft tissue sarcoma (STS) refers to a rare and heterogeneous group of malignant tumors comprising more than 50 histologic subtypes that are derived from connective tissues and other cells of mesenchymal origin (e.g., fat, smooth or striated muscle, blood vessels, nerve sheath, subcutaneous tissue, and visceral connective tissue). Soft tissue sarcomas account for approximately 1% of all incident malignancies [1], with an estimated 23,574 new cases in Europe annually [2]. e estimated 5-year relative survival of patients with STS of any stage in Europe is 58%, and 5-year overall survival is approximately 50% [2,3]. However, this rate is dependent on a number of factors, including the stage at diagnosis, histologic subtype, primary site, and presence of metastases [3].
An estimated 40% to 50% of patients with STS either present initially with metastatic or unresectable locally advanced disease (collectively, "advanced STS") or present initially with a more limited extent of disease and subsequently develop advanced STS [4]. Although there are curative surgical options for early-stage STS, treatment goals for advanced STS are more limited. Chemotherapy (e.g., doxorubicin or ifosfamide, alone or in combination with each other or other agents) is most commonly used to treat patients with advanced STS [3,5,6]. e intent of these treatments generally is palliative rather than curative, and response rates are low (typically in the range of 10%-25%) [3]. e objective of this study was to evaluate treatment patterns and survival outcomes among patients with advanced STS not amenable to surgery or radiotherapy in realworld clinical settings in the United Kingdom (UK), Spain, Germany, and France in order to provide context for the evaluation of emerging therapies for this population.

Study Design.
is study was a retrospective review of medical records of patients in the UK, Spain, Germany, and France treated with systemic therapy for advanced, histologically diagnosed STS (excluding Kaposi's sarcoma and gastrointestinal stromal tumor). Physicians abstracted anonymized data from medical records directly into a webbased data collection form. e relevant national competent authorities reviewed and approved the study on ethical grounds in all four countries. In Germany and Spain, the ethics committee at the study site of the principal investigator in each country reviewed and approved the study, as required by national authorities; this site-level approval applied to all participating sites within each country.

Physician Selection.
Physicians specializing in oncology, who had been in practice for 3 to 35 years, and who personally treated at least 3 patients with advanced STS in the past year were recruited to provide patient data. Physician recruitment was done in partnership with countryspecific agencies using in-house physician databases and physician directories maintained by local medical associations. Physicians were sampled from geographic regions throughout each country (5 regions

Patient Selection.
Eligible patients were diagnosed with histologically confirmed metastatic or unresectable locally advanced STS (either at presentation or after progression from limited disease), between 1 January 2005 and 18 months before the initiation of record abstraction in the patient's country of residence. is end date for eligibility was chosen to allow the opportunity for a minimum duration of follow-up information unless a patient died sooner. Patients were required to be aged 18 years or older on the date of initial diagnosis of STS, to have started at least one line of systemic therapy for advanced STS, and to have records that were accessible to the abstracting physician for at least 1 year from diagnosis of advanced STS unless the patient died or refused further treatment.
Patients were excluded if they had received treatment with curative intent after diagnosis of advanced disease; received treatment with doxorubicin, daunorubicin, idarubicin, or other anthracyclines or anthracenediones (e.g., mitoxantrone) at any time before initiating first-line treatment for advanced STS; had evidence of concurrent malignancy, except adequately treated nonmelanoma skin cancer or in situ neoplasm; or had received first-line treatment within an interventional clinical trial (however, those who participated in post-firstline interventional trials or any noninterventional studies were eligible).
To reduce the possibility for physicians to select convenient patients, they were asked to select medical charts for patients whose last name begins with a randomly generated letter. If no patient whose last name began with that letter met the study criteria, physicians selected a patient whose last name began with the next letter in alphabetical order. To ensure internal consistency of the abstracted data, data checks for illogical or unusual patterns in physicians' responses on the data collection form were conducted.

Study Measures.
Patient characteristics abstracted from the medical records included age, gender, race/ethnicity (except in France, where collection of such information is prohibited), comorbidities, and performance status. Tumor characteristics including American Joint Committee on Cancer cancer stage [7], histology, primary site, and metastatic sites were abstracted. For the analysis, specific histologic subtypes reported in the medical records (see Supplementary Appendix A) were grouped into broader categories based on the World Health Organization (WHO) Classification of Tumours of Soft Tissue and Bone [8]. Table 1 links the histologic categories used in the analyses to the specific histologic subtypes reported in medical records and to the broader histologic categories (chapter headings) in the WHO Classification of Tumours of Soft Tissue and Bone [8].
Data on STS treatments were collected from the time of diagnosis of advanced disease until the end of the most recent medical record data available. ese included numbers and percentages of patients receiving cancerdirected treatments, detailed usage patterns for chemotherapy and targeted therapy (i.e., crizotinib, everolimus, imatinib, pazopanib, sirolimus, sorafenib, sunitinib, tamoxifen, and toremifene), and numbers and percentages of patients receiving supportive care during each line of systemic therapy and after stopping active treatment. Details of systemic therapy were collected, including regimens prescribed, dates and duration of therapy, and reasons for discontinuation. Performance status was recorded (if available) at the start of each treatment line. Tumor response and progression, according to Response Evaluation Criteria in Solid Tumors (RECIST) as interpreted by the physician, were also evaluated.
Mortality data, including death during follow-up and cause of death (i.e., STS related or not), were collected from medical records. Overall survival was estimated using the Kaplan-Meier method, which accounts for right censoring for patients whose death was not recorded during the observation period; the date of censoring was the date of the last available medical encounter before the data abstraction.

Physician Characteristics.
A total of 130 physicians participated in the study (21 in the UK, 34 in Spain, 40 in Germany, and 35 in France). A majority (≥65%) of treating physicians practiced in a cancer center or academic/teaching hospital as their primary practice facility type. In the previous year, the participating physicians treated an average of 45 patients (standard deviation (SD), 37) with advanced STS.

Patient Characteristics.
Overall, 807 patients were included in the study sample (199 in the UK, 203 in Spain, 204 in Germany, and 201 in France). e patients were 58.9% male (UK, 63.3%; Spain, 60.6%; Germany, 63.7%; France, 47.8%), and 92.7% of patients were white (UK, 89.5%; Spain, 99.0%; Germany, 89.7%; race/ethnicity not reported in France) ( Table 2). Mean age at the time of advanced diagnosis was 57.1 years (SD, 12.3). Across countries, the most common histologic categories were leiomyosarcoma (28.4%), liposarcoma (13.0%), rhabdomyosarcoma (10.7%), vascular sarcoma (9.8%), fibroblastic/myofibroblastic sarcoma (9.7%), and synovial sarcoma (5.8%). In each country, leiomyosarcoma was the most commonly observed histologic category (UK, 30.7%; Spain, 22.2%; Germany, 32.4%; France, 28.4%). e mean observed follow-up time (from diagnosis of advanced STS to death or the last medical encounter before data abstraction) was 24.7 months (SD, 20.9). Most patients were initially diagnosed with stage IV STS (n � 557, 69.0% overall; 78.9% in the UK; 68.5% in Spain; 77.0% in Germany; 51.7% in France). Among patients initially diagnosed with limited-stage disease (n � 188; 23.3% of the sample), the mean (SD) time from initial diagnosis of STS to development of advanced disease was 23.1 (20.6) months. Eight percent of patients were reported to have limited-stage disease at diagnosis but had the same dates of initial STS diagnosis and advanced STS diagnosis; these patients were considered to have advanced disease at diagnosis for the purposes of the analysis.

Overall Advanced STS Treatment.
All patients received at least one line of systemic therapy for advanced STS as an eligibility criterion. e mean (SD) total number of lines of therapy received was 1.6 (0.8) (Table 3). Overall, 56.5% of patients received only one line of therapy, 32.5% received two lines of therapy, and 11.0% received at least three lines of therapy. Conventional chemotherapy alone was received by 84.7% of all patients, while targeted therapy alone was received by 2.5%; 12.8% received at least one drug in each class (data not shown).

Overall Survival.
In total, 545 patients (67.5%) died during observed follow-up, and most (94.9%) of these deaths were STS related. e median estimated overall survival time from the start of first-line therapy was 17.6 months (95% confidence interval (CI), 15.6-19.0 months) (Figure 1). In successive subgroups of the study population receiving additional lines of treatment, the median overall survival was 15.6 months (95% CI, 13.3-18.3 months) from the start of second-line therapy and 11.2 months (95% CI, 8.1-14.6 months) from the start of third-line therapy ( Figure 1).
Median overall survival varied considerably by histologic category, ranging from 15.2 months (95% CI, 10.4-22.0 months) for patients with vascular sarcoma to 23.8 months (95% CI, 20.3-27.4 months) for those with leiomyosarcoma (Table 8). By disease stage at initial diagnosis, median overall survival from the time that advanced disease developed was 29.9 months (95% CI, 23.3-38.5 months) among patients with stage I or II disease at diagnosis, 20.5 months (95% CI, 16.9-34.0 months) among those with stage III disease, and 16.2 months (95% CI, 14.9-18.2 months) among those with stage IV disease at initial diagnosis ( Figure 2).

Discussion
We have presented a descriptive review of treatment patterns among patients who received at least one line of systemic therapy for advanced STS in the UK, Spain, Germany, and France, based on a retrospective evaluation of their medical records. Owing to the observational nature of this study and the limited information available to NA: not applicable; UK: United Kingdom; a mesna was always coadministered with ifosfamide; b may have been observed in the overall or country-specific sample but was not among the five most commonly prescribed first-line regimens.
6 Sarcoma control potential confounding, the results presented here are descriptive rather than suggestive of causal relationships between treatments and outcomes. Among the countries studied, there are differences in the organization of medical care and other cultural factors that could affect the treatment and survival of patients with advanced STS. e study provides recent data on characteristics, realworld treatment patterns, and survival from several European countries for a relatively large group of patients with advanced STS who received at least one line of systemic therapy.
e clinical characteristics of the study population are generally unsurprising. Consistent with other published information [2,10], we found that leiomyosarcoma and liposarcoma were generally the most frequently reported histologic categories across study countries. As has been reported in literature [11,12], our findings confirm that the lung is the most common location of metastases among patients undergoing systemic treatment for advanced STS. In our study, most primary tumors at the time of diagnosis of advanced STS were >5 cm and classified as "deep." e relatively high proportion of patients with metastases at the time of initial diagnosis (relative to the population-based epidemiology of all patients diagnosed with STS) is likely due to the selection of patients for this study who ultimately had advanced disease and who underwent at least one line of systemic therapy.
Our study reported first-line treatment with doxorubicin alone or in combination as the most frequently used agent across the four countries we studied. Treatments were generally similar across histologic categories; however, some exceptions (in line with clinical guidelines) were noted, including the frequent first-line use of paclitaxel for vascular sarcoma and docetaxel plus gemcitabine for uterine leiomyosarcoma.      [13][14][15]. Results reported here are generally consistent with these guidelines in that standard first-line chemotherapy should be anthracycline based, with ifosfamide as a first-line alternative for patients who are not able to receive an anthracycline. Multiagent chemotherapy (gemcitabine plus docetaxel) has not been shown to provide superior results to single-agent doxorubicin [16]. Despite greater toxicity with doxorubicin plus ifosfamide than with doxorubicin alone, a higher response rate and longer progression-free survival with doxorubicin plus ifosfamide makes this combination a treatment of choice when tumor response is a high priority in patients with good performance status [17].
Ifosfamide (both as monotherapy and in combination with doxorubicin and/or dacarbazine) was among the five most commonly prescribed first-line treatments across all countries. Patients who received doxorubicin plus ifosfamide in the first line were also more likely to receive a second-line treatment than patients who received other first-line regimens.  Other real-world studies reported similar treatment patterns to what we present here. For example, Leahy et al. [18] conducted a retrospective medical record review of patients with metastatic STS across nine countries, including France, Germany, Spain, and the UK and found that the most common first-line regimens were doxorubicin monotherapy and anthracycline plus ifosfamide. Similarly, Nersesyan et al. [19] analyzed data for the European Union Five (France, Germany, Italy, Spain, and the UK) from country-specific cancer registries, published scientific studies, and proprietary surveys of 76 physicians conducted in March 2015.
ey concluded that doxorubicin plus ifosfamide is the most commonly used treatment in first line across the study countries, while trabectedin and pazopanib are used frequently in second-and third-line treatment (except in France, where first-line use of trabectedin among patients with leiomyosarcoma was not uncommon). Guest et al. [20] posit that "many clinicians do not initiate chemotherapy with ifosfamide monotherapy" at least in part because ifosfamide is associated with increased risks of toxicities at high doses (with high doses recommended after failure of first-line chemotherapy). Our data present somewhat conflicting findings in this regard, with ifosfamide monotherapy being one of the five most frequently used first-line therapies in both Spain and the UK (although not in France or Germany).
As Schöffski et al. [3] pointed out, " e heterogeneity of this disease poses a challenge to the physician caring for patients with sarcoma, as the prognosis and potential response to treatment in a given subtype of a mesenchymal malignancy is difficult to predict." Evaluating the efficacy of treatments can be challenging given different histologic types, each of which may exhibit differential chemosensitivity [14] and response to treatments [21]. For example, in a recent study of patients in Japan with STS who were treated with pazopanib, Nakamura et al. [22] found significant differences in median progression-free survival across histologic subtypes, ranging from 8 weeks for liposarcoma to 18.6 weeks for leiomyosarcoma.
Previous studies have reported limited survival for patients with advanced STS [4,23,24], which varies somewhat by histologic category [25]. Our study confirms these findings, with median overall survival estimates of less than 2 years across histologic categories. It should be noted that in medical record reviews such as this, where the ability to control potential confounding is limited, it cannot be inferred that any observed survival differences among subgroups treated with different therapies are a result of treatment as opposed to differences in prognostic factors among patients who are prescribed particular therapies ("channeling bias"). It is apparent from our data, as well as from registry data and previously published studies, that survival rates for advanced STS are poor, indicating a continuing, critical need for advances in therapeutic options for this condition. With the exception of eribulin, which has demonstrated improved overall survival as a third-line treatment for STS [26], no new treatment has shown an overall survival improvement over the past 20 years.
A number of prior studies have evaluated symptom burden and medication use among adult patients with STS. In evaluating the literature, Chan et al. [27] found that pain, dyspnea, nausea, and vomiting were the most frequent symptoms experienced. Kuo et al. [28] conducted a study of an STS population in a UK sarcoma unit and found that more than half of assessed patients reported pain. Finally, Gough et al. [29] conducted a medical record abstraction in the UK and deduced that patients with metastatic STS have high symptom burden with pain being a significant problem. Our results are consistent with the literature in that more than half of all patients in our study (regardless of the country) utilized some form of pain medication.
Our results are subject to several limitations inherent to many retrospective medical record review studies. First, the patients selected for study inclusion represented a convenience sample in that the records were obtained from physicians who were willing to participate in the study. e extent to which physician self-selection for participation in studies such as this could influence results is unknown. Both the sampling procedures and the patient eligibility criteria limit the generalizability of this study. In particular, some types of STS are more common in younger patients (who were intentionally excluded from our study), and patients who were surgically cured and never developed advanced disease were excluded from this study. us, the study sample may not be completely representative of all patients in each country studied, and our findings may not be fully generalizable to the overall population of adult patients with advanced STS who undergo systemic treatment in these countries. Data were entered directly by the treating physicians based on medical records available at the time of data entry, and therefore the data are potentially subject to data entry errors and other limitations of retrospective data capture. Response rates observed in this study should be interpreted with caution. Response was subject to physicians' interpretation and may not have routinely been assessed radiographically; RECIST criteria may not have been as rigorously applied in routine practice settings as in clinical trials. Further, trabectedin and pazopanib were not commercially available during the full study period, and their utilization may have been underestimated.

Conclusions
is is one of the first large-scale, real-world retrospective medical record reviews to evaluate current treatment patterns and survival in a multinational population of adult patients with advanced STS. Patients in the UK, France, Germany, and Spain were generally treated consistently with established treatment guidelines. Variations in treatment patterns across countries were evident, but most involved use of anthracyclines; doxorubicin, alone or in combination with ifosfamide, was the most common firstline therapy. Median overall survival across histologic categories was less than 2 years, indicating a continuing, critical need for advances in therapeutic options for this difficult-totreat disease.

Conflicts of Interest
is study was performed under a research contract between RTI Health Solutions and Eli Lilly and Company and was funded by Eli Lilly and Company. Sean D. Candrilli, James A. Kaye, and Saurabh P. Nagar are salaried employees of RTI Health Solutions. Yulia D'yachkova, Maria Lorenzo, and Daniel S. Mytelka are salaried employees of Eli Lilly and Company.