Prognosis of Patients with Metastatic Soft Tissue Sarcoma: Advances in Recent Years

Soft tissue sarcomas (STS) are a group of rare and heterogeneous tumors arising from mesenchymal tissue. 23,500 new cases are being diagnosed in Europe each year [1]. Up to 50% of STS patients will develop metastatic disease, which highly impairs prognosis and remains challenging for the treating physician [2, 3]. The most common metastatic sites are the lungs, followed by the liver and the peritoneum, while lymphogenic metastases are relatively rare [4, 5].

Therapy for localized STS mainly consists of surgery and subsequent chemotherapy or radiation. In the metastatic disease state, palliative systemic chemotherapy is administered [6]; selected cases receive metastasectomy [7]. Besides conventional treatment options of doxorubicin and combinations with ifosfamide or dacarbazine, several new agents have entered the therapeutic armamentarium of STS in recent years, showing benefit in terms of overall survival (OS), foremost in second-line treatment for metastatic disease (pazopanib, trabectedin, and eribulin) [8-10].

Data on survival of localized STS have been published regularly, whereas data restricted to patients suffering from metastatic disease are limited. The prognosis of those patients has risen from as low as 1 year in the 1980s and 1990s up to 18 months in clinical studies in the last 2 decades [11-13]. Recent clinical trials showed even longer durations of median OS, indicating a further increase [14, 15]. Given the presumably rising trend, the therapeutic advances, and the long period of time passed since the last publication focusing on the prognosis of metastatic STS patients, an up-to-date analysis of real-life data of patients being treated in a high-volume sarcoma reference center seems reasonable [12].

The objective of this study was to analyze the prognosis of patients with advanced metastatic STS under current best practice of care in order to challenge former published data on prognosis. The analysis was undertaken in the Sarcoma Unit at the Mannheim University Medical Center which represents one of Germany’s certified high-volume STS reference centers. Furthermore, we evaluated predictive factors related to survival outcome, such as size and location of the primary tumor, histological grade and subtype, local relapse, disease-free interval (DFI), and metastatic site.

After approval by the ethical review committee of the Medical Faculty of Mannheim (MFM), all patients (age >14 years) with a histologically confirmed STS who received at least a part of the management of their disease at the Sarcoma Unit were screened. Among these, all patients who developed metastatic disease between January 1, 2010, and December 31, 2016, were detected. A total of 212 patients comprised the group of this analysis.

Median OS as well as 1-, 2-, and 5-year OS rates were the primary endpoints of the study. We defined OS as the time span between first evidence of metastases and recorded death of any cause. Follow-up data were either collected via clinical patients’ records, by federal registries, or by contacting other involved physicians. Predictive factors were evaluated. The following information was retrospectively obtained from medical records: gender and age at diagnosis, size and location of the primary tumor, and histological grade and subtype. The number of organs which were affected by metastases and DFI were also analyzed. DFI was defined as the time span between resection of the primary tumor and diagnosis of metastatic disease. The number of metastatic sites was defined as the number present at the time of first metastasis diagnosis.

The nonparametric method of Kaplan and Meier was used for calculating estimations of OS of the population as well as various subgroups [16]. Patients were either censored at the time of death or last follow-up. For comparison of survival between curves we performed the log-rank test [17]. Multivariate analysis with a Cox proportional hazards model was used to test whether one of the above-mentioned characteristics was associated with survival and to calculate relative risks together with 95% CI [18]. Significance of all statistical analyses was defined as p < 0.05. Analyses were performed using R software for ^©macOS Mojave (Release 3.5.1, ^©R foundation for statistical computing).

Demographic and tumor-related characteristics of the study population of the 212 patients are summarized in Table 1.

A Kaplan-Meier plot for survival after first detection of metastases is depicted in Figure 1. Median OS of the whole population was 24 months (95% CI 21–33). 1-, 2-, and 5-year OS rates were 70.0% (95% CI 64–77), 49.9% (95% CI 43–58), and 24.8% (95% CI 19–33), respectively. During follow-up, patients’ death was determined in 137 cases (64.6%), while 75 cases (35.4%) were censored at the date of last recorded follow-up. Median duration of follow-up was 20 months.

The analyses of clinicopathological factors which were investigated for prognostic significance are shown in Tables 2 and 3. Characteristics which showed prognostic significance in the multivariate analysis appeared to be male gender (p = 0.0005), age >55 years (p = 0.003), location of the primary tumor, histologic subtype, as well as synchronous metastasis (p = 0.001). Regarding the location of the primary tumor, patients with STS of the internal trunk reached a median survival of only 11 months (95% CI 9–23) and a 1-year OS of 44.1% (95% CI 29–65). It is notable that those 31 patients constituted a wide range of sites, containing intrapelvic, intraperitoneal, as well as intrathoracic tumors. A Kaplan-Meier plot is shown in Figure 2. Concerning DFI, patients with synchronous metastasis showed a median OS of only 17 months (95% CI 12–25), a 1-year OS rate of 57.9% (95% CI 48–70), and a close relationship with the group of short DFI. A Kaplan-Meier plot is shown in Figure 3. Three histological subtypes independently appeared to be of prognostic relevance for survival: Malignant peripheral nerve sheath tumor (MPNST), sarcoma not otherwise specified, and synovial sarcoma. MPNST patients formed the group with the lowest median OS (11 months; 95% CI 4 to NA) and a 1-year OS rate of 46.3% (95% CI 25–88). A Kaplan-Meier plot is shown in Figure 4. With regard to the size of the primary tumor, a size of ≥5 cm was a significant risk factor for lower OS (p = 0.04). Regarding histologic grade, multivariate analysis revealed no significant prognostic value for higher grading, missing significance level by only a small margin (p = 0.05). The number of metastatic sites showed no significant difference in OS as well as presence of local relapse.

Over the last decades, numerous studies on prognosis of STS patients with advanced disease have been published. However, recent data are limited. Italiano et al. [12] published an analysis in 2011 comprising 1,024 French patients over several time periods, the last one ranging from 2002 to 2006. This study revealed an OS of 18 months for the latter and a 2-year OS rate of 39%. Two retrospective studies by Ferguson et al. [4] and Verschoor et al. [19] both revealed low results, either a 5-year OS rate of 17% and a median OS of 13 months. Their impact is limited though by the long durations of their observation periods which are ranging back to the 1980s. Nagar et al. [13] published another retrospective study in 2018 analyzing a population treated between 2005 and 2014 (n = 557) with advanced disease including only patients who were not eligible for surgery. Perhaps thereby selecting patients with a worse outcome, the reported median OS was 16 months only. On the other hand, first-line clinical studies on olaratumab and evofosfamide revealed promising results with up to 20 months’ median OS in the standard doxorubicin arm [14, 15]. However, the use of clinical trials for the assessment of prognosis is limited, since trials include selected cohorts of patients, therefore limiting external validity. In the present study, we retrospectively analyzed a population of 212 patients with a wide spectrum of histological subtypes at our institution with a median follow-up of 20 months. A median OS of 24 months and a 2-year OS rate of 49.9% were observed.

This study demonstrated an improvement of the prognosis of patients with metastatic STS compared to previously published data exceeding the highest previously reported median OS by 6 months [12]. Since therapy of metastatic STS largely depends on systemic treatment, improvements in drug development might have been a decisive factor. Over the last decade, several new drugs have been added to the therapeutic armamentarium of STS, mainly second-line therapies such as gemcitabine, paclitaxel, pazopanib, trabectedin, eribulin, and various tyrosine kinase inhibitors. Of the active substances mentioned, however, only gemcitabine in the second-line setting was able to demonstrate a significant OS advantage (when compared with dacarbazine) in a randomized controlled trial [20]. For eribulin, a significantly prolonged OS compared to dacarbazine in the second-line setting was published in 2016 [10]. Therefore, the introduction of new compounds in systemic therapy over the last 10–20 year does not completely explain the OS increase. However, the combination and sequencing of systemic treatments is associated with prolonged OS and thereby might have caused a survival advantage [21]. Recent studies show that 30% of STS patients in Europe receive a combination therapy and up to 50% a sequence of 3 or more therapy lines, also including newer active compounds [13, 21].

Other possibly relevant factors are improvements in quality of care, such as experienced personal and drug care supporting the ongoing oncological treatments. The present analysis is the first analysis reporting data on metastatic STS patients treated in a specialized high-volume reference center in Germany, whereas comparable studies were carried out in the US or other European countries [4, 12, 13]. However, the European care of STS patients shows great differences, especially with regard to the definition of sarcoma reference centers [22]. The latter have increasingly taken over the treatment of STS patients in recent years [23, 24]. Those patients more often receive guideline-oriented therapy, radio- and chemotherapy, access to clinical studies, and have a prolonged life expectancy [25, 26]. The establishment of STS centers is also critical for the conduct of clinical studies on STS and the access to new therapies for STS patients [27].

Several prognostic variables have been analyzed. As commonly reported, male sex (p = 0.0005) and higher age (p = 0.003) were confirmed as independent risk factors for decreased OS [12, 28]. The anatomic location of the initial primary tumor appeared to be a significant prognostic factor as well. Multivariate analysis demonstrated that patients with a primary tumor located in the internal trunk did worse with a median OS of only 11 months (p < 0.001). Tumors of the trunk wall had the best outcome and, therefore, were used as a reference. These results could be due to surgical reasons, but also due to the bigger size of those tumors and the shorter interval between resection and metastases. Short DFI in the literature has been reported several times as an independent predictor for higher mortality, which might be due to tumor biology and the ability for more aggressive spread [11, 12, 19]. In the present analysis, presence of synchronous metastasis was predictive of lower OS (p = 0.001). Concerning histological subtype, patients with metastatic MPNST, synovial sarcoma, and NOS sarcoma showed the lowest OS. Those findings are comparable with several reports, despite major classification changes over recent years [12, 13]. No significant differences in OS could be found between patients with metastases at one site and those with metastases at more sites. Due to largely overlapping groups, no risk analysis for those sites was undertaken.

The authors would like to thank all patients and staff who participated in this study.

Prior to the start of the study, ethical approval was obtained from the Ethics Committee II of the University of Heidelberg (Medical Faculty of Mannheim, Registry No. 2015-827R-MA). Identifiable data was removed.

The authors have no conflicts of interest to declare.

There are no funding sources to declare.

J.L. developed the study design, collected and analyzed the data, and wrote the manuscript. B.K. conceived and developed the study and supervised the work throughout the process. F.M., N.V., and P.H. helped to gain access to the medical data and took important part in the treatment of the observed cohort. All authors read the manuscript and approved its final version.