Clinical characteristics of sarcoma patients: a population-based data analysis from a German clinical cancer registry

For the present study, data from the clinical cancer registry of Saxony-Anhalt were analyzed. This population-based registry is regulated by German federal and state law and incorporates data that are transferred from healthcare facilities in Saxony-Anhalt.

Among other information, data sets include structured information on tumor, node, metastasis (TNM)-stage, grading and histology, date of birth, cause and date of death, date of diagnosis (month as smallest temporal unit in each date variable) and treatment. Furthermore, information on treatment procedures such as administered radiation dose and fractionation or number of surgeries was included. Additionally, the TNM-stage referred to in this data set, to the clinical or pathological stage (if an operation was performed). Pathological stages were favored if both ratings differed. Some cases showed incomplete information. If information on subclassification of T-stages (e.g., T1a, T1b) was not available, cases were classified as subgroups T1–4. Likewise, Union for International Cancer Control (UICC)-stages were defined as I–IV. Analogous to former study designs, age groups were defined as follows: <  = 14 years, 15–29 years, 0–44 years, 45–59 years and older than 60 years (Fukushima et al. 2021; Xu et al. 2021).

Regarding a high number of histological subtypes, the cutoff for inclusion in the study was set at more than five documented cases for each subgroup. Furthermore, soft tissue sarcomas of the extremity and trunk not included in other subgroups were categorized as “soft tissue sarcoma.” Remaining cases were summarized as “other histological subtype.”

Tumor localization was defined based on International Statistical Classification of Diseases and Related Health Problems (ICD)-codes for sarcoma cases. All ICD-codes C49.0–C49.9 were used to assign tumor locations for all recorded sarcoma cases. We did not define tumor locations for organ-related sarcomas because of differing ICD-codes and classification criteria (e.g., pleural mesothelioma, mesothelioma, gastrointestinal stromal tumors). Further, tumor size was estimated by T-stage in accordance with the American Joint Committee on Cancer (AJCC) staging criteria (Cates 2018). Histological grades were defined as low, intermediate and high grade sarcoma according to the “Arbeitsgemeinschaft Deutscher Tumorzentren, ADT” (Stegmaier et al. 2019). For illustrative purposes, the 10 most frequent histological subtypes were illustrated with Kaplan–Meier curves.

All primary therapies within 365 days after diagnosis were considered in our analysis.

Cases diagnosed between 2005 and 2022 (most recent data with sufficient quality) were included for this analysis. Cases were censored at October 2022 (latest complete recording of death) or after 60 months to avoid a bias due to cases that died in more recent years, but whose changed survival status had not yet been considered in the data.

We used proportional hazard Cox regression models to assess the association of cancer-related parameters with mortality and computed hazard ratios (HR) with 95% confidence intervals (CIs).

Furthermore, we computed univariate and multivariate Cox regression models. All models were adjusted for histological grade (as defined above), age at diagnosis, T-stage, N-stage, M-stage, tumor location based on ICD-codes as described above, histological subtype, tumor side, patient sex and UICC-stage. For illustrative purposes, Kaplan–Meier curves were created for all included risk factors. Our primary endpoints were overall survival (OS) and progression-free survival (PFS). Additionally, we computed median OS and PFS survival rates.

A significance level of 5% was used. All statistical analyses were performed using RStudio Version 1.4.1717 (RStudio 2020, Integrated Development for R. RStudio, PBC, Boston, MA, USA). All graphics were computed with the “gtsummary” package in RStudio (Sjoberg et al. 2021).

The total number of UICC stage I–VI sarcoma cases diagnosed between 2005 and 2022, provided by the cancer registry of Saxony-Anhalt, was 1529 (Cases were reported in accordance with the German Manual of Cancer Registration (Stegmaier et al. 2019)). In order to meet the specific aim of this study, we excluded subgroups with different tumor biology and low number of cases and focused on soft tissue sarcoma patients. Consequently, all patients diagnosed with gastrointestinal stromal tumors, Ewing sarcoma, (pleural) mesothelioma, chondrosarcoma, Kaposi sarcoma and schwannoma were excluded from further analysis. Incomplete data with undefined histological subtypes or lacking ICD-codes were removed from the study as well, leading to a total number of 752 excluded cases. This resulted in a total number of 787 cases included for analysis (Fig. 1).

All patient characteristics can be viewed in Table 1. The majority of analyzed patients were men (838 male, female). Based on higher sarcoma prevalence rates in children and young adults, a subgroup analysis of patients younger than 40 years was performed, in accordance with former studies (Avila et al. 2018). 6.6% (n = 102) of included patients were younger than 40 years. On average, patients were 68 years old when they were diagnosed. The majority of sarcoma patients was diagnosed with leiomyosarcoma (12%), liposarcoma (11%), pleural mesothelioma (5.4%) and angiosarcoma (5.0%). In the younger age group, rhabdomyosarcoma (5.9%) and synovial sarcoma (5.9%) had a higher prevalence compared with patients > 40 years. One-third of all patients had a grade 3 (27%) or grade 4 (5.2%) sarcoma. In contrast, younger patients were more likely to have a low-grade sarcoma (G1: n = 11, 19%; G2: n = 16, 28%). In a high percentage of cases, information on histological grading was not determined (13%) or unknown (13%). The most frequent T-stages were stage 2 (15%) and 2b (18%). T-stages could be assessed in 34% percent of sarcoma cases. Furthermore, the vast majority of cases had no lymphonodal metastases (N-stage = N0, 63%). Likewise, 63% of the included cases were not diagnosed with distant metastases. Analogous to measured T-stages, N-stage data were missing in 32% of collected cases. M-status data could not be obtained in 11% of sarcoma cases. Twenty-seven percentage of the patients were classified as UICC-stage IV. Moreover, in the younger patient group, almost half of included cases were staged as UICC IV patients (n = 20, 44%). This high proportion of advanced cancer stages could be explained by the higher distant metastasis rate in younger patients (n = 17, 35%) compared to older patients (n = 152, 21%). Regarding tumor localization, the majority of sarcomas were located in the lower extremity (32%), abdomen (16%) and in the head, face and neck region (11%). The upper extremity region was affected in only 9.5% of analyzed cases. Younger patients had more sarcomas in the trunk (n = 8, 9.8%) and upper extremity (n = 12, 15%) region compared to patients older than 40 years. In contrast, thoracic sarcomas occurred more often in older patients (patients > 40 years: n = 132, 12%; patients < 40 years: n = 2, 2.4%). Concerning tumor side, left- and right-sided tumors were equally distributed (right side: n = 440, 45%; left side: n = 513, 52%). Midline and both-sided tumors were comparatively rare (midline: n = 24, 2.4%; both sides: n = 7, 0.7%). Most patients were treated with surgery (n = 484, 36%). Radiation therapy was performed after surgery in 287 (21%) cases. A large proportion of younger patients received additional systemic therapy after surgery and radiation therapy (n = 22, 26%). For not documented reasons, 15% of all included sarcoma patients did not receive any treatment (n = 201). However, in the younger patients’ group, 9 patients (10%) remained untreated.

Univariate and multivariate Cox regression models based on overall survival are viewed in Table 2. We computed univariate and multivariate Cox regression models for OS and PFS. In our univariate survival models, we could not find a survival benefit regarding sex. Female patients had a slightly higher mortality risk compared to men (HR 1.13, 95% CI 0.92–1.37). Younger age (15–29 years) was related to a worse survival compared to older patients > 60 years (HR 1.34, 95% CI 0.63–2.83). Sarcomas identified in more than one location had a higher mortality risk compared to sarcomas located at a single site (HR 7.10, 95% CI 2.20–22.9). Upper extremity sarcomas were associated with a better overall survival (HR 0.84, 95% CI 0.55–1.26) compared to lower extremity sarcomas (HR 1.44, 95% CI 1.05–1.96). Patients diagnosed with pelvic and thoracic sarcomas had a slightly higher mortality risk (Pelvis: HR 1.27, 95% CI 0.86–1.89; Thorax: HR 1.24, 95% CI 0.84–1.82) compared to patients with sarcomas localized in the head and neck or trunk region (Head, Face and Neck: HR 1.35, 95% CI 0.89–2.06; Trunk: HR: 1.10, 95% CI 0.58–2.09).

Regarding adjusted covariates, lower UICC-stages were associated with a longer survival in our univariate Cox regression models. Further, higher histological grades and TNM-stages were related to a higher mortality risk. In our regression models, tumor side had no significant influence on overall survival. Left- and right-sided tumors had a statistically non-significant worse survival outcome related to both-sided sarcomas (left side: HR 1.12, 95% CI 0.28–4.55; right side: HR 1.36, 95% CI 0.33–5.50). Moreover, differing histological subtypes had significant influence on overall survival. Patients diagnosed with fibromyxoid sarcoma, rhabdomyosarcoma and angiosarcoma were related to higher mortality risk compared to other histological subtypes (fibromyxoid sarcoma: HR 5.2, 95% CI 0.71–38.1; rhabdomyosarcoma: HR 2.93, 95% CI 1.44–6.00; angiosarcoma: HR 1.07, 95% CI 0.53–2.18). In our data, liposarcoma, leiomyosarcoma and myxofibrosarcoma diagnosis were beneficial in terms of survival (liposarcoma: HR 0.55, 95% CI 0.39–0.79; leiomyosarcoma: HR 0.59, 95% CI 0.42–0.82; myxofibrosarcoma: HR 0.42, 95% CI 0.23–0.79) (see Table 2).

In our multivariate regression models, female patients had a higher mortality risk compared to men (HR 1.13, 95% CI 0.92–1.37). Differing to our univariate regression models, a younger age at diagnosis was not a strong risk factor in terms of survival compared with patients > 60 years (30–44 years: HR 0.23, 95% CI 0.02–2.28). In contrast, tumor localization had no statistically significant effect on survival. Multilocular tumor localizations, tumors localized in head, face and neck region as well as lower extremity sarcomas were associated with increased mortality risk (head, face and neck: HR 2.05, 95% CI 0.40–10.4; more than one location: HR 3.74, 95% CI 0.18–79.9; lower extremity: HR 1.47, 95% CI 0.55–3.88). Similar to our univariate regression results, worse histological grades and higher TNM-stages (T-status) were related to a higher mortality risk. Nodal and organ-related metastases were associated with higher mortality. Lower UICC-stages were associated with a better survival in our multivariate Cox regression models. Right-sided tumors had a slightly better survival compared to both-sided tumors (HR 0.92, 95% CI 0.47–1.80). Angiosarcoma, fibromyxoid sarcoma, synovial sarcoma and rhabdomyosarcoma were related to a higher mortality risk compared to other histological subtypes (angiosarcoma: HR 9.76, 95% CI 0.86–110; rhabdomyosarcoma: HR 21.6, 95% CI 0.79–591; fibromyxoid sarcoma: HR 5.20, 95% CI 1.13–287; synovial sarcoma: HR 4.14, 95% CI 0.90–19.1). In comparison, liposarcoma and pleomorphic sarcoma and myxofibrosarcoma patients had a survival benefit related to other subtypes (liposarcoma: HR 0.81, 95% CI 0.27–2.47; pleomorphic sarcoma: HR 0.64, 95% CI 0.10–4.27; HR myxofibrosarcoma: 0.86, 95% CI 0.18–4.07) (Table 2). Figures 2, 3, 4, 5, 6 show Kaplan–Meier curves of all included prognostic factors based on overall survival.

The median overall survival (OS) of all patients was 24 months (95% CI 16–22 months). Median survival rates differed between histological subtypes. Myxofibrosarcoma, spindle cell sarcoma, liposarcoma and leiomyosarcoma had the best median survival (myxofibrosarcoma: 56 months; spindle cell sarcoma: 50 months; liposarcoma: 32 months; Leiomyosarcoma: 27 months). Fibromyxoid sarcoma and rhabdomyosarcoma had the worst median survival rates (fibromyxoid sarcoma: 5 months; rhabdomyosarcoma: 6 months) (Table 3).

In addition, median progression-free survival was computed. The median progression-free survival was 15 months (95% CI 12–19 months). The histological subtypes with the best median non-progression rates were myxofibrosarcoma, leiomyosarcoma and fibrosarcoma (myxofibrosarcoma: 40 months; leiomyosarcoma: 25 months; fibrosarcoma: 22 months). Rhabdomyosarcoma, angiosarcoma and MPNST were the histological subtypes with the earliest progression rates (rhabdomyosarcoma: 2.5 months; MPNST: 9 months; angiosarcoma: 7 months). All extremity and trunk soft tissue sarcomas also had an early median progression at 9 months (95% CI 6–21 months).

Regarding our second endpoint PFS (Table 4), women had a progression-free survival benefit in the univariate regression model. The multivariate model showed an association of female sex and worse survival (univariate model female: HR 0.96, 95% CI 0.73–1.27; multivariate model female: HR 1.15, 95% CI 0.42–3.17).

Younger patients (15–29 years) had a higher risk of tumor progression compared to patients > 60 years (univariate model 15–29 years: HR 0.96, 95% CI 0.59–2.48; multivariate model 15–29 years: HR 60.5, 95% CI 2.51–1458). Tumor localization was not a prognostic factor regarding PFS in univariate and multivariate Cox regression models. Better histological grading was associated with a higher chance of non-progression in our univariate regression models (low grade: HR 0.62, 95% CI 0.42–0.89). Moreover, rhabdomyosarcoma had a higher risk in progression compared to other histological subtypes (rhabdomyosarcoma: HR 8.96, 95% CI 1.99–40.3). In comparison, liposarcoma and pleomorphic sarcoma and myxofibrosarcoma patients had a non-progression benefit related to other subtypes in our univariate regression model (liposarcoma: HR 0.45, 95% CI 0.27–0.73; leiomyosarcoma sarcoma: H 0.48, 95% CI 0.29–0.79; HR: myxofibrosarcoma: 0.29, 95% CI 0.14–0.64). Analogous to tumor localization, tumor side and tumor size had no prognostic impact on PFS in our regression models. Lower TNM-stages as well as lower UICC-stages were related to a non-progression (Table 3). Figures  7, 8, 9, 10, 11 show Kaplan–Meier curves of all included prognostic factors based on PFS.