Characteristics of the TARPSWG and CR cohort are shown in Table 1. While the gender distribution was comparable in both cohorts, strong age differences were observed with a median age of 64 years in the CR and 58 years in TARPS cohort (p < 0.0001). While 23.9% in the CR cohort were 75 years or older, the proportion was only 8.7% in the TAPRS cohort.

The CR cohort had fewer Grade 1/2 than TARPSWG (51.9% vs. 72.6%), and a correspondingly greater proportion of Grade 3 (48.1% than 27.4%). This difference remained significant after adjustment for age and sex (p < 0.0001). In terms of histological subtypes, the proportion of DDLS (37.1% CR vs. 37% TARPSWG) and LMS (20.1% vs. 19.2%) was comparable between both cohorts. By contrast, there were differences in the proportion of WDLS (15.1% vs. 26.1%), MPNST (0.8% vs. 3.3%), SFT (0.5% vs. 5.9%) and the group of “other” histological subtypes (22.8% vs. 6.3%). Histological distributions remained different after adjustment (p < 0.0001). The resection margin status was similar in both cohorts (adjusted p = 0.7608).

There was no significant difference in the proportion of patients who received chemo- and/or radiation therapy between the two cohorts (adjusted comparison: p = 0.7175 for chemotherapy, p = 0.6208 for radiotherapy). With respect to the timing of therapy, planned preoperative (neoadjuvant) therapy was given less often in the CR cohort than the TARPSWG cohort (chemotherapy: 7.1% vs. 15.1%, p = 0.0001; radiotherapy: 14% vs. 21.7%, p = 0.0016). This pattern remained significant after adjustment for age and sex. Postoperative adjuvant chemotherapy was given more commonly in the CR cohort (9.3% vs. 3.9%, p = 0.0001). After adjustment, post-operative radiotherapy was also given more commonly in the CR cohort (p = 0.0087).

As shown in Fig. 2a, overall survival (OS) at 5 years following resection was lower in the CR than in the TARPSWG cohort (56.3% vs. 67.9%, p = 0.0015). After adjustment for age, sex, histotype and r-status the hazard ratio (HR) was 0.99 (0.77–1.29).

Overall survival curves CR vs. TARPSWG cohorts

Table 2 shows the comparison between patients with WDLS in CR vs. TARPSWG cohorts. In terms of grade, surgical margins, receipt of adjuvant RT and 5-year LR and DM rates the cohorts were comparable. In the CR cohort, age was on average 5 years older, and there was less frequent receipt of chemotherapy (1.8% vs. 5.7%) or radiotherapy (12.7% vs. 20.7%), in particular less neoadjuvant chemotherapy (0% vs. 5.4%) and/or neoadjuvant radiotherapy (7.3% vs. 15.7%). Age- and sex-adjusted analyses on treatment showed the same direction of the association but no significance and could sometimes not be conducted due to small case numbers. The 5-year OS for patients with WDLS was lower in the CR compared to the TARPSWG cohort (71.2 vs. 89.9%, HR 2.42).

Compared to the TARPSWG cohort, patients with DDLS in the CR cohort were less often female (31.1% vs. 42.4%, adjusted OR: 0.64), were on average six years older, had less often grade 2 (34.9% vs. 60.4%, OR: 0.33) but more often grade 3 and had less often an R2 status (2.9% vs. 6.8%, OR: 0.42, Table 2). While receipt of chemo- and radiotherapy were overall comparable in both cohorts, the setting was more often post-surgery in the CR cohort (chemotherapy: 10.4% vs. 3%, OR 4.38; radiotherapy: 15.6% vs. 10.3%; OR 2.05). After adjustment for sex and age, there was a tendency to higher 5-year OS (0.88) and CS (0.86) rates for patients in the CR. Figure 3 shows the crude cumulative incidence function of LR. The incidence of LR is higher in the CR cohort after 1 year, even after adjustment for sex and age (HR 1.69), but comparable after 5-years.

Crude cumulative incidence of local recurrence in patients with DDLS, CR vs. TARPS cohorts

Table 2 shows the comparison between patients with LMS in the CR vs. TARPSWG cohorts. In terms of sex, grade, surgical margins and 5-year LR rate, the cohorts were comparable. In the CR cohort, age was on average 10 years older, and patients were less likely to receive chemotherapy (12.3% vs. 28.6%) and/or radiotherapy (24.7% vs. 37.5%), particularly neoadjuvant chemotherapy (6.8% vs. 23.4%) and/or radiotherapy (9.6% vs. 27.6%), than in the TARPSWG cohort. After adjustment for sex and age, 5-year OS, CS, and cumulative incidence of local recurrences were comparable. While there was overall no significant difference in the cumulative incidence of distant metastasis within 5 years (crude: p = 0.3084, Fig. 4, adjusted HR: 0.81), five to 58 months following resection the incidence was slightly lower in the CR than the TARPSWG cohort.

Crude cumulative incidence of distant metastasis in patients with LMS