Purpose: To explore patterns of failure and postrelapse outcome of patients with retroperitoneal sarcoma primarily treated by extended resection.
Methods: All consecutive patients with primary retroperitoneal sarcoma, treated between January 2002 and December 2011 at two European sarcoma centers were included. Five-year overall survival (OS) and crude cumulative incidence (CCI) of local recurrence (LR) and distant metastases (DM) were calculated. Multivariate analyses for OS and CCI of LR and DM were performed. Postrelapse OS was investigated.
Results: A total of 377 patients were identified. Median follow-up from the time of primary surgery was 44 months [interquartile range (IQR) 27-82]. Five-year OS was 64 % [95 % confidence interval (CI) 0.588, 0710]. CCI of LR and DM were 23.6 % (95 % CI 18.9, 29.4) and 21.9 % (95 % CI 17.6, 27.3), respectively. OS, CCI of LR and DM were 87, 18 % and 0 for well-differentiated liposarcoma; 54, 44 and 9 % for grade II dedifferentiated liposarcoma; 41,33, and 44 % for grade III dedifferentiated liposarcoma; 58, 5, and 55 % for leiomyosarcoma; and 85, 4, and 17 % for solitary fibrous tumor, respectively. Seventy-six patients developed LR. Median postrelapse follow-up was 27 months (IQR 10-58). Twenty-one patients (27 %) underwent a second surgical resection (complete in 18), while 55 (73 %) did not (22 multifocal, 17 inoperable, 16 other causes). Median postrelapse OS was 17 months (IQR 7-31). Well-differentiated liposarcoma histology and disease-free interval predicted postrelapse OS, while surgical resection did not.
Conclusions: When primary extended surgery limits LR, histologic subtype and grade determine the outcome. At recurrence, a second surgery is of limited benefit.