Early recurrence and prognosis in patients undergoing resection of primary retroperitoneal sarcoma: an international, retrospective cohort study.

Abstract

Complete surgical resection of retroperitoneal sarcomas (RPS) remains the cornerstone of potentially curative treatment. However, recurrence is common and timing of relapse is thought to reflect tumour biology. We aimed to determine the impact of early recurrence on prognosis in patients undergoing resection of RPS, and to identify preoperative predictors of early recurrence. In this international, retrospective cohort study, consecutive patients (aged 19-90 years) with primary, non-metastatic RPS treated between March 2012 and October 2019 at two high-volume referral centres in Milan, Italy and Toronto, Canada were identified from prospectively maintained databases. Patients with Ewing sarcoma, alveolar/embryonal rhabdomyosarcoma, desmoid tumour, gynaecologic sarcoma, or GIST were not eligible for inclusion. The primary outcome of interest was overall survival (OS) from diagnosis. Patients were grouped by disease-free interval (DFI) following resection: 0-6, 7-12, 13-18, 19-24, and >24 months. Patients who did not undergo resection served as a reference group. Multivariable analyses evaluated predictors of OS. Existing prognostic tools, such as the Sarculator nomogram and the Inflammatory Biomarkers Prognostic Index (IBPI), were tested as potential preoperative predictors of early recurrence (DFI 0-6 months). 651 patients (median age 62.7 years) were included; of whom 566 (87%) underwent resection and 85 (13%) did not. Median follow-up from time of diagnosis was 75.9 months (IQR 58.6-99.5). Of the 566 patients who were resected, 259 (46%) developed recurrence, including 49 (19%) within 6 months. Patients with DFI 0-6 months had similar OS to that of unresected patients (2-year OS: 51.6% vs 42.7%, p = 0.32), while later recurrence was associated with improved OS (p < 0.001). On multivariable analysis, DFI was independently associated with OS, but patients with early recurrence had a hazard ratio for death of 0.96 compared with unresected patients. No preoperative clinical variable, nor Sarculator or IBPI, reliably predicted early recurrence. These descriptive findings show that nearly one in five patients who recur after resection of primary RPS do so within 6 months, with OS comparable to those not undergoing surgery. Whilst acknowledging the limitations of this retrospective design and that resection is a time-dependent and clinically selected decision, these findings raise questions about the clinical benefit of resection in this subgroup of patients. Improved tools are needed to guide preoperative patient selection and optimise treatment strategies. Further research is warranted. This work was partially funded by Italian Ministry of Health - "Ricerca Corrente" funds.