Preoperative or postoperative radiotherapy versus surgery alone for retroperitoneal sarcoma: a case-control, propensity score-matched analysis of a nationwide clinical oncology database

Recruitment into clinical trials for retroperitoneal sarcoma has been challenging, resulting in termination of the only randomised multicentre trial in the USA investigating perioperative radiotherapy. Nonetheless, use of radiotherapy for retroperitoneal sarcoma has increased over the past decade, substantiated primarily by its established role in extremity sarcoma. In this study, we used a nationwide clinical oncology database to separately compare overall survival for patients with retroperitoneal sarcoma who had surgery and preoperative radiotherapy or surgery and postoperative radiotherapy versus surgery alone.

We did two case-control, propensity score-matched analyses of the National Cancer Data Base, which included adult patients with retroperitoneal sarcoma who were diagnosed from 2003 to 2011. Patients were included if they had localised, primary retroperitoneal sarcoma. Patients were classified into three groups based on use of radiotherapy: preoperative radiotherapy, postoperative radiotherapy, and no radiotherapy (surgery alone). Patients were excluded if they received both preoperative radiotherapy and postoperative radiotherapy, or if they received intraoperative radiotherapy. Parallel propensity score-matched datasets were created for patients who received preoperative radiotherapy versus those who received no radiotherapy and for patients who received postoperative therapy versus those who received no radiotherapy. Propensity scores were calculated with logistic regression, with multiple imputation and backwards elimination, with a significance level to stay of 0·05. Matching was done with a nearest-neighbour algorithm and matched 1:2 for the preoperative radiotherapy dataset and 1:1 for the postoperative radiotherapy dataset. The primary objective of interest was overall survival for patients who received preoperative radiotherapy or postoperative radiotherapy compared with those who received no radiotherapy within the propensity score-matched datasets.

9068 patients were included in this analysis: 563 in the preoperative radiotherapy group, 2215 in the postoperative radiotherapy group, and 6290 in the no radiotherapy group. Matching resulted in two comparison groups (preoperative radiotherapy vs no radiotherapy, and postoperative radiotherapy vs no radiotherapy) with negligible differences in all demographic, clinicopathological, and treatment-level variables. In the matched case-control analysis for preoperative radiotherapy median follow-up time was 42 months (IQR 27-70) for the preoperative radiotherapy group versus 43 months (25-64) for the no radiotherapy group; median overall survival was 110 months (95% CI 75-not estimable) versus 66 months (61-76), respectively. In the matched case-control analysis for postoperative radiotherapy median follow-up time was 54 months (IQR 32-79) for patients in the postoperative radiotherapy group and 47 months (26-72) for patients in the no radiotherapy group; median overall survival was 89 months (95% CI 79-100) versus 64 months (59-69), respectively. Both preoperative radiotherapy (HR 0·70, 95% CI 0·59-0·82; p<0·0001) and postoperative radiotherapy (HR 0·78, 0·71-0·85; p<0·0001) were significantly associated with improved overall survival compared with surgery alone.

To the best of our knowledge, this is the largest study to date of the effect of radiotherapy on overall survival in patients with retroperitoneal sarcoma. Radiotherapy was associated with improved overall survival compared with surgery alone when delivered as either preoperative radiotherapy or postoperative radiotherapy. Together with the results from the ongoing randomised EORTC trial (62092-22092; NCT01344018) investigating preoperative radiotherapy for retroperitoneal sarcoma pending, these data might provide additional support for the increasing use of radiotherapy for patients with retroperitoneal sarcoma undergoing surgical resection.

Department of Surgery, Duke University School of Medicine.