11513 Background: Surgery is the mainstay of treatment for patients with retroperitoneal sarcoma (RPS), but this can be challenging, and recurrence rates are high. Novel treatment approaches are needed. In this study, we sought to 1) determine the frequency and potential predictors of radiologic tumor response and 2) assess clinical outcomes in patients with primary high risk RPS who were treated at sarcoma referral centers with neoadjuvant systemic therapy followed by surgery. Methods: Clinicopathologic data was retrospectively collected for eligible patients treated from 2008-2018 at 13 institutions within the Transatlantic Australasian Retroperitoneal Sarcoma Working Group (TARPSWG). For each patient, preoperative objective response (RECIST1.1) was reported by each institution. Univariable and multivariable logistic models were performed to determine predictors of response. Kaplan-Meier plots were constructed for overall survival (OS) and cumulative incidences of local recurrence (LR) and distant metastasis (DM). Results: In total, 158 RPS patients were included in this study. A median of 3 cycles (IQ range 2-4) of neoadjuvant systemic therapy were given. No complete responses were observed. Partial response (PR) was seen in 37 patients (23%), stable disease (SD) in 88 (56%) and progressive disease (PD) in 33 (21%). Subtype-specific differences were seen including PR in 5 out of 11 (45%) patients with undifferentiated pleomorphic sarcoma. Overall, higher number of cycles given was positively associated with PR (p = 0.005). No other factors including receipt of neoadjuvant radiation therapy were predictive of PR. All patients underwent complete (R0/R1) resection with a major complication (Clavien-Dindo ≥3) rate of 23%. Differences in OS were observed based on preoperative response type (p = 0.005). In grade 3 dedifferentiated liposarcoma, patients who received adriamycin-ifosfamide versus another regimen had decreased LR and improved OS. In leiomyosarcoma, patients who received adriamycin-DTIC versus another regimen had a higher PR rate (37% vs. 16%), decreased LR, DM and improved OS. Limited by low numbers, these subtype-specific data did not reach statistical significance. Conclusions: In patients with high risk RPS, response to neoadjuvant systemic therapy is overall modest and may be regarded as an indicator of disease biology to predict survival after surgery. Subtype-specific regimens should be further validated and incorporated into prospective trials of neoadjuvant systemic therapy in RPS (e.g. STRASS2).
