Impact of Pathological Stratification on the Clinical Outcomes of Advanced Well-Differentiated/Dedifferentiated Liposarcoma Treated with Trabectedin

SIMPLE SUMMARY: The treatment options for advanced well differentiated liposarcoma (WDLPS) and dedifferentiated liposarcoma (DDLPS) are scant, especially after the failure of first-line, anthracycline-based treatment, when trabectedin is one of the most active agents currently approved. Thus, the identification of biological characteristics allowing an efficient patients stratification in this setting appears mandatory. Here, applying a grading-based stratification, we showed that trabectedin may be more active against low-grade LPS (i.e., WDLPS and low-grade DDLPS) than in high-grade LPS (i.e., high-grade DDLPS). If confirmed, our data might allow the implementation of grading as a tool patients’ stratification in this setting. ABSTRACT: Background. We previously showed that grading can prognosticate the outcome of retroperitoneal liposarcoma (LPS). In the present study, we aimed to explore the impact of pathological stratification using grading on the clinical outcomes of patients with advanced well-differentiated LPS (WDLPS) and dedifferentiated LPS (DDLPS) treated with trabectedin. Patients: We included patients with advanced WDLPS and DDLPS treated with trabectedin at the Fondazione IRCCS Istituto Nazionale dei Tumori between April 2003 and November 2019. Tumors were categorized in WDLPS, low-grade DDLPS, and high-grade DDLPS according to the 2020 WHO classification. Patients were divided in two cohorts: Low-grade (WDLPS/low-grade DDLPS) and high-grade (high-grade DDLPS). Results: A total of 49 patients were included: 17 (35%) in the low-grade cohort and 32 (65%) in the high-grade cohort. Response rate was 47% in the low-grade cohort versus 9.4% in the high-grade cohort (logistic regression p = 0.006). Median progression-free survival (PFS) was 13.7 months in the low-grade cohort and 3.2 months in the high-grade cohort. Grading was confirmed as an independent predictor of PFS in the Cox proportional-hazards regression multivariable model (adjusted hazard ratio low-grade vs. high-grade: 0.45, 95% confidence interval: 0.22–0.94; adjusted p = 0.035). Conclusions: In this retrospective case series, sensitivity to trabectedin was higher in WDLPS/low-grade DDLPS than in high-grade DDLPS. If confirmed in larger series, grading could represent an effective tool to personalize the treatment with trabectedin in patients with advanced LPS.