Comparison of published risk models for prediction of outcome in patients with extrameningeal solitary fibrous tumour

Aims Solitary fibrous tumours (SFTs) are fibroblastic mesenchymal tumours with a 10-30% metastatic rate. Several risk models have been proposed for extrameningeal SFT, but they have not been evaluated in direct comparison with each other. The aim of this study is to compare the utility of published risk models in a multi-institutional SFT cohort. Methods and results Clinicopathological data were evaluated for a cohort of extrameningeal SFTs, and used to stratify tumours by the use of five proposed risk models designed for soft tissue and/or pleural SFT [modified Demicco, Pasquali, Salas overall survival (OS), Salas metastasis, and Salas local recurrence (LR)]. Kaplan-Meier and Cox proportional hazards models were used to assess OS, time to first metastasis, time to first LR, and recurrence-free survival (RFS). The study included 303 patients (109 from a referral cancer treatment centre; previously described in the original Demicco model) and an independent cohort from two large hospitals (n = 194). The median patient age was 54 years, and the median clinical follow-up (available for 220 patients) was 37 months. The independent cohort had a 13% risk of metastasis at 5 years and a 16% risk of metastasis at 10 years. In this cohort, the modified Demicco, Salas OS, and Salas metastasis models predicted metastasis and RFS, whereas the Pasquali model had the best correlation with OS. Conclusions Multivariate risk models that include mitotic rate and patient age can more accurately predict aggressive behaviour in SFTs, with the modified Demicco and Salas OS risk models showing the best correlation with metastasis and RFS.