A comparison of the risk prediction models sarculator and PERSARC in patients with localized soft tissue sarcoma of the extremities and trunk wall

BackgroundDifferent risk classification criteria are used to select patients with localized soft tissue sarcoma (STS) of the extremities and trunk wall for perioperative chemotherapy. The most frequently used risk classification methods are based on Sarculator and PERSARC prediction models. The aim of this study was to evaluate and compare these two methods.MethodsThe study cohort consisted of 664 patients treated at Oslo University Hospital from 1998 to 2017. Predicted probabilities of distant metastasis (DM) and overall survival (OS) were calculated and risk classification was performed according to previously defined thresholds. Hazard ratios (HRs) were estimated using Cox models. Interaction between perioperative chemotherapy and risk groups was included to investigate the effect of chemotherapy according to risk group.ResultsA high correlation was found between Sarculator and PERSARC in predicted 5-year probability of DM (r=0.849; 95% CI 0.825-0.871) and 5-year OS (r=0.908; 95% CI 0.892-0.921). 215 of 664 (32.4%) and 221 of 569 (38.8%) patients were classified as high-risk according to Sarculator and PERSARC, respectively, with agreement found in 511 of 569 patients (89.8%). Estimated 5-year OS for Sarculator high-risk was 43% and for PERSARC high-risk 44%. Patients classified as high-risk by only one method had similar metastasis-free survival (MFS; HR 0.83; 95% CI 0.55-1.24) and OS (HR 1.11; 95% CI 0.79-1.55) as patients who were high-risk using both methods. Perioperative chemotherapy was associated with improved OS in all high-risk groups. In Sarculator high-risk HR was 0.34 (95% CI 0.20-0.58), in PERSARC high-risk 0.32 (95% CI 0.19-0.55) and in patients classified as high-risk by at least one model 0.33 (95% CI 0.19-0.55). Similar results were obtained for MFS (data not shown).ConclusionsA high degree of agreement between Sarculator and PERSARC was observed. Patients classified as high-risk by only one method had similar outcome as those who were high-risk using both. Chemotherapy was associated with improved outcome in all high-risk groups. We propose that patients classified as high-risk by at least one method should be defined as high-risk.Legal entity responsible for the studyOslo University Hospital.FundingSouth-Eastern Norway Regional Health Authority.DisclosureA. Gronchi: Financial Interests, Personal, Advisory Board: Novartis, Pfizer, Bayer, Lilly, PharmaMar, SpringWorks, Nanobiotix, Boehringer Ingelheim; Financial Interests, Personal, Invited Speaker: Lilly, PharmaMar, Deciphera; Financial Interests, Institutional, Research Grant: PharmaMar, Nanobiotix. K. Boye: Financial Interests, Institutional, Advisory Board, Expert testimony on national applications to regulatory authorities: Bayer; Financial Interests, Institutional, Invited Speaker: Eli Lilly; Financial Interests, Personal, Advisory Board: Bayer, GSK, Incyte, NEC Oncoimmunity, Deciphera; Financial Interests, Personal, Invited Speaker: Deciphera; Financial Interests, Institutional, Research Grant: Eli Lilly; Non-Financial Interests, Principal Investigator: Deciphera, Novartis, Boehringer Ingelheim; Non-Financial Interests, Institutional, Product Samples: Merck. All other authors have declared no conflicts of interest.