(P 169) WHAT IS THE MOSTAPPROPRIATE STUDY ENDPOINT FOR PHASE IICLINICAL TRIALS IN SOFT TISSUE SARCOMA? - CHARACTERISTIC OF THE CURRENT LANDSCAPE OF PHASE II TRIALS

Objective: Current therapeutic options for advanced and metastatic soft tissue sarcomas (STS) are still limited, and the prognosis is unfavorable. Many clinical trials, mainly phase II, assess potential new therapies. However, there is high variability and no standardization of phase II designs. Selection of a homogenous study population and appropriate endpoints to prove drug activity and improvement in patients' survival are crucial steps in the trial design, also allowing for minimizing risk of false-positive and false-negative results. We aimed to analyze the current landscape of phase II studies in STS and evaluate how their design can affect the results.

Methods: We have analyzed all phase II studies published between 2010 and 2021. Only studies including patients with STS were included; studies including bone sarcoma were excluded. Detailed information about study design, endpoints, population, and outcomes were extracted. Outcomes of interest included all primary endpoints, progression-free survival (PFS), overall survival (OS), objective response rate (ORR), disease control rate (DCR), and 3- and 6-month PFS rates (3mPFS; 6mPFS). For multi-arm/multicohort studies, each arm/cohort was analyzed separately; randomized comparative trials (RCTs) and single-arm/non-comparative randomized trials were analyzed separately. Weight linear regression was used to assess surrogacy with OS. All cohorts/arms from RCTs and single-arm studies (except those receiving placebo) that reported OS and at least one other endpoint were included in the surrogacy analysis. Additional modeling of study results with 3mPFS or 6mPFS primary endpoints, as recommended by EORTC, has been performed.

Results: We have identified 91 published studies that included 23 RCTs and 108 cohorts from single-arm trials. 56.5% of RCTs were open-label, 52.1% included mixed histotypes, and 30.4% focused on a single STS type. The majority allowed patients in >1 line (56.5%), while 21.7% each was performed in 1 line or ≥1 line. PFS was the most common primary endpoint (73.9%). 34.8% of trials met their primary endpoints. An alternative model using 3mPFS or 6mPFS as a primary endpoint showed that 2 of 5 evaluable negative trials could meet the endpoint, while all 5 positive trials would remain positive.
Most single-arm studies were performed in specific single STS subtypes (58.3%), mainly in leiomyosarcoma or liposarcoma. Regarding the line of therapy, 51.8% enrolled patients in all lines, 32.4% >1 line, and 15.8% in 1 line only. Over 20 different primary endpoints were used, with ORR as the most common (28.7%), followed by 3mPFS (22.2%) and 6mPFS (15.7%). 38.8% (42) of trials met their primary endpoints. Among positive trials, 40.4% were based on 3mPFS as a primary endpoint, 19.0% on 6mPFS, and 7.1% on ORR. From another perspective, only 9.7% of trials designed with ORR as the primary endpoint has met it. Among 58 trials with available data for modeling, all positive trials remained positive, while 61.1% of primarily negative (11/18) would meet the 3mPFS/6mPFS endpoint.
One hundred eleven arms were included in the surrogacy analysis. Among all trials, PFS (R=0.637) and 6mPFS (R=0.694) showed the highest level of surrogacy with OS, while ORR had the lowest (R=0.540). When trials were divided according to the class of tested agents (chemotherapy-based vs. non-chemotherapy) (Tab. 1) or line of therapy (1st vs.>1 vs. all lines) (Tab. 2), differences in surrogacy level were observed, mainly for ORR.

Conclusion: There is high heterogeneity in sarcoma trial designs, mainly in selecting the primary endpoint and hypotheses used for size calculation. ORR is the most commonly used primary endpoint in single-arm trials; however, its surrogacy with OS varies depending on the line of therapy and class of tested drugs. 3- or 6-month PFS rates are associated with similar or higher surrogacy than ORR in most settings, especially in trials with targeted agents or immunotherapy. Overall, there is an unmet need for standardization of trial design that will incorporate factors associated with the rarity and heterogeneity of the disease, historical data from clinical trials and real-life studies, and specific characteristics of new therapeutic agents.

tab1_Ctos.docx

Tab2_ctos.docx