This multicenter, single-arm, open-label, phase 3 study (ClinicalTrials.gov identifier: NCT04488822) was conducted in nine hospitals from the Chinese Mainland and Taiwan. Patients received pexidartinib 400 mg (2 × 200 mg capsules) twice daily (every 12 h) for a total daily dose of 800 mg. Pexidartinib was taken on an empty stomach at approximately the same time each day. Patients continued to receive pexidartinib until disease progression, unacceptable toxicity, patient or investigator decision, surgery, or pregnancy.

All study documents were approved by the sites’ independent ethics committees or institutional review boards (provided in Supplementary File 1). The study was conducted in compliance with the ethical principles of the Declaration of Helsinki, the International Council for Harmonisation of Technical Requirements for Pharmaceuticals for Human Use, and Good Clinical Practice guidelines. All patients provided written informed consent.

Eligible patients were ≥ 18 years of age (≥ 20 years of age in Taiwan) and had a diagnosis of TGCT that was histologically confirmed by a pathologist and associated with severe morbidity or functional limitation and not amenable to improvement with surgery, as determined by two or more surgeons or a multidisciplinary tumor board. Patients were required to have measurable disease, defined by Response Evaluation Criteria in Solid Tumors (RECIST) v1.1, with a minimal size of 2 cm, as assessed by magnetic resonance imaging (MRI) and read by a central radiologist. The patient’s prescription analgesic regimen was required to be stable for 2 weeks prior to enrollment. Patients were excluded if they had metastatic TGCT, had used an investigational drug or device within 28 days of enrollment, or had previously received pexidartinib or any biologic treatment targeting CSF1 or the CSF1 receptor. Prior use of oral tyrosine kinase inhibitors (e.g., imatinib or nilotinib) was allowed.

The primary efficacy endpoint was ORR, defined as complete response (CR) or partial response (PR) by RECIST v1.1 (sum of diameters) at week 25, as determined by centrally read MRI scans, with readers blinded to patient information. Determination of ORR for each time point was based on the combination of responses for target lesions and the presence/absence of ≥ 1 new lesion. To be considered a response, tumors had to meet the criteria for response and have no documented progression at week 25. This study did not require confirmation of response.

Secondary endpoints included ORR by tumor volume score (TVS; based on 10% increments in estimated tumor volume), with CR indicating complete disappearance of lesions and PR indicating ≥ 50% decrease in volume score relative to baseline by centrally reviewed MRI at week 25; mean (standard error [SE]) change in ROM of the affected joint from baseline at week 25, measured in degrees using a goniometer by a qualified assessor (e.g., orthopedic surgeon or physical therapist); best overall response (BOR), defined as CR or PR by RECIST v1.1 and TVS recorded from the start of treatment until the last radiographic assessment; and duration of response (CR or PR) by RECIST v1.1 and TVS, defined as the time from the first recorded response to the first documented disease progression. The effect of pexidartinib on limb function was assessed as the mean change from baseline in the PROMIS Physical Function Scale score at week 25. ORR by RECIST v1.1 and TVS at week 37 were also assessed in an ad hoc analysis.

Safety assessments included vital signs, 12-lead electrocardiograms, laboratory tests, and treatment-emergent adverse events (TEAEs) graded according to the National Cancer Institute Common Terminology Criteria for Adverse Events version 5.0. Liver function tests included alkaline phosphatase, alanine aminotransferase (ALT), aspartate aminotransferase (AST), total and direct bilirubin, and gamma-glutamyl transpeptidase (GGT), which were assessed weekly for the first 8 weeks, then every 2 weeks for the next month and every month thereafter.

The planned sample size was approximately 35 patients, which provided > 95% power to detect a significant difference against 5% ORR, with an expected ORR of 37.7%. The data cutoff for the primary analysis was 27 October 2021, after all patients completed week 25, and 21 December 2021, for an ad hoc analysis at week 37.

Efficacy analyses were conducted on the full analysis set (defined as all patients who received ≥ 1 dose of pexidartinib). Patients who were missing response data were considered nonresponders. The safety analysis set included all patients who received ≥ 1 dose of pexidartinib. The full analysis set and safety analysis set were identical in this study. The proportion of patients meeting the primary and secondary efficacy endpoints (except duration of response), point estimates, and accompanying 95% confidence intervals (CIs) were calculated on the basis of the Clopper–Pearson method. The duration of response was determined using the Kaplan–Meier product limit methodology with median and 25th and 75th percentiles. Change from baseline to posttreatment assessments was determined for patients with values at both time points. ROM and PROMIS Physical Function Scale scores were analyzed using a mixed model for repeated measurements with joint type and extremity type as factors, respectively, and reported as the least squares mean (LSM) with 95% CIs. All analyses were performed using SAS statistical software version 9.4.

Anonymized individual participant data on completed studies and applicable supporting clinical study documents may be available upon request at https://vivli.org. In cases where clinical study data and supporting documents are provided pursuant to our company policies and procedures, Daiichi Sankyo will continue to protect the privacy of the company and the clinical study subjects. Details on data-sharing criteria and the procedure for requesting access can be found at this web address: https://vivli.org/ourmember/daiichi-sankyo.