Background: TGCT is a locally aggressive neoplasm of joint and tendon sheath synovia that may cause pain, limit joint function and destroy bone and local tissues. Measuring TGCT with RECIST is a challenge due to irregular shape and asymmetrical growth, and local tissue damage is not assessed. We reported earlier results of a longitudinal trial of pexidartinib, a selective CSF1R kinase inhibitor, using RECIST as well as novel TVS, modified RECIST and TDS. Here we examine concordance of these MRI measures with PROs. Methods: Patients (pts) with progressive TGCT in a single-arm, multi-center trial of pexidartinib (1000 mg po daily) were assessed by MRI every 2 months by 2 central radiologists (blind to visit order). For RECIST, longest measurable dimensions of up to 2 tumors per joint or tendon sheath were summed (SLD). Modified RECIST summed short axis dimensions (SSD). TVS was based on 10% increments of the estimated maximally distended normal synovial cavity or tendon sheath. TDS scored bone erosion (ERO), cartilage loss (CAR) and bone marrow edema (BME) in multiple regions of each joint. The relationship with PROs (Worst Pain numerical rating scale [NRS] and Worst Stiffness NRS) was assessed. Results: 15 pts (7 knees, 3 hips, 2 ankles, 1 elbow, 1 wrist, 1 thigh) with PRO data and evaluable MRI scans at baseline and Month 7 were assessed. All SLD, SSD and TVS scores improved with respective median changes of -25%, -39% and -50%. Baseline ERO, CAR, and BME ranged 0-19, 0-34, and 0-15, respectively. Median change for each was 0%: ERO worsened in 1 pt, CAR did not change, and BME improved in 4 and worsened in 2. Worst Pain NRS and Worst Stiffness NRS improved in 11 and 9 pts, respectively. Conclusions: TVS demonstrated the greatest pexidartinib effect size, followed by SSD and then conventional RECIST. All had good concordance with PROs. Clinical trial information: NCT01004861