Background: Cabozantinib (C) is a c-met and vascular endothelial growth factor receptor inhibitor that demonstrated remarkable responses on bone scan, but failed to demonstrate overall survival benefit. We studied the dynamics of biomarker changes with imaging and biopsies pre- and post-therapy with C to explore predictors of efficacy. Methods: Eligibility included mCRPC patients (pts) with normal organ function and performance status 0-2. C 60 mg orally was administered daily. Pretherapy and 2 weeks post, ^99mTc-labeled bone scans (BS), positron emission tomography with ¹⁸F-sodium fluoride (NaF-PET) and ¹⁸F-(1-(2′-deoxy-2′-fluoro-β-D-arabinofuranosyl) thymine (FMAU-PET) scans were conducted. The maximum tumor standardized uptake values (SUVmax) were measured for the 5 most active lesions in at least 3 skeletal areas with each tracer. Tumor biopsies were conducted, and bone markers were measured. Results: 20 evaluable pts, 5 African American and 15 Caucasian were enrolled. Median age was 69 years, 11 had received prior chemotherapy. 16 had received prior abiraterone/enzalutamide or both.5 pts had visceral disease. 11 and 6 pts required dose reduction to the 40 mg and 20 mg daily dose respectively.8 patients had a PSA decline, of which 2 had a decline > 50%. Median PFS and OS for all 20 pts were 4.1 and 11.2 months, respectively and 3 pts were on therapy for 8, 10 and 13 months. Circulating tumor cell count changes did not correlate with PFS or OS. The NaF-PET demonstrated a median change in SUVmax of -56% (range -85 to -5%, n = 11) and -41% (range -60 to -25%, n = 9) for the pts who were clinically stable and remained on therapy for ≥ 4 or < 4 cycles, respectively. The FMAU-PET demonstrated a median change in SUVmax of -44% (-60 to -14%) and -42% (-63 to -23%) for these groups. A 50% decrease in serum BSAP was associated with a hazard ratio (HR) for progression of 0.91 (90% confidence interval [CI]: 0.71 – 1.16). For serum NTx and urine NTx, the analogous HR’s were 0.79 (90% CI: 0.61 – 1.01) and 1.07 (90% CI: 0.66 – 1.73), respectively. Conclusions: Early changes in imaging and bone markers were not predictive of clinical benefit with C. Tissue C-met testing is in process and will be correlated with PFS and OS.