Determine if skeletal tumor burden on 18F-fluoride PET/CT (fluoride PET/CT) predicts the risk of bone marrow failure (BMF) after 223Ra dichloride therapy (223Ra). Methods Forty-one metastatic prostate cancer patients (43-89 years old; mean, 71 ± 9 years.) underwent fluoride PET/CT prior to 223Ra. Bone marrow failure was the primary end point and was defined as (1) development of hematologic toxicity (World Health Organization grade 3 or 4) associated with no recovery after 6 weeks or (2) death due to BMF after the last 223Ra dose. Bone marrow failure was correlated to fluoride PET/CT skeletal tumor burden (TLF10 [total lesion on fluoride PET/CT with SUVmax of 10 or greater]), use of chemotherapy, serum hemoglobin concentration, serum ALP, and serum prostate-specific antigen. Results The number of 223Ra cycles ranged from 2 to 6 (mean, 5). Of the 41 patients, 16 developed BMF (G3 = 12; G4 = 4). A significantly increased risk of developing BMF was observed in patients with TLF10 of 12,000 or greater (hazard ratio [HR], 11.09; P < 0.0001), hemoglobin of less than 10 g/dL (HR, 7.35; P = 0.0002), and AP > 146 UI/L (HR, 4.52; P = 0.0100). Neither concomitant (HR, 0.91; P = 0.88) nor subsequent use of chemotherapy (HR, 0.14; P = 0.84) increased the risk of BMF, nor was prostate-specific antigen greater than 10 μg/L (HR, 0.90; P = 0.86). Moreover, in a multivariable analysis, TLF10 was the only independent predictor of BMF (HR, 6.66; P = 0.0237). Conclusions 223Ra was beneficial and reduced the risk of death even in patients with a high skeletal tumor burden. Fluoride PET/CT is able to determine which patients will benefit from 223Ra and which will develop BMF Ver menos