Ronnie Sebro1,2,3,4, S Sharon Ashok5. 1. Department of Radiology, University of Pennsylvania, 3400 Spruce Street, Philadelphia, PA, 19104, USA. rsebro@gmail.com. 2. Department of Orthopaedic Surgery, University of Pennsylvania, 3737 Market Street, Philadelphia, PA, 19104, USA. rsebro@gmail.com. 3. Department of Genetics, University of Pennsylvania, 421 Marie Curie Blvd, Philadelphia, PA, 19104, USA. rsebro@gmail.com. 4. Department of Biostatistics, Epidemiology and Informatics, University of Pennsylvania, 423 Guardian Blvd, Philadelphia, PA, 19104, USA. rsebro@gmail.com. 5. Department of Radiology, University of Pennsylvania, 3400 Spruce Street, Philadelphia, PA, 19104, USA.
Abstract
OBJECTIVES: This study aimed to compare the accuracy of PET/CT parameters with CT parameters for directing bone biopsies. METHODS: The study was an IRB-approved retrospective study of 388 patients who underwent both 2-[18F] FDG PET/CT and CT within 6 weeks before a bone biopsy. Age, sex, cancer type, lesion length, SUVmax, tumor to liver (T/L) ratio, CT attenuation, difference in CT attenuation between the lesion and normal bone (delta CT attenuation), and the absolute delta CT attenuation were used as predictors. T tests and chi-squared tests were used to compare variables. DeLong's test was used to compare receiver operator characteristic (ROC) curves. RESULTS: We reviewed the data from 388 patients. Of these, 295 patients had bone lesion biopsies, and 93 patients had bone marrow aspirations/biopsies. Biopsies of larger bone lesions (p = 0.033) and bone lesions with higher SUVmax (p = 0.005) were more likely to show malignancy. For bone lesions, the ROC curve for the SUVmax (AUC = 0.6827) was better than the ROC curves for delta CT attenuation (AUC = 0.5766, p = 0.032) and absolute delta CT attenuation (AUC = 0.5491, p = 0.006), but not significantly better than the ROC curves for CT attenuation (AUC = 0.5894, p = 0.061) and T/L ratio (AUC = 0.6778, p = 0.774). A threshold SUVmax of 5.25 had an accuracy of 0.713, sensitivity of 0.766, and specificity of 0.549 to predict malignancy in bone lesion biopsies. None of these variables predicted malignancy in bone marrow biopsies (p > 0.05 for all). CONCLUSIONS: Metabolic 2-[18F]FDG PET/CT parameters have more clinical impact for planning bone biopsies as compared to CT parameters. KEY POINTS: • The 2-[18F]FDG PET/CT measurement (SUVmax) has more clinical impact for planning bone biopsies as compared to CT measurements. • Neither the change in CT attenuation of the lesion relative to normal bone nor the absolute value of this change was a significant predictor of malignancy. • 2-[18F]FDG PET/CT may have clinical benefit and an additional role in directing bone biopsies.
OBJECTIVES: This study aimed to compare the accuracy of PET/CT parameters with CT parameters for directing bone biopsies. METHODS: The study was an IRB-approved retrospective study of 388 patients who underwent both 2-[18F] FDG PET/CT and CT within 6 weeks before a bone biopsy. Age, sex, cancer type, lesion length, SUVmax, tumor to liver (T/L) ratio, CT attenuation, difference in CT attenuation between the lesion and normal bone (delta CT attenuation), and the absolute delta CT attenuation were used as predictors. T tests and chi-squared tests were used to compare variables. DeLong's test was used to compare receiver operator characteristic (ROC) curves. RESULTS: We reviewed the data from 388 patients. Of these, 295 patients had bone lesion biopsies, and 93 patients had bone marrow aspirations/biopsies. Biopsies of larger bone lesions (p = 0.033) and bone lesions with higher SUVmax (p = 0.005) were more likely to show malignancy. For bone lesions, the ROC curve for the SUVmax (AUC = 0.6827) was better than the ROC curves for delta CT attenuation (AUC = 0.5766, p = 0.032) and absolute delta CT attenuation (AUC = 0.5491, p = 0.006), but not significantly better than the ROC curves for CT attenuation (AUC = 0.5894, p = 0.061) and T/L ratio (AUC = 0.6778, p = 0.774). A threshold SUVmax of 5.25 had an accuracy of 0.713, sensitivity of 0.766, and specificity of 0.549 to predict malignancy in bone lesion biopsies. None of these variables predicted malignancy in bone marrow biopsies (p > 0.05 for all). CONCLUSIONS: Metabolic 2-[18F]FDG PET/CT parameters have more clinical impact for planning bone biopsies as compared to CT parameters. KEY POINTS: • The 2-[18F]FDG PET/CT measurement (SUVmax) has more clinical impact for planning bone biopsies as compared to CT measurements. • Neither the change in CT attenuation of the lesion relative to normal bone nor the absolute value of this change was a significant predictor of malignancy. • 2-[18F]FDG PET/CT may have clinical benefit and an additional role in directing bone biopsies.
Entities:
Keywords:
Biopsies; Bone marrow biopsies; Neoplasms; Positron emission tomography computed tomography; Sensitivity and specificity