Distinguishing Untreated Osteoblastic Metastases From Enostoses Using CT Attenuation Measurements

Purpose: To determine if CT density thresholds of osteoblastic bone lesions can be used to distinguish untreated osteoblastic metastases from enostoses. Materials and Methods: The study group comprised 62 patients (37 enostoses, 25 untreated osteoblastic metastases) with sclerotic bone lesions found on CT. Etiology of sclerotic lesions was assessed histologically or by clinical and imaging follow-up. None of the patients had prior treatment for metastases. The average and maximum densities in Hounsfield Units (HU) were measured. Receiver operating curve (ROC) analysis was performed to determine sensitivity, specificity, area under the ROC curve (AUC), confidence intervals (CI), and cutoff values of CT densities to differentiate metastases from enostoses. Interreader reproducibility was assessed using intraclass correlation coefficient (ICC) with 95% CI. Results: Mean and maximum CT densities of enostoses were 1190 ± 239 and 1323 ± 234 HU and of osteoblastic metastases were 654 ± 176 and 787 ± 194 HU, respectively. Using a cut-off of 885 HU for average density, AUC was 0.982, sensitivity was 95%, and specificity was 96%. Using a cut-off of 1058 HU for maximum CT density, AUC was 0.976, sensitivity was 95%, and specificity was 96%. Mean density ICC was 0.987 for enostoses and 0.81 for metastases. Maximum density ICC was 0.814 for enostoses and 0.980 for metastases. Conclusion: CT density measurements can be used to distinguish untreated osteoblastic metastases from enostoses. An average density of 885 HU and a maximum density of 1058 HU provide reliable thresholds below which a metastatic lesion is the favored diagnosis.