Diagnostic Performance of Mandibular Bone Density Measurements in Assessing Osteoporotic Status
John Chai, BDS, MS, MJ, DLaw/Anson C. M. Chau, BSc, MPhil, PhD/Frederick C. S. Chu, BDS, MSc, PhD/Tak W. Chow, BDS, MSc, PhD
Purpose: The study evaluated the diagnostic performance of computed tomography (CT) bone density measurements (in Hounsfield units [HU]) in assessing the osteoporotic status of edentulous subjects (21 men, 40 women) scheduled to receive mandibular implant-supported overdentures. Materials and Methods: Density (in HU) of 61 left and 61 right mandibular canine sites was measured. Each subject was also subject to body bone mineral density measurements (dual-energy x-ray absorptiometry). Bone density T-score values of total hip, femoral neck, spine (L1 to L4), and total body were assessed. Pearson correlations were conducted between HU of each mandibular bone site and the subject’s T-score values. Sensitivity/specificity and receiver operating characteristic curve analyses were used to assess the diagnostic performance of HU in detecting subjects who were, based on each of the four T-scores: (1) “not osteonormal”, ie, with a T-score at or below –1.0; or (2) osteoporotic, ie, with a T-score at or below –2.5. Results: HU measurements of mandibular bone sites were found to have modest but significant correlations with each bone density T-score value (total hip, femoral neck, spine, total body). CT mandibular site HU measurement has diagnostic value in detecting a “not osteonormal” condition when total hip, spine, or total body T-scores were used, but not when femoral neck T-score was used. Conclusions: Based on the limited sample size, optimal HU cutoff values of ~530, ~600, and ~640 HU for total hip, spine, or total body T-scores, respectively, were proposed. CT mandibular site HU measurement had diagnostic value in detecting osteoporosis when spine T-score was used but not when total hip, femoral neck, and total body T-scores were used. An optimal HU cutoff value of ~460 HU for spine T-score was proposed.