Affiliations: [a] Department of Radiology, The First Affiliated Hospital of Soochow University, Suzhou, Jiangsu, China
| [b] Institute of Medical Imaging, State Key Laboratory of Radiation Medicine and Protection, Soochow University, Suzhou, Jiangsu, China
Correspondence:
[*]
Corresponding author: Chunhong Hu, MD, PhD Mailing address: Department of Radiology, The First Affiliated Hospital of Soochow University, Suzhou, Jiangsu 215006, China. Tel.: +86 0512 6778 1214; E-mail: sudahuchunhong@163.com.
Note: [1] Yu Zhang and Xing Xiong contributed equally to this work.
Abstract: OBJECTIVE:To explore the potential value of quantitative parameters derived from dual-energy spectral computed tomography (DESCT) as comparing to the parameters derived from magnetic resonance imaging (MRI) in detecting bone marrow (BM) infiltration and distinguishing different patterns of BM infiltration in patients diagnosed with Multiple myeloma (MM). METHODS:This study involved 35MM patients and 15 healthy control subjects who had undergone spinal DESCT and MRI. Pattern assignment was based on visual assessment of MR images, and the regions of interest were defined on both DESCT and apparent diffusion coefficient maps. Quantitative values of DESCT parameters were measured and compared between infiltrated and healthy bone marrow. Receiver operating characteristic (ROC) analysis was performed to determine potential utility of DESCT parameters in identifying BM infiltration and different patterns defined by MRI. Sensitivity and specificity under the optimal thresholds determined by the Youden Index were also calculated. RSULTS:Statistical differences were observed between the DESCT parameters including Ca(Water), Water(Ca), HAP(Fat), Fat(HAP) and Effective atomic number (Eff-Z) but not for the 70-keV CT value between the infiltrated and healthy BM (all P < 0.001). The 70keV CT value and Ca(Water), HAP(Fat) and Eff-Z values were also found to be statistically different in comparing different infiltration patterns (all P < 0.05). Performance of the model-based parameter Ca/Water was superior in differentiating between infiltrated and healthy BM in which the area under ROC curve, AUC = 0.856 [95% CI, 81.4–89.1%] with sensitivity = 0.841 and specificity = 0.768, as well as between MM patients and control subjects (AUC = 0.910 [95% CI, 79.5–97.3%], sensitivity = 0.829 and specificity = 1.000). CONCLUSIONS:Analysis of DESCT offers potential as a quantitative method to detect infiltrated BM and evaluate infiltration patterns of BM in patients diagnosed with MM.
