Dual-Source Dual-Energy Imaging Using Photon-Counting Detector CT for Bone Edema Detection

OBJECTIVES: The aims of the study were to evaluate the spectral performance of an investigational dual-source (DS) scan mode using a tin (Sn) filter on the B-subsystem of a clinical photon-counting detector (PCD) CT system and to demonstrate improved material decomposition performance using clinical examples of bone imaging tasks. MATERIALS AND METHODS: Calcium inserts (Ca 100, 200 and 300 mg/cc) were placed in water phantoms (30-, 40-, and 50-cm lateral diameter) and scanned on clinical PCD-CT (NAEOTOM Alpha, Siemens) using DS spectral scan mode (QuantumPeak). Two tube potential configurations (70/Sn150 kV and 90/Sn150 kV) were used to scan the phantoms (11 mGy to 41 mGy volume CT dose index). The phantoms were also scanned using the single-source (SS) PCD-CT scan mode at 120 kV and 140 kV tube potential, and on a DS energy-integrating detector (EID) CT (SOMATOM Force, Siemens) for quantitative comparison. CT images (from SS-PCD-CT, DS-PCD-CT, and DS-EID-CT) were reconstructed using a quantitative kernel (Qr40) at a 2-mm section thickness using iterative reconstruction strength 1. Spectral separation was quantified using the dual-energy ratio (DER) of Ca inserts and using mean absolute percent error (MAPE) of Ca mass density obtained from Ca/water material decomposition. To demonstrate clinical feasibility, 4 patients were scanned using DS-PCD-CT under an institutional review board-approved study. Bone edema maps were reconstructed from DS-PCD-CT and compared with the corresponding clinical imaging exam of the same patients (MRI or DS-EID-CT). RESULTS: DS-PCD-CT at 70/Sn150 kV for Ca 100 mg/cc showed the highest mean DER (2.49 and 2.56 at 30 and 40 cm, respectively) among all scan configurations. For the 50-cm phantom at Ca 100 mg/cc, DS-PCD-CT at 90/Sn150 kV showed highest mean DER (1.88), followed by DS-EID-CT at 90/Sn 150 kV (1.87) and SS-PCD-CT at 140 kV (1.78). The MAPE values for DS-PCD-CT were consistently lower across all phantom sizes (MAPE max. of 1.44%) compared to SS-PCD-CT (MAPE max. 3.97%) and DS-EID-CT (MAPE max. 3.68%). Qualitatively, patient images illustrated bone edema depiction on DS-PCD-CT comparable to clinical MR images, and more precise edema depiction compared to DS-EID-CT images at the site of fractures and intramedullary lesions, and with fewer artifacts. CONCLUSIONS: DS-PCD-CT showed superior spectral performance for calcium imaging tasks compared to SS-PCD-CT and DS-EID-CT.