Abstract
PURPOSE: Dual-energy computed tomography (DECT) has been proposed for quantification of hepatic iron concentration (IC). However, the lower limit of quantification (LLOQ) has not been established, limiting the clinical adoption of this technology. In this study, we aim to (a) establish the LLOQ using phantoms and (b) investigate the effects of patient size, dose level, energy combination, and reconstruction method. METHODS: Three phantom sizes and eight vials of ferric nitrate solution with IC ranging from 0 to 10 mg/ml were used. DECT scans were performed at 80/140 and 100/140Sn kVp, and using five different levels of CT dose index (CTDI). An image-domain three-material-decomposition algorithm was used to calculate the IC. The LLOQ was determined based on the coefficient of variation from repeated measurements. RESULTS: The measured IC correlated strongly with the true IC in the small and medium phantoms (R(2) of linear regression > 0.99) and moderately in the large phantom (0.8 < R(2) <0.9). The LLOQ improved with increased CTDI. At 30 mGy, the LLOQ was found to be 0.50/1.73/6.25 mg/ml in the small/medium/large phantoms, respectively. 80/140Sn kVp resulted in superior LLOQ for all phantom sizes compared to 100/140Sn kVp, primarily due to the difference in their iron enhancement ratios (1.94 and 1.55, respectively). Iterative reconstruction was found to further improve the LLOQ (by ~ 11%), whereas reconstruction kernel smoothness had negligible effect. The LLOQ of iron was significantly higher than that of iodine due to its lack of a useful k-edge and lower enhancement ratio. CONCLUSION: Iron quantification at clinically important levels was achieved in a small- and a medium-sized phantom using DECT, but proved challenging in a large phantom. Wide spectral separation and accurate calibration were found to be critical to the success of the technology.