Comparative Study of Dual Energy Cone-Beam CT using a Dual-Layer Detector and kVp Switching for Material Decomposition

Abstract

Cone-beam CT (CBCT) is widely used in diagnostic imaging and image-guided procedures, leading to an increasing need for advanced CBCT techniques, such as dual energy (DE) imaging. Previous studies have shown that DE-CBCT can perform quantitative material decomposition, including quantification of contrast agents, electron density, and virtual monoenergetic images. Currently, most CBCT systems perform DE imaging using a kVp switching technique. However, the disadvantages of this method are spatial and temporal misregistration as well as total scan time increase, leading to errors in the material decomposition. DE-CBCT with a dual layer flat panel detector potentially overcomes these limitations by acquiring the dual energy images simultaneously. In this work, we investigate the DE imaging performance of a prototype dual layer detector by evaluating its material decomposition capability and comparing its performance to that of the kVp switching method. Two sets of x-ray spectra were used for kVp switching: 80/120 kVp and 80/120 kVp + 1 mm Cu filtration. Our results show the dual layer detector outperforms kVp switching at 80/120 kVp with matched dose. The performance of kVp switching was better by adding 1 mm copper filtration to the high energy images (80/120 kVp + 1 mm Cu), though the dual layer detector still provided comparable performance for material decomposition tasks. Overall, both the dual layer detector and kVp switching methods provided quantitative material decomposition images in DE-CBCT, with the dual layer detector having additional potential advantages.

1. INTRODUCTION

Dual energy (DE) imaging has demonstrated promising clinical value in CBCT systems equipped with flat panel detectors, with applications in image-guided radiation therapy¹, image-guided interventions², and dedicated breast and extremity systems³. These prior studies have shown that DE imaging provides quantitative CBCT images through its ability to perform material decomposition. Several methods have been developed for obtaining DE data, including fast kVp switching that alternates between low and high tube voltages to separately acquire low energy (LE) and high energy (HE) projections, slow kVp switching that acquires CBCT scans sequentially at different tube voltages, dual source systems, and dual layer detectors that simultaneously capture DE projections in a single scan. For CBCT imaging, the most common way to implement DE scans on a typical scanner is to use kVp switching with sequential acquisitions. Such methods are effective when the imaged objects are stationary, but are limited for moving objects due to the inevitable temporal and spatial misregistration between the DE datasets, and the increased total scan time. These limitations can be overcome by using a dual layer detector, where the LE and HE projections are obtained simultaneously with perfect spatial and temporal registration.

On the other hand, spectral separation between DE spectra is essential to perform material decomposition. The kVp switching technique can achieve good energy separation, which is more challenging for a dual layer detector. In this work, we investigate the material decomposition performance with a prototype dual-layer detector⁴ and compare it to the kVp switching method.

2. METHODS

2.1. Prototype dual layer detector

The prototype detector comprises two 43×43 cm² amorphous-silicon panels, with a top CsI scintillator thickness of 200 μm and a bottom CsI scintillator thickness of 550 μm. The LE projections are generated from the top layer, which preferentially absorbs lower-energy x-rays, while the HE projections are generated from the bottom layer, which tends to stop higher-energy x-rays. A 1 mm Cu filter is placed between the two layers to further improve spectral separation (difference in average energy between the layers).

2.2. Quantitative measurement

A multi-energy CT phantom (Gammex, Middleton, WI) was imaged for quantitative comparison. The phantom is a water-equivalent, 20 cm diameter cylinder and contains several 2.85 cm diameter contrast inserts, including two calcium inserts (300, 100 mg/ml Ca), 3 iodine inserts (10, 5, 2 mg/ml I), 1 adipose insert, and 2 blood/iodine mixtures (4, 2 mg/ml I).

2.2.1. Acquisition parameters for dual layer detector

The phantom was scanned using 120 kVp (inherent filtration of 3 mm Al), 20 mA, 18 ms, resulting in an average spectral separation of 17.1 keV between the two layers. The source-to-axis distance (SAD) and source-to-image distance (SID) were 800 mm and 1300 mm, respectively. The detector was operated at 10 fps with 2×2 binning (0.3 mm pixel pitch), and 500 projections were acquired per rotation. The LE and HE projections were reconstructed to an isotropic voxel size of 0.5 mm with a standard filtered-backprojection (FBP) using Hanning apodization.

2.2.2. Acquisition parameters for kVp switching

Two sets of x-ray spectra were considered for kVp switching: 80/120 kVp and 80/120 kVp + 1 mm Cu filtration (both cases include 3 mm Al inherent filtration), resulting in a detected spectral separation of 10.3 keV and 21 keV, respectively. The latter has a similar spectral separation to that of the dual layer detector. The LE and HE projections were acquired sequentially using a single layer detector (PaxScan 4030CB, Varex Imaging, Salt Lake City, UT) with a 600 μm thick CsI scintillator. Scanning mAs were adjusted to match half the phantom entrance exposure of the scans using the dual layer detector so that total dose was matched between the two techniques. The detector was operated at the same SID and SAD as for the dual layer detector, except a different 2×2 binning pixel pitch of 0.388 mm. For a consistent comparison, the reconstruction parameters were identical to the dual layer detector experiment.

To facilitate the comparison, the two detectors were placed side-by-side on our tabletop system, as shown in Fig. 1, such that the DE acquisition can be switched between the two modes by simply moving the x-ray tube and object along the lateral direction.

Fig. 1.

Layout of the tabletop system and zoomed-in view of the two detectors.

2.3. Material decomposition

We used an empirical dual energy calibration method⁵ to perform the material decomposition. This technique does not require knowledge of the imaging spectrum nor the true attenuation coefficients. As shown in Fig. 2, the decomposition was done in the projection domain, where the basis material line integrals l₁ and l₂ (e.g., water and bone) are obtained as a polynomial function of the two sets of measured line integrals, l_L and l_H, corresponding to data acquired at LE and HE, respectively. The l₁ and l₂ are then reconstructed to obtain images corresponding to the basis materials. In this work, we considered two sets of basis material pairs, i.e., water/bone and water/iodine. The coefficients {a} of the polynomial function were determined using a least square fit of the reconstructed polynomial images to known materials in a calibration phantom:

$$l_W=a_1{l}_L+a_2{l}_H+a_3{l}_L^2+a_4{l}_L{l}_H+a_5{l}_H^2+\cdots .$$ (1)

Fig. 2.

Workflow of material decomposition.

In addition, we created virtual monoenergetic (VM) images by weighting the decomposed material projections by their attenuation coefficients at select energies ranging from 40 to 100 keV. We then determined the optimal VM energy for maximizing contrast-to-noise ratio (CNR) of the 10 mg/ml I and 300 mg/ml Ca inserts with respect to background water.

3. RESULTS

Figure 3 shows an axial view of the reconstructed LE and HE images obtained using the dual layer detector and kVp switching. The comparison of material decomposition between the two techniques is shown in Fig. 4. For both basis material pairs, the decomposition results generated from the dual layer detector and kVp switching with additional copper filtration (80/120 kVp + 1 mm Cu filtration) have similar performance in terms of noise and decomposed material uniformity, while the results from kVp switching without additional filtration (80/120 kVp) tend to be noisier and non-uniform across the images. As shown in Table 1, the quantitative results from the dual layer detector and kVp switching with additional copper filtration perform better than that from kVp switching without additional filtration.

Fig. 3.

DE-CBCT obtained using dual layer detector and kVp switching. Display window: [−300 500] HU.

Fig. 4.

Material decomposition results from the dual layer detector and kVp switching.

Table 1.

Quantitative summary of estimated material densities (mean ± std, mg/ml)

Water/Bone	Dual Layer	80/120 kVp	80/120 kVp + 1 mm Cu
Water (1000 mg/ml)	990.66±53.01	977.56±75.19	989.06±46.66
Bone (300 mg/ml)	294.25±19.17	285.60±32.92	291.93±29.50
Bone (100 mg/ml)	91.91±17.39	90.44±28.07	90.54±21.30
Water/Iodine
Water (1000 mg/ml)	988.67±46.24	977.13±57.53	990.57±37.41
Iodine (10 mg/ml)	8.98±1.48	8.13±1.67	9.04±1.41
Iodine (5 mg/ml)	4.76±1.33	4.90±1.80	5.10±1.54
Iodine (2 mg/ml)	2.11±1.29	2.44±1.69	2.03±1.27

The synthesized VM images from 40, 60, and 80 keV are shown in Fig. 5. In principle, the monoenergetic images should be free of beam hardening artifacts for the entire energy range. However, in practice it is difficult to achieve perfect VM images due to residual scatter, limited spectral difference, and decomposition error. Regardless of the technique used, we found that the VM image at 60 keV achieved the highest image uniformity with least beam hardening artifacts. The comparisons of CNR for 10 mg/ml iodine and 300 mg/ml calcium inserts as a function of VM energies are plotted in Fig. 6. Similarly, the maximum CNR is achieved around 60 keV for both inserts. However, with lower spectral separation, the CNR rapidly drops for energies away from the optimal 60 keV. The lower peak VM CNR of the dual layer detector is in part due to photons lost in the copper filter since quantum efficiency is important for VM images. Other factors include the different energy responses and detector pixel sizes.

Fig. 5.

Comparison of VM images at 40, 60, and 80 keV. Display window: [−300 500] HU.

Fig. 6.

CNRs for 10 mg/ml iodine and 300 mg/ml calcium at different VM energies.

4. DISCUSSION AND CONCLUSION

We have demonstrated that a prototype dual layer detector provides quantitative DE-CBCT with accurate quantification of water, bone, and iodine concentration, and is comparable to kVp switching with similar energy spectral separation. The merits of DE-CBCT with a dual layer detector over kVp switching are perfect spatial and temporal registration between LE and HE images without an additional scan, making it attractive for clinical applications in the future. The material-specific CBCT images can be used to generate VM images. An optimal VM energy at 60 keV that achieved maximum image uniformity and CNR of iodine and calcium inserts was observed for both the dual layer detector and kVp switching. The ability to generate high-quality VM images across a wider energy range would be improved with increased DE spectral separation and increased detector DQE. These tasks will be a focus of future studies to optimize the design of the dual layer detector.