On the improvement of CBCT image quality for soft tissue‐based SRS localization

Abstract Purpose We explore the optimal cone‐beam CT (CBCT) acquisition parameters to improve CBCT image quality to enhance intracranial stereotactic radiosurgery (SRS) localization and also assess the imaging dose levels associated with each imaging protocol. Methods Twenty‐six CBCT acquisition protocols were generated on an Edge® linear accelerator (Varian Medical Systems, Palo Alto, CA) with different x‐ray tube current and potential settings, gantry rotation trajectories, and gantry rotation speeds. To assess image quality, images of the Catphan 504 phantom were analyzed to evaluate the following image quality metrics: uniformity, HU constancy, spatial resolution, low contrast detection, noise level, and contrast‐to‐noise ratio (CNR). To evaluate the imaging dose for each protocol, the cone‐beam dose index (CBDI) was measured. To validate the phantom results, further analysis was performed with an anthropomorphic head phantom as well as image data acquired for a clinical SRS patient. Results The Catphan data indicates that adjusting acquisition parameters had direct effects on the image noise level, low contrast detection, and CNR, but had minimal effects on uniformity, HU constancy, and spatial resolution. The noise level was reduced from 34.5 ± 0.3 to 18.5 ± 0.2 HU with a four‐fold reduction in gantry speed, and to 18.7 ± 0.2 HU with a four‐fold increase in tube current. Overall, the noise level was found to be proportional to inverse square root of imaging dose, and imaging dose was proportional to the product of total tube current‐time product and the cube of the x‐ray potential. Analysis of the anthropomorphic head phantom data and clinical SRS imaging data also indicates that noise is reduced with imaging dose increase. Conclusions Our results indicate that optimization of the imaging protocol, and thereby an increase in the imaging dose, is warranted for improved soft‐tissue visualization for intracranial SRS.


| INTRODUCTION
The use of on-board cone-beam CT (CBCT) has led to significant improvement in localization accuracy for image-guided radiation therapy. However, CBCT image quality generally falls short of helical CT in terms of low contrast visibility. 1 This limits the application of CBCT in many instances to patient setup based on high contrast structures. Although skull matching is sufficient for the majority of intracranial stereotactic radiosurgery (SRS) treatment positioning, for a subset of cases (e.g., when the target abuts a sensitive structure or when deformation between simulation and treatment is more likely), improved soft-tissue contrast is desired for enhancements in intracranial SRS localization. Image quality and imaging dose have been previously studied comparing different machines or existing acquisition CBCT protocols. [2][3][4][5][6][7][8][9][10][11][12][13][14][15][16][17][18][19] Elstrom  images. They reported that dose to the brainstem is 3.7, 0.24, and 0.16 cGy for a pair of MV portal imaging, Trilogy standard head CBCT, and Truebeam ® standard head CBCT scans respectively. 6 This indicates that the Truebeam standard CBCT protocols deliver minimal imaging dose enabling potential to increase dose to improve soft-tissue contrast. The Truebeam ® imaging platform makes it possible for users to adjust multiple acquisition parameters, including xray tube potential, tube current-time product, gantry rotation range, and gantry rotation speed, all of which may affect image quality. 20,21 This study explores sensitivity of image quality to all acquisition parameters and provides suggestions to enhance low contrast visibilities, specifically as it relates to intracranial SRS localization.

| MATERIALS/ METHODS
Twenty-six CBCT acquisition protocols were generated for use on an Edge ® linac (Varian Medical Systems, Palo Alto, CA, USA) as listed in Table 1. All scans were designed utilizing the full-fan bowtie filter.
Imaging protocols with x-ray potential settings of 80, 100, 125, and 140 kVp were used. The x-ray tube current was varied between 15 and 126 mA. All protocols used the same pulse width of 20 ms and tube current-time product was limited to 600 mAs or less except for the 80 kVp series. Gantry rotation speed was varied between 1.5°/s and 6.0°/s for half-rotation trajectory scans (200°total scan angle), corresponding to a total projection number between 2000 and 500, and between 3.0°/s and 6.0°/s for full-rotation trajectory scans (360°t otal scan angle), corresponding to total projection number between 1800 and 900.

2.A | Imaging dose measurement
To evaluate imaging dose, the cone-beam dose index (CBDI) was measured for all CBCT protocols using a 10 cm pencil chamber in a standard CT dose index (CTDI) head phantom (16 cm in diameter) (Computerized Imaging Reference System, Inc., Norfolk, VA, USA). 4,9 Doses at the central and four peripheral positions at 9:00, 12:00, 3:00, and 6:00 o'clock were measured for all half-rotation acquisitions with specified rotation gantry between 20°to 180°E. Peripheral dose was calculated as where D 12 , D 3 , D 9 , and D 6 are the dose values measured at 12:00, 3:00, 9:00, and 6:00 o'clock position respectively. The weighted CBDI (wCBDI) for half rotation protocols were calculated as while D center is the dose at the phantom center. Due to rotational symmetry, the weighted CBDI for full-rotation protocols were: In order to compare wCBDI for CBCT protocols with different tube current-time products, normalized cone-beam dose index (nCBDI) was defined as the wCBDI per 100 mAs.

2.B | Catphan phantom study
To evaluate image quality, a Catphan ® 504 phantom (The Phantom Laboratory, Salem, NY) was scanned using each CBCT protocol multiple times (3-6 image acquisitions for each protocol). The Catphan ® 504 phantom has four test modules: CTP404 for geometry and sensitometry, CTP528 for high resolution, CTP515 for low contrast, and CTP486 for uniformity as described in the website and manual. 22 The CBCT images were reconstructed on the Edge ® treatment console using the following settings: Standard post-processing smoothing filter, Medium ring correction factor, 1 mm slice thickness, 512 × 512 matrix resolution, and 0.51 mm pixel size. All reconstructed Catphan images were analyzed using a commercially available software package, Catphan QA program (Image Owl, Inc., Greenwich, NY). This software quantitatively evaluates the following imaging metrics:  HU Constancy ¼ maxfjI n À I expect n jg; n ¼ 1 to 3: This phantom is constructed of tissue-equivalent materials to simulate soft tissues and bones. A water tube and an acrylic rod were inserted into the phantom as shown in Fig. 1. Three cylinders (10 mm diameter and 25 mm length) were contoured in the water tube, acrylic rod, and background. The noise level and CNR were calculated from the ROI in the water and acrylic inserts for each set of images.

3.A | Imaging dose results
The wCBDI measurement results are listed in Table 2. As expected, the imaging dose increases with tube potential, as illustrated in

3.B | Catphan phantom study results
Image quality results of the Catphan are summarized in Table 2.
There is no clear correlation between the scan acquisition protocol settings and geometric distortion, spatial resolution, uniformity, or   As shown in Fig. 4, image noise was directly correlated with imaging dose (wCBDI) with a standard deviation of 1.4 HU. Consequently, a higher tube potential setting resulted in an increase in imaging dose with a corresponding decrease in image noise. The relationship between noise and wCBDI was best fitted by an inverse square root function (Noise~1= ffiffiffiffiffiffiffiffiffiffiffiffiffiffi wCBDI p ). The normalized cross correlation coefficient between noise and 1= ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi wCBDI p was 0.990.
The contrast-to-noise ratio is plotted as a function of the wCBDI in Fig. 5. The CNR increases with imaging dose, and protocols of 80 and 100 kV were found to yield the largest CNR.

3.C | STEEV phantom study results
Noise values for three different regions of interest in STEEV phantom images scanned utilizing the 12 different protocols are illustrated in Fig. 6 as functions of the imaging dose. Contrast-Noise-Ratio was calculated for Acrylic and Water volumes compared with the reference volume for every scan. CNR results for both Acrylic and Water are shown in Fig. 7 for different x-ray tube potential settings.  Table 3).

3.D | Patient data results
F I G . 5. Catphan contrast-to-noise ratio as a function of weighted cone-beam dose index (wCBDI).  Our results indicate that the imaging dose is the single largest determinant of image noise. Quantitatively, the CBCT imaging noise is proportional to the inverse square root of the imaging dose (wCBDI). Increasing the tube potential leads to less imaging noise; however, this will also result in increased imaging dose and less contrast between different tissue types at the same time, thereby potentially compromising the contrast-to-noise ratio. Therefore, our results support the use of lower tube potential settings (80 or 100 kVp) as the preferred technique for CBCT imaging of the brain.
To maintain acquisition efficiency, increasing the x-ray tube currenttime product is more promising as compared to increasing the number of projections acquired to increase soft tissue contrast. The selection of a CBCT imaging technique protocol is a balance between imaging dose and localization accuracy. Default manufacturer CBCT acquisition protocols were designed with minimal patient dose in mind. SRS patients will benefit from better quality CBCT imaging contrast afforded with slightly higher imaging dose. Clinically, this will improve visual detection of soft tissues necessary for accurate visualization and localization. Other improvements associated with better soft tissue contrast include contouring, dose calculation, and deformable image registration, which may facilitate online adaptive radiation therapy in SRS treatment.

| CONCLUSIONS
Better soft-tissue visualization in the context of intracranial SRS can be achieved through optimization of CBCT imaging protocols, with a moderate increase in the imaging dose relative to standard manufacturer settings.
ACKNOWLEDGMENT This project has been partially supported by a research grant from Varian Medical Systems, Palo Alto, CA.

CONF LICT OF I NTERESTS
None declared.