Quantification of Imaging Doses from Cone Beam Computed Tomography System at Steve Biko Academic Hospital ()
1. Introduction
Image-guided radiotherapy (IGRT) is a critical process in radiotherapy, in which patient alignment is meticulously verified using both kilo-voltage (kV) and megavoltage (MV) X-ray imaging systems throughout the treatment course. These IGRT techniques encompass planar imaging (two-dimensional, or 2D) and volumetric imaging (three-dimensional, or 3D) [1]. One noteworthy example of a contemporary linear accelerator (Linac) used for external radiation therapy is the Elekta Versa HD. These Linac units have been designed to integrate both kV and MV X-ray imaging systems, enabling cone-beam computed tomography (CBCT) imaging, mainly for IGRT [2]. It is noteworthy to emphasize that among these imaging technologies, the kV-CBCT is important for IGRT verification. This prominence is attributed to the heightened soft-tissue contrast achieved through the photoelectric effect in images produced by the kV imaging system, in contrast to the diminished contrast observed in MV imaging due to Compton scatter [3] [4]. Nowadays, kV and MV CBCT systems are used for routine treatment to identify the treatment target and volumes of organs at risk, ensuring precise and accurate treatment setup [5]. While CBCT provides valuable information about internal anatomical structures, it has the drawback of using ionizing radiation. To address this concern, the International Commission on Radiological Protection (ICRP 103) has issued guidelines to protect against the dangers associated with ionizing radiation. These guidelines emphasize keeping medical and public radiation exposures as low as reasonably achievable (ALARA principle) [6]. Patients must remain in the same position during radiotherapy pre-treatment and during treatment delivery. These guarantee that radiotherapy treatment is directed to the tumour while sparing healthy tissues or organs at risk [7]. This is because treating organs at risk can cause detrimental effects later. These CBCT techniques deliver excessive doses to patients; hence, it is vital to know the magnitude of these doses to minimize their effects. Minimizing imaging dose is important, as it may affect image quality. Methods of minimizing imaging dose may include adjusting clinical imaging protocols for a particular acquisition.
At Steve Biko Academic Hospital (SBAH), the radiotherapy centre uses CBCT for patient setup and position verification before treatment delivery. These CBCT techniques may deliver a significant additional radiation dose to the prescribed treatment dose due to their frequent use. With setup imaging being done numerous times during treatment, these doses may increase significantly. The increase in radiation dose may contribute to an elevated risk of late radiation effects, including the potential development of radiation-induced secondary malignancies [8] [9]. Optimized dose assessments will provide a clearer insight of the radiation doses being delivered, particularly in paediatric patients, which remains a major concern at SBAH. Due to the ongoing development of children’s organs and tissues, exposure to radiation at a young age may lead to damage and, in the long term, result in the potential development of radiation-induced cancer many years down the road [6]. It is therefore vital to quantify the doses for each CBCT protocol in radiation therapy, which helps optimize imaging protocols for verifying patient positioning and delivering the minimum imaging dose possible. The study investigated the doses received by patients during treatment setup imaging for each protocol.
2. Materials and Methods
2.1. Preparation of the Scanning
The study was conducted in the radiation oncology at SBAH. Research Ethics Committee at the University of Pretoria granted ethical clearance to conduct the study (Ethics Reference No.: 2220/2022). The investigation was conducted using phantoms scanned with a CT scanner, after which the image sets were transferred to the treatment planning system (TPS). Treatment plans were then generated based on these scans and sent to the Elekta Versa HD linear accelerator via the record-and-verify system (Mosaiq) for image quality assessment and dose measurements. The study evaluated different imaging protocol settings for the head-and-neck, pelvis, and prostate regions. Additionally, CTDI measurements were performed, and the effective doses for the various clinical protocols were calculated.
This study focused on quantifying doses for each clinical protocol without compromising image quality. Therefore, the evaluation was conducted on two fronts: image quality evaluation and dose measurements. Image quality was assessed using the Catphan 503 (Salem, NY, USA) to ensure the system met the required quality standards. For dose assessment, a head-and-body phantom with slots for inserting a calibrated pencil ionization chamber connected to an electrometer was used to measure dose for different imaging protocols, including treatment areas such as the head, pelvis, and prostate.
2.2. KV Image Quality
The Catphan 503 phantom was used to evaluate XVI CBCT image quality. The phantom was placed in its box using the dedicated alignment marks and positioned on the treatment couch, as shown in Figure 1. It was then aligned to the treatment isocenter using a three-dimensional laser system consisting of anterior, left, and right lateral lasers. The alignment was verified by ensuring that the lasers intersected precisely at the white reference markers indicated on the phantom. Measurements were performed using manufacturer-standard protocol settings, including an S10 collimator cassette (used to define the X-ray beam size and shape) and an F0 filter cassette (a standard beam filtration setting). The imaging was conducted at 120 kV, with the kV imaging panel configured to the small field-of-view position for image acquisition. Volume view images were acquired for uniformity, high resolution, geometric consistency, and low-contrast visibility on the XVI system.
![]()
Figure 1. A Catphan 503 phantom setup using room lasers (Courtesy of Steve Biko Academic Hospital).
2.3. Dose Measurement during CBCT Imaging
Measurements of the delivered doses were conducted by employing a CTDI body and head phantom combined with a 100 mm PTW pencil ionization chamber linked to a PTW Unidos Freiburg electrometer. Both these phantoms are 15 cm in length. Both acrylic cylindrical phantoms have five holes for inserting the 100 mm long cylindrical pencil ionization chamber for center and peripheral CTDI measurements.
The Head and body CTDI phantoms were set up separately at the Linac by positioning them on the treatment couch, as in a CT scanner, using lasers, as shown in Figure 2(a) and Figure 2(b), respectively. Before the CBCT measurement, the CTDI scan image from the Monaco TPS was first loaded on the XVI software as the reference image. Then, CBCT was acquired to assess setup accuracy by superimposing the CT image onto the CBCT. Three electrometer charge readings were collected in each of the five measurement positions in the CTDI phantom.
Figure 2. (a) CTDI head phantom setup shown. (b) CTDI body phantom setup shown (Courtesy of Steve Biko Academic Hospital).
The CBCT dose was measured using various manufacturer-standard XVI clinical protocols. The clinical protocols investigated in this study, together with the associated setup parameters—including kilovoltage (kV), milliampere-seconds (mAs), field of view (FOV), filter, frames and the computed tomography dose index measured in both clockwise (CW) and counterclockwise (CCW) directions—are presented in Table 1.
2.4. Computed Tomography Dose Index (CTDI) Calculations
CTDI100 was measured at all five positions, and the mean peripheral CTDI100 was obtained by averaging the four peripheral measurements. Subsequently, the
Table 1. Displays the manufacturer-standard XVI clinical protocols with the exposure parameters.
Clinical protocol |
kV |
mAs projected |
FOV |
Filter |
Collimator |
Frames |
CTDI projected (mGy) |
Head & neck cw S10 |
100 |
36.6 |
small |
F0 |
S10 |
366 |
1.0 |
Head & neck ccw S10 |
100 |
36.6 |
small |
F0 |
S10 |
366 |
1.0 |
Head & neck cw S20 |
100 |
36.6 |
small |
F0 |
S20 |
366 |
0.5 |
Head & neck ccw S20 |
100 |
36.6 |
small |
F0 |
S20 |
366 |
0.5 |
Fast head & neck cw S20 |
100 |
18.3 |
small |
F0 |
S20 |
183 |
0.5 |
Fast head & neck ccw S20 |
100 |
18.3 |
small |
F0 |
S20 |
183 |
0.5 |
Pelvis M15 cw |
120 |
1056.0 |
Medium |
F1 |
M15 |
660 |
15.3 |
Pelvis M15 ccw |
120 |
1056.0 |
Medium |
F1 |
M15 |
660 |
15.3 |
Pelvis M20 cw |
120 |
1056.0 |
Medium |
F1 |
M20 |
660 |
15.5 |
Pelvis M20 ccw |
120 |
1056.0 |
Medium |
F1 |
M20 |
660 |
15.5 |
Pelvis fast M20 enhanced cw |
120 |
422.4 |
Medium |
F1 |
M20 |
330 |
15.5 |
Pelvis fast M20 enhanced ccw |
120 |
422.4 |
Medium |
F1 |
M20 |
330 |
15.5 |
Pelvis L20 cw |
120 |
844.8 |
Large |
F1 |
L20 |
330 |
18.8 |
Pelvis L20 ccw |
120 |
844.8 |
Large |
F1 |
L20 |
330 |
18.8 |
Prostate M10 cw |
120 |
1689.6 |
Medium |
F1 |
M10 |
660 |
24.5 |
Prostate M10 ccw |
120 |
1689.6 |
Medium |
F1 |
M10 |
660 |
2.5 |
Prostate M15 cw |
120 |
1689.6 |
Medium |
F1 |
M15 |
660 |
24.2 |
Prostate M15 ccw |
120 |
1689.6 |
Medium |
F1 |
M15 |
660 |
24.2 |
Prostate fast M20 enhanced cw |
120 |
844.8 |
Medium |
F1 |
M20 |
330 |
24.2 |
Prostate fast M20 enhanced ccw |
120 |
844.8 |
Medium |
F1 |
M20 |
330 |
24.2 |
Prostate fast M20 low dose cw |
120 |
422.4 |
Medium |
F1 |
M20 |
330 |
15.5 |
Prostate fast M20 low dose ccw |
120 |
422.4 |
Medium |
F1 |
M20 |
330 |
15.5 |
weighted CTDI (CTDIw) was determined using Equation (2), where one-third of the central CTDI100 and two-thirds of the average peripheral CTDI100 were combined. The CTDIvol was then calculated by dividing CTDIw by the pitch factor (Equation (3)). The DLP was obtained by multiplying CTDIvol by the scan length (Equation (4)). Finally, the effective dose was estimated by multiplying the DLP by the appropriate anatomical-region conversion coefficient (k-factor) (Equation (5)). The final reported dose values were calculated as the average of the measurements obtained from the CW and CCW acquisitions.
The CTDI was calculated as follows:
(1)
where D(z) is the dose profile along the scan direction, N is the number of tomographic slices acquired, T is the slice width, the product NT is the nominal beam width, ±L/2 represents integration limits (L shows the length of the active detector volume) and D(z) represents dose integral [10].
The weighted CTDI (CTDIw) was calculated by using the center and peripheral measurements with 1/3 and 2/3 weighting, respectively [11]. Weighted CTDI is expressed as follows:
(2)
The volume CTDI (CTDIvol) was subsequently determined by dividing CTDIw by the pitch factor as
(3)
The dose-length product (DLP) was calculated by multiplying CTDIvol by the scan length [11].
(4)
Effective dose (ED) was estimated from the DLP using the appropriate anatomical-region conversion coefficient (k-factor),
(5)
3. Results
3.1. CBCT Results for 100 KV & 120 KV Adult Patients
The CBCT findings were measured while utilizing the manufacturer’s default settings XVI clinical protocols. The measurements were conducted at 100 kV and 120 kV for adult patients, involving both a 16 cm diameter head & neck phantom and a 32 cm diameter body phantom. Each scan was performed on three separate occasions to ensure consistency.
Table 2 presents the experimental mean dose measurements for the clinical protocols used to estimate the effective dose. DLP was determined from the measured CTDIvol and calculated by multiplying the CTDIvol by the scanned length of 10 cm. Effective dose was subsequently calculated by multiplying the DLP values by conversion factors for adult patients from AAPM report 96 [12]. Conversion coefficients of 0.0031 mSv·mGy−1·cm−1 and 0.015 mSv·mGy−1·cm−1 were used for head-and-neck and pelvis/prostate examinations, respectively [12].
Table 2. Experimental mean dose measurements for the clinical protocols using an adult-sized phantom.
Clinical protocol |
CTDIvol (mGy) |
DLP (mGy.cm) |
Effective dose (mSv) |
Head & neck |
0.33 ± 0.11 |
3.24 ± 1.15 |
0.01 ± 0.00 |
Pelvis |
4.63 ± 1.15 |
46.34 ± 11.53 |
0.70 ± 0.17 |
Prostate |
8.43 ± 3.54 |
84.34 ± 35.44 |
1.26 ± 0.53 |
The CTDIvol for the head and neck ranged from 0.23 to 0.52 mGy, with a mean value of 0.33 ± 0.11 mGy at a 95% confidence interval. For the pelvis, CTDIvol values ranged from 3.1 to 6.44 mGy, with a mean of 4.63 ± 1.15 mGy at a 95% confidence interval. In the prostate region, CTDIvol ranged from 3.9 to 13.46 mGy, with a mean of 8.43 ± 3.54 mGy at a 95% confidence interval. The effective dose for the head and neck ranged from 0.01 to 0.02 mSv, with a mean of 0.01 ± 0.00 mSv. For the pelvis, the effective dose varies between 0.47 and 0.97 mSv, with a mean of 0.70 ± 0.17 mSv. In the prostate region, the effective dose ranges from 0.58 to 2.02 mSv, with a mean value of 1.26 ± 0.53 mSv. Figure 3 illustrates the mean effective doses in mSv for clinical protocols.
Figure 3. Mean effective dose (mSv) measured for clinical CBCT imaging protocols using both a 16 cm diameter head & neck phantom and a 32 cm diameter body phantom on the Elekta XVI system.
3.2. Cone Beam Computed Tomography Doses Findings for Paediatric Patients
The paediatric dosage outcomes were obtained using a 16 cm diameter head and neck phantom, while a 16 cm diameter phantom was also used for the body. The 16 cm diameter phantom was selected to represent paediatric body measurements, as it better reflects smaller body dimensions. Effective dose calculations were performed using conversion coefficients for a 5-year-old patient, considered representative of the midpoint of the selected paediatric age range (0 - 10 years), based on values extracted from the AAPM Report [12].
Table 3 presents the experimental mean dose measurements for the clinical protocols used to estimate the effective dose. The CTDIvol, DLP, and effective dose were obtained using the same procedures defined in Table 2. However, conversion coefficients of 0.0057 mSv·mGy−1·cm−1 and 0.020 mSv·mGy−1·cm−1 were applied for head-and-neck and pelvis/abdomen examinations, respectively.
The CTDIvol for the head and neck ranged from 0.19 to 0.40 mGy, with a mean of 0.26 ± 0.08 mGy at a 95% confidence interval. For the pelvis, CTDIvol values
Table 3. Experimental mean dose measurements for the clinical protocols using a Paediatric-sized phantom.
Clinical protocol |
CTDIvol (mGy) |
DLP (mGy∙cm) |
Effective dose (mSv) |
Head & neck |
0.26 ± 0.08 |
2.56 ± 0.80 |
0.01 ± 0.01 |
Pelvis |
8.33 ± 2.40 |
83.28 ± 25.00 |
1.67 ± 0.45 |
Abdomen |
14.77 ± 6.00 |
147.72 ± 70.00 |
2.95 ± 1.30 |
ranged from 6.00 to 12.10 mGy, with a mean of 8.33 ± 2.40 mGy at a 95% confidence interval. In the abdomen region, CTDIvol ranged between 6.00 and 24.00 mGy, with a mean of 14.77 ± 6.00 mGy at a 95% confidence interval. The effective dose for the head and neck ranges from 0.01 to 0.02 mSv, with a mean of 0.01 ± 0.01 mSv. For the pelvis, the effective dose varies between 1.20 and 2.50 mSv, with a mean of 1.67 ± 0.45 mSv. In the abdomen region, the effective dose ranges from 1.20 to 5.00 mSv, with a mean value of 2.95 ± 1.30 mSv. Figure 4 illustrates the mean effective doses in mSv for clinical protocols.
Figure 4. Effective dose (mSv) measured for clinical CBCT imaging protocols using a 16 cm diameter head and neck phantom, while a 16 cm diameter phantom was also used for the body on the Elekta XVI system.
4. Discussion
For kV imaging quality, regarding image uniformity, the calculation of the percentage difference for the average HU values across the five regions of interest (ROIs) produced a value of 0.33%, which was within the acceptable limit of 1.5% of the manufacturer and limit of 10% for SASQART. Spatial resolution exceeded 10 (lp/cm) by 12 lp/cm for both the manufacturer and SASQART. The low contrast visibility evaluation resulted in a percentage of 2.07%, well below the specified threshold of 3% for the manufacturer and 10% for SASQART. In terms of geometry, the measured vertical and horizontal distances between inserts were 116.70 mm and 116.70 mm, respectively, while the expected value was 117 mm. Both measurements displayed high accuracy, falling within the 1 mm tolerance range for the manufacturer and the 2 mm limit for SASQART. The results of the image quality tests presented here align with the specified tolerances outlined by both the manufacturer and SASQART.
Numerous studies have investigated CBCT dosimetry; however, most researchers [13]-[15] have based their assessments on the weighted computed tomography dose index (CTDIw) and have not extended their analyses to include conversion into effective dose. This study indicates that CTDIvol values (equal to CTDIw, since the pitch is 1) range from 0.1 to 14 mGy, with observed differences between the projected and delivered mAs. For the head-and-neck protocol, the machine was intended to deliver 36.6 mAs according to the manufacturer’s default presets, but the actual measurements averaged 19 mAs. Similarly, the manufacturer’s default presets set the intended mAs to 1056 for the pelvis protocol, but the actual measurements averaged 550 mAs. The manufacturer planned to deliver 1689.6 mAs for the prostate protocol, yet these measurements indicated an average delivery of 890 mAs. This study found that the manufacturer’s default presets for the XVI clinical protocols delivered approximately half of the intended mAs and CTDI dose. If these mAs were delivering the entire mAs, the CTDIvol would be in the range of 0.8 to 24.6 mGy, a range similar to that reported by other researchers [13]-[15] and in line with the expected dose from the manufacturer’s default settings for the XVI clinical protocols.
Song WY et al. (2008) have done a study using XVI acquisition default settings for all the clinical sites on an Elekta Synergy XVI system similar to ours, and dose measurements were conducted utilizing two cylindrical acrylic phantoms, one with an 18 cm diameter (head phantom) and the other with a 30 cm diameter (body phantom) [13]. They found that the average CBCT dose (CTDIw) ranged from 1 to 35 mGy, with total mAs of 36.1 for head and neck, 643 mAs for pelvis, and 1028.8 mAs for prostate. The body phantom recorded the highest measured dose under the prostate protocol [13]. Also, Isam MK et al. (2006) conducted measurements within the dose range of 1 - 23 mGy, while Spezi E et al. (2011) measured doses within the range of 0.1 - 30 mGy using Monte Carlo calculation, both utilizing similar XVI acquisition settings to ours using Elekta Synergy [14] [15].
Differences between the selected settings and the actual delivered values highlight discrepancies in CTDI measurements compared with previous studies [13]-[15]. These variations are further influenced by factors such as phantom diameter, acquisition parameters, and the dosimetry methodologies used in previous studies, all of which impact the reported dose magnitude. Delivered mAs values from the Elekta Versa HD were consistently lower than the preset protocol values. This is due to the system’s reporting methodology, whereby the displayed delivered mAs corresponds to the exposure recorded for the completed CBCT scan only.
The CTDIvol for the head and neck is generally much lower, while the prostate protocol displays a higher dose than the pelvis protocol. This is mainly due to the prostate protocol’s use of high mAs settings necessary to achieve better image quality. This is particularly crucial in areas such as the prostate, where detailed visualization of soft tissues is essential. If a prostate patient undergoes daily imaging, they may receive approximately 14 mGy over a 30-fraction treatment course. This results in a cumulative imaging dose of up to 420 mGy, which is relatively small compared with the treatment dose. Therefore, incorporating the cumulative imaging dose into the prescribed dose may not be necessary, particularly for adult patients.
The findings regarding CTDIvol within paediatric clinical protocols. Notably, the head and neck scans display lower doses than the pelvis and abdomen scans. The results indicated that, per fraction, paediatric patients in the abdominal region may receive a maximum CTDIvol of about 24 mGy. Over an average 25-fraction treatment course, this could accumulate to a total imaging dose of up to 600 mGy. This is significant, as repeated imaging throughout treatment may increase the risk of stochastic radiation effects. The fast M20 low dose CBCT preset evaluated in this study resulted in lower radiation dose and may be suitable for IGRT applications where reduced patient exposure is a priority. However, image quality was not assessed as part of this study, and some of the evaluated protocols are not routinely used clinically at our institution. Therefore, protocol selection should consider both radiation dose and the image quality requirements of the intended clinical application.
Figure 3 graphically presents the findings on effective doses for adult clinical protocols, while Figure 4 illustrates the effective dose findings for paediatric clinical protocols. Comparison of the graphical data between adult and paediatric clinical protocols shows that paediatric doses are higher than adult doses. The results also indicate a strong dependence on patient size, with a more pronounced increase in dose for smaller phantom sizes.
5. Conclusions
This study shows that CBCT doses used during verification procedures are generally low for regions such as the head and neck. However, for clinical protocols like the prostate, the doses are relatively higher, although still small compared with the treatment dose. Imaging dose increases with higher energy and mAs settings. The findings indicate that the modern linear accelerator, the Elekta Versa HD CBCT imaging system, reduces patient radiation exposure due to lowered mAs settings compared with the manufacturer’s default presets, resulting in a lower delivered dose than originally intended.
For paediatric patients who require daily CBCT scans for precise treatment setup, additional precautions are essential. The findings show that significant radiation doses can be delivered to paediatric patients in the abdominal region, especially those with smaller bodies. This study concludes that there is a potential for a significant dose to be delivered to paediatric patients during image-guided radiotherapy.
This study has several limitations. Due to resource constraints, doses from the CBCT systems were not measured directly using thermoluminescent dosimeters (TLDs). Instead, dose assessments were based on CTDI measurements obtained using standard CTDI phantoms, and effective doses were estimated using established conversion coefficients. Consequently, the reported CTDI and effective dose values do not represent patient-specific organ doses and do not account for variations in patient anatomy, size, age, or positioning. The calculated effective doses should therefore be considered standardized metrics for comparing imaging protocols rather than precise estimates of individual patient risk. Future studies incorporating direct TLD measurements, patient-specific dosimetry, or organ dose calculations would provide a more comprehensive assessment of radiation exposure and associated risks.