Four-dimensional computed tomography (4DCT) which is an advanced technique to acquire a sequence of 3DCT with respect to respiration signal which could be used to monitor the lesion motion in patients has been widely utilized in the radiation therapy as well as diagnostic arena [
With the Institutional Ethnics Committee approval (CRTOG1601) and patient consent, radiotherapy treatment planning (RTP) CT images, organ contours, age, and gender of 102 cancer patients (51 males and 51 females) treated from the year of 2007 to 2017 in our institution were used in this retrospective study. The average age of patients at diagnosis was 65 (range, 6-93). The volumes of OARs were segmented using the Pinnacle3 RTP system (Philips Healthcare, Best, Netherlands). Patient effective diameter was computed at the nipple level from body contours using DICOMan software [
A commercially available 16-slice Brilliance Big Bore CT scanner (Philips Medical System) in our clinic was used. Patients were set up in supine position and immobilized with Body Pro-Lok immobilization device (CIVCO Medical Solutions, Coralville, IA, USA) during data acquisition. For each patient, thoracic 4D helical and 3D axial scans were acquired with the collimation of 16 × 0.75 mm, 16 × 1.5mm, and 8 × 3 mm. Varian real-time position management system v1.7.5 (RPM, Varian Medical Systems, Palo Alto, CA, USA) was used in the 4DCT acquisition. The scan protocol was set as 120 kV and 100 mAs. The pitch of helical mode is 0.059 and rotation time is 0.44s.
The volumes of OARs within the primary beam were segmented using the Pinnacle3 RTP by one experienced radiation oncologist and confirmed by another experienced radiation oncologist. The following OARs were defined: heart, bilateral lungs, spinal cord, trachea, and esophagus. Averaged intensity projection (AveIP) was used in OAR definition in the 4DCT. With the aid of DICOMan, the images and structures of the patients were converted into EGS4 CT phantoms based on scanner-specific Hounsfield units to density conversion.
The Monte Carlo method was used to simulate the kV X-ray beam of 3DCT and 4DCT with the thoracic protocols. The source model of Philips Brilliance Big Bore 16-slice CT scanner was calibrated and validated through measurement in our previous work [
In the axial scan mode of 3DCT, a series of 12 coplanar fields around the gantry rotation axis with an interval of 30° were simulated to mimic the axial mode of CT acquisition. After each 360° gantry rotation, the phantom isocenter moved by a certain distance equivalent to the table movement. In the helical scan mode of 4DCT, the pitch value was considered in computing the incremental table movement (i.e., isocenter movement) between two consecutive gantry fields.
To convert Monte Carlo simulation into absolute dose, absorbed doses were first measured at the isocenter of a CTDI phantom (16 cm in diameter) following the AAPM TG-61 protocol with a calibrated EXRADIN A12 ionization chamber (Standard Imaging, Middleton, WI, USA) for the thoracic and abdominal scan protocols in 3DCT and 4DCT. The CTDI phantom was also scanned and the images were imported into treatment planning system and converted into digital phantom for Monte Carlo simulation with MCSIM. Monte Carlo simulation was then performed to a chamber volume inside the phantom with the same beam setups using the proposed source model. The ratios of Monte Carlo simulation values and measured absorbed doses of the same phantom yielded conversion factors, which were used in the absolute dose calculations in the patient’s anatomy.
We calculated organ doses for 102 patients from the 3DCT and cine 4DCT by using Monte Carlo simulation. Doses delivered to the OARs were calculated as the mean doses of all voxels within the defined structures. Then the doses from 3DCT and 4DCT were fitted against patient effective diameter to investigate the relationship between the patient size and imaging dose from CT scanner.
One set of patient data was used for estimating dose difference calculated with different phantoms generated from maximum intensity projection (MIP) and AveIP. The OARs were contoured and confirmed by the same radiation oncologists following the same criterion on both MIP and AveIP.
In this study, the function of estimated relative risk (ERR) in BEIR VII models was used in calculation of the cancer risk in female lung cancer and male lung cancer [
To depict the dose distribution in patients from 3DCT and 4DCT scans, a pediatric patient was exemplified with different isodose lines in Figure
Dose distribution of a pediatric patient delivered by one scan of (a) 3DCT (120kV, 100mAs) and (b) 4DCT (120kV, 100mAs).
In the 102 patients, the average of mean dose to heart, bilateral lungs, spinal cord, esophagus, trachea, and skin was 0.8 (±0.25), 0.71(±0.24), 0.74 (±0.21), 0.79 (±0.24), 0.85 (±0.34), and 0.64 (±0.17) cGy in one3DCT scan, while that in one 4DCT scan was 10.3 (±3.03), 9.46 (±2.44), 9.72 (±2.54), 10.37 (±3.13), 11.84 (±3.22), and 7.82 (±1.58) cGy. The mean dose delivered to whole body per 4DCT scan was 12.8-fold higher as compared with that of per 3DCT scan.
The differences of estimated dose between AveIP- and MIP-based simulations are shown in Table
Differences of estimated dose between AveIP- and MIP-based simulations.
AveIP | MIP | ||||
---|---|---|---|---|---|
Volume (cc) | Dose (cGy) | Volume (cc) | Dose (cGy) | Ratio | |
Heart | 524.28 | 7.02 | 558.03 | 7.87 | 1.12 |
Lungs | 1682.19 | 7.38 | 1680.46 | 8.25 | 1.12 |
Spinal Cord | 9.16 | 6.67 | 8.81 | 7.65 | 1.15 |
AveIP = averaged intensity projection; MIP = maximum intensity projection.
For 3DCT and 4DCT scans, the mean doses deposited to the various organs decreased with the increasing patients’ effective diameters precalculated as shown in Figure
The mean doses to (a) heart, (b) lungs, (c) esophagus, (d) trachea, (e) spinal cord, and (f) skin decreased monotonically with increasing patients’ effective diameter for 3DCT and 4DCT scans.
Heart
Lungs
Esophagus
Trachea
Spinal cord
Skin
The estimated relative risk that incorporates the magnitude of radiation exposure, sex, and patient age at the time of exposure in BEIR VII report was computed. Sex-specific estimated relative risk of lung cancer was shown in Figure
Estimated relative risks for (a) male and (b) female lung cancer from 3DCT and 4DCT scans. The upper and lower bars indicate the 95% confidence intervals (CI).
4DCT has been used widely in radiation oncology for RT planning which helps reduce the chances of having a geographic miss and increases the chances of local control. 4DCT-scan protocols are designed to produce a highly oversampled CT data set with assistance of a low pitch scan to cover the patients’ respiratory motion, which would involve a steep increase in radiation dose. This work quantifies the dose distribution and demonstrates the relationship between the patient size and organ dose.
With the parameter settings, the thoracic organ doses for 4DCT were much higher than that for 3DCT (7.82-11.84 cGy versus 0.64-0.85 cGy after normalization to 100 mAs). In helical mode, 4DCT-scan protocols attempt to account for the patient’s respiratory motion by utilizing a low pitch scan (e.g., pitch = 0.1) to produce a highly oversampled CT data set [
Other studies on 4DCT reported that the radiation dose depends intensively on the setting of different protocols. Effective doses measured in adult anthropomorphic phantoms were 6.14 cGy for lung (3) in cine mode, 5.72 cGy for lung, and 5.05 cGy for esophagus (4) in helical mode with the pitch of 0.125, 5.34 cGy for lung, and 8.73 for esophagus (5) in helical mode with the pitch of 0.516. The scan setups were 120 kV and 100-120 mA. All these effective doses were measured in anthropomorphic phantoms (54-74 kg) by metal oxide semiconductor field effect transistor (MOSFET) dosimeters or thermoluminescence dosimeters (TLD). Although these effective doses did not reflect the accurate dose distributions in the real patients’ body, it was concluded that lower pitch could involve the more dose delivered to the OARs in helical 4DCT. For dose estimation, DeMarco
The converted phantoms from patients CTs in this study could not be used to mimic the dynamic respiratory motion. The differences of estimated dose resulting from AveIP and MIP were in consistent with the one from a research of MV-beam treatment [
The helical 4DCT with low pitch could result in 10% speed-up in scanning but 92% dose efficiency and the broadening of slice sensitivity profile from 1.25 to 2.3 mm on the 16-slice system [
It should be noted that dose from 4DCT or 3DCT scans is much less than the treatment dose delivered in radiation therapy to lung tumor (generally in the range of 45-70 Gy) and the normal tissue dose constrains in conventionally fractionated radiotherapy (mean dose of less than 20-26 Gy) [
This study with data of 102 patients demonstrated a strong inverse correlation between patient effective diameter and mean organ dose to the thoracic organs. The average dose delivered per 4DCT scan was 12.8-fold higher than that per 3DCT scan. Substantial increase in lung cancer risk is associated with higher radiation dose from 4DCT and smaller patients’ size.
The data used to support the findings of this study are included within the article.
There are no conflicts of interest associated with this publication.
The authors would like to thank Dr. Jun Deng (Department of Therapeutic Radiology, Yale University School of Medicine, USA) for his help in the Monte Carlo simulation. This work was supported by The Science & Technology Development Fund of Tianjin Education Commission for Higher Education