Association of Radiation Doses and Cancer Risks from CT Pulmonary Angiography Examinations in Relation to Body Effective Diameter

This study aims to evaluate the patient’s dose exposure from Computed Tomography Pulmonary Angiography (CTPA) examination and to estimates the cancer risk induced from the examinations based on the patient’s size. One hundred patients were recruited, and data information was collected retrospectively. A multi-detector (MDCT) (Philips Brilliance 128, USA) scanner were utilized for the CTPA examination, and dose data were obtained from the system. The effective diameter of each subjects’ image was measured for the Size-specific dose estimates (SSDEs). All subjects divided into Group 1 (19 – 25 cm), Group 2 (25-28 cm) and Group 3 (28-38 cm), where the association between gender were analysed. Effective dose (E), SSDE, organ dose and cancer risk of each group were evaluated and compared statistically using independent t-test and one-way ANOVA. The range of mean CTDI vol , DLP and E values were (6.44 – 17.42 mGy), (239 – 631 mGy), (5.19 – 13.90 mSv), respectively. In respective with the patient’s effective diameter, the mean SSDE value for Group 1, Group 2 and Group 3 were 9.93 ± 3.89 mGy, 13.70 ± 9.04 mGy and 22.29 ± 7.35


Introduction
Computed tomography (CT) scan has become the most preferred imaging technique, and its utilization has been increasing significantlyas the numbers of the examination using this technique is consistently growing. Albeit its good diagnostic value, in the year 2001, the International Commission on Radiological Protection (ICRP) has expressed concern over the higher usage of CT in medicine and over its remarks that may induced cancer risk when compared to other imaging modalities (Foley et al., 2012).
CT Pulmonary Angiography (CTPA) is one of the imaging techniques that allow the visualization of pulmonary arteries to facilitate diagnosis and treatment of pulmonary embolism (PE). PE is considered as a severe medical condition associated with high mortality, which requires fast and accurate detection.
Thanks to the advancement of CT technology, more than 90% of acceptable examinations can be achieved for detecting PE (Sauter et al., 2018). However, one single CTPA examination may contribute up to 10 mSv of effective dose which increases the possibility of radiation-induced cancer risk to the populations (Halid et al., 2018;Sauter et al., 2018).
Volumetric Computed Tomography Dose Index (CTDIvol) usually represent radiation dose for CT scanner, which based on a dose estimation of a standardized phantom with consideration of pitch value (Daudelin et al., 2018). Mili-gray is the SI unit and visualizes the mean over the volume of the scan. DLP is the product of CTDIvol and the scan length that describes the total radiation dose delivered to the patient in mili-gray and centimetres. DLP becomes the gold standard as a set reference in the CT examination due to radiation risk-related and represents the full radiation dose per examination (Najafi et al., 2015). CTDIvol however, has many potential errors as they do not consider the size of the body, which varies between patients, especially children.
Hence, the size-specific dose estimates (SSDEs) was introduced by the American Physicians Association (AAPM) in 2011, which incorporates individual patient factors into the CTDIvol calculation (Burton and Szczykutowicz, 2018). Instead of basing dose calculation solely on a phantom, SSDE requires the input of individual patient size in the CT scanner (Pourjabbar et al., 2014). It was found that the ratio of SSDE with CTDIvol is inversely proportional to the size of the patient (Anam et al., 2017).
The k-factor is defined as the proportionality constant between the effective dose and the DLP. In the clinical setting, the DLP is understood to be the product of the CTDIvol and the scan length: where the k is k-factor (mSv mGy-1 cm-1), E is effective dose (mSv), DLP is dose-length product (mGy cm), CTDIvol is in mGy and l is the scan length (cm) for different scan types (Zhou et al., 2017). The k-factors used to date were initially developed by Shrimpton et al. using computational phantoms (Shrimpton, PC, Hillier MC, 2014). The k-factors for extended scan regions were published in the updated EG, which was adopted in the American Association of Physicists in Medicine (AAPM) Report 96 (Pourjabbar et al., 2014). It has been acknowledged that radiation exposure has become crucial issues due to increasing the risk of inducing cancer, especially to young ages patients (O'Daniel et al., 2005). More than 2% of populations receive a significant amount of doses that have the potential to develop cancer risk where younger patients are the main contributor (Sauter et al., 2018). Hence, dose reduction strategies to achieve ''as low as reasonably achievable'' (ALARA) principle is a must while preserving good image quality of CT examination.
The most promising assessment is estimating the organ dose and cancer risk according to the patient's body habitus. Both assessments varied in different conditions, dependency on age, sex and population studied (Karim et al., 2017). The limitation of radiation dose estimation may overcome by these assessments tailored to an individual patient exposure rather than a general population. The 7 th report on Biological Effects of Ionizing Radiation (BEIR VII Phase 2) by the United States National Academy of Sciences popularly cited in estimating the cancer risk-based in radiological scans.
The previous study reported that the detrimental effects of radiation exposure are higher in susceptible patient populations, especially young women and children (Brenner and Hall, 2007). As such, it is crucial to calculate an accurate organ-equivalent dose before obtaining an estimation of cancer risk. At the moment there is no research has been found that evaluated the organ dose and cancer risk meticulously in CTPA examination elsewhere (Laqmani et al., 2016(Laqmani et al., , 2014Sabel et al., 2016;Sauter et al., 2018). This study aims to evaluate the dose exposure and to estimate cancer risk attributes from CTPA examinations based on the size and habitus of the patients.

CT parameter measurements
The research protocol was approved by the ethics committee of the Ministry of Health (Malaysia), which waived patient consent form for the retrospective analysis with an approval ID: NMRR-18-3088-44138. The records of 100 adults who underwent CTPA were retrieved from a Hospital Kuala Lumpur, Malaysia. The subjects were above age 18, and they experienced the procedure between September 2018 to February 2019. All subjects were scanned using a Philips Brilliance 128 multi-detector CT (MDCT) scanner, and the images were reconstructed using the DICOM software.
CTPA was performed using 40 to 70 ml of iodinated contrast medium, followed by 50 ml saline chaser, which was intravenously injected into the subjects at a flow rate of 5 ml/s. The bolus tracker technique was performed by placing the region of interest (ROI) on the main pulmonary trunk. The scan was started after 3 to 14 seconds, with a threshold of 70 HU. The z-dom modulation was activated before each scan. The scan was carried out craniocaudally with either 100 kVp or 120 kVp, a pitch of 0.798 and a 0.625 x 40 mm detector configuration. Images were reconstructed with 1 mm slice thickness and 512 x 512 matrix size. The iDose 4 level 4 iterative reconstructive technique was used for post-processing of images.
The radiographers performing the scan were well experienced in this protocol at least three years.
Scanning acquisition parameters and data such as subjects' gender, anteroposterior (AP) and lateral (LAT) body lengths, tube voltage (kV), tube current(mA), rotation time, pitch factor, CTDIvol and doselength-product (DLP) were recorded from scanner console. The AP and LAT lengths of each subjects' image were measured using digital callipers on the scanner console, at the mid-slice location of the transverse CT images as illustrated in Figure 1. Only the scan data of pulmonary embolism scan were included, while cases with incomplete details and modified protocols were excluded.

Radiation dose
CTDIvol, AP and LAT lengths obtained from the scanner console were used to calculate SSDE based on the American Association of Physicists in Medicine (AAPM) Report 204. For comparison purposes, the SSDE were also estimated using CT-EXPO Ver 2.3.1 (SASCRAD, Bucholz, der Norheide, Germany) and Monte Carlo simulation based on scanning parameter used in each examination. Derivation of SSDE, according to AAPM, report 204 began with the calculation of the effective diameter of the subjects' body, as the following equation: The effective diameters (body size) were normalized to the tabulated body size-dependent conversion factor (f-size) stated in the AAPM report 204. SSDE was calculated by multiplying the normalized f-size with the CTDIvol displayed on the scanner console as in equation below: In this study, the E and organ dose was estimated only by CT-EXPO software. This software offers automatic output calculation of radiation exposure to the organs based on the detailed scanner model, manufacturer and scanning parameters. The software utilized a Monte Carlo simulation model and estimates the radiation dose based on radiation transport attributable organ dose on the adult phantom.

Risk assessments
The cancer risk for selected organs was determined using the following equation: where rT is the nominal risk factor attained from the International Commission on Radiological Protection (ICRP) Publication 103 (ICRP 2007), and the HT is the organ-specific equivalent dose obtained by CT-EXPO in breast, lung and liver respectively. The cancer risk (R) per procedure was estimated by multiplying the organ dose with the risk factor (rT).

Statistical analysis
All data were analysed by using SPSS V25.0 (SPSS, version 25.0 for Windows, Chicago, Illinois, USA) for statistical study. A p-value of <0.05 was considered to indicate statistically significant differences.
Since the data were normally distributed, the independent t-test and one-way ANOVA were used in this study. Table 1 shows the mean baseline characteristics of study subjects compromising 42 men and 58

Patient characteristic and radiation dose
women. The effective diameters were ranged from 20.14 to 37.48 cm in males and 19.71 -32.25 in females.
The calculation of CTDIvol, DLP, E and organ dose were grouped according to effective diameters and in total, as indicated in Table 2 Table 3 and Figure 2.
The subjects size varied along the Z-axis of the scan due to changes in the thickness and composition of the subjects' habitus. As expected, the variation of the subjects' effective diameter was contributed to the SSDE in line with a previous study (Sabel et al., 2016). Hence, the CTDIvol calculated from the console was observed to be undervalued compared to SSDE, especially in small-sized subjects.
Overall, the reference phantom based CTDIvol values underestimated the radiation dose received by the subjects as comparison by SSDE approach. The small variations in SSDE to CTDIvol ratio generated by CT-EXPO in different subject sizes group was expected since the software's calculations were based on a constant mathematical phantom (Karim et al., 2017).
However, the ratio generated by the AAPM report method was wider in a small-sized group compare to bigger-sized as the f-size increased in decreasing subjects' body size. A previous study had observed that when the automated exposure control (AEC) system was deployed, both CTDIvol and SSDE values were higher in large-sized subjects than those who were standard-sized. When the AEC system was turned off, there seems to be no significant difference between the radiation doses in different subject sizes (Kim et al., 2018). This observation was aligned with this study, where the AEC was deployed.
Another study reported that reducing the tube voltage in CT examination involve contrast media could maximize the photoelectric effect, as the applied voltage was closer to the K-edge of iodine (33.2 keV) (Dane et al., 2018). It could enhance the performance of image quality as well as to reduce the radiation dose. Most studies applied the AEC system to modulate the tube current and reduce unnecessary radiation exposure to their subjects (Isa et al., 2019;Kalra et al., 2004;Karim et al., 2017;Smith-Bindman et al., 2009;Sookpeng et al., 2013). However, a different approach to select both pitch factor and beam collimation showed different radiation exposure-outcome with various institutions and CT-Scanner types.

Organ dose assessment
It is clearly shown that the breast and lungs received the highest radiation dose exposure as these organs cover all in the primary beam. Both also had an equally high risk of developing cancer. On the other hand, the liver attained the lowest values in organ dose and cancer risk. It was mainly because, in a CTPA procedure, the liver would only be partially scanned as it was not entirely within the region of interest. This observation was in line with the BEIR VII report, which stated the dose exposure and cancer risk were dependent on the location of the organ from the primary beam, as well as their sensitivity to radiation.
The higher tube current required to scan subjects with increasing body effective diameter was the main factor that caused the significant differences in organ doses and their cancer risk, as observed in Table   2 and 4. The AEC system automatically modulated the tube current according to patient size and habitus (Bashier and Suliman, 2018;Karim et al., 2019). Thus, subjects with large body sizes were at higher risk when undergoing CTPA.

Cancer risk assessment
The cancer risk in a million procedures with different organ and effective diameter are shown in Table 4. The breast seemed to receive the highest organ dose in total, hence resulting in the cancer risk per million procedures. However, in the largest effective diameter, the lung had slightly higher dose exposure with similarly high cancer risk than the breast. The liver had the least dose exposure even though the values increased with effective diameter. However, its risk factor was extremely low-more than three times lower compared to the breast and lung. This breakdown of results between gender is also provided. Table 5 shows that the E, organ doses and cancer risk were all higher in females than males, but only the E was significant (P = 0.04). Figure 2 shows the mean radiation doses and cancer risk.
Unfortunately, the risk estimation in this study was not comparable with other studies, which were presented with various methodologies and different cases (Karim et al., 2016, Hashim ;Lahham, ALMasri, & Kameel, 2018). Although not significant, the result of this study supported previous research that found a higher cancer risk in females. The overall lethality risk for females was approximately 35 % higher in comparison with males, as illustrated in Table 5 (Karim et al., 2017).
There were some limitations to this study. Firstly, the subjects in this study not involved with pediatrics and adolescents hence further investigation needed to evaluate the radiation dose and selected organ risk between these group of ages. Secondly, the SSDE values derived using CT-EXPO software was not normalized to the effective diameters of each patient. Thus, it not accurately estimated since the CT-EXPO software only utilized a fixed size mathematical phantom. However, this study overcomes this limitation with another evaluation by AAPM report 204 methods. Lastly, no alternative dose descriptor software beside CT-EXPO used in this study to calculate the SSDE, E, organ dose and cancer risk. Thus, no comprehensive comparison of the value of these variables had been made in this study.

Conclusion
In conclusion, SSDE is more promising to evaluate sectional radiation dose compare to CTDIvol accurately. Other radiation dose descriptor such as DLP, E and organ dose found dramatically increase with increasing patient's effective diameter. It is observed that the cancer risk was significantly different between patients' effective diameters and slightly lower by a male instead by a female. Further study could focus on another group ages of subjects scan and additional dose descriptor software to compare by current CT dosimetry software.