Development of Institutional Reference Level of Radiation Dose for Radiotherapy Computed Tomography Simulation Procedures.
By
A.D.A.D.Athukorala (D/RTP/19/0008)
This Dissertation submitted to
GENERAL SIR JOHN KOTELAWALA DEFENCE UNIVERSITY SRI LANKA
In partial fulfillment of the requirement for the award of the degree of
Bachelor of Science (Hons) in Radiography
March 2023
Contents
Abstract
Diagnostic reference level (DRL) is used as an indicator in medical imaging quality control procedures to identify abnormally higher radiation exposures. The present study aimed to establish an institutional reference level of radiation dose for radiotherapy computed tomography simulation procedures in National Cancer Institute (Apeksha Hoapital), Maharagama, Sri Lanka. CT dose index (CTDIvol) , dose length product (DLP) and other details including total number of images, exposure(mAs), total scan time and total scan length collected for 350 patients who are randomly selected within three month.
01. Introduction
The use of ionizing radiation for cancer treatment has undergone extraordinary development during the past hundred years. The advancement of medical imaging has been critical in helping to achieve this change. The invention of computed tomography (CT) was pivotal in the development of treatment planning. Despite some disadvantages, CT remains the only three-dimensional imaging modality used for dose calculation.
Radiotherapy (RT) CT simulation determines the position and the extent of the disease as accurately as possible. Only one scan is generally required. The imaging position during the CT scan represents the position during the radiation treatment, that is why the preciseness, and the reproducibility of the patient position is fundamental. The key role of CT simulation is also the precise contouring of the target volumes and normal tissues, so called organs at risk (OAR) which contributes to more accurate and precise RT treatment process. This requires the provision of good quality images.
The CT scan range includes the tumor volume and normal tissue volumes superiorly and inferiorly. Radiosensitive organs are also included in the scan. Since CT is a relatively high dose imaging technique with a subsequent high risk of carcinogenesis and radiosensitive organs in the primary field of irradiation relative to other diagnostic imaging modalities, the optimization of dose is essential. There is not much information available on radiation exposures during CT radiotherapy simulations due to patient exposures not being a matter of concern so far. Despite that CT localization doses are considered to be irrelevant in comparison to therapeutic doses, they are not insignificant. A cancer patients may need to go through many CT procedures in the process from diagnosis to treatment, so the exposure to radiation accumulates. If patients recover from the primary cancer, they may still have a long life expectancy, so the additional imaging doses should be as low as reasonably achievable. The CT dose should be optimized to minimize risk to patients. The radiation dose delivered during CT simulation can vary depending on several factors, including patient size, imaging protocols, and equipment used.
The word “Diagnostic Reference Level” was first introduced by the International Commission on Radiological Protection (ICRP) in 1996 in publication 73. After developing the concept further, the practical guidance was published in 2001. DRLs do not represent a dividing line between good and poor medical practice, they are neither dose limits. Its purpose is to prevent excessive radiation dose that does not contribute additional clinical information. DRLs are used as a tool to discover the need for exposure optimization for standardized patients.
To establish a DRL value, Computed Tomography Dose Index (CTDI) value and Dose Length Product (DLP) are used in each examination. The CTDI, measured in mGy, is the area under the radiation dose profile for a single rotation and fixed table position along the axial direction of the scanner. It divided by the total nominal scan width or beam collimation. The DLP value is used as a surrogate measure for patient dose in a computed tomography scan of length. DLP reported in the units of mGy cm and is the product of the energy deposited in a single slice multiplied by the scan length. Therefore, these data can be collected directly by CT machine parameters.
The present study aimed to establish an institutional reference level of radiation dose for radiotherapy computed tomography simulation procedures in national cancer institute Maharagama, Sri Lanka. The machine using there for CT simulation is the Toshiba aquilion large bore CT scanner. Usually around 20 CT simulations have been done in radiotherapy department in NCI Maharagama per day and around total 400 + procedures done during a month. Thus, the availability of institutional DRL for optimize radiation dose delivered to the patient during CT simulation is considerable. But no attempt has been made to establish a DRL value for CT simulator in NCI Maharagama .
The main objective of the study is to establish an institutional diagnostic reference level (DRL) for CT simulation procedures in National Cancer Institute Maharagama. The established institutional DRLs are further compared with the national DRLS of Sri Lanka and National DRLs of other countries, to confirm whether the radiation dose exposed to the patients are under the acceptable level.
02. Review of Literature
The use of computed tomography (CT) simulation is a crucial step in radiation therapy treatment planning. However, CT imaging results in radiation exposure to patients, and it is essential to ensure that the doses remain within safe limits. Therefore, the development of institutional diagnostic reference levels (DRLs) for CT simulation is necessary to optimize radiation doses and reduce patient exposure. The use of computed tomography (CT) simulation is a crucial step in radiation therapy treatment planning. However, CT imaging results in radiation exposure to patients, and it is essential to ensure that the doses remain within safe limits. Therefore, the development of institutional diagnostic reference levels (DRLs) for CT simulation is necessary to optimize radiation doses and reduce patient exposure. According to the International Commission on Radiological Protection (ICRP) 135 guidelines, CTDIvol and DLP are the two parameters that should be used to determine DRLs for CT simulation in radiation therapy. CTDIvol is the radiation dose measured in a specific region, while DLP is the total radiation dose delivered to the patient. Several studies have investigated the use of CTDIvol and DLP to establish DRLs for CT simulation in radiation therapy.
Rao et al. conducted a study on the topic of “Establishment of diagnostic reference level and radiation dose variation in head and neck and pelvis treatment planning in radiation therapy computed tomography.” in Karnataka, India. The purpose of this study was to create a DRL for radiation therapy computed tomography in head and neck and pelvis regions and compare the created DRL with other regional, national and international DRL values so the current radiation therapy computed tomography procedures could be adjusted in an efficient way. For this study, 120 head and neck cancer patients and 90 pelvic cancer patients who were prescribed for RTCT were selected as the research population. The research tool was the advanced Philips 16 slice big bore CT. As the research parameters, they have collected dose length product and CTDI volume to calculate the third quartile to get the DRL. 790.65 mGy.cm, 2.45 mSv were the DLP and effective dose for head & neck and 999.7 mGy.cm, 15.48 mSv were the DLP and effective dose for pelvis. As per the conclusion of the study, they stated that the evaluation of the dose and the optimizations can be made to get better quality images with lower doses. Also, some variations to the older reference levels also spotted(Rao, Ms. S. et al. 2022).
Zalokar et al. conducted a study to inaugurate a national level diagnostic reference level for computed tomography simulations in radiotherapy procedures in Slovenia. This research was done because the doses of radiotherapy localization of CT was not insignificant and also the doses from the CT in simulations are much higher than the usual diagnostic CT doses. Therefore, the point was to be proved that the reference value for simulations were higher for 70% comparison to diagnostic CT. As the study population, they collected 1631 data from patients for the period of four months. They collected the data parameters such as the sex of the patient, type of the examination, Dose length product (DLP) in total and CTDI volume. For the data differentiation, they chose two CT centers. As the data processing, they calculated, mean, median and 3rd quartile of CTDI volume and DLP and they compared the processed calculations between the two selected CT centers. As per the result, 3rd quartiles of dose length products for specific regions such as abdomen, breast, chest, head and neck, pelvis and spine were 1116.2, 606.6, 832.4, 1942.4, 677.1, and 1042.4 mGy.cm respectively. 3rd quartile CTDIvol values for the same sequence of procedures were 18.7, 13.3, 19.2, 76.9, 22.6, 17.9 and 22.2 mGy, respectively. From the findings of the study, they concluded that the exposure parameters should be revised and augmented in order to get a better-quality scan with lower number of doses (Zalokar et al. 2020).
The study of Ria et al., 2019 had been performed to spread out the idea of diagnostic reference levels which can help to reduce the CT noises and to improve the image quality of CT with the proper ideas of diagnostic reference levels in CT. For this research, they collected data from the institutional performance monitoring system (METIS, Duke University) with the total number of 2851 CT examinations of adults. For data variation, they used two different CT machines from two different manufacturers and two different protocols of scans which were abdominal CT examinations with and without intravenous contrast media administration. As the required research data, they collected the CT protocol type, diameter of the patient, dose index of CT volume and image noise magnitudes which were extracted from each and every CT scans with the help of an informatics system. They classified the data as the ranges of patient size, noise reference level, noise reference image, dose reference level and ranges of dose reference. As per their result, the noise reference level and the noise reference ranges for chest CT were 11.5-12.9 HU and 11.0-14.0 HU and for abdominal were 10.1-12.1 HU and 9.4-13.7 HU respectively. With these findings, they have founded the dose differentiation ranges with the population. They had founded that these quantifiable findings can be help to obtain different types of high quality images with dose reduction methods as per the requests(Ria et al. 2019a).
Clerkin et al. (2018) conducted a study to investigate if there was variation in the radiation dose delivered during head and neck (HN) localization computed tomography (CT) imaging in radiation therapy (RT) in Ireland and a sample of European departments. The aim of the study was to propose national diagnostic reference levels (DRLs) for this procedure and to compare the national DRL to that of the European sample. The study surveyed all RT departments in Ireland and a sample of European departments to acquire data on CT dose-length product (DLP), dose index volume (CTDIvol), current-time product, tube voltage, and scan length for ten average-sized HN patients from each department. The mean DLP, CTDIvol, and scan lengths were compared across the Irish departments and with the European sample to determine if there was significant variation in the radiation dose delivered during HN localization CT imaging. The results showed significant variation in the mean DLP, CTDIvol, and scan lengths across the Irish departments. The authors proposed national DRLs of 882 mGy.cm and 21 mGy for DLP and CTDIvol, respectively, using the rounded 75th percentile of the distribution of the means. The authors also found that the European sample had lower DRLs than the proposed Irish DRLs, with a DLP DRL of 816 mGy.cm and a CTDIvol DRL of 21 mGy.The study by Clerkin et al. (2018) highlights the importance of establishing national DRLs for radiation dose in HN localization CT imaging in radiation therapy. The proposed DRLs can be used for comparison purposes to help optimize radiation dose delivery during the procedure, ensuring that patients receive adequate radiation dose while minimizing unnecessary radiation exposure. The study also emphasizes the need for standardization of radiation dose delivery in HN localization CT imaging across different departments to ensure consistent and safe treatment of patients.
Coombs et al. (2013) conducted a multicenter retrospective study to develop diagnostic reference ranges (DRRs) and a method for calculating site-specific reference doses for computed tomographic (CT) scans of the abdomen or abdomen and pelvis in children based on body width (BW). The study aimed to address the balance between the patient’s risk (radiation dose) and benefit (diagnostic image quality) by establishing DRRs. The study analyzed CT doses in 939 pediatric patients, with doses from 954 scans obtained from September 1 to December 1, 2009, through Quality Improvement Registry for CT Scans in Children within the National Radiology Data Registry, American College of Radiology. The size-specific dose estimate (SSDE), a dose estimate based on BW, CT dose index, dose-length product, and effective dose were analyzed. The BW measurement was obtained with electronic calipers from the axial image at the splenic vein level after completion of the CT scan. An adult-sized patient was defined as a patient with a BW of 34 cm. To establish DRRs, image quality was reviewed on a subset of CT scans through comparison with a five-point visual reference scale with increments of added simulated quantum mottle, and DRRs were determined to establish lower and upper bounds for each range. The study developed appropriate dose ranges for each DRR, and for 954 scans, DRRs (SSDEs) were 5.8–12.0, 7.3–12.2, 7.6–13.4, 9.8–16.4, and 13.1–19.0 mGy for BWs less than 15, 15–19, 20–24, 25–29, and 30 cm or greater, respectively. The fractions of adult doses, adult SSDEs, used within the consortium for patients with BWs of 10, 14, 18, 22, 26, and 30 cm were 0.4, 0.5, 0.6, 0.7, 0.8, and 0.9, respectively. The authors concluded that DRRs provide a tool to help manage pediatric patient radiation doses by providing a balance between radiation dose and diagnostic image quality. Calculation of reference doses as a function of BW for an individual practice provides a way to develop site-specific CT protocols, which can help manage pediatric patient radiation doses. The study provides valuable information for healthcare professionals involved in pediatric CT scanning to optimize radiation dose and image quality.
A study conducted by Sukumar et al (2023) aimed to conduct an extensive literature review to understand the current situation regarding Diagnostic Reference Levels (DRLs) based on clinical indications for Radiation Therapy Computed Tomography (RTCT). The objective was to review established DRLs and methodologies for establishing Diagnostic Reference Levels in radiation therapy planning computed tomography (CT). Eligibility criteria for the study included cohort studies reporting DRLs in adult patients undergoing CT for radiation therapy for the region head and neck or pelvis. Comprehensive literature searches for relevant studies published between 2000 and 2021 were performed using PubMed, Scopus, CINHAL, Web of Science, and ProQuest. Out of 356 articles identified through the literature search, 68 duplicate reports were removed. The title and abstract of 288 studies were assessed, and 16 of 288 articles were selected for full-text screening (studies conducted between 2000 and 2021). Finally, five articles were included in the review after the full-text screening. The conclusion drawn from the study was that a globally approved standard protocol that includes scanning techniques, dose measurement method, and DRL percentile needs to be established to make a valuable and accurate comparison with international DRLs.
General Objective
To Establish a local diagnostic reference level of radiation dose for radiotherapy CT simulation procedures at National Cancer Institute, Maharagama- Sri Lanka.
Special Objectives
- To estimate the patient radiation dose during the radiotherapy CT simulation procedure.
- To compare the established DRLs with National DRLs of other countries
- To optimize the radiation dose in radiotherapy CT simulation procedures in national cancer institute Sri Lanka.
03. Materials and methods
The study was conducted at the CT simulation unit of the Department of Radiotherapy in National Hospital Maharagama, Sri Lanka, using a Toshiba Aquilion Large Bore CT scanner. The data collection period spanned two months, from October to November 2022, during which patient data was collected randomly from 350 patients between the ages of 18 to 90 years. Patients under 18 years of age were excluded due to their significantly lower CT radiation dose and variable sizes. The procedures considered in the study were non-contrast and focused on the most commonly requested simulation body sites, including the head, chest, abdomen, and pelvis.
The collected data was in the form of dose reports stored as DICOM images. The study could not compare radiation dose based on patient weight groups, as weight data was not available. However, the study aimed to collect sufficient data from a large sample per procedure as recommended by the International Commission on Radiological Protection (ICRP). The study collected more than 30 scan data points per category in an anatomical region to ensure that the institutional diagnostic reference levels (DRLs) were established by considering the median. In total, 349 CT simulation procedures were performed.
The main parameters used to set up institutional DRLs were computed tomography dose index (CTDI) and dose-length product (DLP). Additionally, data on the total number of images, exposure (mAs), and total scan time were collected to further analyze the relationships between these parameters. The collected data were recorded separately based on the anatomical region.
To establish the DRLs, the 25th, 50th, and 75th percentiles of DLP and CTDI values were calculated for each anatomical region separately. The 75th percentile was considered as the newly established DRL value. The data analysis was performed using the Statistical Package for Social Sciences (SPSS) software.
In summary, the study aimed to establish institutional DRLs for radiotherapy CT simulation procedures based on a large sample of patient data collected over two months. The study focused on the most commonly requested simulation body sites and used CTDI and DLP parameters to establish DRLs. The newly established DRLs were based on the 75th percentile values of DLP and CTDI separately for each anatomical region.
04. Results
The dataset included 350 valid age values, with a range from 19 to 87 years old. The mean age of all patients in the dataset was 58.96, which indicates that the average age of patients receiving CT scans was just under 60 years old. The standard deviation of 12.719 suggests that there was some variation in the age of patients, with some patients being significantly older or younger than the mean age. The median age of 61 suggests that the distribution of age was roughly symmetric, with an equal number of patients being older and younger than 61. The interquartile range (IQR) of 16 indicates that the middle 50% of patients fell within an age range of 52 to 68.
Table 1; Statistics
| Statistics | ||
| Age | ||
| N | Valid | 350 |
| Missing | 0 | |
| Mean | 58.96 | |
| Median | 61.00 | |
| Mode | 56a | |
| Std. Deviation | 12.719 | |
| Range | 68 | |
| Minimum | 19 | |
| Maximum | 87 | |
| Percentiles | 25 | 52.00 |
| 50 | 61.00 | |
| 75 | 68.00 | |
| Multiple modes exist. The smallest value is shown |
Figure 1: Histogram of age
The study consisted of 50.3% female patients and 49.7% male patients, which suggests that there was a relatively equal distribution of genders in the patient population.
| Gender | |||||
| Frequency | Percent | Valid Percent | Cumulative Percent | ||
| Valid | F | 176 | 50.3 | 50.3 | 50.3 |
| M | 174 | 49.7 | 49.7 | 100.0 | |
| Total | 350 | 100.0 | 100.0 |
Table 2 : Gender
| Body part | |||||
| Frequency | Percent | Valid Percent | Cumulative Percent | ||
| Valid | abdomen | 31 | 8.9 | 8.9 | 8.9 |
| chest | 84 | 24.0 | 24.0 | 32.9 | |
| head | 102 | 29.1 | 29.1 | 62.0 | |
| pelvis | 133 | 38.0 | 38.0 | 100.0 | |
| Total | 350 | 100.0 | 100.0 |
| Table 3: Body Part |
The
research is focused on developing a diagnostic reference level (DRL) for radiation dose for radiotherapy CT simulation procedures. The established DRL value is the 75th percentile. The study has collected data on the computed tomography dose index (CTDI) volume and dose-length product (DLP) for different body parts such as abdomen, chest, head, and pelvis. The CTDI volume values for the abdomen range from 5.60 to 13.00 mGy, with the 75th percentile at 13.00 mGy, while the DLP values range from 256.00 to 548.00 mGy·cm, with the 75th percentile at 548.00 mGy·cm. Similarly, for the chest, the CTDI volume values range from 11.10 to 22.30 mGy, with the 75th percentile at 22.30 mGy, and the DLP values range from 312.50 to 649.00 mGy·cm, with the 75th percentile at 649.00 mGy·cm. For the head, the CTDI volume values range from 58.20 to 63.50 mGy, with the 75th percentile at 63.50 mGy, while the DLP values range from 2269.00 to 3161.00 mGy·cm, with the 75th percentile at 3161.00 mGy·cm. Finally, for the pelvis, the CTDI volume values range from 6.80 to 15.50 mGy, with the 75th percentile at 15.50 mGy, and the DLP values range from 241.00 to 555.00 mGy·cm, with the 75th percentile at 555.00 mGy·cm.
| CTDI volume | DLP | ||||||
| Percentile 25 | Median | Percentile 75 | Percentile 25 | Median | Percentile 75 | ||
| Body part | abdomen | 5.60 | 7.40 | 13.00 | 256.00 | 394.00 | 548.00 |
| chest | 11.10 | 15.50 | 22.30 | 312.50 | 454.50 | 649.00 | |
| head | 58.20 | 58.20 | 63.50 | 2269.00 | 2632.00 | 3161.00 | |
| pelvis | 6.80 | 9.30 | 15.50 | 241.00 | 339.00 | 555.00 |
Table 4: CTDI voumes & DLPs
The comparison of the established Diagnostic Reference Levels (DRLs) for radiation dose in CT simulation procedures for radiotherapy at the National Cancer Institute (NCI) in Sri Lanka with DRLs established in other countries and institutions can be conducted using two metrics, namely, the CT Dose Index (CTDI) volume and the Dose-Length Product (DLP).
When comparing the CTDI volume values of NCI Sri Lanka with other countries, it can be observed that the values for the abdomen and pelvis regions in NCI Sri Lanka are lower than those established in Slovenia and Croatia. However, the value for the chest region in NCI Sri Lanka is higher than that in Slovenia and similar to that in Croatia. For the head region, the value in NCI Sri Lanka is lower than that in Slovenia but higher than that in Kasturba Hospital, Manipal, Karnataka and Croatia.
Similarly, when comparing the DLP values, it can be seen that the values for the abdomen and pelvis regions in NCI Sri Lanka are lower than those established in Slovenia and Croatia. The value for the chest region in NCI Sri Lanka is higher than that in Slovenia and similar to that in Croatia. For the head region, the value in NCI Sri Lanka is lower than that in Slovenia but higher than that in Kasturba Hospital, Manipal, Karnataka and Croatia.
In general, the comparison indicates that NCI Sri Lanka has established relatively lower CTDI volume and DLP values for most body regions compared to other countries and institutions, except for the chest region where the values are higher. Further evaluation is necessary to determine the appropriateness of these lower CTDI volume and DLP values in maintaining patient safety and minimizing radiation exposure during radiotherapy CT simulation procedures.
| Country/ Institute | Body region | CTDI volume(mGy) | DLP(mGy.cm) |
| NCI Sri Lanka | abdomen | 13.00 | 584.00 |
| NCI Sri Lanka | chest | 22.30 | 649.00 |
| NCI Sri Lanka | head | 63.50 | 3161.00 |
| NCI Sri Lanka | pelvis | 15.50 | 555.00 |
| Slovenia | abdomen | 18.70 | 1116.20 |
| Slovenia | chest | 19.20 | 832.40 |
| Slovenia | head | 76.90 | 1942.40 |
| Slovenia | pelvis | 17.90 | 677.10 |
| Kasturba hospital, Manipal, Karnataka | head | 17.60 | 790.65 |
| Kasturba hospital, Manipal, Karnataka | pelvis | 17.60 | 999.70 |
| Croatia | head | 35.00 | 1444.00 |
| Croatia | chest | 17.00 | 865.00 |
| Croatia | pelvis | 20.00 | 1133.00 |
Analyzing the collected data further, the mean number of images was 164.61, with a range from 68 to 1008 images. The standard deviation of 61.569 indicates that there was some variability in the number of images obtained across the dataset, with some scans having many more or fewer images than the mean.
For exposure, which is a measure of the amount of radiation received during the CT scan. The mean exposure was 4622.0488, with a range from 68.09 to 12375.00. The standard deviation of 2457.92381 indicates that there was a large variability in the amount of radiation exposure across the dataset, with some scans receiving significantly more or less radiation than the mean.
scan time, which is the length of time it took to complete the CT scan. The mean scan time was 43.8178 seconds, with a range from 17 to 3587 seconds. The standard deviation of 190.25489 suggests that there was a significant variability in scan time across the dataset, with some scans being much longer or shorter than the mean.
CTDI volume, which is a measure of the radiation dose delivered during the CT scan. The mean CTDI volume was 27.1977, with a range from 2.5 to 116.4. The standard deviation of 23.01101 suggests that there was some variability in radiation dose across the dataset, with some scans receiving significantly more or less radiation than the mean.
DLP, which is a measure of the total amount of radiation received during the CT scan. The mean DLP was 1135.9343, with a range from 23.00 to 5560.00. The standard deviation of 1132.63167 indicates that there was a large variability in the amount of radiation received across the dataset, with some scans receiving significantly more or less radiation than the mean.
scan length, which is the distance covered by the CT scan. The mean scan length was 40.5110, with a range from 0.40 to 272.55. The standard deviation of 17.68602 suggests that there was some variability in scan length across the dataset, with some scans being significantly longer or shorter than the mean.
According to the statistics, the body part that had the highest values for most of the parameters was the head. The head had the highest mean and median values for exposure, scan time, CTDI volume, DLP, and scan length. It also had the highest maximum values for exposure, CTDI volume, and DLP.
On the other hand, the body part with the lowest values for most of the parameters was the chest. The chest had the lowest mean and median values for exposure, CTDI volume, DLP, and scan length. It also had the lowest maximum values for exposure, CTDI volume, DLP, and scan length.
| Statistics | |||||||
| Total number of images | Exposure | Scan time | CTDI volume | DLP | Scan length | ||
| N | Valid | 350 | 350 | 350 | 350 | 350 | 350 |
| Missing | 0 | 0 | 0 | 0 | 0 | 0 | |
| Mean | 164.61 | 4622.0488 | 43.8178 | 27.1977 | 1135.9343 | 40.5110 | |
| Median | 161.00 | 3839.0000 | 31.3250 | 17.3000 | 540.5000 | 38.2892 | |
| Mode | 201 | 6853.00 | 33.10 | 58.20 | 2513.00 | 42.89 | |
| Std. Deviation | 61.569 | 2457.92381 | 190.25489 | 23.01101 | 1132.63167 | 17.68602 | |
| Minimum | 68 | 68.09 | 17.00 | 2.50 | 23.00 | .40 | |
| Maximum | 1008 | 12375.00 | 3587.00 | 116.40 | 5560.00 | 272.55 |
Table 5: Statistics
| Body part | |||||
| abdomen | chest | head | pelvis | ||
| Total number of images | Mean | 215 | 150 | 140 | 181 |
| Median | 213 | 136 | 134 | 182 | |
| Minimum | 103 | 82 | 68 | 106 | |
| Maximum | 331 | 1008 | 245 | 249 | |
| Standard Deviation | 51 | 101 | 31 | 29 | |
| Exposure | Mean | 3795.26 | 3756.40 | 7337.62 | 3278.87 |
| Median | 3280.00 | 3327.50 | 6960.00 | 2701.00 | |
| Minimum | 1866.00 | 1547.00 | 4320.00 | 68.09 | |
| Maximum | 8508.00 | 9573.00 | 12375.00 | 9980.00 | |
| Standard Deviation | 1902.85 | 1840.68 | 1498.40 | 1774.34 | |
| Scan time | Mean | 49.09 | 72.68 | 32.31 | 33.19 |
| Median | 42.73 | 26.67 | 31.28 | 31.80 | |
| Minimum | 17.00 | 22.00 | 21.75 | 25.80 | |
| Maximum | 118.50 | 3587.00 | 48.75 | 88.50 | |
| Standard Deviation | 23.39 | 388.18 | 5.85 | 7.14 | |
| CTDI volume | Mean | 10.05 | 17.70 | 59.75 | 12.23 |
| Median | 7.40 | 15.50 | 58.20 | 9.30 | |
| Minimum | 2.80 | 2.50 | 40.90 | 2.50 | |
| Maximum | 29.70 | 57.50 | 116.40 | 50.10 | |
| Standard Deviation | 7.20 | 10.52 | 9.10 | 9.18 | |
| DLP | Mean | 488.35 | 530.42 | 2669.86 | 492.91 |
| Median | 394.00 | 454.50 | 2632.00 | 339.00 | |
| Minimum | 158.00 | 102.00 | 23.00 | 126.00 | |
| Maximum | 1411.00 | 1524.00 | 4625.00 | 5560.00 | |
| Standard Deviation | 353.72 | 332.05 | 744.69 | 550.55 | |
| Scan length | Mean | 51.09 | 31.81 | 44.33 | 40.62 |
| Median | 47.62 | 29.94 | 43.18 | 37.10 | |
| Minimum | 31.76 | 3.75 | .40 | 20.89 | |
| Maximum | 124.29 | 87.23 | 67.78 | 272.55 | |
| Standard Deviation | 17.44 | 11.16 | 9.61 | 22.87 |
Table 6: Statistics for the body parts
According to the correlation coefficients and associated p-values,
the first variable, age, shows a weak positive correlation with the total number of images (r = 0.107, p = 0.046), but weak negative correlations with exposure (r = -0.113, p = 0.034), scantime (r = -0.083, p = 0.120), CTDI volume (r = -0.111, p = 0.039), DLP (r = -0.113, p = 0.035), and scan length (r = -0.044, p = 0.409).The second variable, total number of images, exhibits a weak positive correlation with scan length (r = 0.241, p = 0.000), and a weak negative correlation with CTDI volume (r = -0.257, p = 0.000) and DLP (r = -0.138, p = 0.010).The third variable, exposure, displays a strong positive correlation with CTDI volume (r = 0.878, p = 0.000) and DLP (r = 0.870, p = 0.000), and a weak positive correlation with scan length (r = 0.211, p = 0.000).The fourth variable, scan time, shows no significant correlation with any of the other variables. The fifth variable, CTDI volume, exhibits a strong positive correlation with DLP (r = 0.921, p = 0.000), and weak negative correlations with age (r = -0.111, p = 0.039) and total number of images (r = -0.257, p = 0.000).The sixth variable, DLP, displays a strong positive correlation with CTDI volume (r = 0.921, p = 0.000), and weak negative correlations with age (r = -0.113, p = 0.035) and total number of images (r = -0.138, p = 0.010).The seventh variable, scan length, shows weak positive correlations with the total number of images (r = 0.241, p = 0.000) and exposure (r = 0.211, p = 0.000).
Overall, the correlation analysis indicates that exposure, CTDI volume, and DLP are strongly correlated with each other, while age and total number of images have weaker correlations with these variables. Scan time does not show any significant correlation with the other variables, and scan length has weak positive correlations with the total number of images and exposure.
| Correlations | ||||||||
| Age | Total number of images | Exposure | Scan time | CTDI volume | DLP | Scan length | ||
| Age | Pearson Correlation | 1 | .107* | -.113* | -.083 | -.111* | -.113* | -.044 |
| Sig. (2-tailed) | .046 | .034 | .120 | .039 | .035 | .409 | ||
| N | 350 | 350 | 350 | 350 | 350 | 350 | 350 | |
| Total number of images | Pearson Correlation | .107* | 1 | -.073 | -.009 | -.257** | -.138** | .241** |
| Sig. (2-tailed) | .046 | .174 | .865 | .000 | .010 | .000 | ||
| N | 350 | 350 | 350 | 350 | 350 | 350 | 350 | |
| Exposure | Pearson Correlation | -.113* | -.073 | 1 | -.035 | .878** | .870** | .211** |
| Sig. (2-tailed) | .034 | .174 | .516 | .000 | .000 | .000 | ||
| N | 350 | 350 | 350 | 350 | 350 | 350 | 350 | |
| Scan time | Pearson Correlation | -.083 | -.009 | -.035 | 1 | -.048 | -.037 | .016 |
| Sig. (2-tailed) | .120 | .865 | .516 | .373 | .495 | .762 | ||
| N | 350 | 350 | 350 | 350 | 350 | 350 | 350 | |
| CTDI volume | Pearson Correlation | -.111* | -.257** | .878** | -.048 | 1 | .921** | .084 |
| Sig. (2-tailed) | .039 | .000 | .000 | .373 | .000 | .116 | ||
| N | 350 | 350 | 350 | 350 | 350 | 350 | 350 | |
| DLP | Pearson Correlation | -.113* | -.138** | .870** | -.037 | .921** | 1 | .397** |
| Sig. (2-tailed) | .035 | .010 | .000 | .495 | .000 | .000 | ||
| N | 350 | 350 | 350 | 350 | 350 | 350 | 350 | |
| Scan length | Pearson Correlation | -.044 | .241** | .211** | .016 | .084 | .397** | 1 |
| Sig. (2-tailed) | .409 | .000 | .000 | .762 | .116 | .000 | ||
| N | 350 | 350 | 350 | 350 | 350 | 350 | 350 | |
| *. Correlation is significant at the 0.05 level (2-tailed). | ||||||||
| **. Correlation is significant at the 0.01 level (2-tailed). |
Table 7: Correlations
The regression analysis aims to predict CTDI volume using a set of predictors, including Total number of images, Exposure, Scan time, Scan length, Gender, Body part, and a constant. The results show that the model is significant (p < 0.05) and accounts for 81.7% of the variance in CTDI volume.
The standardized regression coefficients show that Exposure has the strongest positive effect on CTDI volume, while Total number of images has a negative effect. Gender, Scan length, and Body part have weak effects on CTDI volume.
The collinearity diagnostics indicate that there is some multicollinearity between the Total number of images and Exposure predictors, as their variance proportions are relatively high. However, their tolerance and VIF values suggest that this is not a significant issue.
Overall, the results suggest that Exposure and Total number of images are important predictors of CTDI volume, while Gender, Scan length, and Body part have weaker effects
| Model Summary | |||||||||
| Model | R | R Square | Adjusted R Square | Std. Error of the Estimate | Change Statistics | ||||
| R Square Change | F Change | df1 | df2 | Sig. F Change | |||||
| 1 | .904a | .817 | .814 | 9.92012 | .817 | 255.810 | 6 | 343 | .000 |
| a. Predictors: (Constant), Body part numeric, Total number of images, Scan time, Exposure, Gender , Scan length | |||||||||
| b. Dependent Variable: CTDI volume |
Table 8: Model Summery
| ANOVA | ||||||
| Model | Sum of Squares | df | Mean Square | F | Sig. | |
| 1 | Regression | 151043.607 | 6 | 25173.934 | 255.810 | .000a |
| Residual | 33754.211 | 343 | 98.409 | |||
| Total | 184797.818 | 349 | ||||
| a. Predictors: (Constant), Body part numeric, Total number of images, Scan time, Exposure, Gender , Scan length | ||||||
| b. Dependent Variable: CTDI volume |
Table 9: ANOVA
| Coefficients | ||||||||
| Model | ||||||||
| 1 | ||||||||
| (Constant) | Total number of images | Exposure | Scan time | Scan length | Gender | Body part numeric | ||
| Unstandardized Coefficients | B | 6.597 | -.069 | .008 | -.002 | -.090 | -2.984 | .095 |
| Std. Error | 3.509 | .009 | .000 | .003 | .032 | 1.136 | .054 | |
| Standardized Coefficients | Beta | -.184 | .863 | -.013 | -.069 | -.065 | .041 | |
| t | 1.880 | -7.656 | 34.833 | -.560 | -2.795 | -2.627 | 1.760 | |
| Sig. | .061 | .000 | .000 | .576 | .005 | .009 | .079 | |
| 95.0% Confidence Interval for B | Lower Bound | -.304 | -.087 | .008 | -.007 | -.154 | -5.217 | -.011 |
| Upper Bound | 13.498 | -.051 | .009 | .004 | -.027 | -.750 | .201 | |
| Correlations | Zero-order | -.257 | .878 | -.048 | .084 | -.284 | -.016 | |
| Partial | -.382 | .883 | -.030 | -.149 | -.140 | .095 | ||
| Part | -.177 | .804 | -.013 | -.064 | -.061 | .041 | ||
| Collinearity Statistics | Tolerance | .920 | .868 | .992 | .863 | .872 | .989 | |
| VIF | 1.087 | 1.152 | 1.008 | 1.159 | 1.147 | 1.011 |
Table 10: Coefficients
| Collinearity Diagnostics | |||||||||||
| Model | Dimension | Eigenvalue | Condition Index | Variance Proportions | |||||||
| (Constant) | Total number of images | Exposure | Scan time | Scan length | Gender | Body part numeric | |||||
| 1 | 1 | 5.459 | 1.000 | .00 | .00 | .01 | .00 | .00 | .00 | .00 | |
| 2 | .943 | 2.406 | .00 | .00 | .00 | .98 | .00 | .00 | .00 | ||
| 3 | .234 | 4.833 | .00 | .02 | .52 | .00 | .03 | .08 | .02 | ||
| 4 | .155 | 5.936 | .00 | .15 | .19 | .00 | .37 | .10 | .02 | ||
| 5 | .098 | 7.454 | .00 | .35 | .06 | .00 | .12 | .06 | .51 | ||
| 6 | .093 | 7.666 | .00 | .32 | .02 | .01 | .41 | .24 | .21 | ||
| 7 | .018 | 17.316 | 1.00 | .16 | .21 | .00 | .07 | .51 | .24 |
Table 11: Collinearity Diagnostics
| Residuals Statistics | |||||
| Minimum | Maximum | Mean | Std. Deviation | N | |
| Predicted Value | -56.3100 | 85.8704 | 27.1977 | 20.80360 | 350 |
| Residual | -33.57513 | 58.81004 | .00000 | 9.83448 | 350 |
| Std. Predicted Value | -4.014 | 2.820 | .000 | 1.000 | 350 |
| Std. Residual | -3.385 | 5.928 | .000 | .991 | 350 |
| a. Dependent Variable: CTDI volume |
Table 12: Residuals Statistics
Figure 3
Based on the regression analysis, the model shows a strong relationship between the predictors and the dependent variable (DLP). The R-squared value of 0.829 indicates that the model explains 82.9% of the variance in the dependent variable.
Among the predictors, exposure has the highest standardized coefficient (0.795), indicating that it has the strongest effect on DLP. Scan length (0.252), body part (0.062), and total number of images (-0.147) also have significant effects on DLP. Gender and scan time have relatively weaker effects on DLP.
The collinearity diagnostics show that there is low multicollinearity among the predictors, indicating that the model is reliable.
Overall, the regression analysis suggests that exposure, scan length, body part, and total number of images are significant predictors of DLP, and these factors should be considered when performing CT scans to minimize radiation exposure.Top of Form
| Model Summary | |||||||||
| Model | R | R Square | Adjusted R Square | Std. Error of the Estimate | Change Statistics | ||||
| R Square Change | F Change | df1 | df2 | Sig. F Change | |||||
| 1 | .911a | .829 | .826 | 471.84209 | .829 | 277.998 | 6 | 343 | .000 |
| a. Predictors: (Constant), Body part , Total number of images, Scan time, Exposure, Gender, Scan length | |||||||||
| b. Dependent Variable: DLP |
Table 13: Model Summary
| ANOVA | ||||||
| Model | Sum of Squares | df | Mean Square | F | Sig. | |
| 1 | Regression | 3.714E8 | 6 | 6.189E7 | 277.998 | .000a |
| Residual | 7.636E7 | 343 | 222634.956 | |||
| Total | 4.477E8 | 349 | ||||
| a. Predictors: (Constant), Body part, Total number of images, Scan time, Exposure, Gender, Scan length | ||||||
| b. Dependent Variable: DLP |
Table 14: ANOVA
| Coefficients | ||||||||
| Model | ||||||||
| 1 | ||||||||
| (Constant) | Total number of images | Exposure | Scan time | Scan length | Gender | Body part | ||
| Unstandardized Coefficients | B | -803.647 | -2.696 | .366 | -.049 | 16.128 | -114.541 | 7.134 |
| Std. Error | 166.881 | .428 | .011 | .133 | 1.537 | 54.013 | 2.571 | |
| Standardized Coefficients | Beta | -.147 | .795 | -.008 | .252 | -.051 | .062 | |
| t | -4.816 | -6.304 | 33.216 | -.371 | 10.491 | -2.121 | 2.775 | |
| Sig. | .000 | .000 | .000 | .711 | .000 | .035 | .006 | |
| 95.0% Confidence Interval for B | Lower Bound | -1131.885 | -3.537 | .345 | -.312 | 13.104 | -220.780 | 2.078 |
| Upper Bound | -475.409 | -1.855 | .388 | .213 | 19.152 | -8.302 | 12.190 | |
| Correlations | Zero-order | -.138 | .870 | -.037 | .397 | -.332 | .014 | |
| Partial | -.322 | .873 | -.020 | .493 | -.114 | .148 | ||
| Part | -.141 | .741 | -.008 | .234 | -.047 | .062 | ||
| Collinearity Statistics | Tolerance | .920 | .868 | .992 | .863 | .872 | .989 | |
| VIF | 1.087 | 1.152 | 1.008 | 1.159 | 1.147 | 1.011 |
| Table 15: Coefficients |
| Collinearity Diagnostics | |||||||||||
| Model | Dimension | Eigenvalue | Condition Index | Variance Proportions | |||||||
| (Constant) | Total number of images | Exposure | Scan time | Scan length | Gender | Body part | |||||
| 1 | 1 | 5.459 | 1.000 | .00 | .00 | .01 | .00 | .00 | .00 | .00 | |
| 2 | .943 | 2.406 | .00 | .00 | .00 | .98 | .00 | .00 | .00 | ||
| 3 | .234 | 4.833 | .00 | .02 | .52 | .00 | .03 | .08 | .02 | ||
| 4 | .155 | 5.936 | .00 | .15 | .19 | .00 | .37 | .10 | .02 | ||
| 5 | .098 | 7.454 | .00 | .35 | .06 | .00 | .12 | .06 | .51 | ||
| 6 | .093 | 7.666 | .00 | .32 | .02 | .01 | .41 | .24 | .21 | ||
| 7 | .018 | 17.316 | 1.00 | .16 | .21 | .00 | .07 | .51 | .24 |
Table 16: Collinearity Diagnostics
| Residuals Statistics | |||||
| Minimum | Maximum | Mean | Std. Deviation | N | |
| Predicted Value | -2141.5264 | 4657.6221 | 1135.9343 | 1031.52657 | 350 |
| Residual | -2061.62476 | 2263.52637 | .00000 | 467.76856 | 350 |
| Std. Predicted Value | -3.177 | 3.414 | .000 | 1.000 | 350 |
| Std. Residual | -4.369 | 4.797 | .000 | .991 | 350 |
| a. Dependent Variable: DLP |
Table 17: Residuals Statistics
Bottom of Form