Supervisors info:
Τζων Καλέφ-Εζρά, Καθηγητής, Πανεπιστήμιο Ιωαννίνων
Σωτήρης Οικονομίδης, Ακτινοφυσικός, ΕΕΑΕ
Γεώργιος Σημαντηράκης, Ακτινοφυσικός, ΕΕΑΕ
Summary:
Introduction and Purpose: Computed Tomography (CT) is one of the most significant applications of Diagnostic Radiology due to (a) the high radial burden of the patient required per examination and (b) the growth rate of the number of CT examinations. The purpose of the current study is the estimation of the radial burden of patients from CT examinations, so that optimization can be taken into account, if it is appropriate. For each CT examination, the lowest possible dose should be achieved while diagnostic accuracy of the image remains satisfactory.
Method: Features such as age, gender and somatometric characteristics of 364 adult patients were obtained. Furthermore, data of dosimetric interest were registered from 471 scans performed on these patients in the Computational Tomography Unit of the Radiological Laboratory of the University General Hospital of Ioannina during 25th of October and 19th of December 2016. Information were obtained by using 16-slice Philips Mx 8000 IDT computed tomography scanner and included eleven of the most common CT examinations at the UGH of Ioannina. Moreover, dosimetric scanning data of single-slice CTs was used. This information was recorded in the archive of the Greek Atomic Energy Commission (G.A.E.C.) and was used to estimate the typical dose levels in CT examinations in UGHI and in order to compare them with National Diagnostic Reference Levels (DRLs). Dosimetric indicators CTDIw, CTDIvol and DLP, as well as the effective dose per scan were calculated and compared between two different software packages, ImPACT 1.0.4 and CT-Expo v. 2.2.
Results: The typical dose levels estimated for the University General Hospital of Ioannina are in angreement with the national DRL values as proposed by G.A.E.C. and other countries (United Kingdom, Sweden, Ireland, Switzerland, France). The only exception is the examination of the skull / sinus viscera in which the value of dosimetric indicators is higher than DRL values reported in other countries' studies (18% for Europe, 115% for Ireland and 41% for Switzerland). This is due to the common scanning protocol used in UGHI for all patients for the anatomical regions of lithoidal bone and paranasal sinuses. Also, in the upper abdominal and abdominal&pelvic examinations, the DLP value reported in single acquisitions is 10% higher than the corresponding value of the United Kingdom, which refers to a complete examination. This may result due to the different total scanning length.
The comparison of dosimetric values between two different software packages, ImPACT and CT-Expo, results in a statistically significant difference in the dose estimation. Percentage differences of the coefficient of variance for the effective dose, vary from 6% to 112% (6% of the abdominal, 10% of the brain, 20% of the neck and 31% of the pulmonary angiography), with a significant divergence for sinuses (73%) and carotid angiography 112%. This is due to the different algorithm used between the two software packages, the phantoms used for each software and the options offered by CT-EXPO for corrections a) of scanner-specific influences (for the effect of longitudinal (z-) dose modulation using mAs profiles), b) of overbeaming effects, and c) of over-ranging effects in spiral mode.
Studying chest examinations, the result obtained is the decreasing DLP value for effective diameter less than 25 cm and stability for larger diameters. For abdominal examinations, the CTDIvol value in patients having BMI between 20-25 kg/m2 remains stable and increases for patients having a higher BMI value, above 25 kg/m2. In addition, for pulmonary angiography examinations, a higher CTDIvol value is observed for older patients, due to the increase in the mass observed at these ages and consequently the exposure data, indicating how age may affect the dose of the patient in CT examinations.
Percentage differences in the value of dosimetric indices between patients examined with the same examination protocol range from 2.5% - 44.3%, something that can be explained to the logitutinal (z) dose modulation which depends on the body type of each patient and the anatomical area being examined. Additionally, an increase in mA value was displayed, up to 40% for abdominal examinations and 30% for chest examinations in patients with higher BMI values. Large percentage differences in the neck examination protocol (~ 23.2%) were attributed to the use of a different type of scanning (axial, helical) applied to some examinations and to the variation of the mA value due to heterogeneity in the anatomy of the area (neck, shoulders). Furthermore, large discrepancies in brain angiography (carotid&willis) are due to the use of a different scan protocol (carotid angiography protocol) used for the same examination. The greatest percentage difference observed in effective dose values in all protocols (30%) compared to CTDIvol (5%) and DLP (7%) because of the gender of the patients which was taken into account in the calculations with CT-EXPO software.
In addition, a wide range of effective dose value (5-11 mSv) was observed, even between same type systems (Toshiba Asteion), attributed not only to the dosimetric output of the systems but also, to the choice of exposure data for the clinical protocols used.
In agreement with other studies, single-slice CT scanners have displayed lower dose values in the exams being studied, compared to corresponding values on multi-slice CT scanners with an embedded automatic exposure control system on them.
One of the limitations in this study was the small number of patients used in the sample in studying dose correlation with BMI (15 patients with BMI<=35kg/m2), with patient characteristics (effective diameter) and the effect of the contrast administration in CTDIvol value (10 patients).
Conclusions: This work’s results have emerged the need of further optimization in the radiation protection, making appropriate modifications of the clinical protocols used, without degradation of diagnostic image quality. This can be achieved with careful selecting protocols and adapting them to each patient group, considering the size of each patient, the body of the recommendation and the clinical need. Patient dose estimation according to body type can be achieved by using a standardized approach related to patient’s size. This approach should express radiation attenuation in the patient's body and allow Size Specific Dose Estimate (SSDE) evaluation for all the patients regardless of age and body type for the most accurate dose estimation.
In addition, in the relative dose estimation, DLP of the completed examination should be used and not its value of the standard scan, so that the evaluation of patient’s dose is more accurate. All patient dose-related data should be available to doctors and radiologists in order to take into account the patients’ dose from all CT examinations. Finally, DRL values need to change depending on patient's pathology and be established for examinations such as neck, chest high-resolution, pulmonary angiography, carotid angiography, and brain carotid (carotid & willis).
