Because it is not possible to put thermoluminescent dosimeter (TLD) chips inside the body to directly measure the ovarian dose, the study was performed in two steps (in-vitro and in-vivo). The results of the in-vitro study were used to calculate ovarian dose in the in-vivo study. The TLDs used for this study were LiF: TLD-100, manufactured by Bircon/Harshow (city/country), packed in groups of three. Reading of the TLD-100 chips was carried out by the TLD reader system model Harshaw-3500.
3.1. In-Vitro Study
In this step, the radiation attenuation curves were derived in an anthropomorphic phantom for both conventional DSA (Advantx, GE Medical Systems, Illinois, USA) and flat panel (Innova 4100, GE Medical Systems, Illinois, USA) angiography systems used in this study for calculating the depth dose according to entrance surface dose (ESD). Ten batches containing TLD-100 chips were put on the posterior surface (entrance of X-rays) and in different depths of the phantom at the right and left ovarian regions and mid-point of the phantom. Each batch containing three chips of TLD-10 was used in each location to improve the accuracy and the mean of three TLD chip readings was used to measure the radiation dose of that location. The configuration of angiography systems such as tube-table distance and patient to image receptor distance was set for routine UAE procedures.
As a fixed combination of these two parameters, we set the focal spot-detector distance for both systems at 95 cm. Medium size field of view (FOV) was used in this study for in-vitro and in-vivo steps that was fixed as 30 cm in the conventional DSA and 32 cm in the flat panel system.
The high beam filter used consisted of 1 mm aluminum and 0.1 mm copper for conventional DSA system. For the flat panel system, the high beam filter used was 0.2-0.3 mm aluminum and was selected automatically by the system. The exposure parameters were similar to the routine UAE procedure (pelvic adult program in the DSA system and abdomen pelvic adult program in the flat panel system). The phantom and TLDs were exposed by 20 minutes fluoroscopy and 40 spot images. Then the TLD chips were removed and read by the TLD-reader system. This step was repeated for each location of the right and left ovaries and for the midpoint of the phantom and for both angiography systems separately to achieve the radiation attenuation function for each angiography system. The dose-depth relation was calculated by Excel program using data of ESD and different depth doses. We used the exponential curve fitting and formulation; the depth was considered as the independent variable and the measured dose was considered as the dependent variable. For all fittings, R
2 of model was calculated. The equations are as follows: For the DSA system;
Dx=D20 e-0.19x + 4.07 with R
2= 0.978 and for the flat panel system Dx=D20 e-0.09x+1.87 with R 2 = 0.971.
X=the depth of organ or tissue of interest from the posterior surface.
Dx = Absorbed Dose (mrad) at depth x
D20 = Absorbed Dose (mrad) at AP surface of patients or the phantom
3.2. In-Vivo Study
Thirty patients (fifteen patients for each system) who were referred for UAE to the medical imaging center affiliated to our university hospital and were approved as patients requiring UAE according to guidelines for the procedure (
11) were enrolled in this parallel designed clinical trial and were randomly assigned to each of the angiography systems. For sample size, considering type I and II statistical errors equal to 5% and 10%, respectively and conducting a pilot study, we calculated the sample size of 11 patients for each group and finally we considered 15 patients for each group. Using block randomization method the patients were assigned to the groups. Only the study coordinator was aware of the allocation sequence and assigned each patient to her group after enrollment to the study by the care provider physician. The physicist who performed dosimetry and the statistical analyzer was blinded to the patient’s group.
The study was approved by the Institutional Review Board and informed consent was obtained from each patient before the study. The additional costs of this project were funded by the research department of Tehran University of Medical Sciences.
Patient information, including age, height, weight, anterio-posterior thickness of the pelvis in the supine position, depth of the right and left ovaries from the posterior surface of the body (measured by MRI axial images) were filled in a data collection form before the study.
Before the study, we performed the quality control of our devices with KVP meter and milliampere meter. KVP were defined manually according to the antero-posterior (AP) thickness of the pelvis of each patient and the image view (AP or oblique view) that was in the range of 75-95 KV. The milliampere (mA) tube filament for fluoroscopy and milliampere-second (mAs) for each exposure of spot images were automatically set by the angiography system.
All patients underwent catheterization through a right femoral approach. Aortography was performed with a pigtail catheter before pelvic arteriography. The catheter was placed in the abdominal aorta at the level of the renal arteries and selective catheterization of the uterine arteries was performed.
Then a 4-French cobra-shaped catheter was positioned beyond the junction of the descending and horizontal portions of each uterine artery. Embolization was performed by injection of PVA particles under fluoroscopic control. To avoid retrograde reflux of the particles and infiltration to other internal iliac artery side branches, the injection was stopped when the arterial flow ceased. When an anastomosis was encountered between the uterine and ovarian arteries, the catheter tip was placed in a position distal to the anastomosis. Postembolization angiography was performed for evaluation of redistribution.
3.4. In-Vivo Dosimetry
As mentioned earlier, exposed dose to the ovaries was the main outcome measure of the study. Four dosimeter batches (each containing 3 TLD-100 chips) were used for each patient, two batches on the posterior surface ESD and two batches on the anterior surface of the pelvic region at the level of the right and left ovary. TLD batches were sealed by a waterproof and radiolucent cover and marked by a piece of wire and RA, LA, RP, LP markers (refers to right, left, anterior and posterior) to make their location detectable during the procedure (
Markers showing TLD location in the pelvic cavity
Configuration of the systems such as tube-table and patient to image receptor distance was set similar to what was used in the in-vitro study, although we did not interfere with the routine procedure and other parameters that were set by the internationist such as using magnifying and oblique views. The parameters of the procedure including KV and mA of fluoroscopy, frame rate, fluoroscopy time, KV and mAs and the number of spot images were filled during the procedure. After the procedure, TLD batches were removed and read by TLD-reader and the mean values of three TLD readings in each batch were used for calculating the radiation dose at their location. The depths of the ovaries from the posterior surface of the body were used individually for calculating the ovarian dose in each patient’s ovaries by using functions achieved in the in-vitro study (dose at the depth of interest). Since during some procedures, especially during magnifying and oblique views, the posterior TLD batches moved outside the radiation field making their results inaccurate, we had to use the result of anterior TLD-batches exit dose in the functions to calculate the depth dose.
3.5. Statistical Analysis
We used SPSS ver. 11.5 (SPSS Inc., Chicago, Il, USA) for statistical analysis. Comparison of the variables between the two groups was done by t-test or U-Mann Whitney after normality assessment of the data in the two groups. In addition, we compared the data of the two sides (right and left) considering the angiography system. All P values less than 0.05 were considered statistically significant.