Radiation metrics for vascular and interventional radiology procedures in a tertiary care institution

Objectives: To evaluate the radiation metrics from frequently carried out vascular and interventional radiology (VIR) procedures at a tertiary care institution and compare them to international diagnostic reference levels (DRLs). Methods: A retrospective study of the radiation metrics of VIR-procedures carried out from January 2015 to December 2019. The collected data included age, gender, height and weight, reference point air kerma (mGy), dose area product (DAP; Gy.cm2), and ﬂuoroscopy time (min.) The body mass index (BMI) and peak skin dose were calculated. The study cohort included 8942 adult patients (54.4% male, 45.6% female) with a mean age of 56.96 years and mean BMI of 26.86. Results: Transjugular intrahepatic portosystemic shunt (TIPS) creation recorded the highest mean ﬂuoroscopy time of 69.41 min., followed by central venous recanalization 39.67min. TIPS creation had the highest mean DAP (1161.16 Gy.cm2), followed by trans arterial chemoembolisation (TACE) (500.63Gy.cm2). TIPS creation was associated with the highest peak skin dose (2766.81mGy), followed by TACE (1588.29mGy). Compared to other studies, TIPS creation and TACE are associated with significantly higher DAP. Conclusion: Majority of VIR-procedures demonstrate no significant institutional variations in dosimetry compared to other studies. Using these studied values as reference levels may help identifying procedures that need quality control to minimize unnecessary exposures.

F luoroscopy is increasingly used to guide minimally invasive diagnostic or therapeutic procedures. This exposure to ionizing radiation has inherent potential deterministic or stochastic effects on patients and operators. 1 Dose optimization can be met effectively by the concept of diagnostic reference levels )DRLs(, which was introduced by the International Commission for Radiation Protection )ICRP( as practice guidelines and advised to identify these values in each country or region. 2 It has been mandatory since 1997 in the European Union, in addition to the new European directive for 2013/59/Euratom. 3 Diagnostic reference level is defined as radiation dose levels for typical X-ray examinations for standard-sized patients using standard equipment. National DRLs are set at 75 th percentile of the median values of an appropriate quantity in a representative sample of healthcare facilities. Exceeding the DRL for a certain procedure should trigger an investigation of the equipment, procedure protocol, or the operator's technique. Applying DRLs to interventional procedures is challenging due to multiple factors, including the patient's size, the complexity of the procedure, and the operator's experience. Active dosimetry is not only an audit tool, but also optimizes occupational protection for each procedure. The ICRP recommends gathering regional or national dosimetric data for every case of a procedure from a large number of facilities, or alternatively, use a local data set of the dosimetric data of every case of the same procedure carried out at the local facility. 2,4 Apart from interventional cardiology, there is limited DRLs data for body and neuro interventional procedures, and a paucity of DRLs for pediatric interventional procedures. 4,5 Recently, the Saudi Food and Drug Authority and Saudi Health Council issued national DRLs for a computed tomography )CT(, based on dosimetric data sets from 10 different hospitals in Saudi Arabia. More recently, a local study reported that the radiation dose levels for uterine artery embolization procedures and compared it to published data. 6 The aim of this study is to participate in establishing the national DRL, reporting the dosimetric data of the most frequently carried out procedures in a tertiary care center and compare it to the reported international data from current literature, the radiation doses in interventional radiology procedure )RAD-IR( study, a radiation doses measurements study carried out for DRLs in 2003, and a recently published study with similar procedure categories to our sample. 7,8 Methods. This retrospective review included adult vascular and interventional radiology )VIR( procedures carried out over a period of 5 years from January 2015 to December 2019 at the Department of Medical Imaging, King Abdulaziz Medical City, National Guard Health Affairs, Riyadh, Kingdom of Saudi Arabia. The radiation doses were captured using the dose management software )DoseWatch; GE Healthcare System, Buc, France( that is integrated in the Radiology Information System. The anthropometric data included the height, weight, and body mass index )BMI( in addition to the patients' age )>14 years( and gender. The dosimetric data included the reference point air kerma )mGy( )known as reference dose, cumulative dose, or cumulative air kerma(, dose area product )DAP( )Gy. cm 2 ( )known as Kerma area product [KAP](, and fluoroscopy time )minute(. The peak skin dose )PSD( was calculated using the reference point air kerma and the following formula: PSD )mGy(=206+0.513*Ka,r [mGy](. 9 During the study period, 18068 fluoroscopy-guided interventions were carried out by several consultant interventional radiologists or supervised residents/ fellows in training. This analysis did not include neurointerventional )n=972( and pediatric )n=2250( procedures, which will be evaluated separately. Duplicate procedures accessions, procedures with missing dosimetric or demographic data were excluded. Low count procedures )<15( were also excluded, including procedures such as prostate artery embolization, prophylactic bilateral internal iliac artery occlusion, and pulmonary embolectomy ) Figure 1(. This study protocol was approved by the Institutional Review Board at King Abdullah International Medical Research Center. The study was carried out in accordance with the principles of the Helsinki Declaration.
Procedures of a similar nature were grouped in one category. For example, the tunneled catheter placement category included new tunneled hemodialysis catheter insertion, in addition to exchange +/-simple fibrin sheath angioplasty, chest port placement, and chemotherapy tunneled catheter insertion. Axial embolization procedures included embolization in the thorax, abdomen or pelvis, and could vary between simple embolization procedures, such as posttraumatic bleeding to a complex embolization of arteriovenous malformation. Arteriovenous dialysis access interventions vary between simple angioplasty to thrombectomy with balloon angioplasty +/-stent Disclosure. Authors have no conflict of interests, and the work was not supported or funded by any drug company. placement. Hepatic mapping may include vessel embolization procedures that were not counted under axial embolization. Post liver transplant vascular interventions included hepatic artery, hepatic or portal veins interventions with thrombolysis, and balloon angioplasty with or without stent placement.
All procedures were carried out on one of 5 angiography machines )AlluraXper, Philips Medical systems, Netherlands(. All machines are regularly reviewed by the radiation safety officer, and all staff attend a mandatory annual radiation safety course.
The final analysis included 8942 procedures with 4863 )54.4%( male patients and 4079 )45.6%( female patients, with a mean age of 56.96 years )range: 15-117(. The mean height was 1.6 meters )range: 1.2-1.91(. The mean weight was 68.9 kg )range: 16-214(. The analysis included 22 different VIR-procedures. Peripherally inserted central catheter placement was the most frequently carried out procedure )n=2712, 30.3%(, followed by tunneled catheter placement )n=2151, 24.1%(, gastrostomy )n=1000, 11.2%(, and arteriovenous dialysis access interventions )n=455, 5.1%(. Statistical analysis. The categorical data are presented as percentage and frequency, and the continuous variables as mean, median, standard deviation, 25 th and 75 th percentile, and range. Two independent sample t-test was used to compare the mean values from the current study with the RAD-IR study and Contemporary Interventional Radiology Dosimetry )CIRD( study. 7,8 Statistical significance was defined as a p-value of <0.05. The statistical analysis was carried out on Statistical Analysis System, version 9.4 )Cary, NC,  Compared to the CIRD and RAD-IR studies, vascular access procedures had comparable fluoroscopy time, DAP, and reference point air kerma. However, TIPS-creation and trans arterial chemoembolisation )TACE( were associated with significantly higher fluoroscopy time, DAP, and reference point air kerma in our institution. The mean fluoroscopy time for nephrostomy and gastrostomy procedures was significantly shorter in this study compared to the reported values. Dose area product was also significantly lower for nephrostomy, inferior vena cava filter placement, and gastrostomy procedures. Other differences and variations in radiation exposure parameters are illustrated in Table 5.
There was very strong positive correlation between reference point air kerma and DAP )R=0.93; p<0.0001(. There was moderate positive correlation between fluoroscopy time and DAP )R=0.60; p<0.0001(, reference point air kerma )R=0.66; p<0.0001(, and PSD )R=0.60; p<0.0001(. There was a very weak positive correlation between BMI with DAP )R=0.08; p<0.0001(, reference point air kerma )R=0.08; p<0.0001(, and fluoroscopy time )R=0.04; p<0.0001(.     Interventional Radiology Dosimetry, RAD-IR: radiation doses in interventional radiology procedure N: number, SD: standard deviation mean reference levels compared to the CIRD and RAD-IR studies. 7,8,13 This may be explained by the low volume and higher complexity of TIPS procedures in our training institution. The TACE procedures were carried out as super selective as possible, and frequently mandated a cone-beam CT for lesion localization, which could have contributed to the additional exposure. The large standard deviation for DRLs of TIPS and TACE indicates the significant variability in complexity and required radiation exposure.
Comparing our data with the local study for uterine artery embolization )UAE(, the median DAP value was higher at 347 Gy.cm2 for the UAE procedure, compared to our median DAP at 156.14 Gy.cm 2 . 6 The SIR guidelines suggested DRL thresholds for intra procedural notifications and set the following thresholds to prompt patient follow-up after VIRprocedures: a PSD of 3000 mGy, reference point air kerma of 5000 mGy, KAP of 500 Gy.cm 2 , or a fluoroscopy time more than 60 minutes. 9 Although these values are not restrictive and should not be regarded as endpoints for procedure termination, the currently reported values may be used by peer institutions and authorities with a similar population, and may be incorporated in quality improvement measures and dose management software.
Study limitations. This study is limited to an adult population, excluding pediatric patients and neuro interventions, unlike the CIRD and RAD-IR studies. 7,8,13 These will be reported separately given the inherent differences in interventions and patients as well as the operators. Although there was a very weak positive correlation between the BMI and the reported DRLs, the adult population was largely overweight with a mean BMI of 27 kg/m 2 , which requires additional subgroup analysis to determine the correlation between BMI and DRLs for individual procedures. This study also reports the PSD, which is regarded as the main tissue of concern in interventional procedures and can be used as a threshold to trigger patient follow-up.
In conclusion, the majority of these procedures demonstrate no significant institutional variation in dosimetry, when compared to other studies. However, optimization of radiation precautions is paramount to maintain the exposure as low as reasonably achievable, particularly in TIPS-creation and embolization procedures that are associated with the highest radiation exposure.