Radiation exposure of patients during endourological procedures: IAEA-SEGUR study

Ionising radiation is an essential part of urological practice, used preoperatively for diagnosis, intraoperatively to guide treatment and postoperatively for follow-up. Endourology is a subspecialty area in urology using minimaly invasive techniques to treat renal stones and other abnormalities in the urinary tract with the help of intracoreal scopes and instrumentation. Percutaneous nephrolithotripsy (PCNL) and ureterorenoscopy (URS) are the two main procedures performed worldwide to treat urolithiasis (stones in the urinary tract). During PCNL, the operating physician gains access to the renal tract through a percutaneous needle puncture, under the fluoroscopic or ultrasound guidance. After successful entrance, the percutaneous tract is dilated and the stone is comminuted with instruments using laser or mechanical energy. During URS, physician enters the urinary tract through the urethra and after identifying ureteral orifice, several instruments and guidewires are used under fluoroscopic guidance to reach the level of stone/pathology and manage it accordingly, usually with laser energy [1]. Fluoroscopy, along with ultrasound, is used to guide safe access to kidneys and ureters, while it accounts for about 18% of all fluoroscopically guided interventional procedures [2]. The advantages of endourology procedures over open surgery include a minimally invasive management option, the elimination of prolonged pain after surgery, and faster recovery [1]. Despite its benefits, the use of ionising radiation poses health risks for patients and operative room staff, thus knowledge should be enriched in order to establish optimised protocols.

During a fluoroscopy guided procedures, both operating team and patients are exposed to radiation that may cause two types of health effects: deterministic (tissue reactions) and stochastic (cancer or hereditary effects) [3, 4]. To limit the radiation exposure to an acceptable level for occupational exposure of staff working with ionising radiation, the International Commission for Radiological Protection (ICRP) and corresponding international safety standards prescribe annual dose limits, and actions to keep the exposure as low as reasanably achievable [3, 5, 6]. At the same time, because radiation is used intentionally for the patient direct benefit, applying dose limits for medical exposure of patients is considered inappropriate. Instead, the efforts are to limit unnecessary exposure by applying radiation protection principles of justification and optimisation [3, 4, 6]. Patient dose assessment and comparison of typical dose values between departments performing the same diagnostic or interventional procedure are considered to be an important benchmarking tool and a basis for optimisation of radiation protection of patients [3, 5 – 7]. Typical dose value is defined as the median of the distribution of the data for a commonly and easily measured or determined modality-specific radiation dose metric that is obtained from a local survey or a review of local data in a radiological facility [5, 7]. For fluoroscopy guided procedures, such a metric includes the kerma area product and the air kerma at the patient entrance reference point [7].

The amount of radiation reaching patient during a fluoroscopically-guided procedure depends on the total fluoroscopy time (FT), exposure factors (tube voltage, tube current and beam filtration), fluoroscopy mode (combination of settings of the automatic dose rate control (ADRC) system), x-ray beam geometry, including beam colimation, and the number (N) of the acquired radiographic images and their exposure settings [2]. Practicing urologist is called to increase benefit to risk ratio by minimising radiation dose to patient when appropriate, while achieving the best result regarding treatment of underlying health condition. Another important goal is to avoid tissue reactions, namely skin injuries for patients by keeping dose to the patient skin below the dose threshold for such as a tissue reaction [2, 4]. Although observed in other complex fluoroscopy guided interventional procedures such as in cardiology or neuroradiology, no skin injuries have been reported in the scientific literature from endourology [2]. Minimising radiation exposure to patient contributes to reducing the exposure to surgical team members, which can be further lowered by using personal and in-room protective shielding and increasing distance from patient to the extend possible [2, 8].

It is well known that patient doses from fluoroscopy guided procedures vary significantly depending on the type of procedure, its complexity, the fluoroscopy equipment specification and its operation. Data regarding patient radiation exposure in endourology are scant and come from a single center [2, 9 – 13]. Limited data of patient exposure levels are reported in few publications, which main focus is staff exposure [14 – 20]. To fill the gap and study the potential for optimisation of patient exposure during endourology procedures, the South-Eastern European Group for Urolithiasis Research (SEGUR), in cooperation with the International Atomic Energy Agency (IAEA), conducted a multi-center study in seven South-Eastern European hospitals. The aim was to investigate radiation exposure of patients during fluoroscopy guided endourological procedures in correlation with affecting factors. In parallel, doses to the endourology staff were measured and analysed, the results of this part of the study were presented in a separate paper [21].

2.1. Study design

2.2. Data collection and analyses

The responses received from the endourology centers are presented with their unique ID number, U1–U7. Centers U1, U2, U4, U6, U7 used mobile C-arm fluoroscopy systems, two of Siemens Healthcare, Germany, one of General Electric, USA, two of General Medical Merate, Italy, one of Gilardoni, Italy. U3 used a fixed C-arm (Lithoskop, Siemens Healthcare), and U5 a fixed fluoroscopy system with over-couch x-ray tube (Uroskop, Siemens Healthcare). All with C-arms declared to place the x-ray tube under the table. System U1 had flat panel detector while others used image intensifier. All systems were equipped with ADRC, digital acquisition mode and LIH function. All except U4 and U6 used pulsed fluoroscopy with either 12.5, 25/30 p s^-1. Except U7, all had dose readings in P_KA displayed after the procedure, but display of K_a,r was available in only systems U1–U3 and U5. Dose displays were verified by independent measurements in U4, U5 and U6 and the reported results were found to be within the tolerable limits, while in U1, U2 and U3 no verification was performed due to the lack of access to qualified medical physicist and unadequate quality control procedures No dose data were provided from the center U7 that contributed with a limited data for one procedure only.

The annual workload differed significantly between centers, the highest in U6 with 865 procedures in 2017, followed by U1 with 750 and U5 with 707 procedures, U3 with 450 procedures and much lower in U2 and U4, with correspondingly 185 and 173 procedures.

Procedures of 315 patients in total were recorded; their distribution per type and center is presented in table 1 for PCNL (standard and mini) and RIRS and in table 2 for URS (semi-rigid and flexible).

Majority of patients were treated with semi-rigid URS (40.0%), followed by standard PCNL (32.1%), RIRS (13.7%), mini-PCNL (9.5%) and flexible URS (5.1%). Different case-mix was observed: centers U2, U3 and U7 performed mostly PCNL, while U4 performed more URS and U5 and U6 had the whole spectrum. For URS, U1, U4 and U5 used semi-rigid scope, while U2 and U3 performed only flexible URS, and U6 had 44.4% semi-rigid and 22.2% flexible. All centers performed PCNL in prone position, except U4 routinely using supine.

Tables 1 and 2 present the patient weight, stone location, stone SA, as well as the total procedure time (PT), FT, N, P_KA, K_a,r, and fluoroscopy pulse rate (p/s), per procedure and per clinical center. Data are presented with their mean and medical values, and the range (minimum–maximum).

The stone SA varied from 13 to 1570 mm² in ureter and from 16 to 6005 mm² in kidney (tables 1 and 2). The largest kidney stones were treated with standard PCNL (101 patients), followed by mini-PCNL (30 patients) and RIRS (43 patients). For ureteral stones, semi-rigid URS was used in majority of cases (125) versus flexible URS (16).

All patients were adult, 214 (68%) males and 101 (32%) females. All samples had a mean patient weight within 80 ± 5 kg, except two, in which the mean weight was 73.5 kg and 96 kg correspondingly. This indicates that the patient weight might not be a factor when comparing dose values between centers.

The correlation coefficient r² between the stone SA and the total PT varied between 0.001 for the semi-rigid URS in U1 (26 patients) and 0.409 for the standard PCNL in U5 (36 patients). The correlation coefficient r² between SA and FT was between 0.001 for RIRS in U5 (29 patients) and 0.17 for the standard PCNL in U1 (10 patients). The correlation coefficient r² between P_KA and FT varied from 0.63 to 0.80 for the standard PCNL performed in centers U1, U2 and U5, from 0.75 to 0.90 for mini PCNL performed in U1, U2 and U5, from 0.20 to 0.99 for RIRS performed in U1 and U5, and from 0.08 to 0.52 for both semi-rigid and flexible URS performed in U1, U4, U5 and U6.

Table 3 presents the results for all type of procedures of the nonparametric ANOVA using the Kruskal–Wallis test of the significance of the difference between values of FT and P_KA. A p-value less than 0.05 indicates that there is a significant difference somewhere among the various samples. The difference between FT values for the same procedure performed in different centers was statistically significant for the standard PCNL (p = 0.0125) and semi-rigid URS (p = 0.0226), while no significant difference was observed for the other procedures with less statistical data like mini PCNL (p =0.4894), RIRS (p = 0.7477) and flexible URS (p = 0.1896). The difference between P_KA values for the same procedure performed in different centers was statistically significant for all procedures (table 3, last column).

Figures 1 and 2 show the median values of P_KA and K_a,r , for all procedures performed in different clinical centers. Medians are selected for this presentation because of the non-normal distribution of some samples.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Table 4 presents the findings of other published studies in terms of P_KA and FT for PCNL and URS.

Urolithiasis is a common health issue affecting about 13% of male and 7% of female population and represents a major cause of emergency department visits and a great economic health burden [24]. Minimally invasive procedures rapidly replace open surgery and become the most effective treatment option even for larger stones in anatomically difficult systems. Safe entrance into urinary tract and fragmentation-removal of stones is accomplished with fluoroscopic guidance. To our best knowledge, this is the first multicentric study of radiation exposure associated with endourological procedures and related factors.

The total PT varied from 10 to 225 min, with the higher values for PCNL (12–225 min) and lower for URS (10–151 min). Weak or no correlation was found between the stone SA and the total PT (r² from 0.001 to maximum 0.4), indicating that there are other factors affecting the procedure duration, such as the stone composition and location.

Fluoroscopy was used in a small part of the procedure; FT varied from 0.1 min to 14 min for different patients, with the maximum for standard PCNL. FT was similar to data reported in other studies (table 4), except the study of Vano et al where FT for PCNL was higher [9 – 11, 13, 17 – 20].

Median FT, which better represents the typical value in samples with non-normal distribution, varied 5-fold (from 1.1 to 5.4 min) for the standard PCNL, and 6-fold (from 0.1 to 0.6 min) for the semi-rigid URS (tables 1 and 2). This might be influenced by the different number of cases in each center, but considering that for both procedures the Kruskal–Wallis test found statistical differences of FT between different centers, variations might be mostly related to a different practice of use of fluoroscopy and different complexity of the procedures due to the stone size, structure and location.

No correlation was found between stone SA and FT (the maximum r² was 0.17), showing a potential impact of confounding factors, also discussed by Safak et al [13]. Specifically, the lithotripsy phase does not require fluoroscopy guidance and exposure depends on body habitus and morphology of the urinary tract that can make more difficult the percutaneous access, or the passage of the scope requiring repeated fluoroscopic checks. For example, the largest kidney stone with a SA of 6015 mm² in U5 was treated with the standard PCNL that needed 120 min total procedural time, of which only in 2.1 min fluoroscopy was used. Much smaller kidney stone of SA of 236 mm² required 7.5 min FT due to the more difficult access to the kidney and only 60 min to finish the procedure because of the easier lithotripsy phase.

High variations were found in the number of acquired images—from zero in U4 and U6, up to 156 for standard PCNL and 89 for mini-PCNL in U2, 180 for RIRS and 10 for semi rigid URS in U1 and 136 for flexible URS in U3. It is well known that the use of radiography increases patient dose, and whenever available, LIH has to be used instead [2].

The differences in FT, N and system settings resulted in significant variations of doses between individual patients and between centers. The lowest individual P_KA of 0.01 Gy cm² was recorded for a semi rigid URS case, and the maximum 72.30 Gy cm² for a standard PCNL case. The values of K_a,r varied between < 1 mGy in short RIRS and semi-rigid URS, to 377 mGy for PCNL. The highest P_KA and K_a,r values were observed for the same patient case in U2, treating with PCNL by multiple access a kidney stone with SA of 942 mm², that resulted in the highest FT of 14 min and 156 radiography images. In addition, pulsed fluoroscopy with 30 p s^-1 was routinely used in this center, indicating a potential for dose reduction if lowering pulse rate and reducing the number of acquired radiography images.

Median P_KA (tables 1 and 2, and figure 1) varied 2 fold for RIRS (0.80–1.79 Gy cm²), 7.1 fold for mini-PCNL (1.39–9.90 Gy cm²), 7.3 fold for PCNL (2.40–17.50 Gy cm²), 19 fold (0.13–2.51 Gy cm²) for semi-rigid URS and 29 fold for flexible URS (0.10–2.90 Gy cm²). The difference between median P_KA for the same procedure in different centers was statistically significant for all procedure types (table 3). Higher median P_KA (from 1.4 to 17.5 Gy cm²) were observed for the PCNL procedures compared to the URS procedures (from 0.1 to 2.9 Gy cm²).

Good correlation was found in most departments and procedures between P_KA and FT (r² between 0.5 and 0.9 in most samples) that is reasonable since FT is the main factor affecting patient dose along with N and pulse rate. For example, in U5, for 71% of cases 12.5 p s^-1 was used and 25 p s^-1 in the rest, which explains the relatively lower doses compared to U2 where 30 p s^-1 was always used for all procedures, and U4 and U6 where continuous fluoroscopy (CF) is used. Much higher values of P_KA for PCNL in U4 might be also due to the use of supine position.

The effect of pulse rate was studied further for the center U5 by comparing two normally distributed samples of PCNL with the same mean FT = 2.5 min performed with different pulse rate. The use of 12.5 p s^-1 resulted in a significantly lower P_KA (mean = 1.78 Gy cm², SD = 1.17 Gy cm²) compared to the use of 25 p s^-1 (mean = 3.25 Gy cm², SD = 1.34 Gy cm²), t(33) = −3.5, p = 0.002. A similar significant decrease was found for K_a,r from PCNL performed with 12.5 p s^-1 (mean = 25.0 mGy, SD = 18.3 mGy) compared to the use of 25 p s^-1 (mean = 42.2 mGy, SD = 23.4 mGy), t(33) = −2.41, p = 0.022.

As reported in the first paper from this study [21], which presents the measured doses to the whole body and eye lens of the operating room staff, a strong correlation (r² between 0.91 and 0.99) was found between the measured eye lens dose and the total P_KA of procedures performed by the corresponding urologist. This confirms that the optimal use of fluoroscopy has a positive impact on not only patient dose, but staff dose as well. As an example, although the fluoroscopy system in U5 has an x-ray tube over the table, because of the use of lower FT, N, pulsed fluoroscopy of 12.5 p s^-1 and good collimation, doses to eye lens of staff and patient doses per procedure were lower in this center compared to U4 and U6 [21].

Although not studied separately, some higher P_KA values can be associated also with an inappropriate beam collimation—when keeping other exposure factors unchanged, decreasing the field area linearly decreases P_KA. Additional factors not studied are the distance of x-ray tube and image detector relative to patient, as well as the tube voltage and tube filtration defined by ADRC [2].

The maximum K_a,r value of 377 mGy for a patient treated with PCNL in U2 is much lower than the dose threshold of around 2 Gy that is needed to observe radiation induced skin erythema [2, 4], confirming the findings of Safak et al [13]. The typical (median) values of K_a,r varied between 0.6 mGy and 51.7 mGy.

P_KA and K_a,r values for RIRS were significantly lower compared to PCNL performed in the same centers. Considering also its higher safety and repeatability, RIRS is a valuable option for renal stone treatment [1, 24].

There are few published studies of patient doses in endourology (table 4). Median P_KA for PCNL in our study ranges between 2.40 and 17.50 Gy cm², which is comparable to the findings of Hristova-Popova et al [9, 17] and Medici et al [18], and lower than the value reported by Vano et al who also observed higher FT [20]. For URS, median P_KA is comparable to the values reported by Hristova-Popova et al [9, 17] and lower compared to Medici et al [18].

The study demonstrated variations in the endourology practice between urology centers, that results in high variations in patient doses, as well as variation in staff doses [21]. This is mainly related to the use of different fluoroscopy systems (C-arm with x-ray tube under the table and fixed system with tube over the table), use of CF versus pulsed fluoroscopy with different pulse rate, different patient positioning for PCNL (supine or prone), and different pattern of recording radiography images.

No direct correlation was found between the concrement size and total duration of procedure, FT or dose parameters kerma area product and air kerma at the patient entrance reference point that indicates that there are other concurring factors, e.g. patient condition, concernment structure, stone position, etc, which have not been studied here. Further studies might be needed to find a better descriptor of the complexity of the endouroly procedures and link it to dose data.

Skin doses were significantly lower the threshold and no skin injury can be expected for patients in endourology.

The study showed potential for patient dose reduction by paying attention to the following factors: use of low-dose pulsed fluoroscopy (12.5 p s^-1 or lower when available), better beam collimation to the area of interest, no or less use of electronic magnification, reduction of FT and number of recorded images and use of LIH instead of radiography whenever possible. Additional reduction can be achieved by keeping the x-ray tube further and image detector closer to the patient body [2, 25]. Endourologists must understand and optimally use these factors to reduce simultaneously patient and staff exposure and optimise the benefit-risk ratio. Radiation protection needs to be an integral part of the training programs of endourologists, with an emphasis on the practical dose management and optimisation [2, 26]. The study demonstrated the need to further strentgen quality control program and involve qualified medical physicist to advise on dose reduction approaches.

The study was organised by the South-Eastern European Group for Urolithiasis Research (SEGUR) and was supported under the IAEA Technical cooperation projects RER9135 'Strengthening Radiation Protection of Patients and Medical Exposure Control'. The authors acknowledge the contribution of Ognyan Gatsev, Stefan Hristoforov and Petar Petrov, all from the Department of Urology and Nephrology, Military Medical Academy, Sofia, Bulgaria.