Interventional Cardiology-Eight Years of Practice: Do Increased Experience and Technological Evolution Lead to Undertaking More Difficult Cases and Higher Patient Doses?

Several studies have been published, in the field of Interventional Cardiology (IC) procedures, focused on factors that may influence patient radiation exposure [1-10]. Radiation involves a probability of carcinogenic effects (the risk of stochastic effects), the actual probability increasing with the magnitude of the dose. Effective dose (E) is a dose quantity used to roughly estimate the radiogenic risk to an individual (radiation induced cancer is an example). The stochastic risk to an average member of an irradiated population is expressed in terms of sieverts (Sv). Effective dose also enables medical examinations or techniques to be compared in terms of radiation dose.


Introduction
Several studies have been published, in the field of Interventional Cardiology (IC) procedures, focused on factors that may influence patient radiation exposure [1][2][3][4][5][6][7][8][9][10]. Radiation involves a probability of carcinogenic effects (the risk of stochastic effects), the actual probability increasing with the magnitude of the dose. Effective dose (E) is a dose quantity used to roughly estimate the radiogenic risk to an individual (radiation induced cancer is an example). The stochastic risk to an average member of an irradiated population is expressed in terms of sieverts (Sv). Effective dose also enables medical examinations or techniques to be compared in terms of radiation dose.
Radiation also involves detrimental health effect (i.e., skin injury, hair loss, cataracts) for which the severity increases with the dose of radiation. The effect is not observed unless the threshold is exceeded, although the threshold dose is subject to biologic variation [11]. In interventional fluoroscopy procedures, the tissue of concern is the skin, although the lens of the eye is another consideration. The skin at the site where radiation enters the body receives the highest radiation dose of any body tissue. Once the threshold dose is exceeded, the injury becomes progressively more severe with increasing dose, although the true severity of major injuries will only become apparent weeks to months after the procedure. Peak Skin Dose (PSD) is a dose quantity estimating the highest dose at any portion of a patient's skin during a procedure. Peak skin dose includes contributions from both the primary X-ray beam and from scatter. Peak skin dose is measured in grays (to soft tissue).
The prime objective of radiation protection is not only to minimize the stochastic risks but also to avoid deterministic injuries, by keeping radiation dose as low as reasonably achievable. However, the potential risk of patient radiation damage must be viewed in the context of the general benefit of these procedures and the likelihood of greater trauma associated with heart surgical interventions and perhaps with imminent death if the intervention was not performed. Radiation dose associated with IC procedures is a complex function of a large number of variables including the X-ray tube voltage (kVp) and the product of tube current and exposure time (mAs), the X-ray field size, the number of projections, focus-skin distance as well as practical skill of the operator [12]. Therefore, for optimizing radiation exposure in such procedures, it is essential to measure patient doses in a given set-up of cardiac catheterization laboratory. During IC procedures, both PSD and E need to be determined. However, neither PSD nor E is measured on a routine basis, because the patient dose monitors commonly available to the operator are kerma-area product (P KA ), fluoroscopy time (FT) and reference air-kerma (K a,r ), which can be recorded in real time from the control console of the IC equipment. The kerma-area product, P KA , is defined as the integral of dose across the X-ray beam (Gy•cm 2 ) and provides information about patient dose as well as the irradiated skin area. It is a unit historically known as dose-area product (DAP) and currently named kerma-area product (KAP, current notation P KA ). P KA provides a good index for estimating stochastic risk

Materials and Methods
The IC procedures were performed by two interventional cardiologists with more than 10 years of experience. Coronary Angiography (CA) and Percutaneous Coronary Intervention (PCI) were included in the sample. The procedures were carried out with a Philips Integris Allura Xper FD20 fully digital monoplane machine with flat detector (FD) (Philips Medical Systems, Best, The Netherlands) in a C-arm configuration.
The radiation dose for each patient was provided by the X-ray system in terms of kerma-area product (P KA ) and reference point air kerma (K a,r ). Patient radiation metrics in terms of patients age, kermaarea product (P KA ) in Gy•cm 2 , reference air-kerma (K a,r ) in mGy, fluoroscopy time (FT) in min and total number of frames (F) were analyzed from a pool of data ranging from January 2009 to May 2016.

Conversion factors and equations
Effective dose: Effective dose (E) is estimated by adding the products of the dose in an organ or tissue and the specific weighting factor for that tissue. The weighting factors are values that express the sensitivity of each particular tissue or organ to radiation. Each weighting factor relates to the risk associated with stochastic effects and has specific value for every organ. E is expressed in millisieverts (mSv). In practice, it is very difficult to determine E, since the radiation doses in 12 organs would have to be measured during each cardiological procedure. Therefore, the use of a special conversion factor provides a practical way to estimate E. Modern fluoroscopy machines calculate kerma-area product P KA using generator and collimator settings. P KA does not depend on the distance of the measuring plane from the X-ray source because dose decreases according to the inverse square law and the area of the field increases with the square of the distance. This keeps the P KA value constant at any distance. P KA represents the total energy incident on the patent. P KA is combined with a coefficient depending on the irradiated portion of the body and protocol (irradiated organs) to estimate E. The coefficients range from 0.028 to 0.29 (mSv/Gy•cm 2 ) depending on type of X-ray procedure with regards to the beam geometry and beam quality. They are derived from Monte-Carlo simulations using anthropomorphic digital phantoms. For coronary angiographies, the most recent studies determine E from P KA using a conversion factor of 0.185 mSv/Gy•cm 2 for normal adult patient, although the precise factor value depended on whether an additional copper filtration was available during cine mode [22].

Peak skin dose:
The estimation of absorbed dose at the surface of the skin by fluoroscopy, to account for deterministic effects, is a difficult task since the examination is conducted with changing regularly the incidence X-ray beam. However, it is possible to have an idea of the dose to the skin using a particular operational quantity, called the cumulative dose. This quantity estimates the dose that would have but is not directly useful for estimating tissue reactions. It is also a wellaccepted dosimetric parameter for comparison between patient doses during different interventional procedures and is also used to assess E [13]. Air Kerma is the energy extracted from an X-ray beam per unit mass of air in a small irradiated air volume. Air kerma is measured in grays. The reference air-kerma (K a,r ) is the air kerma accumulated at a reference point in space which is located along the central ray of the X-ray beam of the C-arm fluoroscopic system at a distance of 15 cm away from the isocenter toward the X-ray tube focal spot and is provided by the manufacturer. K a,r was introduced by the International Electrotechnical Commission for standardization purposes and is similar to the Cumulative Dose (CD) used in the past [14]. Reference point air kerma does not include backscatter and is measured in grays (Gy). Reference point air kerma is sometimes referred to as reference dose, or cumulative air kerma. FT is required by the US Food and Drug Administration on all fluoroscopy equipment manufactured. However, it lacks information regarding X-ray beam area, thickness of patient and technique employed [15].
It is desirable to measure the PSD to map skin doses in a given IC set-up. However, this does not reflect the entire dose delivered during the course of the procedure. So, the information about the peak as well as overall dose to the patient during IC procedure can be obtained in terms of PSD as well as other dose metrics such as P KA , FT and K a,r . Each patient record should contain P KA and/or reference point air kerma (K a,r ).
The number of fluoroscopy-guided procedures in cardiology is increasing over time, a fact that led European countries to investigate closely this field and to gather information in the attempt to define Dose Reference Levels (DRLs) [16][17][18][19]. The concept of DRLs was first introduced by the International Commission on Radiological Protection in Publication 60 [20] and further expanded in Publication 73 [21]. The results of patient dose surveys could then be compared with the corresponding DRL value, to discover which IC departments have doses above the reference values. An audit process could then be initiated to determine the underlying cause of higher doses and an action plan developed to improve radiological techniques for dose reduction purposes. In practice, DRL values may be regarded as an optimization tool for the reduction of patient doses. In order to establish DRLs, the third quartile of value from survey data distribution is calculated. The median values of the distribution must also be assessed in order to estimate the variation in values.
The purpose of this study was to evaluate patient radiation doses during coronary angiography (CA) and Percutaneous Coronary Intervention (PCI) procedures in a dedicated catheterization laboratory and to analyze annual patient radiation data performed the last 8 years, investigate trends and compare with national and international literature and DRL values. The dosimetry approach was to determine patient doses in terms of three dose metrics: P KA , FT and K a,r . The PSD was calculated to take into consideration the likelihood of deterministic skin effects and to isolate procedures involving PSDs higher than 2 Gy. Radiation induced skin effects are deterministic in nature, with a generally accepted threshold dose of 2 Gy [11]. A Belgian study has proposed two P KA action levels for skin dose [13,19,22]. A first P KA action level of 125 Gy•cm 2 corresponds to 2 Gy which is the threshold dose for erythema. A P KA value >125 Gy•cm 2 would imply an optional radio pathological follow-up depending on the cardiologist's decision. The second action level of 250 Gy•cm 2 corresponds to 3 Gy skin dose and would imply a systematic follow-up. To cover the stochastic risk associated with IC procedures, the effective dose to patients was also evaluated by using proposed factors for conversion of P KA to E [22]. received the skin if the geometry was kept unchanged throughout the procedure. It may be considered that the cumulative dose indicated by the facility in cardiology in many (but not all cases) overestimates the dose to the skin by a factor of 2-3 since several tube-detector incidences are used that distributes the exposure of the skin. If this operational quantity is not available, the skin dose could be estimated from the P KA [25]. So, kerma-area product P KA correlates with peak skin dose (PSD) as it expresses the total X-ray flux in the beam. Reference point air kerma (K a,r ) correlates better with peak skin dose (PSD) and possible radiation injuries, as this quantity indicates the cumulative air kerma at the reference point during the procedure. PSD has been implemented as a function of the highest radiation dose at any point of the patient's skin to stand for the likelihood and severity of radiation-induced skin injury to the patient. K a,r is clinically useful as a real-time safety indicator. It does not take into account the scattered radiation and, therefore, is not the actual dose to the patient's skin. For this reason, the latest guidelines of the Society of Interventional Radiology (SIR) Safety and Health Committee presented a conversion formula to help operators estimate the PSD from the K a,r shown on the X-ray machine at the end of interventional procedure [11]: PSD = 206 + 0.513 × K a,r (mGy) (for K a,r >500 mGy) It should be underlined that these are broad estimations that on the other hand provide an immediate estimate of patient PSD easily and quickly in every day routine practice. Of course, accurate estimation of PSD must account for gantry motion, patient size and patient location relative to the gantry but is not possible without dedicated expensive software purchased additionally to the angiography machine.

Results
Our data resulted in 8286 IC procedures (diagnostic and therapeutic interventions) within the years 2009-2016. There were 5469 (66%) diagnostic Coronary Angiographies (CA) and 2817 (34%) Percutaneous Coronary Interventions (PCI). The range of kVp used in these procedures was 50-125 kVp. Patient demographic data collected included age (63±12, 20-92 years), weight (84±18 kgr) and height (168±15 cm). The patient exposure-related parameters recorded for each IC procedure were number of images F, FT, K a,r and P KA , as they are presented in Table 1.
For CA procedures, median values of P KA , K a,r, FT, and F ranged from 24.5-30.5 Gy•cm 2 , 365-447 mGy, 2.3-3.6 min and 436-514 respectively. The P KA, K a,r and FT values reached 639 Gy•cm 2 (Figure 1), 6.8 Gy and 103 min respectively. The max number of frames was 3326. On the other hand, for PCI procedures, median values of P KA , K a,r, FT, and F ranged from 86-104 Gy•cm 2 , 1532-1942 mGy, 12.7-14.8 min and 1082-1294 respectively. The P KA values reached 818 Gy•cm 2 ( Figure  2), K a,r max was 13.5 Gy, FT reached 99 min and the max number of frames was 6535. Looking closer to these results, for CA cases, the concentration of higher P KA values is noted to be almost stable over time. On the other hand, for PCI cases, this concentration seems to be increased for the second half of the total number. PSD value was calculated by applying equation (1). Median PSD ranged from 393 to 435 and from 992 to 1202 mGy for CA and PCI procedures respectively (Table 2). During CA, PSD max was found 5003 mGy (Figure 3), whereas during PCI, PSD max was found 7134 mGy ( Figure 4). Looking closer to these results, for CA cases, the concentration of higher PSD values is noted to be almost stable over time. On the other hand, for PCI cases, this concentration seems to be increased for the second half of the total number. The effective dose E was also calculated from the corresponding value of P KA by using the factor 0.185 mSv/Gy•cm 2 , for converting P KA to E [17,21]. Median E effective ranged from 4.5 to 5.7 mSv and from 15.9 to 19.1 mSv for CA and PCI procedures respectively ( Table 2). During CA, E max was found 118 mSv ( Figure 5), whereas during PCI, E max was found 151 mSv ( Figure 6). In Table 2 mean values and SD of F, K a,r , PSD , P KA and E, per year, per procedure, are also presented for comparison reasons (Discussion section).
Concerning P KA , for 719 CA cases (13%) P KA value was higher than 55 Gy•cm 2 (Greek DRL) and for 15 cases (0.3%) P KA value was higher than 250 Gy•cm 2 (Table 3 and Figure 7). Among 2817 PCI cases, 885 (31%) P KA value was higher than 130 Gy•cm 2 (Greek DRL) and for 172 cases (6%) P KA value was higher than 250 Gy•cm 2 (Table 3 and Figure 7).   In accordance with the P KA values, the same shape is followed by the effective dose E values, as can be seen in Figures 5 and 6. It is calculated that 24 CA cases (0.4%) received more than 40 mSv, whereas 25 PCI cases (0.9%) received more than 80 mSv.
Concerning the time FT, for 1049 CA cases (19.2%) FT value was higher than 6 min (Greek DRL) and for 471 CA cases (8.6%) FT value was higher than 10 min (Table 3 and Figure 7). Among 2817 PCI cases, 899 (32%) FT value was higher than 18 min (Greek DRL), for 320 cases (11%) FT value was higher than 30 min and for 137 cases (5%) FT value was higher than 40 min (Table 3 and Figure 7).

Discussion
The mean and 3 rd quartile P KA values recorded during both CA and PCI procedures in the present study were compared with those published in the literature (Figures 8 and 9). Mean values for patient doses are not essential, as they do not represent the dominant daily practice in every laboratory. The IC procedures present wide variation in patient dose resulting from complexity. Median and 3 rd quartile values have been established as they are identified to be more representative, the radiation dose data being not normally distributed. It can be seen however that in both figures, current study's CA P KA value is lower than half of the presented laboratories, whereas current study's PCI P KA value is the 4 th (13 centers) and 3 rd (12 centers) higher respectively. It is generally accepted that the large variability in the recorded values of P KA , especially during PCI procedures, is mainly attributed to different levels of procedure complexity, patient thickness and skill of the cardiologist.
Concerning mean effective dose, for CA procedures current study has the 4 th higher value (8 centers) and for PCI procedures it has the 2 nd higher value ( Figure 10). It has to be highlighted however that slight differences were noticed between the values of conversion factor calculated by Uniyal et al. [4] and Betsou et al., which were 0.183 mSv/ Gy•cm 2 . Nevertheless, although E is derived from P KA by using a single multiplicative factor, it provides a convenient way to judge patient doses [26][27][28][29][30][31][32][33][34][35].
Concerning mean PSD value, for CA procedures current study has the 2 nd higher value (7 centers) and for PCI procedures it has the 3 rd higher value (Figure 11). Although the mean PSD (median values are even lower) for CA and PCI procedures in this study were well below the deterministic threshold of skin injury, 13%, 3% and 0.4% of patients received PSDs during PCI exceeding 2, 3 and 5 Gy, respectively (Table 3 and Figure 7). Dose Reference Levels are the guide to what is achievable with current good practice, rather than optimum performance, and are dose neither limits nor thresholds that define competent performance of the operator or the equipment. A mean dose for a procedure that is less than the RL does not guarantee that the procedure is being performed optimally. On the other hand, even with optimal technique,     Pantos [12] Ying [31] Hansson [32] Trianni [33] Giordano [34] Uniyal [4] Silva [35] mean PSD (mGy) CA PTCA  Betsou [28] Pantos [12] Katritsis [29] Toossi [30] Uniyal [4] Bogaert [13] Signorotto [8] mean Effective Dose (mSv) CA PTCA

Conclusion
The data of this study resulted in 8286 IC procedures within the years 2009-2016. There were 5469 (66%) Coronary Angiographies (CA) and 2817 (34%) Percutaneous Coronary Interventions (PCI). Concerning the number of frames F, for 99 CA cases (1.8%) F value was higher than 1250 (Greek DRL). Among PCI cases, 1128 (40%) F value was higher than 1300 (Greek DRL), for 302 cases (11%) F value was higher than 2000 and for 37 cases (1%) F value was higher than 3000.
it is not always possible to keep peak skin dose below the threshold for skin effects. This does not necessarily indicate poor operator technique. It is not necessarily a contraindication to performing or continuing a procedure. However, no deterministic injury was reported from any patient during these eight years of practice in this unit. An event that can be explained by the fact that interventionists are well trained and have always in mind to move the X-ray source as often as possible, so as none patient small skin region to receive large portion of the incident radiation. Nevertheless, the operator is always notified when a significant radiation dose has been administered. The patients are also instructed to notify the operator if any signs or symptoms of a possible radiogenic deterministic effect are observed. The operator writes an appropriate note in the patient's medical record. Follow-up of patients who have received a significant radiation dose applies at 10-14 days and 1 month after the procedure, according to the guidelines [11].
The everyday practice of the two well-experienced interventionists, with the help of evolving technology, shows that simpler procedures, like CA, need slight shorter fluoroscopy time, while the most demanding procedures like PCI, appear to require more frames without increasing in fluoroscopy time or patient's skin dose.