Sie sind auf Seite 1von 12

Radiation Physics and Chemistry xxx (xxxx) xxx–xxx

Contents lists available at ScienceDirect

Radiation Physics and Chemistry


journal homepage: www.elsevier.com/locate/radphyschem

Monte Carlo based estimation of organ and effective doses to patients


undergoing hysterosalpingography and retrograde urethrography
fluoroscopy procedures

J.E. Ngailea, , P.K. Msakia, R.R. Kazemab
a
Department of Physics, University of Dar es Salaam, P O Box 35063, Dar es Salaam, Tanzania
b
Department of Radiology, Muhimbili University of Health and Allied Sciences, P O Box 65001, Dar es Salaam, Tanzania

A R T I C L E I N F O A B S T R A C T

Keywords: Contrast investigations of hysterosalpingography (HSG) and retrograde urethrography (RUG) fluoroscopy pro-
Kerma-area product cedures remain the dominant diagnostic tools for the investigation of infertility in females and urethral strictures
Organ dose in males, respectively, owing to the scarcity and high cost of services of alternative diagnostic technologies. In
Effective dose light of the radiological risks associated with contrast based investigations of the genitourinary tract systems,
Hysterosalpingography
there is a need to assess the magnitude of radiation burden imparted to patients undergoing HSG and RUG
Retrograde urethrography
fluoroscopy procedures in Tanzania. The air kerma area product (KAP), fluoroscopy time, number of images,
Contrast fluoroscopy procedures
Monte Carlo simulation organ dose and effective dose to patients undergoing HSG and RUG procedures were obtained from four hos-
pitals. The KAP was measured using a flat transmission ionization chamber, while the organ and effective doses
were estimated using the knowledge of the patient characteristics, patient related exposure parameters, geo-
metry of examination, KAP and Monte Carlo calculations (PCXMC). The median values of KAP for the HSG and
RUG were 2.2 Gy cm2 and 3.3 Gy cm2, respectively. The median organ doses in the present study for the ovaries,
urinary bladder and uterus for the HSG procedures, were 1.0 mGy, 4.0 mGy and 1.6 mGy, respectively, while for
urinary bladder and testes of the RUG were 3.4 mGy and 5.9 mGy, respectively. The median values of effective
doses for the HSG and RUG procedures were 0.65 mSv and 0.59 mSv, respectively. The median values of ef-
fective dose per hospital for the HSG and RUG procedures had a range of 1.6–2.8 mSv and 1.9–5.6 mSv, re-
spectively, while the overall differences between individual effective doses across the four hospitals varied by
factors of up to 22.0 and 46.7, respectively for the HSG and RUG procedures. The proposed diagnostic reference
levels (DRLs) for the HSG and RUG were for KAP 2.8 Gy cm2 and 3.9 Gy cm2, for fluoroscopy time 0.8 min and
0.9 min, and for number of images 5 and 4, respectively. The suggested DRLs for the HSG and RUG procedures
may be used by the radiology departments in Tanzania for management of attained dose levels until the national
DRLs are established.

1. Introduction minimize patient dose associated with these procedures or/and im-
prove image quality, alternative diagnostic technologies (including
Contrast investigations of hysterosalpingography (HSG) and retro- magnetic resonance imaging, hysteroscopy, and laparoscopy) have
grade urethrography (RUG) procedures studied under fluoroscopy continued to replace a sizeable fraction of conventional fluoroscopy
imaging have been important and most frequently radiological proce- procedures of the genitourinary tract system (Merkle et al., 1996;
dures for evaluation of the genitourinary tract diseases and abnormal- Maciejewski and Rourke, 2015; Yoder and Papanicolaou, 1992; Philips
ities (Merkle et al., 1996; Efstathopoulos et al., 2013; Maciejewski and et al., 2010). However, due to high capital and service cost associated
Rourke, 2015). In spite of promising clinical results, the conventional with some of alternative diagnostic technologies, the conventional HSG
HSG and RUG procedures continued to suffer from unavoidable radia- and RUG fluoroscopic procedures are still the standard radiological
tion exposure to patients as a result of the use of a combination of procedures in many countries for the investigations of infertility in
fluoroscopy and a large series of radiographic images of different pro- young women and urethral strictures in males, respectively
jections (Calicchias et al., 1998; Kramer et al., 2006). In an effort to (Efstathopoulos et al., 2013; Maciejewski and Rourke, 2015; Perisinakis


Corresponding author.
E-mail address: jngaile@gmail.com (J.E. Ngaile).

http://dx.doi.org/10.1016/j.radphyschem.2017.10.015
Received 26 February 2017; Received in revised form 23 October 2017; Accepted 28 October 2017
0969-806X/ © 2017 Elsevier Ltd. All rights reserved.

Please cite this article as: Ngaile, J.E., Radiation Physics and Chemistry (2017), http://dx.doi.org/10.1016/j.radphyschem.2017.10.015
J.E. Ngaile et al. Radiation Physics and Chemistry xxx (xxxx) xxx–xxx

Table 1
Details of the patient demographic data and patient-exposure related parameters collected for each patient enrolled in this study.

Patient demographic data Patient related exposure parameters for each radiological projection

Radiographic exposure parameters Fluoroscopy exposure parameters Geometric factors

Gender (Female/Male) Applied potential (kV) Applied potential (kV) Image field size (cm2)
Patient age (years) Exposure setting (mA s) Tube current (mA) Patient exit to image distance (cm)
Patient thickness (cm) Exposure time (s) Fluoroscopy time (min) Tube focus-to-skin distance (cm)
Patient weight (kg) and height (cm) Kerma area product (Gy cm2) Kerma area product (Gy cm2) Tube focus to table top distance (cm)

et al., 2003). In many developing countries like Tanzania alternative RUG fluoroscopy procedures, (b) to evaluate the influence of existing
technologies being scarce and high diagnostic service cost, the con- patient related exposure parameters on radiation dose to patients from
ventional HSG and RUG procedures are still the dominant diagnostic contrast based fluoroscopy procedures (c) propose local diagnostic re-
tools for investigations of these diseases and disorders. ference levels (LDRLs) and (d) compare these doses to those reported in
As a result of the relative high dose to patients associated with the the literature.
HSG and RUG procedures, there have been great concerns of the pos-
sible undesired health effects such as stochastic cancer risk and genetic 2. Materials and methods
hereditary disorders following direct irradiation of pelvic region, in
which some of the most radiosensitive organs including ovaries, uterus 2.1. Description of data sources
and testes are in the primary beam (Merkle et al., 1996; Efstathopoulos
et al., 2013; Maciejewski and Rourke, 2015; Philips et al., 2010; The data used in the present study were collected between February
Perisinakis et al., 2003). Of particular concern due to long life ex- 2014 and February 2015 from four consultant hospitals (i.e., H1, H2,
pectancy are the young male patients undergoing RUG and the young H3, and H4), which experience larger percentage of patients under-
female patients undergoing HSG procedures, which in addition to ra- going HSG and RUG procedures in Tanzania. All four hospitals are
diation induced cancer they are susceptible to inherent risk of genetic equipped with 3 phase 12 pulse Philips fluoroscopic systems (Philips,
hereditary disorders in males due to changes in the sperm and in fe- Duo Diagnostic, Eindhoven, The Netherland) consisting of under coach
males due to changes in oocyte; and descendants (Merkle et al., 1996; image intensifier with three selectable input field diameters of 38, 31
Efstathopoulos et al., 2013; Plecas et al., 2010; ICRP, 2007). Patient and 23 cm. The focus-to-image distance was 110 cm and the equipment
dose from the HSG and RUG procedures are also influenced by use a maximum tube voltage of 150 kV. The anode angle was 13°,
fluoroscopy personnel and institutional dependent factors such as the whereas the total X-ray beams filtration including that by transmission
level of skills and experiences among fluoroscopy personnel, and the ionization chamber amounts to 3.05 mm Al equivalent for the H2, H3,
employment of different procedural protocols among technologists and and H4 and 3.1 mm Al equivalent for the H1.
institutions (Philips et al., 2010; Plecas et al., 2010; Sulieman et al., In order to investigate the influence of patient characteristics and
2008; Tsapaki et al., 2009). These factors attribute significantly to the patient related exposure parameters on patient dose, patient-exposure
wide range of dose among patients observed within and across in- related parameters were collected using a patient dose survey form
stitutions and nations for the same type of fluoroscopy examination. prepared for each patient participated in the present study. The pa-
The relative high dose to patients from these procedures are anticipated tient–exposure related parameters extracted from the patient survey
to be more pronounced in developing countries like Tanzania due to the form are presented in Table 1. The selection of patient was random and
general lack quality assurance programme, poor equipment main- included a minimum of ten adult patients per hospital for each selected
tenance, and inadequate skills on the use of existing dose reduction fluoroscopy procedure. A total of 127 contrast based X-ray fluoroscopy
techniques (Tsapaki et al., 2009; Muhogora et al., 2008). procedures, for 127 patients were studied; of these 61 female patients
In view of these radiological concerns associated with the use of followed HSG procedures and 66 male patients followed RUG proce-
contrast based fluoroscopy procedures for the genitourinary track dures. The average age for patients who underwent HSG was 33 years
system with observation that radiation dose to patients is dependent on with age range of 21–44 years, while for the RUG was 58 years with age
various factors that vary widely across fluoroscopy personnel, institu- range of 22–89 years. Moreover, details of the total fluoroscopy pro-
tions and nations, there is a need to assess how these factors influence cedures (TFP) performed in 2014 were extracted from medical records
patient dose in Tanzania in general and its institutions in particular. of the four hospitals and then analysed.
Organ and effective dose are dosimetric quantities for estimation of For the contrast based fluoroscopy studies, the exposure parameters
radiation stochastic risk to patients (ICRP, 2007; Martin, 2008; (i.e., kV and mA) for all units were selected automatically through the
UNSCEAR United Nations Scientific Committee on the Effects of Atomic automatic brightness control mode. In case of conventional radio-
Radiation, 2010). However, estimation of organ and effective dose to graphy, the exposure parameters (i.e., kV and mA s) were usually se-
patients from complex fluoroscopic procedures has presented con- lected manually by technologists based on anatomical region of interest
siderable difficulties and time consuming owing to the dynamic nature and patient size for all hospitals although at the H2 these exposure
of the procedures and the wide diversity of examination procedures parameters were usually selected automatically. The preference for
(Hart and Wall, 1994; Elbakri, 2014; Yakoumakis et al., 2015). A well- manual selection of radiographic exposure parameters was done in
established method for estimation of organ and effective doses to pa- order to get clinical acceptable images. All hospitals under study did not
tients is to use simple clinical measurements of air kerma area product have written standard protocols for the imaging procedure and hence,
(KAP) and Monte Carlo (MC) simulation techniques (Yakoumakis et al., the technologists used their experiences and skills in the selection of
2015; Eckerman et al., 1996; Ruiz-Cruces et al., 2000). To our knowl- exposure parameters.
edge few studies have been conducted to assess the dose to the radio- To ensure proper function of the X-ray fluoroscopy units in com-
sensitive organs and effective doses to patients from HSG procedures; pliance with the existing standards (IPEM Institute of Physics and
while for the RUG there are quite limited studies on patient dose Engineering in Medicine, 1997; IAEA International Atomic Energy
compared with the importance and the frequency of the procedure. Agency, 2007), dosimetric and image quality assessment were regularly
Thus, the aim of the present study were to (a) assess the magnitude of performed during the study. The quality control tests such as kV re-
radiation burden imparted to individual patients during the HSG and producibility and accuracy, beam quality for each X-ray tube and the

2
J.E. Ngaile et al. Radiation Physics and Chemistry xxx (xxxx) xxx–xxx

Table 2
Projection and technical parameters applicable to the two contrast based fluoroscopy procedures employed by hospitals participated in the study.

Procedure Depicted anatomy Projection # of images FSD (cm) Field size on film (cm2) per hospital

H1 H2 H3 H4

RUG Urethral RAO 1–2 62–84 18 × 36 12 × 15 35 × 43 24 × 30


Urinary bladder RAO 1–2 62–84 18 × 36 12 × 15 35 × 43 24 × 30
Urinary base AP 1–2 62–84 18 × 36 12 × 15 35 × 43 24 × 30
HSG Uterine cavity AP 1–2 67–87 18 × 36 12 × 15 24 × 30 23 × 30
Fallopian tubes AP 1–2 67–87 18 × 36 12 × 15 24 × 30 23 × 30
Pelvic cavity AP oblique 1–2 67–87 18 × 36 24 × 30 24 × 30 23 × 30

AP, anteroposterior; RAO, Right anterior oblique; FSD, tube focus-to-skin distance.

function of the automatic brightness control were performed for all four using MC simulation techniques. However, the organ and effective
fluoroscopy units using Unfors Xi multimeter (Unfors Xi, type No. doses cannot be determined without the knowledge of KAP (IAEA
8202031-H Xi, Ser. No. 181017, Unfors Inc., Billdal, Sweden). The International Atomic Energy Agency, 2007; ICRU International
image quality tests including low and high contrast resolution for all Commission on Radiation Units and Measurements, 2005). KAP
four units were performed using a Leeds test object TOR 18FG (Leeds (Gy cm2) historically known as dose-area product (DAP) is defined as
Test Objects Limited, Borough bridge, UK). the surface integral of the air kerma Ka over the area of the entire X-ray
beam in a plane perpendicular to the beam axis (IAEA International
2.2. Radiological procedures Atomic Energy Agency, 2007; ICRU International Commission on
Radiation Units and Measurements, 2005; IEC International Electro
The examination protocols used for the HSG and RUG procedures in technical Commission, 2010). In this study, as none of the fluoroscopy
four hospitals under investigations are presented in Table 2. The table unit had its own DAP meter, the KAP was measured for all four hos-
summarizes the depicted anatomy being imaged per procedure, the pitals using a single DAP meter (PTW DIAMENTOR E2, type 11033-
radiological projections, field size per projection, the focus–to–skin dis- 03515, PTW-Freiburg, Freiburg, Germany) by using typical exposure
tance (FSD) per projection and the number of radiographic images per parameters specific for hospital and fluoroscopy unit. The DAP meter
projection. A general explanation on how each of these procedures was was connected via a cable to the transmission ionization chamber (PTW
conducted is described as follows: RUG as a radiological procedure (most Diamentor E2, type TA34028-1-12799, PTW-Freiburg, Freiburg, Ger-
typically in males) in evaluation of urethral obstructions, urethral di- many) fitted to an X-ray tube light-beam diaphragm.
verticula or urethral stricture was divided into three different projections The ionization chamber was calibrated using a solid state detector
each of which included fluoroscopy and the corresponding radiography (Unfors Xi, type No. 8202031-H Xi, Ser. No. 181017, Unfors Inc.,
(Merkle et al., 1996; Maciejewski and Rourke, 2015; Aitchison et al., Billdal, Sweden) on each X-ray unit according to the method described
2009). While the patient lies supine on the examination table, the urethra in TRS 457 (IAEA International Atomic Energy Agency, 2007). The
area of the patient was cleaned by using antiseptic solution and a small overall accuracy of the ionization chamber measurements was esti-
Foley balloon catheter (8-F) was inserted into the distal urethra and fixed mated to be ± 5%. The cumulative KAP values displayed by the DAP
at the navicular fossa. A non-ionic water soluble contrast medium (CM) meter during fluoroscopic screening and image acquisition were col-
(type lopamidol Injection USP or Ultravist of approximately 50–150 ml) is lected for each projection in the procedure. Similarly, the technical
introduced into the penile and pushed retrogradely until it reaches the parameters (i.e., kV, mA s) for fluoroscopy screening and radiographic
bladder. The whole procedure is done under fluoroscopic control. Images imaging, displayed in the console were manually recorded into the
to demonstrate the urethra are taken in oblique positions during in- patient survey form. The total KAP value for each study was obtained
troduction of the CM and after removal of the catheter. by summing the individual KAP values from each set of projection. The
HSG as a radiological procedure in visualization of uterine cavity overall uncertainty of measurement of total KAP value was estimated to
and tubal pathologies for investigations of infertility in young female be ± 5% (1 standard deviation) (Martin, 2008; IAEA International
patients was divided into three different projections each of which in- Atomic Energy Agency, 2007; ICRU International Commission on
cluded fluoroscopy and the corresponding radiography (Efstathopoulos Radiation Units and Measurements, 2005). The fluoroscopy time was
et al., 2013; Kramer et al., 2006; Sulieman et al., 2008; Aitchison et al., determined from the differences between the recorded start and end
2009). The patient was placed in the lithotomy position at the end of times of each exposure during a procedure. The irradiation field size
examination couch. While the patient lies supine on the examination (cm2) was displayed on the collimator assembly and was monitored
table and a short fluoroscopic examination, an anteroposterior (AP) continuously during the HSG and RUG procedures.
control radiograph was obtained prior to the administration of the CM
to ensure the correct position of the cannula and patient preparation. 2.4. Calculation of organ and effective doses
Following uterine filling with non-ionic water soluble CM (type lopa-
midol Injection USP or Ultravist of approximately 20 ml), two AP The organ and effective dose of both examinations were determined
radiographic images were obtained; one during the opacification of the using the measurements of KAP as described in the previous section and
uterine cavity up to the fallopian tubes, and the other during the ex- MC simulation. In the present study, a PC-based X-ray Monte Carlo
travasations of the CM to the peritoneal cavity after external pressure is computer program, PCXMC 2.0 code, supplied by the Finnish Centre for
applied. A third AP radiographic image was acquired with the patient in Radiation and Nuclear Safety, STUK (Tapiovaara and Siiskonen, 2008)
erect position in order to examine the opacification of the peritoneal was used for the photon transport simulation and the calculation of the
cavity and document the intraperitoneal spill. organ and effective dose. PCXMC 2.0 uses computational hermaphro-
dite phantom models of Eckerman et al. (1996) to represent the patient
2.3. Clinical measurements of six different ages: new-born, 1, 5, 10, 15 years old and adult patients.
The computational phantom is modelled by a series of geometric shapes
Of interest in this study was to determine patient dose from the HSG such as spheres and ellipsoids to represent the patient organs
and RUG fluoroscopy procedures in terms of organ and effective doses (Tapiovaara and Siiskonen, 2008). The element composition of the

3
J.E. Ngaile et al. Radiation Physics and Chemistry xxx (xxxx) xxx–xxx

phantom tissues include hydrogen (H), carbon (C), nitrogen (N), oxygen International Commission on Radiation Units and Measurements,
(O), phosphorus (P), and calcium (Ca). The sodium (Na), manganese 2005). The annual collective effective dose from the HSG and RUG
(Mg), phosphorus (P), sulphur (S) and chlorine (CI) have been grouped procedures was obtained by multiplying of the estimated effective dose
together and treated as phosphorus, while all elements of atomic per examination type with the corresponding annual frequencies and
numbers from that potassium (K) or higher have been grouped together summation over all types of examinations (UNSCEAR United Nations
and treated them as calcium. The density for skeleton (except new Scientific Committee on the Effects of Atomic Radiation, 2010; IAEA
born), lung tissues and other tissues are 1.40, 0.296 and 1.04 g/cm3, International Atomic Energy Agency, 2007). In order to facilitate the
respectively (Tapiovaara and Siiskonen, 2008). The phantom sizes are comparison between nations, the mean values (for each country) of
adjustable to mimic patients of an arbitrary weight and height; and KAP, fluoroscopy time, number of images, organ dose for selected or-
allows freely adjustment of X-ray projection and other examination gans and effective dose for the HSG and RUG procedures were de-
conditions used in radiology. The X-ray spectra used for every projec- termined using all fluoroscopy investigations per given examination. In
tion are specified in terms of kVp, the anode angle of the tungsten this way, it was possible to compare the KAP, fluoroscopy time, number
target, and filtration. The focus of the X-ray tube was modelled as a of images, organ and effective dose between Tanzania (this study) and
point source emitting a square cone of photons. the published results from the literature.
In order to initiate simulation, patient specific parameters (age,
weight and height) were first entered into the MC code in order to 2.5. Statistical analysis
obtain a mathematical phantom to represent the patient. The image
field size and focus to image distance were then entered into the code in Statistical analyses were performed using R- (version 3.1.2 (2014-
order to attain the X-ray beam dimensions at the patient's skin and the 10-31): R Foundation for statistical Computing, Vienna, Austria) sta-
focus to skin distance taking into account the obtained phantom. For tistical software and Microsoft Excel 2007. The mean, median, standard
each patient model, the focus to image distance was set to 110 cm, deviation (SD), range and third quartile of the distribution of patient
while the patient exit to image distance was set to 10 cm. The beam demographic parameters, technical parameters, dosimetric parameters
direction and the part of the patient being diagnosed were indicated on (i.e., organ dose and effective dose) recorded and measured during the
the obtained phantom. With obtained X-ray projections and patient procedure were determined at each examination per each hospital.
orientations, the code simulated the information provided for patient Since multiple independent variables may influence patient dose var-
dose estimation. The number of starting particles involved in simulation iation, a multivariable regression analysis was used to identify potential
for each projection was 2 × 107 with maximum energy of 150 keV, explanatory variables that had greatest impact on patient dose varia-
keeping statistical error in organ and effective dose calculations as tion. The selection of variables was made by using a stepwise technique.
given by the software down to 2% (Tapiovaara and Siiskonen, 2008). The coefficient of determination (R2) for multiple regressions was used
The kVp, X-ray tube anode angle, X-ray filtration and KAP (mGy × to measure the percentage of the variation in the KAP that has been
cm2) were entered into the simulated information for organ and ef- explained by the variation of the independent variables. Correlations
fective dose calculation. The anode angle was taken at 13°, while total between various dose indications such as KAP and effective dose re-
filtration was taken at 3.05 mm Al for H2, H3, and H4 and 3.1 mm Al ceived by the patients were determined by means of the Pearson cor-
for H1. The X-ray tube voltages were 55–110 kV for all fluoroscopy relation coefficient (r). A confidence interval of 95% was used in all
procedures. For each projection dose was estimated for all of the 29 statistical calculations. Hence, p-values of less than 0.05 were con-
organs and tissues in the PCXMC phantom, even if they were not in the sidered to represent a statistically significant result.
direct path of the X-ray beam. For all calculations, the arms were re-
moved from the phantom to simulate the normal clinical practice. 3. Results and discussions
Following the calculations of organ doses for each projection, the
total organ doses were estimated by summing the organ doses from 3.1. Patient workload and patient-exposure related parameters
individual exposures. Effective doses were estimated using these values
and the ICRP 103 tissue weighting factors (ICRP, 2007; Martin, 2008). The characteristics of patients for the HSG and RUG procedures in
The overall uncertainty in estimation effective dose was ± 36% (1 four hospitals are presented in Table 3. The table summarizes the
standard deviation). The overall uncertainty in estimation of effective number of patients per procedure; the mean and the range for patient
dose was derived from the square root of the sum of the squares of age, weight, height, the thickness of body part investigated, and the
uncertainty of KAP measurements (5%), the statistical and other errors body mass index. The analysis of the result of general fluoroscopy
from MC simulations (5%), the uncertainty in the fluoroscopy settings procedures conducted in 2014 indicated that a total of 930 patients
and positioning of the patients (25%), and the uncertainty due to the (equivalent to 17.5% of TFP) underwent HSG procedures, of which 72,
phantom used in PCXMC and adult Male and Adult Female (25%) 290, 178, and 390 were for the HI, H2, H3, and H4, respectively. A total
(Martin, 2008; IAEA International Atomic Energy Agency, 2007; ICRU of 913 patients (17.2% of the TFP) underwent RUG procedures, of

Table 3
Sample size and mean (range) patient characteristics observed in the two types of fluoroscopy investigations in four hospitals.

Procedure/Hospitals Sample size Age (years) Height (cm) Weight (kg) Thickness (cm) Body mass index (kg m−2)

Hysterosalpingography
H1 17 32 (23–44) 158 (151–168) 68 (42–96) 19 (12–26) 27.1 (17.7–35.8)
H2 20 30 (21–40) 161 (145–179) 62 (48–78) 18 (13–22) 24.2 (18.8–33.3)
H3 14 35 (23–44) 159 (148–164) 70 (50–93) 18 (14–22) 27.6 (19.3–36.8)
H4 10 38 (27–43) 163 (156–168) 78 (63–92) 24.7 (21–27) 29.6 (22.3–36.5)
All 61 33 (21–44) 160 (145–179) 68 (42–96) 19.5 (12–27) 26.7 (17.7–36.8)
Retrograde urethrography
H1 27 64 (29–89) 166 (152–178) 67 (46–87) 23 (18–36) 24.1 (16.9–30.1)
H2 11 47 (35–68) 165 (143–178) 69 (41–95) 19 (16–27) 25.1 (19.7–32.9)
H3 11 57 (32–70) 166 (158–179) 66 (43–80) 19 (14–28) 23.9 (16.8–28.7)
H4 17 57 (22–88) 165 (143–190) 64 (50–88) 21 (17–29) 23.7 (19.2–31.6)
All 66 58 (22–89) 165 (143–190) 66 (41–95) 21 (14–36) 24.1 (16.8–32.9)

4
J.E. Ngaile et al. Radiation Physics and Chemistry xxx (xxxx) xxx–xxx

Table 4
Summary of mean (range) of exposure parameters for each hospital per fluoroscopy procedure.

Hospital Fluoroscopy Radiography

Sample size Applied potential (kV) Tube current (mA) Applied potential (kV) Current time product (mA s)

Hysterosalpingography
H1 17 67–85 1.5–2.3 78 (70–81) 25.9 (20.0–32.0)
H2 20 110–110 3–3 104 (102–117) 16.9 (5.1–54.1)
H3 14 69–95 1.6–2.8 74 (66–77) 24.9 (12.5–32.0)
H4 10 69–84 1.7–2.5 78 (73–85) 42.2 (21.0–90.0)
All 61 67–110 1.5–3.0 86 (66–117) 29.5 (5.1–90.0)
Retrograde urethrography
H1 27 61–87 1.1–2.5 77 (70–90) 25.6 (20.0–40.0)
H2 11 110–110 3–3 103 (102–117) 17.5 (2.5–53.5)
H3 11 56–102 1.0–2.9 75 (70–77) 25.2 (20.0–32.0)
H4 17 65–85 0.7–3.2 73 (70–81) 28.6 (16.0–40.0)
All 66 56–110 0.7–3.2 80 (70–117) 26.4 (2.5–53.5)

which 178, 86, 166 and 488 were for the H1, H2, H3, and H4, re- assurance programs in order to optimize the radiological protection of
spectively. From the analysis, HSG was the second most frequently the patients.
performed fluoroscopy procedures followed by the RUG. The observed wide variations in number of radiographic images,
Tables 4, 5 presents the result of analysis of technical operating fluoroscopy time, mAs and kV within and across the hospitals in-
parameters used during the fluoroscopy and image acquisition for vestigated were largely attributed to the nature of the pathology, the
contrast based fluoroscopy procedures of the RUG and HSG. It is no- skills and experiences of radiographic technologists and the different
ticeable from the table that there is a substantial variations of exposure examination protocols employed among fluoroscopy personnel and
parameters used for a given fluoroscopy procedure within and across hospitals as observed elsewhere (Calicchias et al., 1998; Sulieman et al.,
the four hospitals investigated. For example, the inter-hospital mean 2008; Livingstone et al., 2004). The variation of exposure parameters
values of number of images were in the ranges of 2.3–5.6 and 2.2–5.3 was further influenced by the manual selection of radiographic ex-
for the RUG and HSG, respectively; while for individual patients across posure parameters by personnel without a written standard protocol
the four hospitals investigated, the variations were in the ranges of 2–9 observed to all hospitals under the study. This finding suggests that
and 1–6 for the RUG and HSG, respectively. The inter-hospital mean there is a potential for dose reduction through development of specific
values of fluoroscopy time were in the ranges of 0.3–1.9 min and protocols per given examination type in order to limit the number of
0.1–0.9 min for the RUG and HSG, respectively; while for individual images acquired and duration of screening as observed elsewhere
patients among the four hospitals investigated, the variations were in (Kramer et al., 2006; Livingstone et al., 2004).
the ranges of 0.1–5.6 min and 0.1–5.5 min for the RUG and HSG, re-
spectively. 3.2. Clinical measurements
It is worth to note from the table that the automatic fluoroscopy
exposure parameters (i.e. kV and mA) were very consistent among The median (range), mean and 3rd quartile values of KAP along
hospitals. Although the fluoroscopy exposure parameters (i.e. kV, mA) with fluoroscopy time and number of radiographic images for the HSG
were varied automatically based on the patient size and the anatomical and RUG procedures across the four hospitals are summarized in
area under study, it is evident from Table 4 that this was not the case at Table 5. It is evident from the table that there is considerable variations
the H2 where the applied tube voltage was mostly fixed at 110 kV, of KAP values per given examination within and across the four hos-
while at the other hospitals the kV varied from 55 to 110 kV. Likewise, pitals investigated. For example, the median values of KAP per hospital
the tube current was generally at 3 mA at the H2, while for other for the HSG and RUG procedures were in the ranges of 1.6–2.8 Gy cm2
hospitals it varied from 0.7 to 3 mA. The observed automatic selection and 1.9–5.6 Gy cm2, respectively. The overall differences in KAP per
of the high values of kV and mA at the H2 regardless of patient size exam between individual patients across the four hospitals investigated
variation or attenuation may have been due to improper calibration of were in the ranges of 0.4–6.5 Gy cm2 and 1.0–13 Gy cm2, respectively
the fluoroscopy unit during the installation. This observation suggests for the HSG and RUG procedures. It is important to note that the H3 had
that the hospitals should strengthen the implementation of the quality the highest median value of KAP for the RUG procedures, followed by

Table 5
Summary of median (range), mean, and 3rd quartile values of fluoroscopy time, number of radiographic images and KAP for each hospital and for all fluoroscopy procedures.

Hospital Fluoroscopy time (min) Number of radiographic images KAP (Gy cm2)

Median (range) Mean 3rd Qu Median (range) Mean 3rd Qu Median (range) Mean 3rd Qu

Hysterosalpingography
H1 0.3 (0.1–1.1) 0.4 0.4 2.0 (1–4) 2.2 2.0 1.6 (0.8–3.3) 1.8 2.5
H2 0.4 (0.1–5.5) 0.9 1.03 5.0 (5–6) 5.3 5.0 1.8 (0.4–6.5) 2.2 2.8
H3 0.1 (0.1–0.2) 0.1 0.12 4.0 (3–4) 3.9 4.0 2.7 (1.5–4.4) 2.9 3.5
H4 0.8 (0.3–1.5) 0.8 1.0 3.0 (3–4) 3.4 3.0 2.8 (1.7–5.8) 3.1 3.2
All 0.3 (0.1–5.5) 0.6 0.8 4.0 (1–6) 3.7 5.0 2.2 (0.4–6.5) 2.3 2.8
Retrograde urethrography
H1 0.6 (0.2–1.0) 0.6 0.84 2.3 (2–6) 2.3 2.0 1.9 (1.0–8.2) 2.5 3.0
H2 1.7 (0.1–5.6) 1.9 2.3 5.6 (4–9) 5.6 6.5 3.7 (1.0–9.5) 4.4 5.7
H3 0.3 (0.1–0.7) 0.3 0.5 4.0 (3–6) 4.1 4.0 5.6 (2.4–13) 5.8 6.2
H4 0.5 (0.2–4.3) 0.8 0.9 4.0 (2–4) 3.5 4.0 2.1 (1.5–3.5) 2.3 2.6
All 0.6 (0.1–5.6) 0.8 0.9 4.0 (2–9) 3.5 4.0 3.3 (1.0–13) 3.3 3.9

5
J.E. Ngaile et al. Radiation Physics and Chemistry xxx (xxxx) xxx–xxx

the H2, H4 and then H1. The highest median value of KAP observed at across the four hospitals investigated. For instance, the median organ
the H3 was probably attributed to the use of largest field sizes. The doses per hospital for the ovaries, uterus and urinary bladder for the
relative lowest median values of KAP for the RUG procedures at the H1 HSG were in the ranges of 0.7–1.3 mGy, 1.3–2.3 mGy and 2.3–7.0 mGy,
might be explained by the use of minimum number of projections respectively; while for testes and urinary bladder for the RUG were in
taken, resulting in fewer radiographic images compared to other hos- the ranges of 4.9–26.7 mGy and 2.8–12.7 mGy, respectively. The
pitals. On the other hand, H4 had the highest median values of KAP for overall differences in individual organ doses across the four hospitals
the HSG, followed by the H3. The highest median value of KAP at the for ovaries, uterus, urinary bladder for the HSG procedures were in the
H4 may be attributed to the use of high radiographic exposure setting ranges of 0.2–5.4 mGy, 0.3–11.0 mG and 1.3–28.2 mGy, respectively;
(mAs), as shown in Table 4 and the use of relative larger image field whereas for testicles and urinary bladder for the RUG were in the
size, as shown in Table 2. ranges of 1.7–127.5 mGy and 1.1–54.5 mGy, respectively. The observed
The prediction of potential explanatory variables that have great wide range of organ doses in Figs. 1(a) and 1(b) for all fluoroscopy
impact on patient dose variation was done using stepwise multivariable procedures of the HSG and RUG, respectively, is an indication that
regression analysis described earlier. From stepwise multivariable re- different procedural protocols (i.e., kV, mA s, fluoroscopy time, image
gression analysis, it was revealed that the patient dose variation for all field size) employed among technologists and hospitals have a sig-
fluoroscopy procedures of the HSG was best explained by a four vari- nificance influence on organ dose determination.
ables regression model described by the following equation: As expected, the organs such as ovaries, urinary bladder, uterus,
testes, and prostate which part or the whole of the organ are located
KAP = 8.28 × 10−4 (field size) + 0.03(mA sr ) + 0.03(PW) within the field of view of the primary beam received relatively high
+ 0.29(fluoroscopy time)–1.90 (1) dose. The organs such as stomach, kidney, spleen, lungs, liver, gall
bladder received relative low dose (range: 0.003–0.33 mGy) owing to
where mAsr was tube current time product for radiographic mode, PW
the fact that they are located outside the field of view of the primary
was patient weight (kg). The overall regression model was significant, F
beam. The uncertainty in estimated organ doses per hospital for the
(4, 56) = 21.7, p < 0.001 and R2 = 0.61. In this model, the important
ovaries, urinary bladder, and uterus for the HSG procedures were in
independent variables were image field size, mAs for radiographic
ranges 38–100%, 41–76%, and 33–100%; respectively. The overall
mode, patient weight and fluoroscopy time; and these accounted for up
uncertainties in estimation of individual organ doses across the four
to 61% of the patient dose variations in the HSG. The uncertainties of
hospitals for ovaries, urinary bladder, and uterus for the HSG proce-
the fitting parameters for Eq. (1) are 0.01%, 1.0%, 0.9%, and 13.6% for
dures were 53%, 85% and 81%, respectively. The highest uncertainties
image field size, patient weight, tube current time product (mA s) and
for estimation of organ dose per hospital for the ovaries, urinary
fluoroscopy time, respectively. Similar analysis revealed that the pa-
bladder and uterus for the HSG procedures was observed at the H2,
tient dose variation for the RUG procedures was best explained by a
while the lowest was observed at the H3. On the other hand, the un-
three variable regression model given by the following equation:
certainty in estimated organ doses per hospital for the testicles, urinary
KAP = 9.58 × 10−5 (Field size) + 1.04(Fluoroscopy time) + 0.40(NI) bladder, and prostate for the RUG procedures were in ranges 43–96%,
(2) 44–97%, and 43–93%, respectively. The highest uncertainties for esti-
+ 0.31
mation of organ dose per hospital for the testicles, urinary bladder and
where NI was the number of radiographic images. The overall regres- prostate for the RUG procedures was observed at the H2, while the
sion model was significant, F (3, 62) = 31.95, p < 0.001, and R2 = lowest was observed at the H4. The overall uncertainties in estimation
0.61. In this model, the important independent variables were image of individual organ doses across the four hospitals for the testicles,
field size, fluoroscopy time and number of images; and these accounted urinary bladder, and prostate for the RUG procedures were 176%,
for up to 61% of the patient dose variation for the RUG. The un- 139% and 177%, respectively.
certainties of parameters fitted in Eq. (2) are 0.001%, 20.5%, and On the other hand, the effective doses were estimated as described
12.2%, respectively for image field size, fluoroscopy time and number earlier for the HSG and RUG procedures using patient exposure para-
of radiographic images. meters specific to the hospital and fluoroscopy units used. The results of
It has been revealed from the multivariable regression analysis that the analysis of these estimations are presented in Table 7. From these
the patient dose variations are mostly determined by the image field results, it is evident that wide variation of median values of effective
size, patient weight, mAs, number of radiographic images, and fluoro- dose per given procedure exist within and across the four hospitals
scopy time. Fortunately, to a large extent these parameters can be investigated. For example, the median values of effective dose per
controlled by fluoroscopy personnel to produce a set of exposure hospital for the HSG and RUG procedures were in the ranges of
parameters that produce the lowest possible dose to patient with still 0.41–0.98 and 0.45–1.94 mSv, respectively. The overall differences
being diagnostic images. Despite the fact that only few variables were between individual effective dose across the four hospitals investigated
identified as potential predictors for the KAP variation in the stepwise were in the ranges of 0.21–4.63 mSv and 0.18–8.41 mSv, respectively
multivariable regression analysis, there are other variables (such as for the HSG and RUG procedures. The uncertainties in estimated ef-
patient thickness and radiographic images for the HSG, and mA, kV and fective dose per hospital for the HSG were in range of 37–82%, while
patient weight for RUG), which were found to have significant con- for the RUG was in range 44–92%. The largest uncertainty for the HSG
tribution to the KAP variation alone but in the model with other vari- was observed at the H2 (82%), whereas the lowest was observed at the
ables their contribution was not significant. Besides, it is worth to note H3 (37%). The largest uncertainty for the RUG was observed at the H2
that there are other possible variables such as skills of the personnel, (92%), while the lowest uncertainty was observed at the H3 (44%). The
image quality and lack of written protocols for manual selection of observed random uncertainty across the four hospitals for the HSG and
exposure parameters, which were not included in the analysis but may RUG were 82% and 132%, respectively. The observed high un-
influence significantly KAP variations. certainties in estimation organ and effective doses within and across the
hospitals were mainly attributed to uncertainties due to variations in
3.3. Calculation of organ and effective doses patient doses within and among hospitals, uncertainties in dose mea-
surements, limited sample size, and uncertainties in conversion coeffi-
The results of the estimated median (range) and mean organ doses cients. The other reasons for the observed uncertainties could be the use
that received the highest dose per procedure for each hospital and for of different examination protocols among fluoroscopy personnel and
all fluoroscopy procedures are presented in Table 6. It is evident from the hospitals, the variation in skills of fluoroscopy personnel, variation
the table that substantial variations of organ dose exist within and in patient characteristics, and equipment characteristics.

6
J.E. Ngaile et al. Radiation Physics and Chemistry xxx (xxxx) xxx–xxx

Table 6
Summary of the median (range) and mean organ doses for selected organs for each hospital and for all fluoroscopy procedures.

Hysterosalpingography
Selected organs Ovaries (in mGy) Urinary Bladder (in mGy) Uterus (in mGy)
Hospitals Median (range) Mean SD Median (range) Mean SD Median (range) Mean SD
H1 0.7 (0.4–1.9) 0.9 0.5 2.3 (1.3–5.2) 2.9 1.3 1.3 (0.7–2.9) 1.6 0.7
H2 0.8 (0.2–5.4) 1.5 1.5 7.0 (2.3–28.2) 8.8 6.7 1.6 (0.3–11.0) 2.6 2.6
H3 1.3 (0.6–2.2) 1.3 0.5 4.2 (1.9–6.8) 3.9 1.6 2.3 (1.1–3.3) 2.1 0.7
H4 1.2 (0.8–2.6) 1.4 0.7 3.6 (2.3–8.5) 4.5 2.5 1.9 (1.1–4.1) 2.2 1.2
ALL 1.0 (0.2–5.4) 1.9 1.0 4.0 (1.3–28.2) 5.3 4.8 1.6 (0.3–11.0) 2.1 1.7
Retrograde urethrography
Selected organs Testicles (in mGy) Urinary Bladder (in mGy) Prostate (in mGy)
Hospitals Median (range) Mean SD Median Mean SD Median (range) Mean SD
H1 7.6 (2.7–27.1) 7.6 6.1 3.0 (1.6–14.7) 4.3 3.4 2.2 (1.2–10.9) 3.2 2.6
H2 47.1 (5.9–127.5) 47.1 45.1 12.7 (2.5–54.5) 19.3 18.7 13.1 (3.3–59.4) 22.6 21.1
H3 7.0 (2.9–15.2) 7.0 3.3 4.6 (1.8–9.8) 5.2 2.4 3.3 (1.2–7.3) 3.7 1.7
H4 5.3 (1.7–12.2) 5.3 2.3 2.8 (1.1–7.5) 3.2 1.4 1.0 (0.8–5.1) 2.3 1.0
ALL 11.9 (1.7–127.5) 11.9 20.9 3.4 (1.1–54.5) 6.1 8.5 2.8 (0.8–59.4) 5.6 9.9

Fig. 1. The frequency distribution of absorbed dose to (a) the urinary bladder for 61 fluoroscopy procedures of the HSG and (b) the testicles for 66 fluoroscopy procedures of the RUG.

The distribution of effective dose values for all fluoroscopy proce- for both the HSG and RUG procedures was not the one that had the
dures of the HSG and RUG are shown in Figs. 2(a) and 2(b), respec- highest mean values of organ and effective dose as expected. This
tively. It is evident from both figures that the frequency distributions finding is explained by the fact that KAP values do not necessary reflect
are skewed to the left with tails projecting to the right. It is worth the change in organ and effective dose in some situation. This is largely
noting from Fig. 2(a) that most of the patients undergoing the HSG because the distribution of the absorbed doses with depth within tissue
procedures received effective dose values between 0.2 and 1.5 mSv various with radiation beam quality (i.e., kV and beam filtration)
with a mean of 0.8 mSv and a median of 0.7 mSv. Similarly, for the (Kramer et al., 2006; Plecas et al., 2010; Martin, 2004). This further
RUG (in Fig. 2(b)), majority of the patients received effective dose suggests that since effective dose take into account the position, ra-
values from 0.2 to 1.5 mSv with mean of 1.0 mSv and median of diation weighting factors, tissue weighting factors and depth of organs,
0.6 mSv. then it is appropriate radiological protection quantity for assessment of
As suggested in the previous paragraph, the H2 had the highest total radiation detriment (ICRP, 2007). The collective effective doses
mean values of effective dose for both fluoroscopy procedures. It is from contrast based fluoroscopy procedures in this study (Tanzania)
surprising to note that the H3 that had relative high mean values of KAP were estimated using the overall mean effective doses of the HSG and

Table 7
Summary of the median (range), mean, SD, 3rd quartile of effective dose (mSv) obtained from each hospital and all fluoroscopy procedures per examination type.

Hospital Hysterosalpingography Retrograde urethrography

Median (range) Mean SD 3rd qu Median (range) Mean SD 3rd qu

H1 0.41(0.22–0.91) 0.50 0.23 0.74 0.46 (0.24–2.19) 0.66 0.48 0.72


H2 0.98 (0.21–4.63) 1.14 0.94 1.29 1.94 (0.48–8.41) 3.14 2.89 5.26
H3 0.72 (0.33–1.05) 0.65 0.24 0.81 1.05 (0.37–2.26) 1.09 0.48 1.23
H4 0.58 (0.37–1.29) 0.70 0.37 0.80 0.45 (0.18–1.24) 0.53 0.24 0.57
ALL 0.65 (0.21–4.63) 0.78 0.64 0.93 0.59 (0.18–8.41) 1.02 1.35 1.03

7
J.E. Ngaile et al. Radiation Physics and Chemistry xxx (xxxx) xxx–xxx

Fig. 2. The frequency distribution of effective dose (a) for 61 fluoroscopy procedures of the HSG and (b) for 66 fluoroscopy procedures of the RUG.

RUG given in Table 7. The estimated annual total collective effective such as 5% for KAP and 36% for effective dose were found to be within
dose from the HSG and RUG procedures was approximately acceptable limits (Martin, 2008; IAEA International Atomic Energy
1.66 manSv. The overall uncertainty in estimation of collective effective Agency, 2007; ICRU International Commission on Radiation Units and
dose for fluoroscopy procedures was 45% (1 standard deviation). The Measurements, 2005), the results obtained in this study can be con-
overall uncertainty in estimation collective effective dose was derived sidered to be adequate accuracy.
by adding in quadrature the uncertainty on the frequencies of the In order to assess the influence of KAP values on effective dose, the
procedures, uncertainty in dose estimation and uncertainty of conver- correlation between KAP with effective dose for individual projections
sion coefficients used to extrapolate of the overall cumulative dose from obtained using MC simulation was performed as described earlier using
the four medical centres. linear regression. The results of the linear regression analysis between
The observed significant wide range of patient dose (KAP, organ values of KAP and effective dose for all fluoroscopy procedures of the
dose, and effective dose) for similar fluoroscopy procedures within and HSG and RUG for H2 hospital are shown in Figs. 3(a) and 3(b), re-
across the hospitals, were mainly attributed to the selection of different spectively; while for other hospitals (i.e. H1, H3, and H4) are shown in
examination protocols (i.e., kV, mA s, fluoroscopy time, number of Figs. 4(a) and 4(b), respectively. From Fig. 3(a), it is evident that there
images, image field size) among radiographic technologists and the is a strong positive correlation between KAP and effective dose (r =
hospitals. For instance, the persistently highest median values of organ 0.73, p < 0.001) for H2 fluoroscopy procedures of the HSG, with the
dose for most of selected organs and effective doses for both HSG and KAP accounting up to 54% of the effective dose variation. It is also
RUG procedures observed for the H2 were largely explained by the use evident from Fig. 3(b) that there is a very strong positive correlation
of relatively high tube current and tube potential for fluoroscopy (3 mA between KAP and effective dose (r = 0.99, p < 0.001) for the H2
and 110 kV, respectively) and large number of radiographic images as a fluoroscopy procedures of the RUG, with the KAP contributing up to
result of large number of projections per procedure, as shown in Tables 98% of the effective dose variation. The uncertainties of the parameters
4, 5. The highest median value of KAP observed at the H3 for the RUG providing the linear regression for Fig. 3(a) were 4.1% and 2.0% for the
procedures was mainly explained by the use of largest field sizes (35 × slope and intercept, respectively; while for Fig. 3(b) were 1.8% and
43 cm2), as shown in Table 2. In addition, the considerable variations of 2.5%, respectively for the slope and intercept. On the other hand, it is
patient doses observed within and across the hospitals were largely evident from Fig. 4(a) that there is a very strong positive correlation
explained by the automatic selection of fluoroscopic exposure para- between KAP and effective dose (r = 94, p < 0.001) for all fluoroscopy
meters (i.e., kV and mA) based on patient characteristics and anato- procedures of the RUG from other hospitals (H1, H3 and H4), with the
mical region of interest; and by the complexity of individual fluoro- KAP contributing up to 89% of the effective dose variation. It is also
scopy procedures, which is characterized by individual patient evident from Fig. 4(b) that there is a very strong positive correlation
variation in terms of prolonged fluoroscopy time and number of between KAP and effective dose (r = 94, p < 0.001) for all fluoroscopy
radiography exposures. The observed wide range of patient dose within procedures of the RUG, with the KAP accounting up to 88% of the ef-
and across the hospitals was further influenced by manual selection of fective dose variation. The uncertainties of the parameters providing
radiographic exposure parameters, which was regularly done by the linear regression for Fig. 4(a) were 0.9% and 0.8% for the slope and
fluoroscopy personnel at the four hospitals without reference to the intercept, respectively; while for Fig. 4(b) were 0.6% and 0.9%, re-
written standard imaging protocols per procedures. This was particu- spectively for the slope and intercept. The excellent positive correlation
larly concerning in the manual selection of radiographic exposure between KAP and effective dose largely attributed to the fact that KAP
parameters by newly employed technologists, radiographic students, or takes into account procedural protocols (i.e., kV, mA s, fluoroscopy
residents, who are not familiar with optimal selection of exposure time, image field size, etc.).
parameters in their fluoroscopic units. The observed large uncertainties In view of the fact that there is a very strong positive correlation
in estimation of mean values of organ and effective doses within and between KAP and effective dose, and the relationship between the ef-
across the hospitals were mainly attributed to the uncertainties due to fective dose and stochastic radiation risks is also assumed to be linear
variations in patient dose (KAP, organ dose and effective dose) within (ICRP, 2007), then it can be hypothesised that KAP, is significantly
and across the hospitals. The large uncertainties could be significantly positively correlated with the level of radiation risks of cancer to pa-
reduced by development of standard protocols per procedure and pro- tients from the HSG and RUG procedures. In assumption that regression
vision of training to radiographic technologists on optimal use of equation pass through the origin (zero intercept), the estimated effec-
fluoroscopy machine. Since the uncertainties due to dose measurements tive dose conversion coefficients (ECCs) for the H2 hospital were

8
J.E. Ngaile et al. Radiation Physics and Chemistry xxx (xxxx) xxx–xxx

Fig. 3. The correlation between the KAP and effective dose per projection for (a) 20 fluoroscopy procedures of HSG, R2 = 0.54, (b) for 11 fluoroscopy procedures of RUG, R2 = 0.98 for
H2 hospital. The solid line represents the least-square linear regression and the dotted lines represent the 95% confidence interval.

0.41 mSv (Gy cm2)−1 and 0.61 mSv (Gy cm2)−1 for the HSG and RUG radiographic images, organ doses for selected organs and effective
procedures, respectively; which were relatively higher than reported doses for all fluoroscopy procedures per examination in the present
values of 0.29 mSv (Gy cm2)−1 and 0.18 mSv (Gy cm2)−1 established study and from reported values from the literature for the HSG are
by Hart and Wall (2002) for the HSG and RUG procedures, respectively. presented in Table 8 (Efstathopoulos et al., 2013; Calicchias et al.,
On the other hand, the estimated ECCs for other hospitals were 1998; Plecas et al., 2010; Abdullah et al., 2001; Fernandez et al., 1996;
0.26 mSv (Gy cm2)−1 and 0.19 mSv (Gy cm2)−1 for the HSG and RUG Gyekye et al., 2012), while for the RUG are presented in Table 9
procedures, respectively, which were in good agreement with reported (Merkle et al., 1996; Gyekye et al., 2009). It is noticeably from these
values of 0.29 mSv (Gy cm2)−1 and 0.18 mSv (Gy cm2)−1 established tables that the average values of KAP for the HSG and RUG procedures
by Hart and Wall (2002) for the HSG and RUG procedures, respectively. in the present study were comparable and slightly higher than those
The overall uncertainty in estimation of conversion coefficients was reported by the Merkle et al. (1996) for Germany, Gyekye et al. (2009)
37% (1 standard deviation). The observed high ECCs for both HSG and for Ghana, Gyekye et al. (2012) for Ghana and Efstathopoulos et al.
RUG for the H2 hospital relative to those from other hospitals and those (2013) for Greece. The mean value of KAP for the HSG from the present
established by Hart and Wall (2002) were mainly explained by the use study was higher than those from Ghana and Greece by factors of 4.4
of relatively high exposure parameters (i.e., kVp and mA). and 5.7, respectively; while for the RUG was comparable to those from
Ghana and Germany. The observed higher mean value of KAP in Tan-
zania to that from Ghana for the HSG procedure may be explained by
3.4. Comparison with other studies from the literature the use of fewer radiographic images; while for Greece may be attrib-
uted to the use of digital X-ray fluoroscopy equipment (Efstathopoulos
The mean (range) values of KAP, fluoroscopy time, number of

Fig. 4. The correlation between the KAP and effective dose per projection for (A) 41 fluoroscopy procedures of HSG, R2 = 0.89, (B) for 55 fluoroscopy procedures of RUG, R2 = 0.88 for
other hospitals (H1, H3 and H4). The solid line represents the least-square linear regression and the dotted lines represent the 95% confidence interval.

9
J.E. Ngaile et al. Radiation Physics and Chemistry xxx (xxxx) xxx–xxx

Table 8
Comparison of mean (range) values of KAP, organ dose and effective dose between this study (Tanzania) and other published results from literature for the HSG procedures.

Ref. Sample size Mean KAP Mean fluoroscopy time # of spot D (mGy) for D (mGy) for uterus Mean effective dose
(Gy cm2) (min) images ovaries (mSv)

This work 61 2.3 (0.4–6.5) 0.6 (0.1–5.5) 3.7 (1–6) 1.3 (0.2–4.0) 2.1(0.3–11.0) 0.8 (0.2–4.6)
Fernandez et al. (1996) 41 7.1 (2.5–16.2) (0.1–1.0) 7.0 (7–8) 4.6 3.1 (1.0–8.1)
Calicchias et al. (1998) 37 0.2 (0.1–0.5) 6.5 (6–7) 4.7 6.4 2.0
Abdullah et al. (2001) 5.0 (2.2–13.0) 2.0 (0.7–4.8) 2.0 1.4 (0.6–3.5)
Plecas et al. (2010) 31 3.2 (1.1–4.5) (0.4–0.9) 2.9 (2–3) 1.6 (0.3–2.8) 2.2 (0.4 – 3.8)
Gyekye et al. (2012) 20 0.5(0.4–0.8) 0.8 (0.4–1.2) 2.5 (2–3) 0.4 ± 0.1 0.7 ± 0.1 0.2 ± 0.0
Efstathopoulos et al. 0.4 (0.2–0.9) 0.2 (0.1–0.5) 0.2 (0.1–0.3)
(2013)

et al., 2013; Plecas et al., 2010; Livingstone et al., 2004; Gyekye et al., ovaries lie nearer to the anterior surface, while in PA projection the
2012; Schultz et al., 1998). On the other hand, the mean value of KAP radiosensitive organs including uterus, ovaries and urinary bladder are
for the RUG in the present study was comparable and relatively lower placed in deeper locations and are partially covered by pelvic bones,
than those reported by Fernandez et al. (1996) for Spain; Abdullah et al. which provide natural shielding from radiation beam (Efstathopoulos
(2001) for Malaysia, and Plecas et al. (2010) for Serbia by factors of 3.1, et al., 2013; Kramer et al., 2006; Plecas et al., 2010; Martin, 2004). This
2.1 and 1.4, respectively. The observed significant variations of mean further suggests that there exist potential for reducing radiation dose to
values of KAP between the present study and those reported from Spain, patients in Tanzanian hospitals through minimization of number of
Malaysia, Serbia and Ghana may be explained by the use of long radiographic images, the use of PA projection where possible instead of
fluoroscopy time and the relative large number of radiographic images, AP projection and the use of digital fluoroscopy equipment. In spite of
as shown in Tables 8, 9. the fact that the mean values of organ doses for selected organs and
It is clear from the Table 8 that the average values of organ dose for effective doses were relatively lower than reported values from other
selected organs for the HSG procedure in the present study were rela- studies, the estimated uncertainties for the organ and effective doses of
tively higher than those reported by Gyekye et al. (2012) for Ghana the ranges 43–177% and 82–132%, respectively were relatively high.
and Efstathopoulos et al. (2013) for Greece by factors of up to 3.3 and Moreover, the proposed diagnostic reference levels (DRLs) for the
5.3, respectively. It is clear from the Table 8 that the mean values of HSG and RUG fluoroscopy procedures were obtained from the third
organ doses for the HSG procedures in the present study were com- quartile values of overall distribution of KAP, fluoroscopy time and
parable to those reported by Plecas et al. (2010) for Serbia, while re- number of radiographic images, as shown in Table 5 (Aroua et al.,
latively lower than those reported by the Fernandez et al. (1996) for 2007). The results of the proposed DRLs for the HSG and RUG for the
Spain; and Calicchias et al. (1998) for Italy by factors of up to 4.0 and present study along with the recommended DRLs from the literature for
3.6, respectively. The relative higher mean organ doses in Spain and the UK, Australia and Switzerland are summarized in Table 10 (FOPH
Italy than those from the present study might be explained by the larger Federal office of Public Health, 2008; Hart et al., 2009; Erskine et al.,
number of images, which were higher by factors of up to 1.9. 2014). It is noticeably from these results that the DRLs of KAP and
It is evident from Table 8 that with exception of Ghana and Greece, fluoroscopy time for the HSG procedures in the present study were
the mean effective dose for the HSG procedures in the present study was slightly lower than those reported by Hart et al. (2009) for the UK. The
relatively lower than those reported by Fernandez et al. (1996) for DRLs for KAP, fluoroscopy time and number of images for the RUG in
Spain; Calicchias et al. (1998) for Italy and Abdullah et al. (2001) for the present study were lower than those reported by FOPH Federal
Malaysia by factors of 4.0, 2.5 and 1.7, respectively. The lower mean office of Public Health (2008) for Switzerland and by Erskine et al.
value of effective dose in Tanzania (this study) relative to those from (2014) for the Australia. The proposed DRL may be used by the radi-
Spain and Italy could be attributed to the use of high number of images ology departments in Tanzania to provide guidance on attainable dose
by factors of 1.9 and 1.8, respectively, whereas for Malaysia may be levels during the HSG and RUG procedures until the national DRLs are
attributed to the longer fluoroscopy time by a factor of 3.4. It is also established.
evident from the Table 8 that the mean effective doses for the HSG
procedures in the present study was relatively higher than those re-
3.5. Recommendations for improvement of radiological protection to
ported by Efstathopoulos et al. (2013) for Greece and Gyekye et al.
patients
(2012) for Ghana by factors of 5.2 and 3.9, respectively. As suggested
from previous paragraph, the relative lower values of mean organ doses
In light of the observed root causes of variations in procedural
and effectives doses in Ghana may be explained by the use of fewer
protocols and patient doses within and across the four hospitals in-
images than the present study whereas for Greece could be attributed to
vestigated and similar experience observed elsewhere (Plecas et al.,
the use of digital fluoroscopy equipment. The other possible factor that
2010; Livingstone et al., 2004; Geleijn et al., 1998); it is apparent that
might also contributed to the lower mean organ dose and effective dose
necessary measures are required in order to optimize radiological
in Greece relative to the present study could be the use of posterior-
protection of patients from the HSG and RUG fluoroscopy procedures.
anterior (PA) projection instead of AP. This finding is explained by the
There are number of observed parameters that require optimisation:
fact that in AP projection some of the organs such as bladder and
Firstly, following the observed use of different procedural protocols

Table 9
Comparison of mean (range) values of KAP, organ dose and effective dose (Tanzania) and other published results from literature for the RUG procedures.

Ref. Sample size Mean KAP value Mean fluoroscopy time Mean number of D (mGy) for testes Mean effective dose
(Gy cm2) (min) images (mSv)

This work 66 3.3 (1.0–13.0) 0.8 (0.1–5.6) 3.5 (2–9) 11.9 (1.7–127.5) 1.0 (0.2–8.4)
Gyekye et al. (2009) 12 3.6 ± 1.0 1.9 ± 0.2 5.0 2.0 ± 0.5 0.3 ± 0.1
Merkle et al. (1996) 40 3.2 (0.7–7.7) 4.6 (3–8)

10
J.E. Ngaile et al. Radiation Physics and Chemistry xxx (xxxx) xxx–xxx

Table 10
Comparison of 3rd quartile values of KAP, fluoroscopy time and number of images of all contrast fluoroscopy procedures in the present study and recommended DRLs from the literature.

This study (Tanzania) Hart et al. (2009) (The UK) Erskine et al. (2014) (Australia) FOPH Federal office of Public Health (2008) (Switzerland)

HSG RUG HSG RUG HSG RUG HSG RUG

KAP (Gy cm2) 2.8 ± 1.3 3.9 ± 0.9 2.9 23.6 5


Fluoroscopy time (min) 0.8 ± 0.8 0.9 ± 1.0 1.0 2 1
Number of images 5±1 4±2 7

among the hospitals for similar fluoroscopy procedure, there is an ur- while for the RUG were lower than the DRLs proposed by Australia and
gent need to develop guidelines for the HSG and RUG procedures Switzerland. The comparable patient dose variations between the de-
among hospitals with intention to better standardizes clinical practice veloping and advanced nations tend to suggest that both developing
among the hospitals for similar procedure. Secondly, in view of the and advanced nations need to optimize procedural protocols in order to
observed substantial variations of technical parameters and procedural minimize radiation dose to patients, while keeping the diagnostic in-
protocols among radiographic technologists for a given fluoroscopy formation. The observed wide range of procedural protocols and patient
procedure within hospital, there is a need for each hospital to develop doses within and across the hospitals; call for the need to standardize
written standard protocols for the HSG and RUG procedures based on examination protocols and optimize contrast based fluoroscopy proce-
these guidelines. This will significantly reduce the variation of proce- dures.
dural protocols and patient doses within the hospital.
Thirdly, in view of the observed large differences in image field size Acknowledgements
among hospitals of up to a factor of 8.4 (35 × 43 cm2/12 × 15 cm2) for
the RUG examination, there is a need for the hospitals to reduce the The authors are grateful to the management of the consultant hos-
image field size to 12 × 15 cm2, which is sufficient or are more (partly pitals included in the study for allowing them to use their X-ray
overlapping) images needed to cover the whole range. This will sig- fluoroscopy facilities. Similarly, they would like to sincerely thank the
nificantly reduce unnecessary radiation dose to patients undergoing following radiographic technologists: Messrs P. Masue (H4), F.
RUG procedures. Fourthly, following the observed improper calibration Maxmillian (H4), S. Mtawa (H1), H. Juma (H1), P. Jongole (H2), D.
of automatic brightness control (ABC) of the fluoroscopy unit at the H2, Simba (H2), J. Kulinga (H3), N. Mitti (H3) and E. Mlelwa (H3); and Ms
there is an imperative need for the hospitals to strengthen the im- E. Severine (H3) for their technical support at the X-ray fluoroscopy
plementation of the periodical performance tests of fluoroscopy unit in facilities. They sincerely wish to acknowledge the financial support
order to optimize the radiological protection of the patients. Lastly, from Tanzania Commission for Science and Technology (COSTECH);
considering the observed substantial variations of patient doses (KAP, and equipment support from the Tanzania Atomic Energy Commission
fluoroscopy time, organ doses, effective dose) within and across the (TAEC) and Pacific Diagnostic Limited (Tanzania). They gratefully ac-
hospitals, there is a strong need to examine the potential for lowering knowledge the National Institute for Medical Research (NIMR) for ap-
radiation dose to patients from the HSG and RUG procedures through proval of their research proposal.
provision of training to fluoroscopy personnel on optimal use of
fluoroscopy unit; optimal selection of operation X-ray settings (i.e.; Funding sources
increase patient to focus distance, beam collimation, dose rate and in-
crease FDD, minimize distances between the patient and image receptor This study was financially supported by the Tanzania Commission
detector), optimal selection of technical parameters (i.e., kV, mA, the for Science and Technology (CST/PhD/017/2011).
use of minimal fluoroscopy time, number of projection, and number of
images) based on the indication of the study, anatomical area examined References
and patient size; the use of high kVp technique and use of additional
beam filtration as observed elsewhere (Kramer et al., 2006; Plecas et al., Abdullah, B.J.J., Ng, K.H., Rassiah, P., 2001. A comparison of radiation dose and image
2010; Martin, 2004; Abdullah et al., 2001). quality in hysterosalpingography using conventional and high kilovolt techniques. J.
H. K. Coll. Radiol. 4, 133–136.
Aitchison, F., Chapman, S., Nakielny, R., 2009. A Guide to Radiological Procedures, Fifth
4. Conclusions Edition. Elsevier Limited.
Aroua, A., Rickli, H., Stauffer, J.-C., et al., 2007. How to set up and apply reference levels
in fluoroscopy at a national level. Eur. Radiol. 17, 1621–1633.
The assessment of radiation dose imparted to patients undergoing Calicchias, A., Chiacchiararelli, L., de Felice, C., Gigliotti, T., Indovina, P.L., Mazzei, F.,
hysterosalpingography and retrograde urethrography procedures under Porfiri, M., 1998. Evaluation of effective dose in hysterosalpingography. Radiat. Prot.
fluoroscopic guidance in Tanzania has been reported. Substantial var- Dosim. 80 (1–3), 159–161.
Eckerman, K.F., Cristy, M., Ryman, J.C., 1996. The ORNL mathematical phantom series.
iations of the procedural protocols, technical parameters and patient 〈http://ats.ronl.gov/documents/mird2.pdf〉.
doses (KAP, organ dose and effective dose), within and across the four Efstathopoulos, E.P., Charalambatou, P.P., Tsalafoutas, I.A., Kelekis, A.D., Antonakos,
hospitals investigated were observed. These variations were mainly Brountzos, E., Kelekis, D.A., 2013. Effective and ovarian dose in PA conventional and
rotational 3D hysterosalpingography examinations. Phys. Med. 29, 549–555.
explained by the complexity of individual procedures, the patient size
Elbakri, I.A., 2014. Estimation of dose area product to effective dose conversion factors
variations, the variations in level of skills and working experience of for neonatal radiography using PCXMC. Radiat. Prot. Dosim. 158 (1), 43–50.
fluoroscopy personnel, lack of written standard protocols and the dif- Erskine, B.J., Bradly, Z., Marshall, E.M., 2014. Local diagnostic reference levels for an-
ferent examination protocols employed among technologists and hos- giographic and fluoroscopic procedures: Australian practice. Australas. Phys. Eng.
Sci. Med. 37, 75–82.
pitals. The prominent variables responsible for patient dose variations Fernandez, J.M., Vano, E., Guibelalde, E., 1996. Patient doses in hysterosalpingography.
were the image receptor field size, patient weight, number of radio- Br. J. Radiol. 69, 751–754.
graphic images, tube current time product and fluoroscopy time. The FOPH (Federal office of Public Health), 2008. Diagnostic reference levels for interven-
tional radiology procedures. R-06-05mod.doc, 2nd version. Geneva, Switzerland.
mean values of KAP, organ dose and effective dose in the present study Geleijn, J., Broerse, J.J., Chandie Swaw, M.P., Schultz, F.W., Teeuvsse, W., Van Unnik,
were mostly lower than reported values from Spain, Italy and Malaysia, J.G., Zoetelief, J., 1998. A comparison of patient dose for examinations of the upper
while higher than those reported from Ghana and Greece. The third gastrointestinal track at 11 conventional and digital X-ray units in the Netherland. Br.
J. Radiol. 71, 745–753.
quartiles of the KAP and fluoroscopy time for all fluoroscopy proce- Gyekye, K.P., Emil-Reynolds, G., Boadu, M., Darko, E.O., Yeboah, J., Inkoom, S., Mensah,
dures of the HSG were lower than the recommended DRLs from the UK,

11
J.E. Ngaile et al. Radiation Physics and Chemistry xxx (xxxx) xxx–xxx

C.K., 2012. Cancer incidence risks to patients due to hysterosalpingography. J. Med. Radiol. 80, 639–647.
Phys. 37 (2), 112–116. Merkle, E., Aschoff, A.J., Muller, M., Merk, J., Brambs, H.J., 1996. Computer assisted
Gyekye, P.K., Schondforf, C., Boadu, M., Yeboah, I., Amoako, J.K., 2009. Patient dose determination of the dose–area product in retrograde urethrography. Br. J. Radiol.
assessment due to fluoroscopic exposure for some selected fluoroscopic procedures in 69, 262–265.
Ghana. Radiat. Prot. Dosim. 158 (1), 43–50. Muhogora, W.E., Ahmed, N.A., Almosabihi, A., Alsuwaidi, J.S., Beganovic, A., Ciraj-
Hart, D., Wall, B.F., 1994. Estimation of effective dose from dose–area product mea- Bjelac, O., et al., 2008. Patient doses in radiographic examinations in 12 countries in
surements for barium meals and barium enemas. Br. J. Radiol. 67, 485–489. Asia, Africa, and Eastern Europe: initial results from IAEA projects. Am. J.
Hart, D., Hillier, M.C., Wall, B.F., 2009. National reference doses for common radio- Roentgenol. 190 (6), 1453–1461.
graphic, fluoroscopic and dental X-ray examinations in the UK. Br. J. Radiol. 82, Perisinakis, K., Damilakis, J., Grammatikakis, J., Theocharopoulos, N., Gourtsoyiannis,
1–12. N., 2003. Radiogenic risks from hysterosalpingrography. Eur. Radiol. 13, 1522–1528.
Hart, D., Wall, B.F., 2002. Radiation Exposure of the UK Population from Medical and Philips, J., Cochavi, S., Silberzweig, J.E., 2010. Hysterosalpingography with use of mobile
Dental X-ray Examinations (NRPB-W4). National Radiological Protection Board, c-arm fluoroscopy. Fertil. Steril. 93 (6), 2065–2068.
Oxon. Plecas, D., Zivkovic, M., Ciraj-Bjelac, O., 2010. Radiation dose and risk assessment in
IAEA (International Atomic Energy Agency), 2007. Dosimetry in Diagnostic Radiology. hysterosalpingography. Nucl. Technol. Radiat. Prot. 25 (3), 217–221.
An International Code of Practice. Technical Reports Series No. 457. Ruiz-Cruces, R., Ruiz, F., Perez-Martinez, M., Lopez, J., Tort Ausina, I., Diez de los Rios,
ICRP, 2007. The 2007 recommendations of the international commission on radiological A., 2000. Patient dose from barium procedures. Br. J. Radiol. 73, 752–761.
protection. ICRP Publication 103. Annals of the ICRP. 37 (2–4). Schultz, F., Teeuwisse, W., Broerse, J., Chandle Shaw, M., Geleijns, J., 1998. Dosimetric
ICRU (International Commission on Radiation Units and Measurements), 2005. Patient Consequences of implementation of digital radiology for hysterosalpingography. In:
Dosimetry for X-rays Used in Medical Imaging (ICRU Report 74). 5 Oxford University AAPM Manual Meeting Program, Med. Phys., vol. 25(7) part I, A156.
Press. Sulieman, A., Theodorous, K., Vlychou, M., Topaltzikis, T., Roundas, C., Fezoulidis, I.,
IEC (International Electrotechnical Commission), 2010. Medical Electrical Equipment: Kappas, C., 2008. Radiation dose optimization and risk estimated to patients and staff
Part 2-43- Particular Requirements for the Safety of X-ray Equipment for interven- during hysterosalpingography. Radiat. Prot. Dosim. 128 (2), 217–226.
tional Procedures (Report 60601), 2nd ed. International Electro technical Tapiovaara, M., Siiskonen, A., 2008. PCXMC: A PC Based Monte Carlo Program for
Commission, Geneva, Switzerland. Calculating Patient Doses in Medical X-ray Examinations, Version 2.0. Finnish Centre
IPEM (Institute of Physics and Engineering in Medicine), 1997. Recommended Standard for Radiation and Nuclear Safety, Helsinki, Finland.
for the Routine Performance Testing of Diagnostic X-ray Imaging Systems (Report No. Tsapaki, V., Ahmed, N.A., AlSuwaidi, J.S., Beganovic, A., Benider, A., BenOmrane, L.,
77). IPEM, York. et al., 2009. Radiation exposure to patients during interventional procedures in 20
Kramer, R., Khoury, H.J., Lopes, C., Vieira, J.W., 2006. Equivalent dose to organs and countries: initial IAEA project results. Am. J. Roentgenol. 193 (2), 559–569.
tissues in hysterosalpingography calculated with the FAX (Female Adult voXel) UNSCEAR (United Nations Scientific Committee on the Effects of Atomic Radiation),
phantom. Br. J. Radiol. 79, 893–898. 2010. Sources and effects of ionizing radiation. Reports to the general assembly, New
Livingstone, R.S., Koshy, C.G., Raj, D.V., 2004. Evaluation of work practices and radiation York.
dose during adult micturating cystourethrography examinations performed using a Yakoumakis, E., Dimitriadis, A., Gialousis, G., Makri, T., Karavasilis, E., Yakoumakis, N.,
digital imaging system. Br. J. Radiol. 77, 927–930. Georgio, E., 2015. Evaluation of organ and effective doses during paediatric barium
Maciejewski, C., Rourke, K., 2015. Imaging of urethral stricture disease. Transl. Androl. meal examinations using PCXMC 2.0 Monte Carlo code. Radiat. Prot. Dosim. 163 (2),
Urol. 4 (1), 2–9. 202–209.
Martin, C.J., 2004. A review of factors affecting patient doses for barium enemas and Yoder, I.C., Papanicolaou, N., 1992. Imaging the urethra in men and women. Urol. Radiol.
meals. Br. J. Radiol. 77, 864–868. 14, 24–28.
Martin, C.J., 2008. Effective dose: how should it be applied to medical exposure? Br. J.

12