

ORIGINAL ARTICLE 

Year : 2018  Volume
: 41
 Issue : 2  Page : 8893 


Reader quality control tests, dose algorithm comparison, and signal depletion of optically stimulated luminescence dosimetry systems used for individual monitoring: A case of the dosimetry system of the national individual monitoring service in Gabon
Philippe Ondo Meye^{1}, Cyril Schandorf^{2}, Roger Ondo Ndong^{3}
^{1} General Directorate of Radiation Protection and Nuclear Safety, Ministry of Water and Energy, Libreville, Gabon ^{2} Department of Medical Physics, School of Nuclear and Allied Sciences, Atomic  Accra, Ghana ^{3} Département Sciences Physiques, Ecole Normale Supérieure, Quartier Derrière la Prison, Libreville, Gabon
Date of Submission  22Jan2018 
Date of Decision  16Feb2018 
Date of Acceptance  02Mar2018 
Date of Web Publication  24Aug2018 
Correspondence Address: Mr. Philippe Ondo Meye General Directorate of Radiation Protection and Nuclear Safety, Ministry of Water and Energy, BP 1172 Libreville Gabon
Source of Support: None, Conflict of Interest: None  Check 
DOI: 10.4103/rpe.RPE_6_18
This study was conducted to assess some important performance characteristics of the optically stimulated luminescence dosimetry system of the national individual monitoring service in Gabon, namely, reader performance tests, dose algorithm, and signal depletion. The reader performance tests demonstrated that the reader photomultiplier tube is performing selfconsistently and that there is no abnormality in the counting system. Concerning the dose algorithm, it was shown that the proposed dose algorithm provides a good estimate of the dose compared with the system dose algorithm, which is not known. From the signal depletion study, the depletion rate per reading of 0.4 ± 0.09% (strong stimulation beam) was determined for field dosimeters. Depletion rates per reading of 0.3 ± 0.03% (strong beam) and 0.020 ± 0.001% (weak beam) were also determined for calibration dosimeters. All these values are comparable to those determined in other studies. Keywords: Dose algorithm, individual monitoring, optically stimulated luminescence, signal depletion
How to cite this article: Meye PO, Schandorf C, Ndong RO. Reader quality control tests, dose algorithm comparison, and signal depletion of optically stimulated luminescence dosimetry systems used for individual monitoring: A case of the dosimetry system of the national individual monitoring service in Gabon. Radiat Prot Environ 2018;41:8893 
How to cite this URL: Meye PO, Schandorf C, Ndong RO. Reader quality control tests, dose algorithm comparison, and signal depletion of optically stimulated luminescence dosimetry systems used for individual monitoring: A case of the dosimetry system of the national individual monitoring service in Gabon. Radiat Prot Environ [serial online] 2018 [cited 2021 Jul 30];41:8893. Available from: https://www.rpe.org.in/text.asp?2018/41/2/88/239687 
Introduction   
Occupational exposure is the exposure of workers incurred in the course of their work regardless of the situation of exposure.^{[1]} Practically, workers may be exposed to radiation as a result of various human activities: work associated with the different stages of nuclear fuel cycle, use of radiation in medicine, scientific research, agriculture and industry, and occupations that involve exposure to radionuclides of natural origin. To control this exposure, it is necessary to assess the magnitude of the radiation doses involved. According to the basic safety standards of the International Atomic Energy Agency (IAEA),^{[2]} employers, registrants, and licensees shall be responsible for making arrangements for the assessment of occupational exposure of workers, on the basis of individual monitoring where appropriate, and shall ensure that arrangements are made with authorized or approved dosimetry service providers that operate under a quality management system. One of the means for fulfilling this regulatory requirement, when it comes to external exposure, is to ensure that workers are monitored by individual dosimetry services that use passive dosimeters (e.g., optically stimulated luminescence dosimetries [OSLDs]).^{[3]} In the African region, OSLbased dosimetry systems are more and more used, especially in French speaking countries (Gabon, Cameroon, Chad, Senegal, Mauritania, Morocco, Algeria, and Mauritius). This is the reason why it has been found relevant to undertake a study focusing on performing some important tests related to OSL dosimetry systems. These are reader quality control (QC) tests and signal depletion. A dose algorithm comparison is also presented in this work. The work was limited to the measurement of the quantity Hp (10) in photon fields by the dosimetry system of the individual monitoring service (IMS) of the Ministry of Water and Energy in Gabon. A previous work was carried out on the same dosimetry system where a certain number of performance indicators were assessed.^{[4]} The present study addresses additional performance indicators and proposes a dose algorithm that is compared to the system dose algorithm provided by Landauer.
Materials and Methods   
Materials
The dosimetry system considered here is an InLight dosimetry system (Landauer, Inc., USA). It is designed to assess personnel exposure to beta particles and photons using OSL.^{[5],[6],[7]} The system consists of individual passive dosimeters, a reader, a highintensity light annealer and a software installed on a personal computer.
Reader
The reader under consideration is the portable InLight microStar (32.7 cm width × 23.2 cm depth × 10.9 cm height,^{[5],[7]} serial number 11040681, date of installation, April 2012). The OSL readout of each detector is performed using an array of 38 green lightemitting diodes (LEDs) (532 nm wavelength) operated in continuous waveOSL mode during about 1 s. All the 38 LEDs are used for the stimulation of low doses (strong beam) whereas only 6 LEDs are used for high doses (weak beam). The dose range is evaluated before the measurement using one LED, and the cutoff point is an adjustable parameter.
Dosimeters
The dosimeter considered comprises a case and a slide. The latter contains four detector elements of Al_{2}O_{3}: C cut into round pieces of ~ 5 mm in diameter sandwiched between two layers of polyester for a total thickness of 0.3 mm. The detectors are located in read positions 1 (E1), 2 (E2), 3 (E3), and 4 (E4). When the slide is inside the case, each detector is positioned behind different filters providing different radiation attenuation conditions ^{[6],[8]} [Table 1] and [Figure 1]. The signal from each OSL detector is used in conjunction with a dose algorithm to evaluate different dosimetric quantities (Hp (10), Hp (0.07), and Hp (3)). This dose algorithm inherently uses individual calibration factors. For wider and homogeneous radiation fields, the chest OSLD badge (element E3) may act as a surrogate for eye lens doses.  Table 1: Summary of the recommended ranges of all the parameters that have been assessed in the study
Click here to view 
Methods   
Reader quality control test
There are three reader QC tests – Dark count (DRK), clinical attachment level (CAL), and LED measurements.^{[5]} DRK measures the inherent electronic noise generated by the photomultiplier tube (PMT). The DRK measurement is taken with the filter shutter closed and the LEDs off. CAL measures counts from a small exempted ^{14} C radioactive source embedded in a plastic scintillator. This provides a calibration of the sensitivity of PMT. LED measures counts from the PMT with the filter shutter open and the LEDS on to indicate the beam intensity. These three performance tests are usually performed with no dosimeter under the PMT.
For the purpose of this study, counting statistics was applied to CAL measurement as it is the only QC measurement that involves a radioactive source. Radioactive decay is a phenomenon known to be well described by counting statistics.^{[9]} In the approach followed here, a set of CAL measurements was recorded under conditions in which all aspects of the experiment were held as constant as possible. Due to the influence of statistical fluctuations, the results of these measurements were expected to show some degree of internal variation. The extent of this fluctuation was quantified and compared with predictions of a statistical model (Gaussian distribution). The observed fluctuations should be consistent with predictions otherwise the conclusion would be that the counting system is not functioning normally.
To do this comparison quantitatively, Chisquare test for standard deviations at 95% confidence level was used:
where and are the upper and lower limit values of the Chisquare confidence interval, respectively. N is the number of CAL measurements; s and σ are the sample (experimental) and predicted standard deviation, respectively. The values of and can be found in tables for several different confidence levels and degrees of freedom (N − 1), but due to the large number of measurements implied in this study, they were calculated using the software Matlab ^{®}, version R2013a (The MathWorks, Inc, USA). [Figure 2] shows the procedure that was followed.^{[10]}  Figure 2: The inspection of a set of data for consistency with a statistical model. The diagram is used to compare the experimental (F (x)) and the predicted (P ( x)) distributions. x_ and μ are the sample and predicted means, respectively. σ and s are, respectively, the predicted and experimental standard deviations. In the present work, it was considered that the best estimates of μ and σ were x_ and s, respectively
Click here to view 
Dose algorithm comparison
IEC 62387 standards state that the manufacturer shall state the general form of the model function for the measurement with the dosimeter.^{[11]} However, Landauer ^{®}, the InLight OSL dosimetry system manufacturer and provider, keeps its dose algorithm confidential for commercial reasons. Nevertheless, Landauer can provide information on the model function on request of its clients.^{[12]} Based on information provided, a constructed dose algorithm that takes into account each reading position E_{i} (i = 1, 2, 3, 4) explicitly is proposed in the present work. This could be used, in conjunction with quality control dosimeters, as an additional tool for quality control as it could verify whether the results provided by the system dose algorithm are proper. The dose algorithm, based on Equation 2 and correction factors derived from Gabon data for 2013 and 2016 IAEA regional intercomparison exercises on measurements of the personal dose equivalent Hp (10) in photon fields,^{[3]} was built using a program written in Matlab ^{®} R2013a. The data involved results from linearity response test and energy and angular response test.
where CV is converted values, that is counts corrected for the sensitivity of the detector at position E_{i} (i = 1, 2, 3, 4) and calibration coefficient of the reader; S is the sensitivity of the dosimeter. In fact, it is the sensitivity of each detector at positions E_{i}, i = 1, 2, 3, 4.^{[6]} The calibration coefficient CF is obtained from the system algorithm and is given by the following formula: ^{[5],[7]}
where M = 12 is the number of exposed dosimeters from the reader calibration set. N =4 is the number of reading positions per dosimeters, C_{i} is the counts per reading position, S_{j} is the sensitivity of dosimeter j, and E_{j} is the nominal exposure for dosimeter j. The average background B_{Avg} is given by
where L = 3 is the number of unexposed dosimeters from the reader calibration set.
The proposed dose algorithm is an “if” algorithm with 4 linear branches corresponding to 4 energy ranges. The latter is determined by the quotient of CV_{1} (element detector 1) by CV_{4} (element detector 4). Four ranges of CV_{1}/CV_{4} corresponding to energy intervals centered about 48 keV, 64 keV, 100 keV, and 662 keV, respectively, were determined. For each branch, the dose is given by
Dose = k_{1} CV_{1} + k_{2} CV_{2} + k_{3} CV_{3} + k_{4} CV_{4 }(5)
where k_{1}, k_{2}, k_{3}, and k_{4} are correction factors.
Workplace data were used for the comparison of the proposed and system dose algorithms. These data mainly come not only from the medical sector (4 public hospitals and 5 private hospitals and clinics) but also from the mining sector (1 facility) and the industrial sector (1 facility).
Signal depletion
The purpose of this test was to assess the depletion rate per read out of InLight dosimeters of the national IMS of Gabon and to compare it with values provided by other studies. Three sets of dosimeters were used for this test. The first set was composed of 4 InLight dosimeters of XA case type (i.e., dosimeters having case serial number starting by the letters XA; they are mainly used as field dosimeters) and 10 mSv nominal dose. The other two sets consisted of three dosimeters each of CC case type (i.e., dosimeters for which the case serial number starts with the letters CC; they are mainly used for the calibration of the reader) of 5 mSv and 500 mSv nominal doses.
An approach making use of arithmetic progressions as mathematical model can be followed. The equation to be considered is the following
u_{n} = u_{0} − n × r(6)
where u_{n} is the signal on reading number n, u_{0} is the signal before any reading depletion, and r is the common difference (fraction by which the signal is depleted per reading). This is a linear equation with slope r and intercept u_{0 }=_{ }1.
Results and Discussion   
Reader quality control test
The procedure described in [Figure 2] was applied to 10586 values collected for CAL measurement. [Figure 3] presents the results obtained following this procedure. The experimental parameters (mean and standard deviation) were used as the predicted parameters. It is observed that the shapes of experimental and predicted distributions are similar. [Table 2] shows that the predicted standard deviation is within the required interval. Therefore, the CAL measurement passes the Chisquare test. Finally, it is concluded that the experimental fluctuations are consistent with predictions and that the counting system performs well. However, this needs to be confirmed by the use of control charts and data collected for all the three reader QC measurements for a period much >3 months. It is also worth mentioning that the experimental coefficient of variation was ≈ 4% as the experimental standard deviation and mean values were 36.60 and 975.31 counts, respectively.  Figure 3: Results of counting statistics applied on clinical attachment level measurement. The shapes of the experimental and predicted distributions are similar
Click here to view 
 Table 2: Uncertainty budget of the reference measured dose of the 2013 International Atomic Energy Agency intercomparison exercise. A relative overall uncertainty of ±33.85% ([0.713/2.107]×100) was obtained
Click here to view 
Dose algorithm comparison
Workplace data extracted from the dosimetry system software microStar and recorded from September 2012 to 2017 (3847 measurements) were used for the comparison. [Figure 4] presents the plot of dose estimated by the proposed algorithm against the dose measured by the system algorithm. The first observation is that most of the data points are located in the region of the recording level H_{0 }=_{ }0.05 mSv (the lower limit of detection = 0.042 mSv). Most importantly, [Figure 4] shows that the proposed dose algorithm provides a good estimate of the dose compared with the system dose algorithm. In particular, it is observed that data points are distributed about the line Y = X, where X is the dose measured by the system algorithm.  Figure 4: Plot of the dose estimated by the proposed algorithm against the dose measured by the system algorithm. It is observed that the proposed dose algorithm provides a good estimate of the dose compared with the system dose algorithm
Click here to view 
Signal depletion
From [Figure 5], it was observed that the signal depletion follows an exponential decay law. Each dosimeter was read more than 350 times (meaning that more than 3500 readings were taken for all the 10 dosimeters involved). Another observation is that dosimeters irradiated with high doses (>100 mSv) provide smaller depletion rate per reading than those irradiated with low doses. This is normal as dosimeters irradiated with doses lower than 100 mSv are depleted using the strong stimulation beam (all the 38 LEDs) while dosimeters irradiated with doses above 100 mSv are depleted using the weak beam (only 6 LEDs). The depletion rate values determined in the present study are comparable to those determined in other studies.^{[5],[6],[7],[8],[13]} The value obtained for the strong beam is even clearly better.  Figure 5: Results of depletion signal test for the set composed of dosimeters of CC case type (case serial numbers: CC00015548Z, CC000156073, and CC000156172) of 5 mSv nominal dose. Each data point represents the average of 3 dosimeters
Click here to view 
In [Table 3], the exponential and linear expressions for the depletion rate per reading were in good agreement even though the number of dosimeters read was not the same.  Table 3: Results for photon energy and angle of incidence according to the IEC 62387 standard using data from 2013 International Atomic Energy Agency intercomparison exercise. The required limits are 0.69 and 1.97 and 0.71 and 1.80 for 48 keV and energies ≥65 keV, respectively
Click here to view 
The arithmetic progression model also provides good results. If, for example, the mean value for r (0.004) obtained for the set of dosimeters of XA case type is used, one obtained u_{n}= −0.004n + 1, equation which is in agreement with the corresponding equation given in [Table 3].
Finally, the user manual states that each calibration set should be used up to 10 times, meaning that 98% of the signal is still present in the calibration dosimeters after 10 readings (the depletion rate per reading = ~0.2%). For the present study (depletion rate per reading = 0.3%, strong beam), the number of reading for which 98% of the signal is still present in the calibration dosimeters is approximately 7. Therefore, the calibration set should not be read more than 7 times.
Conclusions   
The reader performance test showed that the PMT of the reader is performing selfconsistently and that there is no abnormality in the counting system. This, however, needs to be confirmed by the use of control charts for all the reader QC measurements. It was also shown that the dose algorithm proposed in this study provides a good estimate of the dose compared with the system dose algorithm. This result could be used for quality control purposes. The signal depletion rate per reading determined in this study for both the strong and weak stimulation beams is comparable to values determined in other studies.
Acknowledgments
The authors would like to thank the General Directorate of Radiation Protection and Nuclear Safety, Ministry of Water and Energy, in Gabon for its operational support.
Financial support and sponsorship
Nil.
Conflicts of interest
There are no conflicts of interest.
References   
1.  International Atomic Energy Agency. Occupational Radiation Protection. Draft Safety Guide DS453. International Atomic Energy Agency; 2014. 
2.  International Atomic Energy Agency. Radiation Protection and Safety of Radiation Sources: International Basic Safety Standards. General Safety Requirements Part 3. IAEA Safety Standards No GSR Part 3. International Atomic Energy Agency; 2014. 
3.  Arib M, Herrati A, Dari F, Ma J, LounisMokrani Z. Intercomparison 2013 on measurements of the personal dose equivalent Hp (10) in photon fields in the African region. Radiat Prot Dosimetry 2014;202:18. 
4.  Ondo Meye P, Schandorf C, Amoako JK, Manteaw PO, Amoatey EA, Adjei DN. Intercomparaison on the measurement of the quantity personal dose equivalent Hp (10) in photon fields. Linearity Dependence, Lower Limit of Detection and Uncertainty in Measurement of Dosimetry Systems of Individual Monitoring Services in Gabon and Ghana. Radiat Prot Dosimetry 2017;38:111. 
5.  Landauer. InLight® Systems – MicroStar® User Manual. Landauer, Inc.; 2009. 
6.  Landauer. InLight Model 2 Dosemeter, Characterization and Uncertainty Analysis for National Dosimetry Services (NDS), Health Canada. Rev. 004. Landauer, Inc.; 2009. 
7.  Yukihara EG, McKeever SW. Optically Stimulated Luminescence – Fundamentals and Applications. Chichester, UK: Wiley; 2011. 
8.  Passmore C, Kirr M. InLightTM OSL Training. Module 1 – Fundamentals. Glenwood, Illinois, USA: Landauer; 2011. 
9.  Turner JE. Atoms, Radiation, and Radiation Protection. 3 ^{rd} ed. Completely Revised and Enlarged Edition. Weinheim, Germany: Wiley; 2007. 
10.  Knoll GF. Radiation Detection and Measurement. 3 ^{rd} ed. New York, USA: Wiley; 2000. 
11.  International Electrotechnical Commission. Radiation Protection Instrumentation – Passive Integrating Dosimetry Systems for Personal and Environmental Monitoring of Photon and Beta Radiation. International Standard IEC 62387. International Electrotechnical Commission; 2012. 
12.  Perks CA, Passmore CN. The InLight Dosemeter – Comparison with IEC 61066. TR01, Revision 2.0. Landauer Inc.; 2006. 
13.  Jursinic PA. Characterization of optically stimulated luminescent dosimeters, OSLDs, for clinical dosimetric measurements. Med Phys 2007;34:4594604. 
[Figure 1], [Figure 2], [Figure 3], [Figure 4], [Figure 5]
[Table 1], [Table 2], [Table 3]
