X-ray imaging dosimeter performance in standard and non-standard radiography radiation fields in terms of air kerma

Original paper X-ray imaging dosimeter performance in standard and non-standard radiography radiation fields in terms of air kerma Nikola Krˇ zanovi´ c a , Paula Toroi b,c,* , Ivana Komatina a,d , Bartel Jansen e , Argiro Boziari f , Margarida Caldeira g , Maysa Costa de Castro h,i , Alessia Ciccotelli j , Ana Fernandes g , Andrea Koji´ c a,k , Kam Lee l , Carita Lindholm b , Stefan Pojtinger h , Erinc Reyhanoglu m , Luigi Rinaldi j,n , Amra ˇ Sabeta o , Elisabeth Salomon p,h , Siarhei Saroka q , Claudia Silvestri j , Vladimir Sochor r , Viivi Valkama b , Jelena Vlahovi´ c a,s , Miloˇ s ˇ Zivanovi´ c a a Vinca Institute of Nuclear Sciences – National Institute of Serbia, University of Belgrade, Belgrade, Serbia b Radiation and Nuclear Safety Authority, Vantaa, Finland c Helsinki University Hospital Diagnostic Center, Radiology, University of Helsinki, Helsinki, Finland d Faculty of Physical Chemistry, University of Belgrade, Belgrade, Serbia e VSL National Metrology Institute, Delft, the Netherlands f Elliniki Epitropi Atomikis Energeias, Athens, Greece g Instituto Superior T´ ecnico, Lisbon, Portugal h Physikalisch-Technische Bundesanstalt, Braunschweig, Germany i Federal University of Pernambuco, Recife, Brazil j Istituto Nazionale di Metrologia delle Radiazioni Ionizzanti - Agenzia Nazionale per le nuove tecnologie, l’energia e lo sviluppo economico sostenibile, Rome, Italy k Faculty of Physics, University of Belgrade, Belgrade, Serbia l Australian Radiation Protection and Nuclear Safety Agency, Victoria, Australia m Türkiye Enerji, Nükleer ve Maden Aras¸tırma Kurumu, Ankara, Turkey n Scientific Institute for Research, Hospitalization and Health Care, Bambino Gesù Children’s Hospital, Rome, Italy o Institut za mjeriteljstvo Bosne i Hercegovine, Sarajevo, Bosnia and Herzegovina p Medical University of Vienna, Center for Medical Physics and Biomedical Engineering, Vienna, Austria q Institutul Nat ¸ional de Metrologie, Chisinau, Moldova r Cesky Metrologicky Institut, Brno, Czech Republic s Faculty of Sciences, University of Novi Sad, Novi Sad, Serbia ARTICLE INFO Keywords: Dosimeter response Ionization chamber Radiography Solid-state detector X-ray multimeter ABSTRACT Introduction: X-ray medical imaging developments have introduced needs for updated dosimetry practices. Methods: Performance of commercially available dosimeters used for air kerma measurements in diagnostic and interventional radiology was examined. Ionization chambers and X-ray multimeters were tested in a wide range of air kerma rates, photon energies (using standard and non-standard radiation qualities), and angles of incidence with different dosimeter orientation and rotation. Stability and repeatability of the measured value, the influence of pulse duration, non-linearity of dosimeter response, energy and angular dependence were studied against the IEC 61674:2024 limits of variation. Energy response was tested using the standard RQR and RQT radiation qualities defined in IEC 61267:2005, as well as non-standard copper-filtered beams with added 0.9 mm Cu filtration. Results: Most dosimeters complied with the IEC 61674:2024 standard limits of variation, for both standard and non-standard radiation fields. In some cases, observed performance was significantly better than the current limits allowing for the introduction of more stringent values. Conclusion: Modification of the performance requirements was proposed, considering differences between reference-class and field-class dosimeters, while introducing more stringent requirements for reference-class dosimeters. * Corresponding author at: Radiation and Nuclear Safety Authority, Vantaa, Finland. E-mail address: [email protected] (P. Toroi). Contents lists available at ScienceDirect Physica Medica journal homepage: www.elsevier.com/locate/ejmp https://doi.org/10.1016/j.ejmp.2025.105687 Received 11 July 2025; Received in revised form 10 October 2025; Accepted 3 December 2025 Physica Medica 141 (2026) 105687 1120-1797/© 2025 Associazione Italiana di Fisica Medica e Sanitaria. Published by Elsevier Ltd. This is an open access article under the CC BY license ( http://creativecommons.org/licenses/by/4.0/ ). 1. Introduction Frequency of medical imaging procedures in diagnostic and interventional radiology has increased with time, leading to a corresponding rise in the cumulative dose for the population. The purpose of diagnostic and interventional procedures is to produce medical images of satisfying quality while delivering patient doses as low as reasonably achievable (ALARA) [1,2]. Many of the quality control (QC) tasks are performed through the measurement of the X-ray tube output in terms of air kerma (rate), either by in-house medical physicists or third-party dosimetry services. These data also provide the medical physicists with means to perform optimization of medical imaging procedures and patient doses. The use of fit for purpose high quality equipment (with reduced sensitivity to influence quantities) would improve accuracy of the acquired dosimetry data, which would have significant clinical implications by improving QC, optimizing diagnostic radiology procedures and reducing patient exposure. The goal of these activities is to ensure that clinical information of sufficient quality can be extracted, and patient doses can be optimized. The incident air kerma is one of metrics used as patient dose indicator in diagnostic radiology [3,4], and it can be further used for patient dose estimations. Ionization chambers (ICs) were commonly used for quality control of X-ray units in different modalities of medical X-ray imaging, including diagnostic and interventional radiography. Currently, quality control procedures related to dosimetry-based quantities and measurements can be performed either with ionization chambers or solid-state detectors, while for other QC tests different instruments are utilized (e.g., collimation/beam alignment test tools). Diagnostic radiology dosimeters which employ vented ionization chambers used for incident air kerma or air kerma rate measurements are usually realized either as spherical, cylindrical or plane-parallel detectors. Historically, plane-parallel ionization chambers were predominantly used for diagnostic radiology measurements (along with passive dosimeters and solid-state detectors), while spherical ionization chambers were usually designed for scattered radiation measurements [5]. Today, both plane-parallel and spherical ionization chambers are suitable for diagnostic radiology measurements due to their relatively constant response to variations in photon energy. Ionization chambers are commonly used with an associated electrometer measuring assembly. X-ray multimeters (XMMs) are based on semiconductor detectors which enable simultaneous measurement of several quantities and parameters of interest, including air kerma, X-ray tube voltage, half-value layer (HVL), total filtration (TF), and other. Due to the possibility of measuring several quantities simultaneously, these devices are being more frequently used in QC procedures. XMMs are realized either as multi-element or single-element detectors, which are connected to a designated electrometer module (EMM) and display unit. These dosimeters can usually be used in a wide range of radiation conditions, since their performance is dependent on the incorporated manufacturer defined algorithm which compensates for energy dependence, with further adjustments of the associated software settings for each individual XMM. The design of XMM detector unit usually includes lead shielding, reducing the contribution of scattered radiation which originates from the objects and equipment in the clinical setup (e.g., patient table). Therefore, XMMs are designed to measure the incident air kerma instead of the entrance surface air kerma (ESAK), which includes backscattered radiation [6]. It is common practice that XMMs are adjusted by the manufacturers, based on the radiation conditions related to clinical needs of medical physicists, which are often not aligned with the standardized radiation fields used in calibration laboratories. Adjustment of a measuring instrument represents a set of operations carried out on the device (e.g., the parameters of the implemented dose calculation algorithm) so that the displayed value corresponds to the measured value of a reference standard [7]. As an addition to manufacturer-specific XMM adjustment, standard dosimetry laboratories (abbreviated in the paper as SDLs, and including both primary standard −PSDLs and secondary standard −SSDLs) provide calibration services for these devices in standard radiation fields. Conditions under which dosimeter calibration is performed do not always correspond to the clinical settings, where the QC is performed. Calibration under reference conditions improves the quality of the dosimetry data by providing traceability to the relevant primary standard, ensuring good long-term stability and performance of the detectors and can help in identifying any malfunctions. Reference conditions for calibration and testing of diagnostic radiology dosimeters are defined in IEC 61267:2005 [8]. This standard was issued by the International Electrotechnical Commission (IEC) subcommittee 62C which deals with equipment used for radiotherapy, nuclear medicine and radiation dosimetry. The reference radiation conditions are described parametrically through X-ray tube voltage, first half-value layer and homogeneity coefficient. As per the IEC 61267:2005 standard, tube voltage and HVL values defined for many reference radiation conditions are non-nominal, so the additional filtration must be modified to achieve these values within stated tolerance (e.g., ratio of air kerma with absorber of HVL-equivalent thickness and with no absorber should be within 0.485 – 0.515) [8]. Reference radiation fields include aluminium filtered and copper filtered radiation qualities in general radiography. Current technological advancements in medical imaging have led to the situation where the present IEC radiation conditions defined for radiography-based procedures are not well representative of the radiation fields encountered in clinical settings. Consequently, traceability needs to be established for non-standard radiation fields. Apart from diagnostic radiology, special consideration should be taken for interventional radiology procedures, due to their significant contribution to patient doses. Previously conducted research addressed this matter through the development of two non-standard copper-filtered radiation qualities [9]. The properties of these radiation fields were determined based on the analysis of the manufacturer specifications of the X-ray generators used for interventional radiology and cardiology procedures. Based on the sample of 180 Radiation Dose Structured Reports, the most frequently added copper filtration thicknesses were 0.1, 0.3 and 0.9 mm Cu. Therefore, two radiation qualities which utilize the extreme value of 0.9 mm copper filtration were developed. XMM performance in these radiation fields was evaluated, relative to the standard radiation quality RQR8 (defined by 100 kV nominal voltage and 3.97 mm Al first HVL). The previously mentioned research highlighted the need for establishing new radiation qualities in calibration laboratories, which correspond to conditions encountered in clinical settings. Calibration of dosimeters in such radiation fields would improve the QC procedures by reducing the measurement uncertainty. Hourdakis et al., [10] have investigated the performance of ICs and XMMs in reference radiation qualities under laboratory conditions and in the clinical radiation fields, with the goal of evaluating the influence of different exposure conditions on their response. They have observed that the IEC compliant dosimeters which are used in line with their manufacturer specifications can be used in clinical conditions, with the dose measurement error less than 3 % if adequate energy dependence corrections are applied in typical clinical conditions. Research performed by [11] and [12] presented evaluation of XMMs and ICs performance across standard radiation quality series, covering a wide range of operating conditions encountered in diagnostic radiology. Versatility outside of manufacturer-stated specifications was investigated, highlighting the potential for use in diverse clinical environments. Some of the previous research was focused on the performance of XMMs in mammography radiation fields, where it was concluded that these devices can be used for quality control in non-standard radiation fields due to observed low energy dependence of response in terms of air kerma for different anode/filtration setups. If these devices were to be N. Krˇ zanovi´ c et al. Physica Medica 141 (2026) 105687 2 used for quality control in conditions other than used for calibration, this should be considered in the measurement uncertainty budget [13–15]. This research has been conducted under the scope of the joint research project 22NRM01 TraMeXI: Traceability in Medical X-ray imaging dosimetry. The aims of this project consist of assessing currently available calibration procedures and the performance of dosimeters used in clinical applications, for the purpose of developing new and updated calibration procedures and implementing them into standards, guidelines and protocols. In this paper the performance of ICs and XMMs for the measurement of air kerma was investigated, in terms of their response to different influence quantities, including energy dependence, angular dependence, non-linearity, dosimeter stabilization and repeatability and pulse duration. A comprehensive study emphasizing the performance of diagnostic dosimeters in standard radiation fields was done, considering the effect of several influence quantities on air kerma measurements with both reference class and field class dosimeters, as well as the performance of some dosimeters in non-standard copper filtered radiation fields. A proposal for changes in the limits of variation for different influence quantities was given, as well as the differentiation between the requirements set for reference class and field class dosimeters was introduced. Implementation of proposed changes would impact the quality of acquired diagnostic dosimetry data, leading to better optimization of diagnostic radiology procedures and reduced patient doses. 2. Materials and methods 2.1. Radiation qualities The IEC 61267:2005 [8] defines four distinct radiation quality series which describe primary X-ray beams in radiography. Aluminium filtered radiation qualities which are commonly used for dosimeter calibration in SDLs are abbreviated as RQR. These radiation qualities cover X-ray tube voltages from 40 to 150 kV with HVLs ranging from 1.42 to 6.57 mm Al (Table 1). Typically, diagnostic radiology dosimeters are calibrated in the RQR 5 and/or RQR 8 radiation qualities in conventional radiography [16] and RQR 5 is usually designated as the reference radiation quality [17]. It is important to stress that medical physicists in hospitals may use XMMs which have undergone certain adjustments by manufacturers, specific for their clinical needs, not complying with the standard reference radiation qualities. Besides the Al-filtered beams, two series of radiation qualities which utilize copper filtration are also defined in the standard. The RQT radiation quality series is defined in the X-ray tube voltage range from 100 kV to 150 kV with HVL values ranging from 6.9 to 10.1 mm Al and nominal added copper filtration ranging from 0.2 to 0.3 mm Cu, respectively (Table 1). The RQT radiation qualities are commonly used in SDL calibration procedures of pencil-type ionization chambers (abbreviated as CT-chambers) used for computerized tomography dosimetry. The RQC radiation quality series utilize large copper filtration (0.5 – 2.0 mm Cu) in the X-ray tube voltage range from 50 kV to 100 kV. These radiation fields are not commonly used for dosimeter calibration since they do not correspond well to the clinically used fields. Both RQT and RQC radiation qualities are established by adding a specific thickness of copper filtration to the corresponding RQR radiation qualities. For example, while both RQT8 and RQC 8 have the same nominal voltage of 100 kV, the additional copper filtration for these radiation qualities is 0.2 and 2.0 mm Cu respectively. The copper filtration is added to the RQR 8 beam quality to realize the RQT 8 or RQC 8 beam qualities. A market analysis of the available X-ray generators used for general and interventional radiography was performed under the scope of TraMeXI project, through the direct contact with the manufacturers and based on the manufacturer specifications, with the goal to propose new radiation qualities which more closely correspond to the clinical conditions. These radiation qualities would incorporate copper filtrations of thicknesses in the range from 0.1 mm Cu up to 0.9 mm Cu, updating the current set of RQC radiation qualities. Three of the new clinically relevant and physical representative radiation qualities (abbreviated as CPRQs) have been included in this study. CPRQs used in this research, CPRQ 5, CPRQ 8 and CPRQ 9, are built upon the RQR 5, RQR 8 and RQR 9 beam qualities, by introducing additional 0.9 mm Cu filtration to the respective RQR radiation qualities. These radiation qualities were established by measuring the first and second HVLs and by comparing these values with the values generated by SpekPy Web (built on SpekPy v.2.0.8) [18–20]. First HVLs for these radiation qualities were determined according to the standard procedure described in the International Atomic Energy Agency Technical Report Series document, IAEA TRS 457 [17]. Comparison of parameters which describe RQR, RQT and RQC radiation quality series with selected CPRQs is presented in Table 1. For mammography applications, reference radiation conditions are defined for molybdenum (Mo) anode/filtration setup in the X-ray tube voltage range from 25 to 35 kV with respective HVL values ranging from 0.28 to 0.36 mm Al [8]. On the other hand, clinically encountered mammography X-ray generators utilize additional different anode/ filtration setups (e.g., W/Ag, W/Rh, and other), expanding the HVL range for the same X-ray tube voltages up to approximately 0.6 mm Al [14,15]. 2.2. Ionization chambers A total of 16 ionization chambers were examined in this research, covering 7 different detector models from different manufacturers. The results were collected and evaluated by the Vinca Institute of Nuclear Sciences. The list of ionization chambers with associated dosimetry laboratories is presented in Table 2. Ionization chambers included in this study are the 3.6 cm 3 spherical ionization chamber Exradin A3 (Standard Imaging) which is commonly used in dosimetry laboratories as a Table 1 Properties of standard RQR, RQT, RQC [8] beam qualities and non-standard copper-filtered radiation qualities (CPRQ) used in this research, where RQ – radiation quality. Nominal values of aluminum and copper filter thickness are given, as well as the corresponding HVL values in terms of mm Al. U (kV) RQR [8] RQT [8] RQC [8] CPRQ (non-standard) RQ HVL (mm Al) RQ HVL (mm Al) RQ HVL (mm Al) RQ HVL (mm Al) 40 RQR 2 1.42       50 RQR 3 1.78   RQC 3 4.5   60 RQR 4 2.19       70 RQR 5 2.58   RQC 5 8.4 CPRQ 5 7.65 80 RQR 6 3.01       90 RQR 7 3.48       100 RQR 8 3.97 RQT 8 6.9 RQC 8 11.5 CPRQ 8 10.1 120 RQR 9 5.00 RQT 9 8.4   CPRQ 9 11.3 150 RQR 10 6.57 RQT 10 10.1     N. Krˇ zanovi´ c et al. Physica Medica 141 (2026) 105687 3 standard instrument for conventional radiography; the 3.46 cm 3 Exradin Magna A650 (Standard Imaging) which is a plane-parallel ionization chamber used for both general radiography and mammography; the 6 cm 3 10X6-6 (Radcal) is a cylindrical ionization chamber used for a wide range of photon energies encompassing the range from 20 keV up to 1.25 MeV. The 6 cm 3 RC6M (Radcal) plane-parallel ionization chamber is designated for mammography-specific low-energy X-ray measurements up to 40 keV. A thin-window plane-parallel 0.2 cm 3 PTW 23344 (PTW) ionization chamber for low-energy X-ray radiotherapy calibrations was also included in the study. Even though this is a radiotherapy ionization chamber, according to the manufacturer specifications the chamber can be used in the energy range from 8 keV up to 35 keV. The other two plane-parallel ionization chambers included in this study are the 6 cm 3 PTW 34069 and the 75 cm 3 PTW 34060, which are used for absolute dosimetry in diagnostic radiology/mammography and diagnostic radiology, respectively. As per manufacturer specifications, these ionization chambers can be used in front of or behind a patientequivalent phantom in diagnostic procedures. 2.3. X-ray multimeters A total of 24 XMMs (9 dosimeter models) from different manufacturers were included in this study, including state-of-the-art dosimeter models, as well as some of the old previous models. In Table 3 the list of XMMs is provided, along with the respective manufacturer specifications in terms of air kerma (rate) measurements, and the calibration dosimetry laboratory. 2.4. Dosimeter response and limits of variation Performance of diagnostic radiology dosimeters is evaluated according to the IEC 61674:2024 standard [21], which defines the performance tests with associated compliance criteria expressed in terms of limits of variation. The dosimeter response is defined as the quotient of the indicated (measured) value by the diagnostic dosimeter under test and the conventional true value determined by the standard instrument (ionization chamber). If the deviation in the dosimeter response caused by variation of certain influence quantities is within certain testdependent limits, compliance of the dosimeter with the standard is achieved. Testing of the effect of a certain influence quantity is performed while maintaining other influence quantities at their respective reference values (in line with the standard defined reference conditions), or by performing necessary corrections accounting for these effects. Relative response at a specific value of an influence quantity is derived by normalization of the response at that value to the response value determined for the reference conditions. During the tests with the XMMs, the reference values were defined by the SDL secondary or primary standard ionization chambers. On the other hand, performance tests of ICs presented in terms of response, were done by comparing the measured value with the reference value determined by the ionization Table 2 Ionization chambers included in the study, along with the respective dosimetry calibration laboratories (SDL). IC model Intended use IC type Active volume IC No. (SDL) Symbol Exradin A3 radiography spherical 3.6 cm 3 IC1 (CMI) IC2 (IMBiH) IC3 (EEAE) IC4 (PTB) IC5 (VSL) Exradin Magna A650 mammography radiography plane-parallel 3.46 cm 3 IC6 (ENEA) Radcal RC6M mammography plane-parallel 6 cm 3 IC7 (CMI) IC8 (ENEA) IC9 (PTB) Radcal 10X6-6 radiography cylindrical 6 cm 3 IC10 (STUK) PTW 34069 mammography plane-parallel 6 cm 3 IC11 (IMBiH) IC12 (TENMAK) IC13 (IST) PTW 34060 radiography plane-parallel 75 cm 3 IC14 (TENMAK) IC15 (ARPANSA) PTW 23344 superficial radiotherapy plane-parallel 0.2 cm 3 IC16 (TENMAK) N. Krˇ zanovi´ c et al. Physica Medica 141 (2026) 105687 4 chambers and free air chambers of the higher order in the metrological traceability chain. 3. Performance tests 3.1. Repeatability and resolution Repeatability of the dosimeter indication was evaluated using the reference radiation quality RQR 5 for both XMMs and ICs. According to the standard [21] if the air kerma rate is equal to or greater than 100 µGy•s −1 in the primary beam, the coefficient of variation should be less than or equal to 3 %. This test was performed for three different situations: at the typical air kerma rate used for calibration; at the lower limit of air kerma rate, based on the capabilities of the SDLs; at the air kerma rate of 100 µGy•s −1 . It should be noted that for the dose rates less than 100 µGy•s −1 in the unattenuated beam, the criterion can be more relaxed, up to 5 % (for e.g., 1 µGy•s −1 ). At these dose rates the limit is dose rate dependent and determined with the equation presented in Table 3 of the standard [21]. Data on resolution of diagnostic dosimeters was collected during the repeatability test, as this property of the dosimeter represents an important limitation regarding measurement uncertainty. Resolution represents the smallest significant change in the dosimeter indication. Requirement on dosimeter resolution set by the standard is that the resolution should not be greater than 1 % over the effective measurement range [21]. 3.2. Dosimeter stabilization The effect of stabilization time on the dosimeter indication was investigated by performing a series of measurement commencing with different time lapse intervals after the dosimeter is switched on. The time lapse intervals used were <1 min, ~15 min, ~30 min, ~45 min and ~60 min. The dosimeter is in compliance with the standard if the relative error is in agreement with the following condition: δ( τ ) = R( τ ) − Rref Rref ≤L(1) where R( τ )represents the value of the dosimeter response after the certain time interval, Rref is the response reference value at 60 min (approximation of steady state), and L is the ±2 % limit of variation, as defined by the IEC standard [21]. 3.3. Linearity of the response Response linearity of ICs and XMMs in terms of air kerma rate measurements was examined in a wide range of dose rate values, where the selected values were dependent on the capabilities of each SDL. The dose rate ranged from 5 µGy•s −1 up to 0.1 Gy•s −1 with reference air kerma rate being 0.5 mGy•s −1 . A diagnostic dosimeter is compliant with the standard if its highest and lowest response values over the tested dose rate range fulfil the following condition: Rmax −Rmin Rmax +Rmin ≤L(2) where Rmax and Rmin represent the maximum and minimum values of the dosimeter response determined over the whole test range, and L represents the limit of variation of ±2 % [21]. 3.4. Variation in response due to photon energy Exposure of diagnostic radiology dosimeters to radiation fields in clinical settings may differ from the standard defined reference radiation qualities established in dosimetry calibration laboratories. The changes in the incident X-ray spectra (caused either by the changes in incident Table 3 X-ray multimeters included in the study, their manufacturer specifications for air kerma (rate) measurements and the respective dosimetry calibration laboratory (SDL). XMM Air kerma (rate) range Uncertainty XMM No. (SDL) Symbol Raysafe X2 R/F 1 nGy•s −1 – 500 mGy•s −1 1 nGy – 10 kGy 5 % XMM1 (ARPANSA) XMM2 (IST) XMM3 (STUK) XMM4 (VINS) XMM5 (UFPE) Raysafe Xi R/F 20 µGy•s −1 – 1 Gy•s −1 10 µGy −10 kGy 5 % XMM6 (STUK) XMM7 (VINS) XMM8 (EEAE) Raysafe ThinX RAD 0.1 mGy•s −1 – 100 mGy•s −1 20 µGy −1 Gy 5 % XMM9 (VINS) RTI Piranha 10 µGy•s −1 – 450 mGy•s −1 0.7 µGy −1 kGy 5 % XMM10 (IST) XMM11 (VINS) XMM12 (UFPE) XMM13 (CMI) XMM14 (EEAE) RTI Mako 1 nGy•s −1 – 500 mGy•s −1 1 nGy – 10 kGy 5 % XMM15 (STUK) PTW Nomex 5 µGy•s −1 – 500 mGy•s −1 50 nGy – 500 Gy 3.5 % XMM16 (ARPANSA) XMM17 (INM) Radcal AGMSDM+/ AGMSD+ 40 nGy•s −1 – 200 mGy•s −1 40 nGy – 100 Gy 5 % XMM18 (STUK) XMM19 (ENEA) XMM20 (UFPE) IBA MagicMax 100 nGy•s −1 – 160 mGy•s −1 150 nGy – 50 Gy 5 % XMM21 (VSL) XMM22 (PTB) Quart didoNEO 0.5 µGy•s −1 – 1 Gy•s −1 1 µGy – 1 Gy 5 % XMM23 (VSL) XMM24 (PTB) N. Krˇ zanovi´ c et al. Physica Medica 141 (2026) 105687 5 beam primary filtration, or the X-ray tube voltage) can lead to variation in dosimeter response. In general, the effect of energy dependence of dosimeter response is more pronounced with solid-state detectors compared to ICs. XMMs, being based on solid-state detectors, usually have sets of filters and manufacturer-defined software corrections to compensate for variations in beam energies. In some cases, user input is needed (e.g. to select tube voltage range, or to select target/filter combination for mammography applications). The energy dependence of ICs and XMMs was examined by determining the relative energy response, normalized to the reference radiation quality RQR 5 for all radiation quality series used in the tests. The test was performed using the RQR and RQT series of radiation qualities [8]. In some of the SDLs the performance of dosimeters was also evaluated in the non-standard radiation fields. The traceability of the reference value was obtained either through direct traceability to the primary standard or by performing HVL-based interpolation of the secondary standard calibration coefficient. Relative energy response of the dosimeters is defined by the following equation: r(E) = R(E) R0(E0)=M(E)/KR(E) M(E0)/KR(E0)(3) where R represents the dosimeter response at a certain radiation quality E, and R0 is the response at the reference radiation quality E 0 . KR represents the reference air kerma, while M is the measured value. The limits of variation for the relative energy response are set to ±5 % for the radiation qualities in the range from RQR 3 (50 kV) to RQR 10 (150 kV) [21], while the RQR 2 is not included in the rated range. For mammography specific ionization chambers (IC6, IC7, IC8 and IC9) the energy dependence of response was evaluated in the X-ray tube voltage range from 25 kV to 35 kV, for each of the anode/filtration setups separately, by normalizing the response to the respective reference response at 28 kV. Available anode/filtration setups are different between dosimetry calibration laboratories. Relative response limits of variation are set to ±5 % for each anode/filtration setup [21]. 3.5. Variation in response due to angle of incidence Angular dependence of dosimeter response was tested in two perpendicular planes (horizontal and vertical dosimeter orientation, as shown in Fig. 1) by performing rotations in both directions, within the rated range, up to the angle of incidence of ±5◦. The variation in response was evaluated by performing normalization to the response at 0◦. Limits of variation of ±3 % are defined in the standard for the minimum rated range [21]. Besides testing in these two dosimeter orientations, the effect of dosimeter rotation was also evaluated. For this test, angles from 0◦to 360◦were used, with the dosimeter facing the X-ray source over the whole duration of the test (Fig. 1). Normalization of dosimeter response relative to the response at 0◦was performed. Dosimeter rotation test is not included in the current version of the standard for air kerma measurement [21]. Although, this effect is considered as one of the tests defined in the standard related to measurement of X-ray tube voltage with non-invasive instruments, IEC 61676:2023 [22]. 3.6. Effect of pulse duration Variation of the dosimeter response due to the effect of pulsed radiation was examined by two SDLs (VSL and PTB). Pulse duration was varied from 200 ms up to 20 s, where the selected reference pulse duration was 10 s [23]. The minimum pulse duration defined in the IEC standard [21] of 1 ms was not included in this study due to the limitations of the laboratory X-ray generators. The limits of variation for this effect are defined in the same manner as for the linearity test (Section 3.3). 4. Results Properties of diagnostic radiology dosimeters for the measurement of air kerma (rate) were evaluated by performing several performance tests following the compliance tests defined by IEC 61674:2024 [21], in terms of limits of variation of dosimeter response. Legends in all figures were listed in Table 2 for ICs and Table 3 for XMMs. It was ensured by each SDL that the requirements on the number of measurements were fulfilled for each test that was performed. Minimum number of measurements needed was determined based on the requirements from IEC 61674 [21], which consider the coefficient of variation with respect to the limits of variation. 4.1. Repeatability and resolution At the typical air kerma rate used for dosimeter calibration in the dosimetry laboratory practice, the maximum values of the coefficients of variation of 0.3 % and 0.9 % were recorded for ICs and XMMs, respectively. For the dose rate of 100 µGy•s −1 , the largest coefficient of variation observed was 0.6 % for ICs and 0.2 % for XMMs. These values fulfil the IEC criteria [21], where the limit for the coefficient of variation for the unattenuated beam is set at 3 %. In the case of the lowest air kerma rate available in the diagnostic radiology calibration facilities, maximum coefficients of variation of 2.3 % (for ICs) and 0.9 % (for Fig. 1. Setup for the angular dependence test in two dosimeter orientations (left and middle) and the effect of rotation (right), where N represents angle of incidence. N. Krˇ zanovi´ c et al. Physica Medica 141 (2026) 105687 6 XMMs) were observed, for the respective air kerma rate of 5 µGy•s −1 and 1 µGy•s −1 . Resolution of the ICs and XMMs depends on the properties of the associated electrometer. For the tested ICs, resolution of 0.5 % or below was observed, based on the used electrometer manufacturer specifications. Selected XMMs were in line with the 1 % requirement prescribed by the standard [21] for resolution in their respective effective ranges of indicated values. 4.2. Dosimeter stabilization Deviation of dosimeter indication related to stabilization time is presented in Fig. 2 for both ICs and XMMs. It can be observed that both dosimeter types exhibit reduction in relative error in comparison to the error at the time τ , where τ <1 min. All the tested devices have deviations of their indication within ±2 %, complying with the IEC criteria [21]. The time stabilization property depends on the unit under test rather than the dosimeter type. Overall, ICs display more converging results compared to XMMs with time. 4.3. Linearity of the response From Fig. 3, it can be observed that the variation of the ionization chamber responses is within ±1 % for a wide range of dose rates. In the case of XMMs (Fig. 4) this effect is more pronounced for some tested units. The majority of XMMs performed well, displaying relative response within ±2 % if the outlier dosimeter units are excluded. Compared to the ICs, XMMs exhibited more pronounced non-linearity over a wide range of air kerma rates. If the maximum and minimum response values for each dosimeter unit are examined against the criterion described in the standard [21], all the ICs in the study fulfilled the criterion defined by Eq. (2) with variation of response lower than 0.5 % (criterion is 2 %). Most of the XMM models fulfilled the requirement prescribed by the standard [21], except for XMM21 and XMM14, whose variation of response is 3.1 % and 5.4 % respectively. 4.3.1. Variation in response due to photon energy Results on dosimeter performance in terms of relative energy response are presented in Fig. 5 and Fig. 6 for ICs and XMMs respectively for RQR and RQT radiation qualities [8]. Responses of dosimeters in the selected CPRQ non-standard radiation fields are presented in Fig. 7. The IEC standard [21] defines the ±5 % limits of variation for each diagnostic modality separately, e.g. relative to the RQR 5 reference radiation quality for RQR series and RQT 9 reference radiation quality for RQT series. For the purpose of dosimeter performance investigation in non-standard copper-filtered radiation qualities, RQR 5 was used as the reference radiation quality for all radiation fields (RQRs, RQTs and CPRQs). Relative to RQR 5, variation in the energy response of ionization chambers was within ±1 % for most ICs over the whole energy range in both RQR and RQT standard radiation quality series (Fig. 5). In addition to the performance of ICs in radiography radiation fields, performance of the mammography ionization chambers was evaluated in radiation fields with different anode/filtration setups. IC7 was tested in reference RQR-M radiation qualities (Mo/Mo setup), IC6 and IC8 were tested in the non-standard W/Mo radiation fields. IC9 was tested in a wide range of anode/filtration setups covering Mo, Rh and W anode target materials paired with Mo, Rh, Al or Ag filtrations to cover the most clinically used mammography radiation fields. The tested ionization chambers displayed a very small variation in response, up to ±1 % in the X-ray tube voltage range from 25 kV to 35 kV for all anode/ filtration setups included in the study. The relative response of most XMMs is within ±5 % represented as a red line on graph, complying with the IEC standard [21] in the entire rated range (Fig. 6). Some of the XMM models have exhibited a pronounced over-response for the lowest energy general radiography radiation quality (RQR2), where XMM21 and XMM22 were the same dosimeter model. Fig. 7 shows the performance of five XMM models and one IC Fig. 2. Relative error of diagnostic dosimeter measured value due to the effect of stabilization time ( τ ) for ionization chambers (a) and X-ray multimeters (b). Measured values were compared with indication measured at τ 0 =60 min. Fig. 3. Variation in ionization chamber relative response as a function of air kerma rate, normalized to the 0.5 mGy⋅s −1 reference air kerma rate, in the RQR5 reference radiation field. N. Krˇ zanovi´ c et al. Physica Medica 141 (2026) 105687 7 evaluated in the non-standard copper filtered radiation qualities (CPRQs), presented in Section 2.1. The 0.9 mm Cu radiation qualities were selected for performance assessment in terms of energy response, as the most heavily filtered primary beams encountered in clinical conditions. As with the RQR and RQT series, the response of the dosimeters was normalized to the RQR 5 radiation quality. The dosimeters tested were XMM4, XMM7 and XMM11 in VINS SDL, and XMM21, XMM23 and IC5 in VSL SDL. The response of IC5 deviated less than 0.6 % for all three CPRQ radiation qualities when compared to reference radiation quality. Measurement uncertainties for only one of the devices examined in PSDL and one of the devices examined in SSDL are displayed for clarity purposes. 4.3.2. Uncertainty assessment in non-standard CPRQ radiation fields The XMMs were examined in one PSDL and one SSDL, where the reference values in the non-standard radiation fields were obtained differently. In case of the PSDL, regular procedures were used and reference values were obtained using a primary standard, resulting in usual measurement uncertainties – 0.92 % (k =2). In case of the SSDL, calibration coefficients for the secondary standard were not available for some non-standard radiation qualities, and they were obtained by interpolation, which introduced additional measurement uncertainty. SSDL stated uncertainty is between 1.4 % (CPRQ9) and 1.6 % (CPRQ6) (k =2). 4.4. Variation in response due to angle of incidence Angular response performance test results are presented in Table 4. Angular dependence of responses in both horizontal and vertical orientation of dosimeters agree with the IEC standard in the minimum rated range, being within ±3 % limits of variation [21]. The maximum deviation of −1.2 % was observed for XMM12 at −5◦, and −1.1 % for XMM18 at +5◦, for horizontal and vertical orientations, respectively. Although, it was observed that most devices displayed angular dependence within ±1 %. Even though the IEC standard for air kerma measurements [21] does not provide any requirements on dosimeter rotation around the beam axis, this effect was examined following the requirements in the IEC 61676:2023 standard for tube voltage measurements [22]. The results of rotation performance test are presented in Table 5. Overall, the XMMs displayed relative error of indication within ±0.5 % for all angles of rotation with respect to the 0◦value. 4.5. Effect of pulse duration Effect of pulsed radiation is presented in Fig. 8. Performance of the diagnostic radiology dosimeters was evaluated in radiation fields of different pulse durations. Observed relative errors for IC4 were 0.14 % at 1 s pulse duration, increasing to 0.41 % at 200 ms pulse, respectively. The reduction in pulse duration led to the increase in variation of indication to >1.5 % and up to 1.7 % for XMM21, XMM23 and XMM24. On the other hand, XMM22 displayed minimal variation in indicated value, less than 0.05 % for all pulse durations. 5. Discussion 5.1. Evaluation of dosimeter performance The performance test results conducted in this research are presented in Sections 4.1–4.6. The IEC 61674:2024 [21] requirements on dosimeter performance are currently set for diagnostic dosimeters with no distinction between field-class (XMMs) and reference-class dosimeters (ICs). Based on the test results it was determined that in most cases the acceptability criteria can be more stringent for reference-class dosimeters and in some cases for field-class dosimeters as well. The largest difference between referenceand field-class dosimeters was found in the energy dependence of the response. Therefore, more detailed discussion on this topic is provided. Fig. 4. Variation in X-ray multimeter relative response as a function of air kerma rate, normalized to the 0.5 mGy⋅s −1 reference air kerma rate, in the RQR5 reference radiation field: a) XMM1 – XMM11; b) XMM13 – XMM24. Fig. 5. Energy dependence of ionization chamber response, in terms of relative response normalized to the reference radiation quality RQR5 [8]. N. Krˇ zanovi´ c et al. Physica Medica 141 (2026) 105687 8 5.2. Energy dependence of response For 9 out of 11 ICs included in this test, the variation in the energy response was within ±1 % over the whole energy range tested as shown in Fig. 5. The relative response of most XMMs in this study were within the ±5 % limit defined by the IEC standard [21] both in standard and non-standard radiation qualities as shown in Fig. 6. The presented Fig. 6. Energy dependence of X-ray multimeter response, in terms of relative response normalized to the reference radiation quality RQR5 [8]: a) XMM1 – XMM6; b) XMM7 – XMM12; c) XMM13 – XMM18; d) XMM19 – XMM24. Fig. 7. Relative energy response of XMMs and IC in the non-standard copperfiltered radiation qualities with added filtration of 0.9 mm Cu to the RQR 5, RQR 8 and RQR 9 radiation qualities, compared with the IEC limit [21]. Table 4 Relative angular response of tested ICs and XMMs in the minimum rated range for vertical (V) and horizontal (H) dosimeter orientation. Dosimeter V (−5◦) V (+5◦) H (−5◦) H (+5◦) Relative deviation [%] IC12 −0.10 −0.01 −0.01 −0.10 IC14 −0.01 −0.12 0.03 0.01 XMM1 / / 0.09 −0.02 XMM3 0.18 −0.14 0.34 0.10 XMM5 / / 0.10 −0.08 XMM6 0.49 0.14 0.03 −0.20 XMM12 / / −1.23 0.15 XMM16 / / 0.05 0.06 XMM17 0.11 0.13 −0.06 −0.10 XMM18 −0.83 −1.12 −0.92 −0.53 XMM20 / / −0.78 −0.21 Table 5 Relative error of indication for rotation of XMMs with respect to the 0◦value. Dosimeter 45◦90◦135◦180◦225◦270◦315◦ Relative error [%] XMM3 0 0.1 0.1 0.2 0.3 0.3 0.2 XMM6 / −0.4 / 0 / 0.5 / XMM17 / −0.1 / 0 / −0.1 / XMM18 −0.1 −0.2 −0.2 −0.3 −0.3 −0.2 0 XMM22 / 0.5 / 0.5 / 0.2 / XMM24 / −0.1 / −0.1 / −0.1 / N. Krˇ zanovi´ c et al. Physica Medica 141 (2026) 105687 9