Acta Oncologica ISSN: 0284-186X (Print) 1651-226X (Online) Journal homepage: https://www.tandfonline.com/loi/ionc20 IAEA methodology for on-site end-to-end IMRT/ VMAT audits: an international pilot study Pavel Kazantsev, Wolfgang Lechner, Eduard Gershkevitsh, Catharine H. Clark, Daniel Venencia, Jacob Van Dyk, Paulina Wesolowska, Victor Hernandez, Nuria Jornet, Milan Tomsej, Tomislav Bokulic & Joanna Izewska To cite this article: Pavel Kazantsev, Wolfgang Lechner, Eduard Gershkevitsh, Catharine H. Clark, Daniel Venencia, Jacob Van Dyk, Paulina Wesolowska, Victor Hernandez, Nuria Jornet, Milan Tomsej, Tomislav Bokulic & Joanna Izewska (2020) IAEA methodology for on-site end- to-end IMRT/VMAT audits: an international pilot study, Acta Oncologica, 59:2, 141-148, DOI: 10.1080/0284186X.2019.1685128 To link to this article: https://doi.org/10.1080/0284186X.2019.1685128 © 2019 International Atomic Energy Agency. View supplementary material Published by Informa UK Limited, trading as Taylor & Francis Group. Published online: 20 Nov 2019. Submit your article to this journal Article views: 3084 View related articles View Crossmark data Citing articles: 6 View citing articles Full Terms & Conditions of access and use can be found at https://www.tandfonline.com/action/journalInformation?journalCode=ionc20 ACTA ONCOLOGICA 2020, VOL. 59, NO. 2, 141–148 https://doi.org/10.1080/0284186X.2019.1685128 ORIGINAL ARTICLE IAEA methodology for on-site end-to-end IMRT/VMAT audits: an international pilot study Pavel Kazantseva , Wolfgang Lechnerb,c , Eduard Gershkevitshd, Catharine H. Clarke,f, Daniel Venenciag, Jacob Van Dykh , Paulina Wesolowskaa, Victor Hernandezi , Nuria Jornetj, Milan Tomsejk, Tomislav Bokulica and Joanna Izewskaa a International Atomic Energy Agency, Vienna, Austria; bDepartment of Radiation Oncology, Division of Medical Physics, Medical University of Vienna, Vienna, Austria; cChristian Doppler Laboratory for Medical Radiation Research for Radiation Oncology, Vienna, Austria; d North Estonia Medical Centre, Tallinn, Estonia; eDepartment of Medical Physics, Royal Surrey County Hospital, Guildford, UK; fMetrology for Medical Physics (MEMPHYS), National Physical Laboratory, Teddington, UK; gInstituto Privado de Radioterapia, Cordoba, Argentina; h Department of Oncology and Medical Biophysics, Western University, London, Canada; iDepartment of Medical Physics, Hospital Sant Joan de Reus, IISPV, Tarragona, Spain; jServei de Radiofisica i Radioproteccio, Hospital de la Santa Creu i Sant Pau, Barcelona, Spain; kCHU Charleroi, Hopital Andre Vesale, Montigny-le-Tilleul, Belgium ABSTRACT ARTICLE HISTORY Background: The IAEA has developed and tested an on-site, end-to-end IMRT/VMAT dosimetry audit Received 4 April 2019 methodology for head and neck cases using an anthropomorphic phantom. The audit methodology is Accepted 21 October 2019 described, and the results of the international pilot testing are presented. Material and methods: The audit utilizes a specially designed, commercially available anthropo- morphic phantom capable of accommodating a small volume ion chamber (IC) in four locations (three in planning target volumes (PTVs) and one in an organ at risk (OAR)) and a Gafchromic film in a cor- onal plane for the absorbed dose to water and two-dimensional dose distribution measurements, respectively. The audit consists of a pre-visit and on-site phases. The pre-visit phase is carried out remotely and includes a treatment planning task and a set of computational exercises. The on-site phase aims at comparing the treatment planning system (TPS) calculations with measurements in the anthropomorphic phantom following an end-to-end approach. Two main aspects were tested in the pilot study: feasibility of the planning constraints and the accuracy of IC and film results in comparison with TPS calculations. Treatment plan quality was scored from 0 to 100. Results: Forty-two treatment plans were submitted by 14 institutions from 10 countries, with 79% of them having a plan quality score over 90. Seventeen sets of IC measurement results were collected, and the average measured to calculated dose ratio was 0.988 ± 0.016 for PTVs and 1.020 ± 0.029 for OAR. For 13 film measurement results, the average gamma passing rate was 94.1% using criteria of 3%/3 mm, 20% threshold and global gamma. Conclusions: The audit methodology was proved to be feasible and ready to be adopted by national dosimetry audit networks for local implementation. Introduction The IAEA has developed several auditing methodologies During the last decades, intensity-modulated radiotherapy which involve measurements of absorbed dose to water for (IMRT) and volumetric-modulated arc therapy (VMAT) have high-energy photon and electron beams in non-reference become widespread around the world. At the same time, conditions including end-to-end audits for 3D conformal and external auditing capabilities for radiotherapy (RT) institu- intensity modulated RT treatments [9–13]. Although these tions in many countries are limited to the measurement of audits are not provided directly by the IAEA as a regular ser- the absorbed dose to water in reference conditions, which is vice, their methodologies are available for implementation at provided by several dosimetry audit networks (DANs) includ- a national level by a local DAN with the IAEA sup- ing the International Atomic Energy Agency (IAEA) in collab- port [14–17]. oration with the World Health Organization (WHO) [1–3]. End-to-end testing is a suitable approach for checking the Meanwhile, implementation of modern RT techniques leads RT chain from imaging through treatment planning to the to optimal patient treatment outcomes when supported by plan delivery using anthropomorphic phantoms [18–20]. It is comprehensive local quality assurance (QA) programs and particularly useful for checking IMRT and VMAT techniques external audits [4–8]. since they largely rely on automatic processes such as CONTACT Pavel Kazantsev p.kazantsev@iaea.org IAEA Laboratories, Dosimetry Laboratory, Friedenstrasse 1, 2444 Seibersdorf, Austria Supplemental data for this article can be accessed here. ß 2019 International Atomic Energy Agency. Published by Informa UK Limited, trading as Taylor & Francis Group. This is an Open Access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives License (http://creativecommons.org/licenses/by-nc-nd/4.0/), which permits non-commercial re-use, distribution, and reproduction in any medium, provided the original work is properly cited, and is not altered, transformed, or built upon in any way. 142 P. KAZANTSEV ET AL. inverse treatment planning, network data transfer, dynamic plastic materials equivalent to soft tissues, bones and teeth multileaf collimator (MLC) based dose delivery with variable (Figure 1). It has four parallel cylindrical hollow channels in gantry speed and dose rate (for VMAT) as well as multiple the cranial-caudal (CC) direction accompanied with a set of handovers between different professionals. On the other sleeves and spacers for positioning of ion chambers (ICs) hand, in case a discrepancy is found, an end-to-end test is of along them. The SHANE phantom may also be disassembled limited use for a root cause analysis and thus should be along the mid-coronal plane to accommodate a film. In add- accompanied by evaluation of the main influencing parame- ition, the caudal part of the phantom contains four cali- ters. These parameters include the beam output in the refer- brated reference plugs (cortical bone, trabecular bone, lung ence conditions, small beam output factors (OFs), small inhale, lung exhale) and a vial with water for obtaining a beam profiles, MLC characteristics and several treatment computed tomography (CT) number to relative electron planning system (TPS) specific features. In particular, it is density (RED) conversion curve (CT-to-RED curve). important to check how heterogeneities are accounted for in A reference set of CT images of the SHANE phantom was TPS calculations since IMRT is performed by the use of super- obtained with 1 mm slice thickness, 120 kV, 545 mAs. PTV position of small beam segments in heterogeneous media, and OAR contours and planning objectives and constraints which currently represents a major dosimetry challenge have been developed for the SHANE phantom by a team of [21,22]. One of the anatomic sites that benefits the most experts in a similar way to RTOG 00-22, RTOG 0615 and from IMRT treatments is head and neck (H&N), which is PARSPORT trials [23–25]. In accordance with clinical practice, highly heterogeneous and contains multiple organs at risk three PTVs are prescribed to 70, 60 and 54 Gy in 30 fractions (OARs) located close to the planning target volumes (PTVs). and defined as PTV_7000, PTVn1_6000 and PTVn2_5400, Furthermore, IMRT allows simultaneous delivery of different respectively. Organs at risk are the parotid glands, brain doses to different PTVs within a single fraction, an approach stem and spinal cord. The brain stem and the spinal cord called simultaneous integrated boost (SIB). Both, the SIB have expanded planning risk volumes with additional dose approach and dose constraints to nearby organs at risk often constraints (Supplementary Table 1). The structure set also cause high degrees of MLC modulation, making it a challeng- contains the delineations of the four IC cavity volumes posi- ing site for RT treatment planning and delivery. tioned in pre-defined measurement locations inside the PTVs Taking into account the considerations above, the IAEA and the spinal cord (namely, IC_PTV_7000, IC_PTVn1_6000, has developed an on-site, end-to-end IMRT audit method- IC_PTVn2_5400 and IC_SpinalCord, Figure 2) in order to cal- ology using a commercial anthropomorphic H&N phantom. culate the dose to a chamber volume and for the dose This work describes the details of the methodology and sum- homogeneity assessment. A small volume (0.125 cm3) PTW marizes the results of the international pilot testing. 31010 Semiflex ion chamber (PTW Freiburg Inc., Breisgau, Germany) was chosen for this audit and therefore its adapter Material and methods is included in the phantom package and its active volume contours in measurement positions are included in the struc- Audit package ture set. However, the IC itself was not a part of the package The ‘Shoulders, Head And Neck, End-to-end’ (SHANE, CIRS in the pilot study. For the consistency of results in national Inc., Norfolk, VA, USA) phantom was designed based on the audits, the national auditing organization uses its reference IAEA specifications during the audit development stage and IC for measurements in all participating institutions. EBT3 is commercially available now. It represents an H&N region Gafchromic film (Ashland Inc., Covington, KY, USA) was of a typical RT patient with complex anatomy modeled using chosen for film measurements. Figure 1. The disassembled SHANE phantom. ACTA ONCOLOGICA 143 Figure 2. Anterior and lateral projections of the SHANE phantom and some of the delineated structures. Location of the IC measurement points inside the SHANE phantom is indicated with square symbols. Overall, the package developed for the on-site, end-to-end distributions exported from the TPS using gamma analysis IMRT/VMAT audit consists of the SHANE phantom, the DICOM [27] with 3%/3 mm criteria, 20% threshold, global gamma, and CT image set, the associated DICOM RT structure set, a list of with normalization to a point within a high-dose low-gradient planning objectives and constraints, a detailed audit method- region in the central part of PTV_7000. ology description, step-by-step instructions for the auditor and As such, the on-site visit is planned for two days. The the institution being audited, data reporting forms, and a set data collected during the on-site visit are analyzed by the of reference data for the tested parameters [26]. auditor and the audit report for the institution is prepared. Audit methodology Audit testing The audit consists of a pre-visit phase and an on-site visit Testing of the audit methodology was initially performed at phase. During the pre-visit phase, the audited institution is the Medical University of Vienna/Vienna General Hospital, asked to prepare a clinically acceptable treatment plan using Vienna, Austria (AKH) and then at the international level with the reference SHANE CT and structure set following the audit the group of the IAEA consultants. Some steps of the meth- guidelines, to perform local pre-treatment QA as well as MLC odology in the preparatory phase such as small field output testing and to undertake several computational exercises. The and profile calculations and MLC QA were tested previously on-site visit takes place only if the pre-visit phase was com- during IAEA-organized Coordinated Research Projects (CRPs) pleted successfully and any potential discrepancies have been [10,11]. Therefore, the main focus of the audit testing was on resolved. During the on-site visit, the audited institution’s staff assessing the suitability of the SHANE phantom for the audit together with auditor performs CT scanning of the SHANE process and the end-to-end methodology. phantom, registers the acquired images, copies the structure set and the pre-visit phase plan onto the new CT images, re- calculates and if needed, re-optimizes the plan, and irradiates Treatment planning review the phantom. The details of the pre-visit and on-site visit A review of the treatment planning part of the audit was activities may be found in the Supplementary Material. organized to ensure that the time needed to create a clinic- The irradiation of the SHANE phantom is performed as if it ally acceptable treatment plan for the audit was representa- were an actual patient. After the phantom is aligned on the tive of a typical planning time for similar cases in the treatment table, its position is verified according to the local participating institution. Participants were requested to pre- set-up procedure using volumetric or planar imaging. Ion pare an IMRT/VMAT plan recording the time needed to cre- chamber measurements are performed at least three times in ate it, ask a local radiation oncologist to check the plan for each of the four measurement points and the average is cal- clinical acceptability, perform the local pre-treatment QA and culated. The measured doses are compared with the mean submit the plan for review to the IAEA. The treatment plans TPS calculated doses in the corresponding IC volumes. Film were analyzed in terms of quality (using PlanIQ, Sun Nuclear irradiation is done with three consecutive fractions per film to Corp., Melbourne, FL, USA) by comparing the calculated plan be in the useful dose range of 2–8 Gy for EBT3 films. Film doses for PTVs and OARs with their corresponding planning measurement results are compared to the 2D coronal dose constraints. Each dose constraint achieved was added on to 144 P. KAZANTSEV ET AL. the total plan score in such a way that the maximum plan methods and software packages used were investigated quality value was 100. The participants did not know about before [28] and considered during the analysis of the aggre- the scoring method beforehand. Technical deliverability of gated pilot study results. The IC and film measurement plans was verified by the local pre-treatment QA procedure. results obtained during the multicenter study were used to In all cases, it involved 2D dose comparison between calcula- establish the audit tolerance levels. Additionally, the relation- tions and measurements using either film or a detector ship between the number of monitor units (MUs) as a plan matrix; 3%/3 mm/global gamma criteria were used. complexity descriptor and IC measurement results was ana- Fourteen institutions from Argentina, Austria, Estonia, lyzed and a dose homogeneity index (HI¼(Dmax – Dmin)/ Hungary, Portugal, Russia, Slovakia, Slovenia, Spain and the Daverage) within IC structures was assessed as these easily UK submitted 42 treatment plans (Supplementary Table 2). In computable metrics can provide useful information for the terms of IMRT delivery modes, 21 plans were for dynamic interpretation of the measurement results [29]. This analysis MLC IMRT (dMLC), 12 for VMAT, six for Step and Shoot IMRT was performed for a subset of results where the required (SnS) and three for Tomotherapy. treatment plan related data were available. Multicenter study Statistical analysis After the initial feasibility testing of the audit methodology, A Wilcoxon signed rank test was used to test whether the a multicenter pilot study with a broad complement of partic- average values of the ratios of the IC measured dose to the ipants having different RT equipment was organized. TPS calculated dose (DIC/DTPS) in four measurement points Participants were supplied with the same SHANE phantom, were equal to 1. The test was carried out separately for clin- instructions and datasheets, and asked to follow the on-site ical and research modalities. visit part of the methodology and thus perform the SHANE A regression analysis was conducted to investigate the phantom CT scanning, treatment planning and IC and film statistical relationship between DIC/DTPS and the number of measurements. In total, six RT institutions from Austria, MUs in the plan, as well as the dose homogeneity index (HI). Belgium, Estonia, Spain and the UK participated in the multi- center pilot study with several combinations of treatment unit, TPS, dose calculation algorithm and IMRT delivery tech- Results nique (Supplementary Table 3). All combinations were com- Treatment planning review missioned for IMRT treatment planning but some of them were used only for research. Therefore, these modalities The results of the treatment planning review are summarized were not covered with a routine clinical QA program, which in Figure 3. The planning time varied from 1 to 12 hours became the reason for separating ‘clinical’ and ‘research’ with the average of 3.5 hours. About 80% of the plans modalities into different groups. This allowed the assessment achieved a quality score >90 with nine plans fully satisfying of the influence of a regular clinical QA on the audit results. the planning constraints (score 100). None of the centers Altogether, 17 sets of IC measurements and 13 film measure- used autoplanning strategies. 95% of plans passed the local ment results were collected. Films were evaluated by the pre-treatment QA while in two cases the tolerance level was institutions participating in the pilot study following their marginally exceeded. The analysis of the data showed that best practices and audit-specific recommendations provided there was no correlation between the time necessary to carry by the IAEA. Differences related to scanners, film analysis out the plan and the plan quality score (p¼.97). Figure 3. Calculated treatment plan scores and corresponding local QA results in terms of gamma passing rates in comparison with the total planning time for all participants. ACTA ONCOLOGICA 145 Multicenter pilot study Initial concerns about achievability of the treatment plan- ning goals within a reasonable time were not supported by In Figure 4, the IC measurement results for seven clinical the results of the treatment planning review. Participants IMRT modalities and 10 research modalities in terms of DIC/ generally reported that the case was similar to those rou- DTPS for the four measurement points are shown and the tinely treated in their institutions and they were able to cre- averaged results are given in Table 1. According to the ate clinically acceptable plans within four hours. At the same Wilcoxon signed rank test, the IC results for clinical IMRT time, during standardized analysis of treatment plans using modalities were not significantly different from 1 and had a PlanIQ, it was found that not all of the submitted plans were standard deviation (SD) ranging from 0.9% to 1.8%. At the close to satisfying all the plan acceptability criteria (i.e., plan same time, in the group of results for the research modal- score >90). In general, some slight differences in the PlanIQ ities, there was a significant difference (p<.05) from unity of score may be attributed to the differences in the dose grid the DIC/DTPS in the IC_SpinalCord measurement point with sizes between the participant’s TPS and the PlanIQ and thus SD of 3.2%. Of 13 film measurement results obtained, eight different dose-volume histogram calculations during the plan belonged to clinical IMRT modalities and the average gamma quality assessment. In some cases, poor plan scores were pass rate for this group constituted 95.9% (1.5% SD) while related to insufficient experience of the planner who was not for research modalities it was 91.1% (1.8% SD). Based on the able to achieve the prescription stated in the methodology results of the multicenter study, the following tolerance lev- (as can be seen in Figure 3, plan number 4). Nevertheless, els were established: 5% in PTVs and 7% in OARs for IC even the poorest quality plans were reported to be clinically measurements, and a 90% gamma pass rate with 3%/3 mm acceptable by the local radiation oncologist and passed the criteria, 20% threshold, global gamma, normalized to a point local pre-treatment QA, which means they could be used for within high-dose low-gradient area in PTV for film measure- treating a hypothetical patient. This raises the issue of clinical ments. 2/17 IC results and 1/13 film results of the multicenter training in IMRT treatment planning and use of independent study were out of the established tolerance levels. treatment planning process QA tools [30]. The regression analysis showed that neither the number As a result of the multicenter study, 5% tolerance level of MUs nor HI have correlation with the IC results DIC/DTPS was adopted for local dose differences between the calcu- lated and the IC measured doses for points inside PTVs. At (p¼.16 for the average of four points and p¼.23 for the spi- the same time, it was necessary to widen it to 7% for the IC nal cord point, respectively). There was also no difference point in the spinal cord due to the noted higher dose gradi- found (p¼.12) between results for clinical and research ent and the increased SD. Although, on average, IC results modalities in the observed range of MUs (Supplementary were in agreement, systematic shifts were noted for the Figure 1). DIC/DTPS ratios were within approximately ±4% as long as the HI was lower than 0.15 (Supplementary Figure 2). Table 1. Summary of the ion chamber results in terms of DIC/DTPS. Clinical IMRT Research IMRT modalities (n¼7) modalities (n¼10) Total Discussion Measurement point Average SD Average SD Average SD The multicenter pilot study together with the treatment IC_PTV_7000 0.977 1.2% 0.984 1.3% 0.982 1.3% planning review allowed the validation of the audit method- IC_PTVn1_6000 1.000 0.9% 0.988 2.5% 0.993 2.0% IC_PTVn2_5400 0.989 1.5% 0.991 1.6% 0.990 1.5% ology and the establishment of the tolerance levels for both IC_SpinalCord 1.006 1.8% 1.030 3.2% 1.020 2.9% IC and film measurement results. Figure 4. Ratios of ion chamber measured doses to TPS calculated doses (DIC/DTPS) in four measurement points located in the PTVs and the spinal cord. Blue/solid symbols – clinical IMRT modalities, red/hollow symbols – research IMRT modalities. Circles – VMAT plans, triangles – dMLC plans, squares – SnS plans. 146 P. KAZANTSEV ET AL. IC_PTV_7000 and IC_SpinalCord points, which also contrib- that external audits greatly complement the hospital QA pro- uted to the establishment of the tolerance levels. It is grams by allowing the detection of systematic errors due to expected that they will be further verified in future national their inherent independence and the different dosimetric studies and might be adjusted if necessary [31]. traceability of the measurement systems used. The heterogeneous composition of the SHANE phantom Although the results of all participants of the multicenter increases the uncertainty of both the dose calculation and study were within the established tolerance levels with their measurements. At the same time, its commercial availability clinically commissioned IMRT modalities, inclusion of the makes it easier to adopt the audit methodology at a national results from the research modalities give an estimate of the level as opposed to using self-developed phantoms. It is potential failure rate of 2/17 IC results and 1/13 film results. therefore challenging to directly compare the proposed tol- It also emphasizes the sensitivity of the methodology to limi- erance levels with other on-site IMRT audits employing differ- tations in TPS calculation accuracy and the requirement for a ent phantoms. For example, the dosimetry audit for the special clinical commissioning of algorithms used for IMRT PARSPORT trial [32] adopted tolerance levels of 3% for PTV treatment planning and their subsequent regular QA. Special and 5% for OAR. The Japanese Clinical Oncology Group on- attention should be paid to modeling of MLC collimated site IMRT audit system in support of clinical trials [33] also small fields including fine-tuning of beam parameters and utilizes 3% tolerance for dose in the PTV. As for remote IMRT careful measurement of OFs as well as the dosimetric leaf audits, point dose measurement tolerance levels are usually gap [43]. The pre-visit activities help to identify these kinds higher than those of on-site audits due to the increased of problems before the audit visit takes place. uncertainty of solid state dosimetry systems compared with The results of this study indicate that the number of MUs IC measurements. In addition, there is an advantage of the as a descriptor of the IMRT plan complexity is not related to auditor presence during the on-site measurement session. its deliverability although a higher spread of IC results with For example, IROC-Houston utilizes 7% tolerance for point the increase number of MUs was observed. Similarly, a dose measurements with thermoluminescent dosimeters higher HI is noted to be a possible predictor of higher IC (TLDs) in PTV for IMRT H&N credentialing [34]; the same tol- measurement deviations. Currently, a series of studies show erance is adopted in the national remote IMRT/VMAT audit contradictory results for different plan complexity metrics pilot run using lithium formate dosimeters in Sweden [35]; and for different number of institutions involved. For 5% tolerance is used for alanine point dose measurements in example, in a single-institutional study a significant correl- the national IMRT audit using a slab phantom in the UK [36]; ation was found between a small aperture score (used to and 5% tolerance is also used for the PTV in the IAEA remote describe the complexity of 122 SnS IMRT beams) and QA audit of IMRT/VMAT head treatments with TLD [13]. results [44]. At the same time, in the study analyzing the Consequently, the tolerance levels of 5% and 7% in PTVs multi-institutional results of IROC-Houston H&N phantom and OAR, respectively, used in the audit methodology here irradiations there was no correlation found between 13 dif- are comparable to those of the existing audit systems using ferent plan complexity metrics and the average dose inaccur- a similar type of audit and phantom. acy in 343 cases employing different IMRT delivery The film measurement tolerance value of 90% points techniques [45]. Different plan complexity metrics will be passing 3%/3 mm gamma criteria, 20% threshold, global tested for this audit methodology as more data gamma, was established in the pilot study. The proposed gamma criteria represent a common practice although there become available. is a concern about the ability to detect errors using them [37,38]. The tolerance proposed here is similar to the ones Conclusions used in PARSPORT trial and the UK national IMRT audit: 95% points passing 4%/3 mm or 4%/4 mm gamma criteria respect- This on-site, end-to-end IMRT audit methodology for H&N ively with 20% threshold [32,36]. According to IROC-Houston cancer treatments was developed by the IAEA and has been analysis of their credentialing results, the 3%/3 mm/90% shown to be feasible. It was successfully tested in an inter- gamma criteria/gamma pass rates adopted for SHANE would national, multicenter pilot study. The audit was found to be result in only 28% institutional pass rate and 42% institu- clinically relevant and can be executed efficiently using the tional pass rate if using 4%/4 mm/95% criteria [39]. Thus, in two-phase process, a pre-visit preparatory phase and an on- the context of large-scale dosimetry audits, this approach site phase. The audit methodology is currently being imple- appears reasonable and even may be considered quite strict. mented at a national level in several countries. Similar to the IC measurements, the acceptance criterion for film dosimetry results may be adjusted when more data Acknowledgments become available. The established IC and film measurement tolerance level The authors acknowledge the input provided by Edgardo Garrigo , values are slightly higher than those described in recommen- Alberto Alarco n, Carola Sanchez, Maria Cristina Pfaff, Fabian Mun~oz, Guisella Rivelli, Maria Jose Almada, Nicolas Andres Piccatto, Yaqueline dations for hospitals performing patient-specific IMRT QA by Schworer, Csilla Pesznyak, Miguel Capela, Josefina Mateus, Dalibor Lojko, the European Society for Radiotherapy and Oncology Gabriel Kralik, Bozidar Casar, Anna Mikhailova, Willy Gomez de Segura, (ESTRO) and the American Association of Physicists in Xavi Nolla, Mariusz Gruda through their participation in the treatment Medicine (AAPM) [40,41]. Nevertheless, it was shown [42] planning review. ACTA ONCOLOGICA 147 Disclosure statement [17] Jurkovic S, Svabic M, Diklic A, et al. Reinforcing of QA/QC pro- grams in radiotherapy departments in Croatia: results of treat- The authors report no conflicts of interest. ment planning system verification. Med Dosim. 2013;38(1): 100–104. [18] Klein EE, Hanley J, Bayouth J, et al. Task Group 142 report: quality ORCID assurance of medical accelerators. Med Phys. 2009;36(9 Part 1): 4197–4212. Pavel Kazantsev http://orcid.org/0000-0001-7662-0992 [19] Ezzell GA, Burmeister JW, Dogan N, et al. IMRT commissioning: Wolfgang Lechner http://orcid.org/0000-0001-9211-7510 multiple institution planning and dosimetry comparisons, a Jacob Van Dyk http://orcid.org/0000-0003-0304-2647 report from AAPM Task Group 119. Med Phys. 2009;36(11): Victor Hernandez http://orcid.org/0000-0003-3770-8486 5359–5373. Tomislav Bokulic http://orcid.org/0000-0003-3974-6662 [20] Letourneau D, McNiven A, Jaffray DA. Multicenter collaborative quality assurance program for the province of Ontario, Canada: first-year results. Int J Radiat Oncol Biol Phys. 2013;86(1):164–169. References [21] International Atomic Energy Agency. Dosimetry of small static [1] Izewska J, Lechner W, Wesolowska P. Global availability of dosim- fields used in external beam radiotherapy. Technical reports ser- etry audits in radiotherapy: the IAEA dosimetry audit networks ies 483. Vienna: IAEA; 2017. database. Phys Imaging Radiat Oncol. 2018;5:1–4. [22] Davidson SE, Popple RA, Ibbott GS, et al. Technical note: hetero- [2] Izewska J, Andreo P, Vatnitsky S. The IAEA/WHO TLD postal dose geneity dose calculation accuracy in IMRT: study of five commer- quality audits for radiotherapy: a perspective of dosimetry practi- cial treatment planning systems using an anthropomorphic ces at hospitals in developing countries. Radiother Oncol. 2003; thorax phantom. Med Phys. 2008;35(12):5434–5439. 69(1):91–97. [23] Eisbruch A, Harris J, Garden AS, et al. Multi-institutional trial of [3] Kry SF, Peterson CB, Howell RM, et al. Remote beam output accelerated hypofractionated intensity-modulated radiation ther- audits: a global assessment of results out of tolerance. Phys apy for early-stage oropharyngeal cancer (RTOG 00-22). Int J Imaging Radiat Oncol. 2018;7:39–44. Radiat Oncol Biol Phys. 2010;76(5):1333–1338. [4] Peters LJ, O’Sullivan B, Giralt J, et al. Critical impact of radiother- [24] Lee NY, Zhang Q, Pfister DG, et al. Addition of bevacizumab to apy protocol compliance and quality in the treatment of standard chemoradiation for locoregionally advanced nasopha- advanced head and neck cancer: results from TROG 02.02. J Clin ryngeal carcinoma (RTOG 0615): a phase 2 multi-institutional trial. Lancet Oncol. 2012;13(2):172–180. Oncol. 2010;28(18):2996–3001. [25] Nutting CM, Morden JP, Harrington KJ, et al. Parotid-sparing [5] Weber DC, Tomsej M, Melidis C, et al. QA makes a clinical trial intensity modulated versus conventional radiotherapy in head stronger: evidence-based medicine in radiation therapy. and neck cancer (PARSPORT): a phase 3 multicentre randomised Radiother Oncol. 2012;105(1):4–8. controlled trial. Lancet Oncol. 2011;12(2):127–136. [6] Izewska J, Coffey M, Scalliet P, et al. Improving the quality of radi- [26] International Atomic Energy Agency. Dosimetry audit networks ation oncology: 10 years’ experience of QUATRO audits in the (DAN) database: on-site ’end-to-end’ dosimetry audits [Internet]; IAEA Europe Region. Radiother Oncol. 2018;126(2):183–190. 2019; [cited 2019 Apr 04]. Available from: https://dosimetry-audit- [7] Clark CH, Jornet N, Muren LP. The role of dosimetry audit in networks.iaea.org/Home/EndToEndAudits achieving high quality radiotherapy. Phys Imaging Radiat Oncol. [27] Low DA, Dempsey JF. Evaluation of the gamma dose distribution 2018;5:85–87. comparison method. Med Phys. 2003;30(9):2455–2464. [8] International Atomic Energy Agency. Accuracy requirements and [28] Almady B, Wesolowska P, Santos T, et al. Gamma analysis: testing uncertainties in radiotherapy. Human health series 31. Vienna: scanners and software tools. Radiother Oncol. 2016;119: IAEA; 2016. S698–S699. [9] Izewska J, Georg D, Bera P, et al. A methodology for TLD postal [29] International Commission on Radiation Units and Measurements. dosimetry audit of high-energy radiotherapy photon beams in Prescribing, recording, and reporting photon-beam intensity- non-reference conditions. Radiother Oncol. 2007;84(1):67–74. modulated radiation therapy (IMRT). ICRU Report 83. J ICRU. [10] Izewska J, Wesolowska P, Azangwe G, et al. Testing the method- 2010;10(1):1–106. ology for dosimetry audit of heterogeneity corrections and small [30] Nelms BE, Robinson G, Markham J, et al. Variation in external MLC-shaped fields: results of IAEA multi-center studies. Acta beam treatment plan quality: an inter-institutional study of plan- Oncol. 2016;55(7):909–916. ners and planning systems. Pract Radiat Oncol. 2012;2(4): [11] Lechner W, Wesolowska P, Azangwe G, et al. A multinational 296–305. audit of small field output factors calculated by treatment plan- [31] Santos T, Lopes MC, Gershkevitsh E, et al. IMRT national audit in ning systems used in radiotherapy. Phys Imaging Radiat Oncol. Portugal. Phys Med. 2019;65:128–136. 2018;5:58–63. [32] Clark CH, Hansen VN, Chantler H, et al. Dosimetry audit for a [12] Gershkevitsh E, Pesznyak C, Petrovic B, et al. Dosimetric inter- multi-centre IMRT head and neck trial. Radiother Oncol. 2009; institutional comparison in European radiotherapy centres: results 93(1):102–108. of IAEA supported treatment planning system audit. Acta Oncol. [33] Nakamura M, Minemura T, Ishikura S, et al. An on-site audit sys- 2014;53(5):628–636. tem for dosimetry credentialing of intensity-modulated radiother- [13] Wesolowska P, Georg D, Lechner W, et al. Testing the method- apy in Japanese Clinical Oncology Group (JCOG) clinical trials. ology for a dosimetric end-to-end audit of IMRT/VMAT: results of Phys Med. 2016;32(8):987–991. IAEA multicentre and national studies. Acta Oncol. 2019:1–9. [34] Molineu A, Hernandez N, Nguyen T, et al. Credentialing results [14] Lopes MC, Cavaco A, Jacob K, et al. Treatment planning systems from IMRT irradiations of an anthropomorphic head and neck dosimetry auditing project in Portugal. Phys Med. 2014;30(1): phantom. Med Phys. 2013;40(2):022101. 96–103. [35] Adolfsson E, Gustafsson H, Lund E, et al. A system for remote [15] Bulski W, Chełmin ski K, Rostkowska J. Dosimetry audit of radio- dosimetry audit of 3D-CRT, IMRT and VMAT based on lithium for- therapy treatment planning systems. Radiat Prot Dosimetry. 2015; mate dosimetry. Radiother Oncol. 2014;113(2):279–282. 165(1–4):472–476. [36] Budgell G, Berresford J, Trainer M, et al. A national dosimetric [16] Rutonjski L, Petrovic B, Baucal M, et al. Dosimetric verification of audit of IMRT. Radiother Oncol. 2011;99(2):246–252. radiotherapy treatment planning systems in Serbia: national [37] Stojadinovic S, Ouyang L, Gu X, et al. Breaking bad IMRT QA audit. Radiat Oncol. 2012;7(1):155–162. practice. J Appl Clin Med Phys. 2015;16(3):154–165. 148 P. KAZANTSEV ET AL. [38] Steers JM, Fraass BA. IMRT QA: selecting gamma criteria based on [42] Jornet N, Carrasco P, Beltran M, et al. Multicentre validation of error detection sensitivity. Med Phys. 2016;43(4):1982–1994. IMRT pre-treatment verification: comparison of in-house and [39] Carson ME, Molineu A, Taylor PA, et al. Examining credentialing external audit. Radiother Oncol. 2014;112(3):381–388. criteria and poor performance indicators for IROC Houston’s [43] Kron T, Clivio A, Vanetti E, et al. Small field segments surrounded by anthropomorphic head and neck phantom. Med Phys. 2016; large areas only shielded by a multileaf collimator: comparison of 43(12):6491–6496. experiments and dose calculation. Med Phys. 2012;39(12):7480–7489. [40] European Society for Radiotherapy and Oncology. Guidelines for [44] Crowe SB, Kairn T, Kenny J, et al. Treatment plan complexity met- the verification of IMRT. ESTRO booklet No. 9. Brussels: Estro; rics for predicting IMRT pre-treatment quality assurance results. 2008. Australas Phys Eng Sci Med. 2014;37(3):475–482. [41] Miften M, Olch A, Mihailidis D, et al. Tolerance limits and methodol- [45] Glenn MC, Hernandez V, Saez J, et al. Treatment plan complexity ogies for IMRT measurement-based verification QA: recommenda- does not predict IROC Houston anthropomorphic head and neck tions of AAPM Task Group No. 218. Med Phys. 2018;45(4):e53–e83. phantom performance. Phys Med Biol. 2018;63(20):205015.