Note MULTIBIODOSE RADIATION EMERGENCY TRIAGE CATEGORIZATION SOFTWARE Elizabeth A. Ainsbury,* Stephen Barnard,* Lleonard Barrios,† Paola Fattibene,‡ Virginie de Gelder,§ Eric Gregoire,** Carita Lindholm,†† David Lloyd,* Inger Nergaard,†† Kai Rothkamm,* Horst Romm,‡‡ Harry Scherthan,§§ Hubert Thierens,§ Charlot Vandevoorde,§ Clemens Woda,*** and Andrzej Wojcik††† Abstract—In this note, the authors describe the MULTIBIODOSE software, which has been created as part of the MULTIBIODOSE project. The software enables doses estimated by networks of laboratories, using up to five retrospective (biological and physical) assays, to be combined to give a single estimate of triage category for each individual potentially exposed to ionizing radiation in a large scale radiation accident or incident. The MULTIBIODOSE software has been created in Java. The usage of the software is based on the MULTIBIODOSE Guidance: the program creates a link to a single SQLite database for each incident, and the database is administered by the lead laboratory. The software has been tested with Java runtime environment 6 and 7 on a number of different Windows, Mac, and Linux systems, using data from a recent intercomparison exercise. The Java program MULTIBIODOSE_1.0.jar is freely available to download from http://www.multibiodose.eu/software or by contacting the software administrator: [email protected]. Health Phys. 107(1):83–89; 2014 Key words: accidents, nuclear; biological indicators; emergency planning; radiation dose
INTRODUCTION IN RECENT years, the retrospective radiation dosimetry community has been particularly concerned with the development of capabilities for fast and reliable triage categorization in mass casualty situations. A number of methods *Public Health England Centre for Radiation, Chemical and Environmental Hazards, Chilton, Didcot, Oxford OX11 0RQ; †Universitat Autònoma de Barcelona, Spain; ‡Istituto Superiore di Sanità, Italy; §Universiteit Gent, Belgium; **Institut de radioprotection et de sûreté nucléaire, France; ††Radiation and Nuclear Safety Authority, Finland; ‡‡Bundesamt fuer Strahlenschutz, Germany; §§Inst. für Radiobiologie der Bundeswehr in Verb. mit der Univ. Ulm, Germany; ***Helmholtz Zentrum München, Germany; †††Stockholm University, Sweden. The authors declare no conflicts of interest. For correspondence contact: Elizabeth A. Ainsbury, Public Health England Centre for Radiation, Chemical and Environmental Hazards, Chilton, Didcot, Oxford OX11 0RQ, or email at [email protected]. (Manuscript accepted 8 October 2013) 0017-9078/14/0 Copyright © 2014 Health Physics Society DOI: 10.1097/HP.0000000000000049
and tools have been developed specifically for triage (e.g., Blakely et al. 2007; Garty et al. 2010). The EU FP7-funded MULTIBIODOSE project has developed a multiparametric approach to retrospective dosimetry with a particular emphasis on triage of large numbers of potentially exposed individuals following accidental exposures (Romm et al. 2013). The objective of the MULTIBIODOSE project was to establish tools for use in such scenarios and to provide information about their advantages and limitations and the laboratories that are equipped to provide dosimetry services using these tools. MULTIBIODOSE represented the first very large-scale international networking project. One of the main aims of MULTIBIODOSE was therefore the optimization for high throughput and validation of seven different assays [the dicentric (Romm et al. 2011, 2013), micronucleus (Willems et al. 2010), and gammaH2AX foci assays (Horn et al. 2011; Rothkamm et al. 2013); electron paramagnetic resonance (EPR) (Trompier et al. 2009) and optically stimulated luminescence (OSL) (Woda et al. 2009); the skin speckle assay (SSA) (Ahmed et al. 2012); and the serum protein assay (SPA) (Guipaud 2013)] operated in triage mode across the large number of participating laboratories against the benchmark of the “gold standard” dicentric assay (IAEA 2011). An additional aim was the development of a statistical software package to facilitate triage categorization. The MULTIBIODOSE software MUTIBIODOSE_1.0. jar has been created to address this final aim, to bring together the results of the developments within the different assays. The software allows dose estimates from different assays and laboratories to be combined to give a single estimate of triage category (low exposure: < 1 Gy; medium exposure: 1–2 Gy; high exposure: > 2 Gy) for each individual potentially exposed to ionizing radiation in a large-scale radiation accident or incident. Five assays are currently included in the software: the dicentric, micronucleus, foci, EPR, and OSL assays, as these were the tools that were found to be useful for triage categorization. 83
www.health-physics.com
Copyright © 2014 Health Physics Society. Unauthorized reproduction of this article is prohibited.
84
Health Physics
Here the authors present details of the software together with a brief summary of the instructions for downloading and using the program—full instructions are provided in the software manual, which is available for download with the software from www.MULTIBIODOSE.eu/software. MATERIALS AND METHODS Software development The MULTIBIODOSE software has been created in Java, using Java Development Kit (JDK) 1.7.0_03 (Oracle 2013a) and NetBeans Integrated Development Environment 7.1 (Oracle 2013b). In addition, a number of freely available libraries were used to support functionality: jcommon‐1.0.17.jar (JFree 2013a); jfreechart‐1.0.14.jar to support graphical representation of distribution of doses (JFree 2013b); junit.jar for testing (JUnit 2013); jxl.jar to allow data to be exported to a file in Microsoft Excel® format (Khan 2013) and sqlite.jar and sqlitejdbcv056.jar to support connectivity between Java and SQLite (Werner 2013). Using the software The usage of the MULTIBIODOSE software is based on the MULTIBIODOSE Guidance (Jaworska et al. 2013), and it is strongly recommended that users obtain and read a copy of the Guidance as well as the manual in advance of attempting to use the software. One SQLite database is created per incident. The database is administered by the laboratory in whose country the incident occurred, or in the event that a retrospective dosimetry laboratory is not present in that country, the laboratory that is assigned the status of “lead laboratory” by the local authority in charge of the incident. The lead laboratory takes responsibility for collection and analysis of physical/biological samples, assigning work to the other laboratories involved in the MULTIBIODOSE consortium or other networks, for instance RENEB (Kulka et al. 2012) or BioDoseNet (Maznyk et al. 2012). The lead laboratory also takes responsibility for maintaining the
July 2014, Volume 107, Number 1
integrity of the data; i.e., for sample coding, etc., and for reporting to the authority in charge. The current version of the software is Version 1.0, MULTIBIODOSE_1.0.jar, which was released on 24 May 2013. MULTIBIODOSE_software_1.0.zip contains the software itself, MULTIBIODOSE_1.0.jar; the manual corresponding to the current release, MULTIBIODOSEsoftware_manual_1.0.pdf; a test database, Test_Incident. db; the list of incident databases, Incident_List; and the lib folder that contains the libraries needed for usage (as described above). These files can all be downloaded together from www.MULTIBIODOSE.eu/software. The zip file should then be unpacked in the desired location, with the file structure maintained. The user will also need Java runtime environment (JRE) 7 or higher (it is possible to check whether the latest version of Java is installed and, if not, download it at http://java.com/en/download/index.jsp). Description of the software The software is started by double-clicking on the jar file (Windows systems) or by right clicking and selecting “Run As”…; JRE 7.0 (or higher; Linux or Mac systems). The main screen will be presented, as in Fig. 1. The user then has two options: first, to look at the data from an existing incident by selecting from the list of incidents contained in the bottom box and pressing “Go,” or second, using the top box to assign a name to a new incident. If a new incident is selected, then the user will be asked a series of questions about the type of scenario, as detailed in Fig. 2. The answers to these questions determine which assays are relied upon in order to calculate the final triage category. The default settings are for “No information available,” as it is anticipated that this is the most likely situation in true incident triage. The decisions are translated from the MULTIBIODOSE Guidance in the form of a series of assay “weights” for a given scenario. The weights, which were chosen by the MULTIBIODOSE consortium and are based on their combined expert judgment as to the most appropriate assays to use in each scenario, are designated as 1 where an assay is unlikely to be
Fig. 1. The main graphic user interface of the MULTIBIODOSE software. www.health-physics.com
Copyright © 2014 Health Physics Society. Unauthorized reproduction of this article is prohibited.
Emergency triage software c E. A. AINSBURY
ET AL.
85
Fig. 2. Triage questions to be answered for a new incident.
useful and 10 where the assay should provide useful dosimetry information in the given scenario. The assay weights for each scenario are given in the software manual and Guidance; however, as an example, if the user describes a suspected whole body exposure with a likely radiation dose < 1 Gy and a blood sampling time > 24 h, the weights for the assays would be: dicentric assay—10; micronucleus assay—10; foci assay—1; EPR assay—1; and OSL assay—10. This reflects the fact that the EPR assay was shown during the course of the project to be less precise for whole body exposures, and all assays apart from the foci assay were shown to be useful at sampling times > 24 h. However, if the user selects a similar dose range and sampling time but for a partial body exposure where it is known that the personal electronic devices were in the field, the weights would be as follows: dicentric assay—10; micronucleus assay—10; foci assay—1; EPR assay—10; and OSL assay—10, as the EPR assay
would then be likely to give a more reliable dose (Jaworska et al. 2013). It should be noted here that the software does not use statistical weighting of individual assay results to combine doses from repeat measurements and/or different assays. The reason for this is that weighting based on estimated dose alone was found to be most successful in terms of triage categorization. This is because, in triage mode, each dose estimate has a very large uncertainty associated with it (Ainsbury et al. 2014). Once the user has answered the scenario questions and assigned a scenario to an incident, the scenario and the weights for the assays cannot be changed. If new information is forthcoming about the scenario, then a new incident should be created. However, the user can enter “notes” about an incident, which will also be stored in the database and can be updated as more information about the incident and exposure conditions become available.
www.health-physics.com
Copyright © 2014 Health Physics Society. Unauthorized reproduction of this article is prohibited.
86
Health Physics
Once the triage questions are answered and “Go” is clicked, the user is then presented with a screen with an empty table, into which the data for the scenario can be entered. When the incident has been created and saved, it should appear on the drop down list of existing incidents the next time the software is started. If an existing incident is selected, then another screen will appear containing all the current data for that incident, as in Fig. 3. The user can then add/delete/amend data in the database, as is described in detail in the manual. When changes are made to the table, for instance when adding new results, the “Update database” button should be pressed to ensure the new results are saved. Once the most up to date data have been added to the database, pressing the “Calculate” button gives the user the results in table format, as illustrated in Fig. 3. The detailed results for each unique Case_ID are displayed in the
July 2014, Volume 107, Number 1
text box below the results table, by clicking on the appropriate row in the results table. In addition to the results of the individual Case_ID, the combined or “group” dose may be of interest. The results for the entire data set, weighted by assay as above, can be accessed by pressing the “Group data” button on the right-hand side of the data entry screen. Pressing the button will cause a dialog to appear with a text box at the top and a chart, comparing all the individual Case_ID doses, as illustrated in Fig. 4. The data and results can be exported to a Microsoft Excel® file (.xls format) by clicking on the “Export data” button at the bottom right-hand corner of the data entry graphic user interface (illustrated in Fig. 3). Copying and pasting within/outside the tables is also available and can be used by right-clicking over the desired selection of cells. A “troubleshooting” section has been provided in the software manual, which has been created following extensive testing by the MULTIBIODOSE project partners.
Fig. 3. The graphic user interface with details of the current data for Test_Incident, after the “Calculate” button has been pressed. www.health-physics.com
Copyright © 2014 Health Physics Society. Unauthorized reproduction of this article is prohibited.
Emergency triage software c E. A. AINSBURY
Fig. 4. The “Group” dosimetry results for the Test_Incident data.
This section can be consulted in the event that the user experiences problems with operation or running of the software. If a problem or error occurs that is not listed in this section, the user is encouraged to contact the software administrator to ask for further help. RESULTS AND DISCUSSION Validation and testing The software and manual were finalized and tested with the assistance of a large number of project participants. At least one representative from each project work package (focused on the MULTIBIODOSE assays) tested the software and, in addition, eight individuals from across the different work packages were selected as “debugging” testers. These individuals tested the software thoroughly with their own data on a number of different platforms and configurations, including Windows, Mac, and Linux systems, and with up to 10,000 data sets. The testers attempted to identify problems with the functionality. A number of issues were identified and resolved, as a result of which the final version of the software is now extremely stable. The ability of the software to assign appropriate triage categories was tested using data from the MULTIBIODOSE biodosimetry exercise that took place in November 2012, the results of which are reported elsewhere (Ainsbury et al. 2014). However, in brief, three biological dosimetry tools (the dicentric, micronucleus, and gamma-H2AX foci assays)
ET AL.
87
were tested, in addition to provision of the triage status results (low exposure: < 1 Gy; medium exposure: 1–2 Gy; high exposure: > 2 Gy) by the MULTIBIODOSE software. The exercise was run in two modes: an initial triage categorization of 124 samples (based on the first dose estimates received from each laboratory) followed by collation of the full set of 253 estimated doses calculated using as many modes of operation as possible of the different assays developed during the project. Simulated acute whole body and partial body exposures were included. The software was able to assign an appropriate category based on the data in all scenarios, and individual case/group doses were also always correct within the limits of the appropriate triage category. The MULTIBIODOSE software described here has been developed as part of an academic research project, and, as such, no formal validation or benchmarking of the software has taken place other than that described above. However, if, as is hoped, the software can be developed further during the course of other ongoing biodosimetry networking projects, the authors plan to move toward formal benchmarking of the methods and techniques used in the software. Discussion of limitations Despite the successful outcome of the MULTIBIODOSE exercise, the software has, to date, only been tested with a limited amount of data, and a number of limitations remain to be addressed. These include the fact that the scheme for assignment of “weights,” which allow the software to provide a combined estimate of dose based only on the appropriate assays for an exposure scenario, is only very simplistic in nature. The decision about the simplicity of the weighting scheme was made by the consortium as a whole, on the basis of the fact that the tools have not yet been tested in the full range of possible exposure scenarios. Another current constraint lies in the fact that the incident parameters (exposure scenario) cannot be changed once an incident has been set up. Again, the reason for this is that it was judged to be likely that, initially, emergency responders will not have a great deal of information about the scenario (i.e., the likely magnitude of the doses received by individuals, even perhaps the timescale of exposure). It is currently recommended that once such information is forthcoming, sub-incidents should be set up by the team managing the incident; data can be copied over using the copy and paste facilities provided. It may also be seen as a limitation that it is not currently possible for separate laboratories involved in the incident response to enter information concurrently into a live incident database. The reason for this is that, in most cases, it was judged that national authorities dealing with a radiological accident or incident would generally
www.health-physics.com
Copyright © 2014 Health Physics Society. Unauthorized reproduction of this article is prohibited.
88
Health Physics
prefer to have a single laboratory managing the retrospective dosimetry response to the incident. As detailed in the Guidance (Jaworska et al. 2013), the managing laboratory should ideally be a MULTIBIODOSE partner laboratory or a member of a current networking project. Additionally, if only one laboratory has access to personal data, this also avoids issues related to data sharing and data protection across the large number of international project partners. Finally, the choice of the Java platform for the software may prove to be an issue for some users. Although any software platform comes with its own challenges, the Java platform was chosen as it is known to be reliable across a wide range of operating systems. Care was taken to test the software on as many different systems as possible within the consortium; however, initial testing did reveal some compatibility issues, particularly between different countries. Although the issues across the consortium have now been resolved, the sheer number of different systems in use around the world means it is likely that some users will still experience compatibility or other problems. In order to address this as far as possible, users experiencing such technical issues are asked to contact the software administrator, and problems will then be resolved as and when they occur. To address all of the above matters, it is intended that further testing of the software will take place as part of future inter-laboratory comparisons, and further development will be carried out in response to the needs of the community. In particular, testing and optimization of the software will need to be carried out with very large numbers of samples, and it is highly likely that amendments to the weighting scheme will become necessary as more data is collated.
CONCLUSION The MULTIBIODOSE software provides a tool that facilitates the combination of doses estimated by networks of laboratories using five retrospective (biological and physical) assays, which are combined to give a single estimate of triage category for each individual potentially exposed to ionizing radiation in a large scale radiation accident or incident. The software has been initially tested and validated as part of the MULTIBIODOSE biodosimetry intercomparison and emergency response exercise, but further testing and validation by the consortium members is intended, and testing by the wider community is also welcomed. The software and all associated files are freely available to download from www.multibiodose.eu/ software or by contacting the software administrator: [email protected].
July 2014, Volume 107, Number 1 Acknowledgments—The research detailed in this manuscript has received funding from the European Union’s Seventh Framework Programme (FP7/ 2007–2013) under grant agreement number 241536.
REFERENCES Ahmed EA, Agay D, Schrock G, Drouet M, Meineke V, Scherthan H. Persistent DNA damage after high dose in vivo gamma exposure of minipig skin. PLoS ONE 7:e39521; 2012. DOI:10.1371/journal.pone.0039521. Ainsbury EA, Al-Hafidh J, Bajinskis A, Barnard S, Barquinero JF, Beinke C, de Gelder V, Gregoire E, Jaworska A, Lindholm C, Lloyd D, Moquet J, Nylund R, Oestreicher U, Roch-Lefevre S, Rothkamm K, Romm H, Scherthan H, Sommer S, Thierens H, Vandevoorde C, Vral A, Wojcik A. Validating a multibiodosimetric approach to triage in a radiation emergency. Int J Radiat Biol 90:193–202; 2014. Blakely WF, Ossetrova NI, Manglapus GL, Salter CA, Levine IH, Jackson WE, Grace MB, Prasanna PGS, Sandgren DJ, Ledney GD. Amylase and blood cell-count hematological radiation-injury biomarkers in a rhesus monkey radiation model—use of multiparameter and integrated biological dosimetry. Radiat Meas 42:1164–1170; 2007. DOI: 10.1016/j. radmeas.2007.05.013. Garty G, Chen Y, Salerno A, Turner H, Zhang J, Lyulko O, Bertucci A, Xu Y, Wang H, Simaan N, Randers-Pehrson G, Yao L, Amundson SA, Brenner DJ. The RABIT: a rapid automated biodosimetry tool for radiological triage. Health Phys 98:209–217; 2010. DOI: 10.1097/HP.0b013e3181ab3cb6. Guipaud O. Serum and plasma proteomics and its possible use as a detector and predictor of radiation diseases. Adv Exp Med Biol 990:61–86; 2013. DOI: 10.1007/978‐94‐ 007‐5896‐4_4. Horn S, Barnard S, Rothkamm K. Gamma-H2AX-based dose estimation for whole and partial body radiation exposure. PLoS ONE 6:e25113; 2011. DOI: 10.1371/journal.pone.0025113. International Atomic Energy Agency. Cytogenetic dosimetry: applications in preparedness for and response to radiation emergencies. Vienna: IAEA; IAEA Emergency Preparedness and Response Series, EPR-Biodosimetry; 2011. JFree. Welcome to JCommon! [online]. 2013a. Available at www.jfree.org/jcommon/. Accessed 4 July 2013. JFree. Welcome To JFreeChart! 2013b. Available at www.jfree. org/jfreechart/. Accessed 4 July 2013. JUnit. Junit, a programmer-oriented testing framework for. 2013. Available at http://junit.org/. Accessed 4 July 2013. Jaworska A, Wojcik A, Ainsbury E, Fattibene P, Lindholm C, Oestreicher U, Rothkamm K, Romm H, Thierens H, Trompier F, Voisin P, Vral A, Woda C. Guidance for using MULTIBIODOSE tools in emergencies, for radiation emergency response organisations in Europe. 2013. Available at www. multibiodose.eu/News/MBD_Guidance_web.pdf. Accessed 4 July 2013. Khan A. Java Excel API—a Java API to read, write and modify Excel spreadsheets [online]. Available at www.andykhan.com/ jexcelapi/. Accessed 4 July 2013. Kulka U, Ainsbury L, Atkinson M, Barquinero JF, Barrios L, Beinke C, Bognar G, Cucu A, Darroudi F, Fattibene P, Gil O, Gregoire E, Hadjidekova V, Haghdoost S, Herranz R, Jaworska A, Lindholm C, Mkacher R, Mörtl S, Montoro A, Moquet J, Moreno M, Ogbazghi A, Oestreicher U, Palitti F, Pantelias G, Popescu I, Prieto MJ, Romm H, Rothkamm K, Sabatier L, Sommer S, Terzoudi G, Testa A, Thierens H, Trompier F, Turai I, Vandersickel V, Vaz P, Voisin P, Vral A, Ugletveit F, Woda C, Wojcik A. Realising the European network of biodosimetry (RENEB). Radiat Protect Dosim 151: 621–624; 2012.
www.health-physics.com
Copyright © 2014 Health Physics Society. Unauthorized reproduction of this article is prohibited.
Emergency triage software c E. A. AINSBURY
Maznyk NA, Wilkins RC, Carr Z, Lloyd DC. The capacity, capabilities and needs of the WHO Biodosenet member laboratories. Radiat Protect Dosim 151:611–621; 2012. Oracle. Java SE 7 downloads. 2013a. Available at www.oracle.com/ technetwork/java/javase/downloads/java-archive-downloadsjavase7-521261.html#jdk-7u3-oth-JPR. Accessed 4 July 2013. Oracle. NetBeans IDE [online]. 2013b. Available at https:// netbeans.org/. Accessed 4 July 2013. Romm H, Ainsbury E, Barnard S, Barrios L, Barquinero JF, Beinke C, Deperas M, Gregoire E, Koivistoinen A, Lindholm C, Moquet J, Oestreicher U, Puig R, Rothkamm K, Sommer S, Thierens H, Vandersickel V, Vral A, Wojcik A. Automatic scoring of dicentric chromosomes as a tool in large scale radiation accidents. Mutat Res 756:174–183; 2013. DOI: 10.1016/j. mrgentox.2013.05.013. Romm H, Wilkins RC, Coleman NC, Lillis-Hearne PK, Pellmar TC, Livingston GK, Awa AA, Jenkins MS, Yoshida MA, Oestreicher U, Prasanna PGS. Biological dosimetry by the triage dicentric chromosome assay: potential implications for treatment of acute radiation syndrome in radiological mass casualties. Radiat Res 175:397–404; 2011. DOI: 10.1667/RR2321.1. Rothkamm K, Barnard S, Ainsbury EA, Al-hafidh J, Barquinero JF, Lindholm C, Moquet J, Perälä M, Roch-Lefèvre S,
ET AL.
89
Scherthan H, Thierens H, Vral A, Vandersickel V. Manual versus automated gamma-H2AX foci analysis across five European laboratories: can this assay be used for rapid biodosimetry in a large scale radiation accident? Mutat Res 756:170–173; 2013. DOI: 10.1016/j.mrgentox.2013.04.012. Trompier F, Bassinet C, Wieser A, De Angelis C, Viscomi D, Fattibene P. Radiation-induced signals analysed by EPR spectrometry applied to fortuitous dosimetry. Ann Ist Super Sanità 45:287–296; 2009. Werner C. SQLite java wrapper/jdbc driver. 2013. Available at www.ch-werner.de/javasqlite/. Accessed 4 July 2013. Willems P, August L, Slabbert J, Romm H, Oestreicher U, Thierens H, Vral A. Automated micronucleus (MN) scoring for population triage in case of large scale radiation events. Int J Radiat Biol 86:2–11; 2010. DOI: 10.3109/09553000903264481. Woda C, Bassinet C, Trompier F, Bortolin E, Della Monaca S, Fattibene P. Radiation-induced damage analysed by luminescence methods in retrospective dosimetry and emergency response. Ann Ist Super Sanità 45:287–296; 2009.
www.health-physics.com
Copyright © 2014 Health Physics Society. Unauthorized reproduction of this article is prohibited.
■■
INTRODUCTION IN RECENT years, the retrospective radiation dosimetry community has been particularly concerned with the development of capabilities for fast and reliable triage categorization in mass casualty situations. A number of methods *Public Health England Centre for Radiation, Chemical and Environmental Hazards, Chilton, Didcot, Oxford OX11 0RQ; †Universitat Autònoma de Barcelona, Spain; ‡Istituto Superiore di Sanità, Italy; §Universiteit Gent, Belgium; **Institut de radioprotection et de sûreté nucléaire, France; ††Radiation and Nuclear Safety Authority, Finland; ‡‡Bundesamt fuer Strahlenschutz, Germany; §§Inst. für Radiobiologie der Bundeswehr in Verb. mit der Univ. Ulm, Germany; ***Helmholtz Zentrum München, Germany; †††Stockholm University, Sweden. The authors declare no conflicts of interest. For correspondence contact: Elizabeth A. Ainsbury, Public Health England Centre for Radiation, Chemical and Environmental Hazards, Chilton, Didcot, Oxford OX11 0RQ, or email at [email protected]. (Manuscript accepted 8 October 2013) 0017-9078/14/0 Copyright © 2014 Health Physics Society DOI: 10.1097/HP.0000000000000049
and tools have been developed specifically for triage (e.g., Blakely et al. 2007; Garty et al. 2010). The EU FP7-funded MULTIBIODOSE project has developed a multiparametric approach to retrospective dosimetry with a particular emphasis on triage of large numbers of potentially exposed individuals following accidental exposures (Romm et al. 2013). The objective of the MULTIBIODOSE project was to establish tools for use in such scenarios and to provide information about their advantages and limitations and the laboratories that are equipped to provide dosimetry services using these tools. MULTIBIODOSE represented the first very large-scale international networking project. One of the main aims of MULTIBIODOSE was therefore the optimization for high throughput and validation of seven different assays [the dicentric (Romm et al. 2011, 2013), micronucleus (Willems et al. 2010), and gammaH2AX foci assays (Horn et al. 2011; Rothkamm et al. 2013); electron paramagnetic resonance (EPR) (Trompier et al. 2009) and optically stimulated luminescence (OSL) (Woda et al. 2009); the skin speckle assay (SSA) (Ahmed et al. 2012); and the serum protein assay (SPA) (Guipaud 2013)] operated in triage mode across the large number of participating laboratories against the benchmark of the “gold standard” dicentric assay (IAEA 2011). An additional aim was the development of a statistical software package to facilitate triage categorization. The MULTIBIODOSE software MUTIBIODOSE_1.0. jar has been created to address this final aim, to bring together the results of the developments within the different assays. The software allows dose estimates from different assays and laboratories to be combined to give a single estimate of triage category (low exposure: < 1 Gy; medium exposure: 1–2 Gy; high exposure: > 2 Gy) for each individual potentially exposed to ionizing radiation in a large-scale radiation accident or incident. Five assays are currently included in the software: the dicentric, micronucleus, foci, EPR, and OSL assays, as these were the tools that were found to be useful for triage categorization. 83
www.health-physics.com
Copyright © 2014 Health Physics Society. Unauthorized reproduction of this article is prohibited.
84
Health Physics
Here the authors present details of the software together with a brief summary of the instructions for downloading and using the program—full instructions are provided in the software manual, which is available for download with the software from www.MULTIBIODOSE.eu/software. MATERIALS AND METHODS Software development The MULTIBIODOSE software has been created in Java, using Java Development Kit (JDK) 1.7.0_03 (Oracle 2013a) and NetBeans Integrated Development Environment 7.1 (Oracle 2013b). In addition, a number of freely available libraries were used to support functionality: jcommon‐1.0.17.jar (JFree 2013a); jfreechart‐1.0.14.jar to support graphical representation of distribution of doses (JFree 2013b); junit.jar for testing (JUnit 2013); jxl.jar to allow data to be exported to a file in Microsoft Excel® format (Khan 2013) and sqlite.jar and sqlitejdbcv056.jar to support connectivity between Java and SQLite (Werner 2013). Using the software The usage of the MULTIBIODOSE software is based on the MULTIBIODOSE Guidance (Jaworska et al. 2013), and it is strongly recommended that users obtain and read a copy of the Guidance as well as the manual in advance of attempting to use the software. One SQLite database is created per incident. The database is administered by the laboratory in whose country the incident occurred, or in the event that a retrospective dosimetry laboratory is not present in that country, the laboratory that is assigned the status of “lead laboratory” by the local authority in charge of the incident. The lead laboratory takes responsibility for collection and analysis of physical/biological samples, assigning work to the other laboratories involved in the MULTIBIODOSE consortium or other networks, for instance RENEB (Kulka et al. 2012) or BioDoseNet (Maznyk et al. 2012). The lead laboratory also takes responsibility for maintaining the
July 2014, Volume 107, Number 1
integrity of the data; i.e., for sample coding, etc., and for reporting to the authority in charge. The current version of the software is Version 1.0, MULTIBIODOSE_1.0.jar, which was released on 24 May 2013. MULTIBIODOSE_software_1.0.zip contains the software itself, MULTIBIODOSE_1.0.jar; the manual corresponding to the current release, MULTIBIODOSEsoftware_manual_1.0.pdf; a test database, Test_Incident. db; the list of incident databases, Incident_List; and the lib folder that contains the libraries needed for usage (as described above). These files can all be downloaded together from www.MULTIBIODOSE.eu/software. The zip file should then be unpacked in the desired location, with the file structure maintained. The user will also need Java runtime environment (JRE) 7 or higher (it is possible to check whether the latest version of Java is installed and, if not, download it at http://java.com/en/download/index.jsp). Description of the software The software is started by double-clicking on the jar file (Windows systems) or by right clicking and selecting “Run As”…; JRE 7.0 (or higher; Linux or Mac systems). The main screen will be presented, as in Fig. 1. The user then has two options: first, to look at the data from an existing incident by selecting from the list of incidents contained in the bottom box and pressing “Go,” or second, using the top box to assign a name to a new incident. If a new incident is selected, then the user will be asked a series of questions about the type of scenario, as detailed in Fig. 2. The answers to these questions determine which assays are relied upon in order to calculate the final triage category. The default settings are for “No information available,” as it is anticipated that this is the most likely situation in true incident triage. The decisions are translated from the MULTIBIODOSE Guidance in the form of a series of assay “weights” for a given scenario. The weights, which were chosen by the MULTIBIODOSE consortium and are based on their combined expert judgment as to the most appropriate assays to use in each scenario, are designated as 1 where an assay is unlikely to be
Fig. 1. The main graphic user interface of the MULTIBIODOSE software. www.health-physics.com
Copyright © 2014 Health Physics Society. Unauthorized reproduction of this article is prohibited.
Emergency triage software c E. A. AINSBURY
ET AL.
85
Fig. 2. Triage questions to be answered for a new incident.
useful and 10 where the assay should provide useful dosimetry information in the given scenario. The assay weights for each scenario are given in the software manual and Guidance; however, as an example, if the user describes a suspected whole body exposure with a likely radiation dose < 1 Gy and a blood sampling time > 24 h, the weights for the assays would be: dicentric assay—10; micronucleus assay—10; foci assay—1; EPR assay—1; and OSL assay—10. This reflects the fact that the EPR assay was shown during the course of the project to be less precise for whole body exposures, and all assays apart from the foci assay were shown to be useful at sampling times > 24 h. However, if the user selects a similar dose range and sampling time but for a partial body exposure where it is known that the personal electronic devices were in the field, the weights would be as follows: dicentric assay—10; micronucleus assay—10; foci assay—1; EPR assay—10; and OSL assay—10, as the EPR assay
would then be likely to give a more reliable dose (Jaworska et al. 2013). It should be noted here that the software does not use statistical weighting of individual assay results to combine doses from repeat measurements and/or different assays. The reason for this is that weighting based on estimated dose alone was found to be most successful in terms of triage categorization. This is because, in triage mode, each dose estimate has a very large uncertainty associated with it (Ainsbury et al. 2014). Once the user has answered the scenario questions and assigned a scenario to an incident, the scenario and the weights for the assays cannot be changed. If new information is forthcoming about the scenario, then a new incident should be created. However, the user can enter “notes” about an incident, which will also be stored in the database and can be updated as more information about the incident and exposure conditions become available.
www.health-physics.com
Copyright © 2014 Health Physics Society. Unauthorized reproduction of this article is prohibited.
86
Health Physics
Once the triage questions are answered and “Go” is clicked, the user is then presented with a screen with an empty table, into which the data for the scenario can be entered. When the incident has been created and saved, it should appear on the drop down list of existing incidents the next time the software is started. If an existing incident is selected, then another screen will appear containing all the current data for that incident, as in Fig. 3. The user can then add/delete/amend data in the database, as is described in detail in the manual. When changes are made to the table, for instance when adding new results, the “Update database” button should be pressed to ensure the new results are saved. Once the most up to date data have been added to the database, pressing the “Calculate” button gives the user the results in table format, as illustrated in Fig. 3. The detailed results for each unique Case_ID are displayed in the
July 2014, Volume 107, Number 1
text box below the results table, by clicking on the appropriate row in the results table. In addition to the results of the individual Case_ID, the combined or “group” dose may be of interest. The results for the entire data set, weighted by assay as above, can be accessed by pressing the “Group data” button on the right-hand side of the data entry screen. Pressing the button will cause a dialog to appear with a text box at the top and a chart, comparing all the individual Case_ID doses, as illustrated in Fig. 4. The data and results can be exported to a Microsoft Excel® file (.xls format) by clicking on the “Export data” button at the bottom right-hand corner of the data entry graphic user interface (illustrated in Fig. 3). Copying and pasting within/outside the tables is also available and can be used by right-clicking over the desired selection of cells. A “troubleshooting” section has been provided in the software manual, which has been created following extensive testing by the MULTIBIODOSE project partners.
Fig. 3. The graphic user interface with details of the current data for Test_Incident, after the “Calculate” button has been pressed. www.health-physics.com
Copyright © 2014 Health Physics Society. Unauthorized reproduction of this article is prohibited.
Emergency triage software c E. A. AINSBURY
Fig. 4. The “Group” dosimetry results for the Test_Incident data.
This section can be consulted in the event that the user experiences problems with operation or running of the software. If a problem or error occurs that is not listed in this section, the user is encouraged to contact the software administrator to ask for further help. RESULTS AND DISCUSSION Validation and testing The software and manual were finalized and tested with the assistance of a large number of project participants. At least one representative from each project work package (focused on the MULTIBIODOSE assays) tested the software and, in addition, eight individuals from across the different work packages were selected as “debugging” testers. These individuals tested the software thoroughly with their own data on a number of different platforms and configurations, including Windows, Mac, and Linux systems, and with up to 10,000 data sets. The testers attempted to identify problems with the functionality. A number of issues were identified and resolved, as a result of which the final version of the software is now extremely stable. The ability of the software to assign appropriate triage categories was tested using data from the MULTIBIODOSE biodosimetry exercise that took place in November 2012, the results of which are reported elsewhere (Ainsbury et al. 2014). However, in brief, three biological dosimetry tools (the dicentric, micronucleus, and gamma-H2AX foci assays)
ET AL.
87
were tested, in addition to provision of the triage status results (low exposure: < 1 Gy; medium exposure: 1–2 Gy; high exposure: > 2 Gy) by the MULTIBIODOSE software. The exercise was run in two modes: an initial triage categorization of 124 samples (based on the first dose estimates received from each laboratory) followed by collation of the full set of 253 estimated doses calculated using as many modes of operation as possible of the different assays developed during the project. Simulated acute whole body and partial body exposures were included. The software was able to assign an appropriate category based on the data in all scenarios, and individual case/group doses were also always correct within the limits of the appropriate triage category. The MULTIBIODOSE software described here has been developed as part of an academic research project, and, as such, no formal validation or benchmarking of the software has taken place other than that described above. However, if, as is hoped, the software can be developed further during the course of other ongoing biodosimetry networking projects, the authors plan to move toward formal benchmarking of the methods and techniques used in the software. Discussion of limitations Despite the successful outcome of the MULTIBIODOSE exercise, the software has, to date, only been tested with a limited amount of data, and a number of limitations remain to be addressed. These include the fact that the scheme for assignment of “weights,” which allow the software to provide a combined estimate of dose based only on the appropriate assays for an exposure scenario, is only very simplistic in nature. The decision about the simplicity of the weighting scheme was made by the consortium as a whole, on the basis of the fact that the tools have not yet been tested in the full range of possible exposure scenarios. Another current constraint lies in the fact that the incident parameters (exposure scenario) cannot be changed once an incident has been set up. Again, the reason for this is that it was judged to be likely that, initially, emergency responders will not have a great deal of information about the scenario (i.e., the likely magnitude of the doses received by individuals, even perhaps the timescale of exposure). It is currently recommended that once such information is forthcoming, sub-incidents should be set up by the team managing the incident; data can be copied over using the copy and paste facilities provided. It may also be seen as a limitation that it is not currently possible for separate laboratories involved in the incident response to enter information concurrently into a live incident database. The reason for this is that, in most cases, it was judged that national authorities dealing with a radiological accident or incident would generally
www.health-physics.com
Copyright © 2014 Health Physics Society. Unauthorized reproduction of this article is prohibited.
88
Health Physics
prefer to have a single laboratory managing the retrospective dosimetry response to the incident. As detailed in the Guidance (Jaworska et al. 2013), the managing laboratory should ideally be a MULTIBIODOSE partner laboratory or a member of a current networking project. Additionally, if only one laboratory has access to personal data, this also avoids issues related to data sharing and data protection across the large number of international project partners. Finally, the choice of the Java platform for the software may prove to be an issue for some users. Although any software platform comes with its own challenges, the Java platform was chosen as it is known to be reliable across a wide range of operating systems. Care was taken to test the software on as many different systems as possible within the consortium; however, initial testing did reveal some compatibility issues, particularly between different countries. Although the issues across the consortium have now been resolved, the sheer number of different systems in use around the world means it is likely that some users will still experience compatibility or other problems. In order to address this as far as possible, users experiencing such technical issues are asked to contact the software administrator, and problems will then be resolved as and when they occur. To address all of the above matters, it is intended that further testing of the software will take place as part of future inter-laboratory comparisons, and further development will be carried out in response to the needs of the community. In particular, testing and optimization of the software will need to be carried out with very large numbers of samples, and it is highly likely that amendments to the weighting scheme will become necessary as more data is collated.
CONCLUSION The MULTIBIODOSE software provides a tool that facilitates the combination of doses estimated by networks of laboratories using five retrospective (biological and physical) assays, which are combined to give a single estimate of triage category for each individual potentially exposed to ionizing radiation in a large scale radiation accident or incident. The software has been initially tested and validated as part of the MULTIBIODOSE biodosimetry intercomparison and emergency response exercise, but further testing and validation by the consortium members is intended, and testing by the wider community is also welcomed. The software and all associated files are freely available to download from www.multibiodose.eu/ software or by contacting the software administrator: [email protected].
July 2014, Volume 107, Number 1 Acknowledgments—The research detailed in this manuscript has received funding from the European Union’s Seventh Framework Programme (FP7/ 2007–2013) under grant agreement number 241536.
REFERENCES Ahmed EA, Agay D, Schrock G, Drouet M, Meineke V, Scherthan H. Persistent DNA damage after high dose in vivo gamma exposure of minipig skin. PLoS ONE 7:e39521; 2012. DOI:10.1371/journal.pone.0039521. Ainsbury EA, Al-Hafidh J, Bajinskis A, Barnard S, Barquinero JF, Beinke C, de Gelder V, Gregoire E, Jaworska A, Lindholm C, Lloyd D, Moquet J, Nylund R, Oestreicher U, Roch-Lefevre S, Rothkamm K, Romm H, Scherthan H, Sommer S, Thierens H, Vandevoorde C, Vral A, Wojcik A. Validating a multibiodosimetric approach to triage in a radiation emergency. Int J Radiat Biol 90:193–202; 2014. Blakely WF, Ossetrova NI, Manglapus GL, Salter CA, Levine IH, Jackson WE, Grace MB, Prasanna PGS, Sandgren DJ, Ledney GD. Amylase and blood cell-count hematological radiation-injury biomarkers in a rhesus monkey radiation model—use of multiparameter and integrated biological dosimetry. Radiat Meas 42:1164–1170; 2007. DOI: 10.1016/j. radmeas.2007.05.013. Garty G, Chen Y, Salerno A, Turner H, Zhang J, Lyulko O, Bertucci A, Xu Y, Wang H, Simaan N, Randers-Pehrson G, Yao L, Amundson SA, Brenner DJ. The RABIT: a rapid automated biodosimetry tool for radiological triage. Health Phys 98:209–217; 2010. DOI: 10.1097/HP.0b013e3181ab3cb6. Guipaud O. Serum and plasma proteomics and its possible use as a detector and predictor of radiation diseases. Adv Exp Med Biol 990:61–86; 2013. DOI: 10.1007/978‐94‐ 007‐5896‐4_4. Horn S, Barnard S, Rothkamm K. Gamma-H2AX-based dose estimation for whole and partial body radiation exposure. PLoS ONE 6:e25113; 2011. DOI: 10.1371/journal.pone.0025113. International Atomic Energy Agency. Cytogenetic dosimetry: applications in preparedness for and response to radiation emergencies. Vienna: IAEA; IAEA Emergency Preparedness and Response Series, EPR-Biodosimetry; 2011. JFree. Welcome to JCommon! [online]. 2013a. Available at www.jfree.org/jcommon/. Accessed 4 July 2013. JFree. Welcome To JFreeChart! 2013b. Available at www.jfree. org/jfreechart/. Accessed 4 July 2013. JUnit. Junit, a programmer-oriented testing framework for. 2013. Available at http://junit.org/. Accessed 4 July 2013. Jaworska A, Wojcik A, Ainsbury E, Fattibene P, Lindholm C, Oestreicher U, Rothkamm K, Romm H, Thierens H, Trompier F, Voisin P, Vral A, Woda C. Guidance for using MULTIBIODOSE tools in emergencies, for radiation emergency response organisations in Europe. 2013. Available at www. multibiodose.eu/News/MBD_Guidance_web.pdf. Accessed 4 July 2013. Khan A. Java Excel API—a Java API to read, write and modify Excel spreadsheets [online]. Available at www.andykhan.com/ jexcelapi/. Accessed 4 July 2013. Kulka U, Ainsbury L, Atkinson M, Barquinero JF, Barrios L, Beinke C, Bognar G, Cucu A, Darroudi F, Fattibene P, Gil O, Gregoire E, Hadjidekova V, Haghdoost S, Herranz R, Jaworska A, Lindholm C, Mkacher R, Mörtl S, Montoro A, Moquet J, Moreno M, Ogbazghi A, Oestreicher U, Palitti F, Pantelias G, Popescu I, Prieto MJ, Romm H, Rothkamm K, Sabatier L, Sommer S, Terzoudi G, Testa A, Thierens H, Trompier F, Turai I, Vandersickel V, Vaz P, Voisin P, Vral A, Ugletveit F, Woda C, Wojcik A. Realising the European network of biodosimetry (RENEB). Radiat Protect Dosim 151: 621–624; 2012.
www.health-physics.com
Copyright © 2014 Health Physics Society. Unauthorized reproduction of this article is prohibited.
Emergency triage software c E. A. AINSBURY
Maznyk NA, Wilkins RC, Carr Z, Lloyd DC. The capacity, capabilities and needs of the WHO Biodosenet member laboratories. Radiat Protect Dosim 151:611–621; 2012. Oracle. Java SE 7 downloads. 2013a. Available at www.oracle.com/ technetwork/java/javase/downloads/java-archive-downloadsjavase7-521261.html#jdk-7u3-oth-JPR. Accessed 4 July 2013. Oracle. NetBeans IDE [online]. 2013b. Available at https:// netbeans.org/. Accessed 4 July 2013. Romm H, Ainsbury E, Barnard S, Barrios L, Barquinero JF, Beinke C, Deperas M, Gregoire E, Koivistoinen A, Lindholm C, Moquet J, Oestreicher U, Puig R, Rothkamm K, Sommer S, Thierens H, Vandersickel V, Vral A, Wojcik A. Automatic scoring of dicentric chromosomes as a tool in large scale radiation accidents. Mutat Res 756:174–183; 2013. DOI: 10.1016/j. mrgentox.2013.05.013. Romm H, Wilkins RC, Coleman NC, Lillis-Hearne PK, Pellmar TC, Livingston GK, Awa AA, Jenkins MS, Yoshida MA, Oestreicher U, Prasanna PGS. Biological dosimetry by the triage dicentric chromosome assay: potential implications for treatment of acute radiation syndrome in radiological mass casualties. Radiat Res 175:397–404; 2011. DOI: 10.1667/RR2321.1. Rothkamm K, Barnard S, Ainsbury EA, Al-hafidh J, Barquinero JF, Lindholm C, Moquet J, Perälä M, Roch-Lefèvre S,
ET AL.
89
Scherthan H, Thierens H, Vral A, Vandersickel V. Manual versus automated gamma-H2AX foci analysis across five European laboratories: can this assay be used for rapid biodosimetry in a large scale radiation accident? Mutat Res 756:170–173; 2013. DOI: 10.1016/j.mrgentox.2013.04.012. Trompier F, Bassinet C, Wieser A, De Angelis C, Viscomi D, Fattibene P. Radiation-induced signals analysed by EPR spectrometry applied to fortuitous dosimetry. Ann Ist Super Sanità 45:287–296; 2009. Werner C. SQLite java wrapper/jdbc driver. 2013. Available at www.ch-werner.de/javasqlite/. Accessed 4 July 2013. Willems P, August L, Slabbert J, Romm H, Oestreicher U, Thierens H, Vral A. Automated micronucleus (MN) scoring for population triage in case of large scale radiation events. Int J Radiat Biol 86:2–11; 2010. DOI: 10.3109/09553000903264481. Woda C, Bassinet C, Trompier F, Bortolin E, Della Monaca S, Fattibene P. Radiation-induced damage analysed by luminescence methods in retrospective dosimetry and emergency response. Ann Ist Super Sanità 45:287–296; 2009.
www.health-physics.com
Copyright © 2014 Health Physics Society. Unauthorized reproduction of this article is prohibited.
■■
