Development of an online automatic computed radiography dose data mining program: A preliminary study
Introduction
Recent studies have reported the computed radiography (CR) dose creep problem and the need to implement dose monitoring process in place in clinical departments [1], [2], [3]. The dose creep problem is the effect of several causes. The wide dynamic range of CR systems enables technologists to produce acceptable or even better quality images in usual overexposure situations. However, underexposure remains undesirable due to quantum mottle (noise). A tendency has occurred to overexpose patients for obtaining acceptable or better images. This problem can propagate because good image quality is always favoured by radiologists but overexposure is difficult to identify when looking at CR images [2], [4], [5]. This problem can be overcome through quality assurance (QA) programming by evaluating exposure indices of CR images regularly [1], [2], [6]. Exposure indices are dose feedback indicators provided by CR manufacturers to show relative radiation amounts that reach image plates such as AGFA's logarithm of median exposure (Lgm), Fuji's sensitivity number (S-number) and Kodak's exposure index (EI) [6], [7], [8], [9]. This information is stored in the Digital Imaging and Communication in Medicine (DICOM) header of an image but in different tags (AGFA's Lgm in tag 0019,1015 [1]; Fuji's S-number in tag 0018,6000 [2], [7] and Kodak's EI in tag 0018,1405 [8]).
In the study conducted by Stewart et al. [2], a dose monitoring program was implemented to monitor Fuji CR dose. A Fuji data mining program running on a Picture Archiving and Communication System (PACS) (Centricity 2.0, GE Healthcare) was executed automatically once per month. This program extracted S-numbers and other relevant information such as accession number, study date, study time, study description and station name from DICOM headers of images archived in the PACS of the previous month. The main weakness of this Fuji program is that it is only able to extract the S-number from the first image of each examination. A manual process was developed to export monthly missing information, i.e. exposure result log from the CR stations to a floppy disk. Then, a visual basic (VB) program was used to integrate this text file with the comma separated values (CSV) file output from the data mining program for generating a dose report using Microsoft Excel manually. In a similar study conducted by Juste et al. [1], an offline executable program (compiled by MathWorks Matlab 7.3) was developed to extract Lgm values from DICOM headers of AGFA CR images exported from archive for dose monitoring purposes. Again, this requires a manual process to accomplish the export task. The objective of this study is to provide a better technological solution to implement a regular dose monitoring process, i.e. to develop an online automatic CR dose data mining program which can be applied to different CR systems including AGFA, Fuji and Kodak with the aim of reinforcing the ‘As Low As Reasonably Achievable’ (ALARA) principle [4], [10].
Section snippets
Materials and methods
At our institution, a freeware PACS, ConQuest DICOM server 1.4.13 [11], running the Microsoft SQL Server 2005 database management system (DBMS) was installed in a Dell PowerEdge 2950 machine with a Microsoft Server 2003 Operating System (OS). A web server, Internet Information Services (IIS) 6.0, bundled with the OS was also installed for a web-based PACS service. The primary purposes of this PACS are for archiving DICOM files from two AGFA CR stations and a VIDAR film digitizer in student
Results
Using the proposed model shown in Fig. 1, an online automatic CR dose data mining program which can be applied to different CR systems including AGFA, Fuji and Kodak, was developed. There is a login page for authentication to secure the information. Other security measures include timestamp to show the last login time and ‘Logout’ once finished using the system. A user is directed to the dose monitoring system through the login page upon successful authentication (Fig. 2). In Fig. 2, the first
Discussion
Evidence to date suggests an online automatic CR dose data mining program can be successfully developed. It addresses the major weaknesses of some existing studies, i.e. involvement of manual procedures and being specific to only a single manufacturer's CR images [1], [2]. It seems the limitation of the latter issue is not significant because a clinical department usually has CR systems supplied by one manufacturer. However, involvement of manual procedures limits the potential of the systems
Conclusion
An online automatic CR dose data mining program which can be applied to different CR systems was developed and has been tested with AGFA and Fuji CR images. Our preliminary study shows that the program addresses the major weaknesses of some existing studies, i.e. involvement of manual procedures and being only specific to a single manufacturer's CR images. This provides an efficient and effective solution to the implementation of a regular CR dose monitoring program in busy clinical
Conflict of interest
None declared.
Acknowledgement
We gratefully acknowledge Dr. Janice C. McKay for the manuscript review.
References (16)
- et al.
Advances in computed radiography systems and their physical imaging characteristics
Clin. Radiol.
(2007) - et al.
Development of a web database portfolio system with PACS connectivity for undergraduate health education and continuing professional development
Comput. Methods Programs Biomed.
(2009) - et al.
Improving pediatric radiation dose management using Agfa digital radiography DICOM header information
- et al.
Computed radiography dose data mining and surveillance as an ongoing quality assurance improvement process
Am. J. Roentgenol.
(2007) - et al.
Correlation of image quality with exposure index and processing protocol in a computed radiography system
Radiat. Prot. Dosim.
(2008) Strategies for dose reduction in ordinary radiographic examinations using CR and DR
Pediatr. Radiol.
(2004)- et al.
Evaluation of exposure index (IgM) in orthopaedic radiography
Radiat. Prot. Dosim.
(2008) Management of pediatric radiation dose using Fuji computed radiography
Pediatr. Radiol.
(2004)
Cited by (12)
Investigating perceptions of ‘dose creep’ amongst student radiographers: A grounded theory study
2021, RadiographyCitation Excerpt :The dose creep phenomenon is widely recognized within the radiographic literature, whereby incremental X-ray exposures fail to impede image quality, yet, increases dose to patients.1
Evaluating the use of exposure indicators in digital x-ray imaging system: Gauteng South Africa
2019, RadiographyCitation Excerpt :The sealed boxes were collected three weeks later from each of the research sites. Open ended questions on the familiarity and use of the EI were coded as perfectly correct, partially correct and incorrect and as incorrect use and correct use as derived from current literature on EI (Table 1).4,11–15,27 Data collected was coded and then analysed using IBM SSPS V.23.
Image quality and dose analysis for a PA chest X-ray: Comparison between AEC mode acquisition and manual mode using the 10kVp 'rule'
2014, RadiographyCitation Excerpt :The problem is confounded further by the lack of guidelines on adapting from analogue to digital systems. Currently no consensus exists in the adjustment of the parameters of exposure.9–12 To optimize E it is necessary that imaging professionals understand the equipment that they are using.
Development of an online automatic diagnostic reference levels management system for digital radiography: A pilot experience
2011, Computer Methods and Programs in Biomedicine
- 1
Tel.: +61 8 92667509; fax: +61 8 92662377.