eduzhai > Helth Sciences > Medical >

Head CT image quality assessment: control chart as a useful tool

  • sky
  • (0) Download
  • 20211031
  • Save
https://www.eduzhai.net American Journal of Biomedical Engineer ing 2013, 3(6A): 1-7 DOI: 10.5923/s.ajbe.201310.01 Image Quality Assessment of Head CT: Control Charts as an Useful Instrument Joana Guiomar1, Luís Pedro Vie ira Ribe iro2,*, Anabe la Magalhães Ribeiro3, António Fernando Lagem Abrantes4, João Pedro Pinheiro5, Rui Pedro Pereira Almeida6, Kevin Barros Azeve do7 1BSc, Algarve, Portugal 2PhD, M ember of the Research Center of Sports and Physical Activity (CIDAF) of Coimbra University, Professor and M ember of the Center for Health Studies (CES) of Algarve´s University Health School (ESSUAlg), Algarve, Portugal 3M Sc, Coordinator Radiographer, Hospitalar Center of BarlaventoAlgarvio, Professor and M ember of the Center for Health Studies (CES) of Algarve´s University Health School (ESSUAlg), Algarve, Portugal 4PhD, M ember of the Research Center of Sociologic Studies of Lisbon´s Nova University (Cesnova), Professor and M ember of the Center for Health Studies (CES) of Algarve´s University Health School (ESSUAlg), Dir ector of the Radiology Department and professor at ESSUAlg, Algarve, Portugal 5Post-graduate, M Sc student at the National Public Health School, Professor of the Radiology Department at Algarve´s University Health School ( ESSUAlg), Algarve, Portugal 6Post-graduate, M ember of the Center for Health Studies (CES), PhD Student at Beira Interior University, Professor and M ember of the Center for Health Studies (CES) of Algarve´s University Health School (ESSUAlg), Algarve, Portugal 7Post-graduate, M ember of the Center for Health Studies (CES), PhD Student at Cranfield University, Professor of the Radiology Department at Algarve´s University Health School (ESSUAlg), Algarve, Portugal Abstract Themain objective of this research was to determine if the diagnostic image, acquired by CT scan, meets the quality criteria previously established for head CT exams. A total of 360 Head Co mputed Tomography exams were analy zed, using a checklist. For data collect ion, quality criteria were created, organized into four criteria groups, consisting of multip le items that must appear in the images studies. After data processing, a large number o f non-conformexaminations were identified in than more than 50% of the sample. We concluded the main causes of these results are: the “incorrect or incomp lete positioning”, the "lack of name of the radiographer and “mot ion artefacts”. Therefore it is essential to implement a checklist for a systematic evaluationof procedures. Keywords Image Quality, Co mputed Tomography, Conformities, Quality Control Charts 1. Introduction The Quality Control of the Co mputed Tomography (CT) image is crit ical, because examinations performed must have good diagnostic quality without resorting to repeated examinations. The relevance of this issue is the fact that there are few studies conducted on this subject and those that arise in the b ibliography are mo re d irected to the Quality Control of Radiology equip ment, not showing the impo rtance of quality in the product (radiologic image). Although equipment co mp lies with quality program, the technique applied by radiographers may not be the most accurate, wh ile performing the imaging study and non-conformit ies occur diminishing the quality of e xa mination. * Corresponding author: lpribeiro@ualg.pt (Luís Pedro Vieira Ribeiro) Published online at https://www.eduzhai.net Copyright © 2013 Scientific & Academic Publishing. All Rights Reserved Our main object ive is to evaluate image quality of Head Co mputed Tomography (HCT), through a benchmark of quality. On the other hand, this goal leads to other mo re specific objectives, such as to determine the image quality of HCT through control charts, performed in aradiology department. To identify the presence of conformities and non-conformit ies in the radiological image,we created corrective actions that may be imp lemented to reduce the incidence of non-conformities and, ultimately, understand the causes of the existence of non-conformities. Therefore, themain objective of this research is to determine if the diagnostic image, acquired by CT scan, meets the quality criteria previously established for head CT e xa ms . This research is supported on a problem that origins various research questions such as: “what is the best control chart that applies the Quality Control o f HCT imag ing?”; “when evaluating HCT images, are they in agree with the quality criteria?”; “what causes the presence of non-conformit ies, including corrective actions to be 2 Joana Guiomar et al.: Image Quality Assessment of Head CT: Control Charts as an Useful Instrument recommendedandeliminate the presence of non-conformities in the HCT images?”. The data obtained in the bibliography on the Quality Control in radio logical imaging is scarce, since the predominancestudies areabout equipment quality. However, it is possible to find studies related to the Quality Control of radiological image in the valences of general radiology and computed tomography[1,2,3,4,5]. 2. Methods HCTstudies are the imaging exams with a higher incidence inRad iology Department, and the majority of patients come fro m the emergency. We performed a case study that employed quantitative methods, containing a sample of 360 HCT studies performed between February to December 2012 and randomly selected in order to determine if the diagnostic image, acquired by CT scan, meets the quality criteria prev iously established for head CT exams, without knowledge of the radiology staff.This sample was organized into 12 subsamples, with 30 HCTscans each. Patients and Radiographers were not in formed about the purpose of the study in order to eliminate bias and permission to collect data was acquired fro m the ethics board. For data collection,quality criteria were created, guidelines for performing and interpreting diagnostic computed tomography, organized into four criteria groups, consisting of mult iple items that must appearin the images studies[6,7]. We considered conform when they are present in the study, in other words, requirements were obtained (quality criteria), previously established, as listed below: • Preparat ion of a correct and co mplete exam: ο Presence of the name of the Radiographer; ο Presenceof the name of the Rad iologist; ο Presence of the Pat ient's name and identification number. • Correct positioning; • Absence of artefacts in the image: ο Motion (patient motion); ο Metals (dental prostheses, and other foreign objects); ο System (artefact problems associated with the operation of the equipment); ο Image noise; • Criteria for successful e xecution: ο FOV (field of v iew) suited to study the structure; ο Reference dose levels appropriate to the study; ο Examination appropriate to the aims of the study. In the absence of at least one of these quality criteria we consider the presence of non-conformities, making the study non-conform. The data collect ion took p lace in March 2013, using a checklist, in which there was the time, day, month and year in which the HCT studies were performed as well as the type of non-conformities identified. Furthermo re, this consisted of the total ofnon-conformand conformstudies in each sample, organized by work shift (morn ing, afternoon and night). In accordance to institutional guidelines, the approval of this study was obtained fro mthe review board and the data of the patients selected were kept confidential. For data analysis, we used the Statistical Package for the Social Sciences (SPSS) V.20 and a control chart with the type attributes p, np and c with Microsoft Office Excel 2007. These charts classify a product as Conform or Non-conform and assess the overall image quality. A Type p Control Chart was made based on the proportion of Non-conformity exams of the samples. A Type np Control Chart was made based on the total number of non-conformity exams of each sample. A Type c Control Chart was made based on the total number of Non-conformit ies of each sample. 3. Results 3.1. Quality Control Charts Chart 1. p-type, with the limits and warning control, compared to CL American Journal of Biomedical Engineer ing 2013, 3(6A): 1-7 3 Chart 2. Type np, with limits and warning control, compared to CL Chart 3. Ttype c, with the Limit and Warning Control, compared to CL The quality control charts for attributes used in this research were the type p, np and c.The first step involved the calculation of the percentage of defective units for each sample, to obtain values for the control limits. The chart p, wh ich illustrates the percentage of non-conformHCT studies, we obtained the fo llo wing values of the Central Limit (CL), Upper Control Limit (UCL) and Lower (LCL): 60.83%, 87%, 57% and 34.09%.There were no samples out of statistical control and consequently warning limits were established, according ISO 8258:1991, which allow study the tendency of the samples distribution. With the following values: 1???????? (69.74%), 2???????? (78.65%), -1???????? (51.92%) e -2???????? (43.01%) (Chart 1). On the other hand, the values of the control limitscollected for the attribute type np (Chart 2), responsible for assessing the number of HCT studies non-conform are: CL (18.25), UCL (26.27) and LCL (10.23). As in p-type chartthere were no samples out of control limits, and it was established the following warn ing limits: 1???????? (20.92), 2???????? (23.6), -1???????? (15.58) e -2???????? (12.9). To study the number of non-conformities that arise in each sample was used c type chart control (Chart 3), and were calculated the fowling Control Limits, CL (24.92), UCL (39.89) and LCL (9.94). Warning Control limits to this chart are 1???????? (29.91), 2???????? (39.9), -1???????? (19.93) e -2???????? (14.93). 3.2. Quality Criteri a In the total sample observed, 219 HCT studies were not conform, since they didn’tsatisfied at least one of the quality criteria item, prev iously defined. Thus, 60.83% of the samp le non-conform. The main explanation for the high percentage of non-conform studiesis the large number of non-conformities detected during the data collection. Table 1 illustrates, through the relative (fi) and absolute (Fi) frequency which g roups of non-conformities that contributed to the presence of not conform studies. 4 Joana Guiomar et al.: Image Quality Assessment of Head CT: Control Charts as an Useful Instrument Table 1. Absolute frequency, fi and Fi, for groups of non-conformities Group of N/C Type of N/C Absence of the Technician Exam pre paration incorre ct or in com plete Radiology Nonexistence of the name of the Radiologist P at ient’s name and identification number missing Incorrect positioning Mot ion Image Artefacts Met allic Sy st em Noise Absence of crite ria for good achie ving Inappropriate FOV NRD increasing Absence of criteria for study Total Amount of N/C 1 81 0 128 54 9 10 0 3 13 0 Absolu t fre quency 82 128 73 16 299 fi 27,42% 42,81% 24,41% 5,35% Table 2. Absolute frequency fi and Fi, for each type of non-conformities Group of N/C Type of N/C Exam pre paration incorrect or in com plete Absence of the Technician Radiology Nonexistence of the name of the Radiologist P at ient’s name and ident ificat ion number missing Incorrect positioning Mot ion Image Artefacts Met allic Sy st em Noise Absence of crite ria for good achie ving Inappropriate FOV NRD increasing Absence of criteria for study Total Amount of N/C 1 81 0 128 54 9 10 0 3 13 0 299 fi 0,33% 27,09% 0% 42,81% 18,06% 3,01% 3,34% 0% 1% 4,35% 0% Fi 27,42% 70,23% 94,65% 100% Fi 0,33% 27,42% 27,42% 70,23% 88,29% 91,30% 94,65% 94,65% 95,65% 100% 100% On the other hand, the types of non-conformit ies, wh ich arise more often, consist of "incorrect positioning" (42.81%), in the "absence of the name of Rad iologist" (27.09%) and "motionartefacts" (18.06%) (Table 2). In addition, we evaluated the number of HCT studies non-conform per shift, in the morn ing it were observed 100 examinations, in the afternoon173and 87 in the night.It was also found that in the afternoon shift the percentage of non-conformwas higher than in the other two shifts, yielding a value of 28.06%, followed by the night shift (17.22%) and the morning shift (15.56%). 4. Discussion 4.1. Interpretation of Control Charts The quality control charts more suited to the study of data collected, are the attributes of type p, np and c, because they allo w tostudy the percentage of non-conformstudies, the number of non-conformities exams and non-conformities in each sample, respectively. For the charts type p and np planning, it was essential to calculate the percentage of non-conform studies per sample, while for charts typec, we calculated the proportion ofnon-conformities per samp le. Subsequently, we determined the values of CL and UCL and LCL for each chart. For the p-type control chart, the values obtained for the central limit, limit control top and bottom, corresponded, respectively, to 60.83%, 87.57% and 34.09%. It was found, with the preparation of Chart 1 that the samples were within normal limits, it is not necessary to remove outliers and identify non-conformities that were out of statistical control, wh ich is the benchmark for the control chart type p in accordance with ISO 8258:19916. Taking into account the values obtained for the control limits of this chart, it is expected that, in each observed sample, appro ximately, 61% of the thirty elements arise on average as non-conform. On the other hand, the p-type control chart imposes limits on the percentage American Journal of Biomedical Engineer ing 2013, 3(6A): 1-7 5 ofnon-conformexams that may arise. In this way, we have a maximu m o f 87.57% and a min imu m o f 34.09%. The values mentioned above, are high, but it can be explained by the high number of non-conformities recorded (Tables 1 and 2). Another justification for the large gap between the UCL and ICLcould be the small number of elements containing in the twelve samples, ie, a small n, conditioning so that the process is always under statistical control, essentially in the control charts typep, but, if the range is reduced (samples with large n), imply ing that the process get out of statistical control. The control charter of nptype, wh ich study the number of non-conformexams in the sample, had as reference for the limits of control, the following values: LC (18.25%); LCS (26.27%) and LCI (10.23%). Not needing adjustments, like the charter of the p-type control, it was added to the warning top and bottom limits (Chart 2), wh ich, according to ISO 8258:1991[8], considers the model obtained as a reference standard for control chart type np. Interpreting this chart, one comes to the conclusion that in this study, a sample of 30 elements, there may be an average of 18 exams which did not conform, but for the same samp le the maximu m HCT studies non-conform is 26 and the minimu m 10. However, these values are strongly influenced by the number of non-conformities identified in each of the thirty elements of the sample. To study the number of non-conformities that arise in each sample, it was constructed the control charter type c, which have as values for CL, UCL and LCL, the ones shown in Chart 3, such as the warning limits used to study the trend of non-conformities identified in 12 samples evaluated, taking into account the ISO 8258:1991. In each samp le, it can be registered, on average, about 25 non-conformities, the maximu m being 39.89 percent non-conformit ies to be identified and the minimu m 9.94. If in a given sample, these values are exceeded, it will consider the process out of statistical control. Table 3. Rules to study the trend of the samples in Control Charts Rule De signation 1 any point out side the limit s of control 2 9 consecut ive point s on the same side of the cent er line 3 6 consecutive points upward or downward 4 14 points alternately increasing and decreasing 5 2 or 3consecutive points in zone A the same side of center line 6 4 or 5 consecutive points in zone B or A on the same side of center lineor A, same side as central line 7 15 consecutive points in zone C 8 8 points on both sides of the central line, no zone C Source:[3] The ISO 8258:1991 defends eight rules to investigate the tendency of samples in each control chart, in order to understand if thereis statistical control (Tab le 3). If there is the presence of, at least one of the rules set out, the process is out of control. In that case, it’s fundamental to detect what is/arethe special cause(s) underlying in order to eliminate it/them to prevent recurrence. Before starting the application of the rules mentioned above, it was taken into account that, in this research, we have twelve samples,because of that the rules 4, 7 and 8 can’t be applied. Thus, it follows that the references patterns in each chart are in statistical control, because there was no occurrence of any of the rules presented. 4.2. Study of Non-conform Exams Given the fact that the total sample of observed HCT studies, we proceeded to calculate the relative frequency to determine the percentage of non-conformHCT studies and found that approximately three fifths of the studiedpopulation arenon-conform testing, and only two-fifths areconform. In other wo rds, more than half o f the elements of the total sample (n = 360) have non-conformities, which may be, in the short and mediu m term, fixed through corrective and appropriate actions targeted for each identified problem, increasing the level of quality fro mthe radiological studies performed in the Imagiology Depart ment. Furthermore, we proceeded to the evaluation of HCT studies non-conform per shift (morn ing, afternoon and night). It is important to note that the number of elements in each round is not uniform, despite the fact that, it was found that the number of non-conform studies prevails mainly during the afternoon shift followed by morning and night shift. The high proportion of non-conform HTC studies is justified by the number of non-conformit ies detected during data collection. 4.3. Data Analysis for Non-conformities Within the four groups previously defined for the study of non-conform, incorrect positioning of the group, was responsible for almost half of all non-conform obtained, followed by the group of Preparation of an incomp lete or incorrect, Artefacts in the image and Absence of good criteria achievement (Tab le 1). The most common type ofnon-conformities werethe group of exam p reparation incomp lete or incorrect, consisted in "Absence Name Rad iologist". However, in addition to the high percentage of artefacts in the image, there is a predominance of " motionartefacts" beyond, the “systemartefacts” and the "metallic" ones (Table 2). On the other hand, a small quantity of non-conformities identified corresponds to the "increase of the diagnostic reference levels" belonging to the group of the “lack of proper performance criteria”. This total non-conformities explains the high number of HCT studies not conform, noting that the main existingnon conformitiesboils down to "incorrect positioning", "motionartefacts" and "nonexistence of the name of the Radio logist". The first two affect image quality, both interpretation and diagnosis, as established in quality criteria, it is essential, fro m this analysis, to create Fishbone diagram (cause-effect) (Figure 1), to realize the source of 6 Joana Guiomar et al.: Image Quality Assessment of Head CT: Control Charts as an Useful Instrument non-conformities to create corrective, objective and appropriate actions, so that they can be applied in order to reduce those undesirable effects aiming to obtain quality imag es [9] . Briefly, it can be noted that the main causes for these three major non-conformities are: unstable or uncooperative patient, elderly , overwork, fat igue, incorrect positioning and wrong immobilizat ion technique, among others. 5. Conclusions Through the literature search performed, we can see that the quality control o f aradiological imageto perform diagnosis is recent and taking the first steps. The termquality is closely linked to the equip ment’s ability to produce ananatomical image of the patient, regardless the subsequent d iag n os is . So, it is equally important to assess the quality of the radiological technique, since the radio logical image is the product obtained in the radiology department and, this, like any other product that may be acquired in any other type of industry, must contain certain requirements. Thus, it is fundamental to perform a control process of image acquisition, with the aid of an appropriate quality criteria targeted to the intended study and to use control charts to see if the process is within statistical control. The ISO 9001:2008 regulates the radiology department, and defines the work instructions for the conduct of all examinations performed[10]. Nevertheless, there isn’t a program of quality control that focuses in all rad iological examinations carried out,whichmay proceed to the construction of cause and effect diagrams. Finding the causes of the problems is the only way to suggest corrective actions that may be applied, also providing training and rewards for good results. A program of quality control, main ly in Rad iology Depart ment, provides a uniform final product and encourages Radiology Technicians to be aware, confident, critical, innovative, demanding and discerning in their work. Having regard to the objectives set for this study, we verified the existence of a large nu mber of non-conform HCT studies, in which the major non-conformities detected were "incorrect positioning", the "nonexistence of the name of the Radiographer" and "motionartefacts." These can be corrected if a service organization develops quality control programs directed to the same goals and to the institution, as well as fo r the implementation of suggestions given in this work. Fi gure 1. Diagram of cause and effect for non-conformit ies "incorrect or incomplet e posit ioning” American Journal of Biomedical Engineer ing 2013, 3(6A): 1-7 7 Co mpared with other studies on the same subject but in the area of General Rad iology and although these are case studies, is not possible to generalize the results to other institutions. We come to the conclusion that, in both studies, the major non-conformities detected were "incorrect positioning" structures to study[1,5]. Nevertheless is important to mention that none of the HCT studies were repeated, becaus e it was possible perform the medical report. This research allowed tocreate a critica l attitude regarding the implementation of HCT, i.e, to identify non-conformities in image and develop solutions that can be applied in most cases, and to increase knowledge of quality control. As our main suggestions for future work, we reco mmend the following: the application of reco mmended corrective actions and making of a new assessment; imp lementation of a framework of control for HCT image; Study extension to other anatomical reg ions; performance o f quality controls in other hospitals and evaluation of the influence of shift work on the quality of radiological imag ing. REFERENCES [1] Cândido, S., Ribeiro, A. M ., Ribeiro, L. P., Abrantes, A. F., Pinheiro, J. P., Azevedo, K. B. et al (2012). Assessment of radiological imaging conformities based in quality control charts.Insights into Imaging, 4 (suppl 1),S1-SS45:384. Vienna, Austria: Springer. doi: 101007/s13244-01-03228-x [2] Choi, B., Choi, D., Huh, K., Yi, W., Heo, M., Choi, S. et al. (2012). Clinical image quality evolution for panoramic radiographiy in korean dental clinics. Imaging Science in Den tistry, 42: 183-90. [3] Exler, R. B. e Lima, C. J. B.(2012).Controlo estatístico de processos (CEP): uma ferramenta para melhoria da qualidade. Journal of Administração e Contabilidade, 4(3): 78-92. [4] Pomerantz, S. M ., Daly, B., Krebs, T. L., NessAiver, M ., Kepes, S. Y., Wong, J. J. et al (1996). Quality assurance for abdominal CT: a rapid, computer-assisted technique. AJR, 167: 1141-1145. [5] Proença, J. A. (2008). A contribuição da implementação do controlo da qualidade da imagem radiológica digital para melhoria contínua da qualidade num serviço de imagiologia.Porto: J. A. Proença. Dissertation submitted to the University of Fernando Pessoa as part of the requirements for the degree of M aster in Quality. [6] ACR practice guideline for performing and interpreting diagnostic computed tomography. In: Practice Guidelines and Technical Standards. Reston, Va: American College of Radiology; 2011:1-8. [7] ACR-ASNR practice guideline for the performance of computed tomography (CT) of the brain. In: Practice Guidelines and Technical Standards. Reston, Va: American College of Radiology; 2010:1-8. [8] Shewhart control charts. ISO 8258:1991. 1991. [9] Nancy R. Tague’s The Quality Toolbox, Second Edition, ASQ Quality Press, 2004, pages 247–249. [10] Norma Portuguesa 9001:2008 (2009). Sistemas de Gestão da Qualidade: Implementação, Lisboa, Instituto Português da Qualidade.

... pages left unread,continue reading

Document pages: 7 pages

Please select stars to rate!

         

0 comments Sign in to leave a comment.

    Data loading, please wait...
×