Achievable Radiation Dose Reduction with Comparable Image Quality in Chest Radiography

Full Article

AYH Wan, MH Shih, BMH Lai, CY Chu, KYK Tang, RTM Chan, JLS Khoo

Hong Kong J Radiol 2014;17:182-8

DOI: 10.12809/hkjr1413198

Objective: Chest radiography is one of the commonest radiological investigations utilised in various medical specialties. Although the radiation dose of a single chest radiography to the patient is relatively low, the contribution of the accumulated dose is substantial due to its frequent use in medical diagnoses. Optimisation of radiation dose and image quality, therefore, remains a challenging area in research and routine practice. The objective of this study was to evaluate the radiation dose in chest radiograph (CXR) taken with a sensitivity value of 400 (as in factory setting) and to perform quality assessment of the diagnostic applicability of dose-reduced CXRs obtained with a higher sensitivity value (of 600).

Methods: CXRs performed in 100 consecutive adult patients attending a regional hospital in Hong Kong were reviewed. Images were taken with a sensitivity value of either 400 or 600 by an Agfa 85X CR System. Fifty patients were allocated to each acquisition technique and the assignment of this grouping was random. No significant mismatch with respect to patient body size was noted in the two groups. Diagnostic reference levels, estimated with dose area product derived from exposure measurements, were used to quantify the radiation dose. The difference in radiation dose between these two sets of images was statistically evaluated by the Student’s t test. The CXRs were reviewed and rated by two independent radiologists in random order. The image quality of CXRs was assessed by 10 criteria as stated in the European Guidelines On Quality Criteria for Diagnostic Radiographic Images. Differences in image quality ratings between the two acquisition techniques were statistically quantified by analysis of variance with repeated measures. The agreement ratio for each criterion between the two radiologists was calculated. Evaluation of the interobserver agreement for the overall mean score of image quality was performed using weighted Cohen’s Kappa statistics.

Results: A statistically significant reduction in radiation dose of 32.8% (p<0.001) by using image acquisition technique with sensitivity 600 (mean dose area product = 10.00 cGycm2) was noted as compared with the CXR taken with sensitivity 400 (mean dose area product = 14.87 cGycm2). Concerning the image quality of the radiographs, a high level of interobserver agreement was demonstrated in most of the criteria (range, 82.7%-100%). The Kappa score for the overall mean score of image quality between the two raters was 0.547, signifying moderate and clinically acceptable interobserver agreement. No statistically significant difference was noted in the overall mean image quality between the two sets of CXRs taken with different sensitivity values and radiation doses (p = 0.106). Both groups of CXRs showed satisfactory diagnostic applicability for medical practice.

Conclusion: The radiation dose in CXRs obtained with sensitivity 600 was shown to be effectively reduced while maintaining sufficient image quality compared with those taken with sensitivity 400. This setting is recommended for future standard CXR acquisition. Extension of similar clinical audit and quality assurance studies to radiographs of other body regions will be helpful to ensure adherence to the principles of radiation protection.







方法:回顧香港一所分區醫院連續100名成年患者的CXR,圖像由Agfa 85X CR系統拍攝,靈敏度值設置為400或600。所有患者隨機分組,50名患者被分配接受兩種拍攝技術之一種。兩組病人在體型方面並無顯著差異。利用曝光測量得出的劑量面積乘積估算診斷參考水平,再以參考水平量化輻射劑量。用學生t測試來評估兩組所釋放的放射劑量的差異。由兩名獨立的放射科醫生分別隨機審查,然後以歐洲指南中診斷性放射圖像質量標準的十項評分標準為CXR評分。利用重複量數變異數分析把兩種設置之間圖像質量的差異統計量化,計算出兩名放射科醫生在每項評分標準的符合率。採用加權Cohen’s Kappa統計評估圖像質量總平均分的觀察者間一致性。

結果:與靈敏度400(即平均劑量面積乘積=14.87 cGycm2)比較,採用靈敏度600(即平均劑量面積乘積=10.00 cGycm2)所釋放的放射劑量明顯減少了32.8%(p<0.001)。平片圖像質量方面,觀察者間在大多數評分標準上達到高一致性(介乎82.7%至100%)。圖像質量總平均分方面,Kappa系數為0.547,代表觀察者間符合率為中等,臨床尚可接受。兩組不同靈敏度設置及放射劑量的CXR的總平均圖像質量並無統計學顯著差異(p=0.106)。兩組CXR均顯示滿意的臨床診斷適用性。