Relationship between the visual evaluation of pathology visibility and the physical measure of low contrast detail detectability in neonatal chest radiography.

INTRODUCTION: The detectability of low contrast detail (LCD) is a method used to assess image quality (IQ) in neonatal radiography; however, there is a lack of data on the relationship between LCD detectability and visual IQ. The study aims at investigating the relationship between the LCD detectability and visual IQ and pathology visibility (PV). METHODS: Several acquisition parameters were employed to obtain a group of images from a neonatal Gammex chest phantom. Three observers applied relative visual grading analysis (VGA) for assessing the IQ and PV. A simulated pneumothorax visibility (PNV) and simulated hyaline membrane disease visibility (HMV) represented PV. Next, a CDRAD 2.0 phantom was radiographed utilising the same acquisition protocols, and several paired images were obtained. With the use of CDRAD analyser software, the detectability of LCD was assessed and expressed by an image quality figure inverse (IQFiinv) metric. The correlation between the IQFinv and each of IQ, PNV and HMV was examined. RESULTS: The physical measure (IQFinv) and the visual assessment of IQ were shown to be strongly correlated (r = 0.95; p < 0.001). Using Pearson's correlation, the IQFinv, PNV, and HMV were found to be strongly correlated (r = 0.94; p < 0.001) and (r = 0.92; p < 0.001), correspondingly. CONCLUSION: Results of the study show that physical measures of LCD detectability utilising the CDRAD 2.0 phantom is strongly corelated with visual IQ and PV (PNV and HMV) and can be used to evaluate IQ when undertaking neonatal chest radiography (CXR). IMPLICATIONS FOR PRACTICE: This study establishes the feasibility of utilising the physical measure (IQFinv) and the CDRAD 2.0 phantom in routine quality assurance and neonatal CXR optimisation studies.

Video rasterstereography of the spine and pelvis in eight erect positions: A reliability study.

INTRODUCTION: To investigate the reliability and variability of Video Rasterstereography (VR) measurements of the spine and pelvis, for eight proposed standing postures, in order to help define an optimal standing position for erect pelvis radiography. METHODS: Surface topography data were collected using the formetic 4D dynamic modelling (Diers) system. 61 healthy participants were recruited; each participant performed eight different standing positions. Four positions were performed with the feet shoulder width apart and parallel, and four positions were performed with the feet shoulder width apart and internally rotated. For the upper extremity, each of the (two sets of) four positions were performed with different arm positions (arms by the sides, arms crossed over the chest, arms 30° flexed and touching the medial end of the clavicle, arms 30° flexed with the hands holding a support). Three sets of surface topography were collected in the eight positions (n = 24). The variability was assessed by calculating standard error of the measurement (SEm) and the coefficient of variation (CV). Reliability was assessed using intra-class correlation coefficients (ICC ± 95% CI). RESULTS: No significant differences in the SEm were found between the three paired measurements for all standing positions (P > 0.05). ICC values demonstrated excellent reliability for all measurements across the eight standing positions (range 0.879-1.00 [95% CI 0.813-1.00]). CONCLUSION: Evaluating eight standing positions radiographically would be unethical as it would involve repeat radiation exposures. Using the formetic 4D dynamic modelling (Diers) system, provides an alternative and has shown that there was only a minimal, non-statistically significant, differences between the eight different standing positions. IMPLICATION FOR PRACTICE: Different standing positions were proposed for erect pelvis radiography.

Impact of body part thickness on AP pelvis radiographic image quality and effective dose.

INTRODUCTION: Within medical imaging variations in patient size can generate challenges, especially when selecting appropriate acquisition parameters. This experiment sought to evaluate the impact of increasing body part thickness on image quality (IQ) and effective dose (E) and identify optimum exposure parameters. METHODS: An anthropomorphic pelvis phantom was imaged with additional layers (1-15 cm) of animal fat as a proxy for increasing body thickness. Acquisitions used the automatic exposure control (AEC), 100 cm source to image distance (SID) and a range of tube potentials (70-110 kVp). IQ was evaluated physically and perceptually. E was estimated using PCXMC software. RESULTS: For all tube potentials, signal to noise ratio (SNR) and contrast to noise ratio (CNR) deceased as body part thickness increased. 70 kVp produced the highest SNR (46.6-22.6); CNR (42.8-17.6). Visual grading showed that the highest IQ scores were achieved using 70 and 75 kVp. As thickness increases, E increased exponentially (r = 0.96; p < 0.001). Correlations were found between visual and physical IQ (SNR r = 0.97, p < 0.001; CNR r = 0.98, p < 0.001). CONCLUSION: To achieve an optimal IQ across the range of thicknesses, lower kVp settings were most effective. This is at variance with professional practice as there is a tendency for radiographers to increase kVp as thickness increases. Dose reductions were experienced at higher kVp settings and are a valid method for optimisation when imaging larger patients.