Development and validation of a psychometric scale for assessing PA chest image quality: A pilot study

Neonatal chest radiography: Influence of standard clinical protocols and radiographic equipment on pathology visibility and radiation dose using a neonatal chest phantom

INTRODUCTION

Little is known about the variations in pathology visibility (PV) and their corresponding radiation dose values for neonatal chest radiography, between and within hospitals. Large variations in PV could influence the diagnostic outcome and the variations in radiation dose could affect the risk to patients. The aim of this study is to compare the PV and radiation dose for standard neonatal chest radiography protocols among a series of public hospitals.

METHODS

A Gammex 610 neonatal chest phantom was used to simulate the chest region of neonates. Radiographic acquisitions were conducted on 17 X-ray machines located in eight hospitals, utilising their current neonatal chest radiography protocols. Six qualified radiographers assessed PV visually using a relative visual grading analysis (VGA). Signal to noise ratios (SNR) and contrast to noise ratios (CNR) were measured as a measure of image quality (IQ). Incident air kerma (IAK) was measured using a solid-state dosimeter.

RESULTS

PV and radiation dose varied substantially between and within hospitals. For PV, the mean (range) VGA scores, between and within the hospitals, were 2.69 (2.00-3.50) and 2.73 (2.33-3.33), respectively. For IAK, the mean (range), between and within the hospitals, were 24.45 (8.11-49.94) μGy and 34.86 (22.26-49.94) μGy, respectively.

CONCLUSION

Between and within participating hospitals there was wide variation in the visibility of simulated pathology and radiation dose (IAK).

IMPLICATIONS FOR PRACTICE

X-ray units with lower PV and higher doses require optimisation of their standard clinical protocols. Institutions which can offer acceptable levels of PV but with lower radiation doses should help facilitate national optimisation processes.

Development and validation of a visual grading scale for assessing image quality of AP pelvis radiographic images

OBJECTIVE

The aim of this article was to apply psychometric theory to develop and validate a visual grading scale for assessing the visual perception of digital image quality anteroposterior (AP) pelvis.

METHODS

Psychometric theory was used to guide scale development. Seven phantom and seven cadaver images of visually and objectively predetermined quality were used to help assess scale reliability and validity. 151 volunteers scored phantom images, and 184 volunteers scored cadaver images. Factor analysis and Cronbach's alpha were used to assess scale validity and reliability.

RESULTS

A 24-item scale was produced. Aggregated mean volunteer scores for each image correlated with the rank order of the visually and objectively predetermined image qualities. Scale items had good interitem correlation (≥0.2) and high factor loadings (≥0.3). Cronbach's alpha (reliability) revealed that the scale has acceptable levels of internal reliability for both phantom and cadaver images (α = 0.8 and 0.9, respectively). Factor analysis suggested that the scale is multidimensional (assessing multiple quality themes).

CONCLUSION

This study represents the first full development and validation of a visual image quality scale using psychometric theory. It is likely that this scale will have clinical, training and research applications.

ADVANCES IN KNOWLEDGE

This article presents data to create and validate visual grading scales for radiographic examinations. The visual grading scale, for AP pelvis examinations, can act as a validated tool for future research, teaching and clinical evaluations of image quality.