Comparison of the performance of computed radiography and direct radiography in glass soft tissue foreign body visualisation

Samantha A. Petri, Curtise K.C. Ng


Purpose: The purpose of the study was to investigate the performance of recent computed radiography (CR) and direct radiography (DR) in glass soft tissue foreign body (FB) visualisation in terms of their overall performance, and effects of sizes and locations of FB, and exposure parameters on the visualisation.

Methods: Eighty anteroposterior (AP) and 80 lateral images of chicken legs with five sizes of FBs inserted into two locations were taken by our CR and DR systems using four exposure parameter combinations. Contrast-to-noise ratio (CNR), and visual grading analysis (VGA), were employed to assess the FB visibility. The CNR and VGA data were analysed using descriptive and inferential statistics.

Results: The mean CNR value, 3.89 and median VGA score, 1 (definitely invisible) of all CR images were statistically significantly lower than the mean CNR value, 9.47 (p<0.001) and median VGA score, 2 (possible visible [but uncertain]) of all DR images (p<0.05) respectively. Despite this, the FBs were visible on CR (median VGA score: 3 - visible but could be shown better) and DR (median VGA score: 4 - definitely visible) lateral images (without FBs overlapping with bone). The smallest FBs visible on CR and DR lateral images were 2 and 1 mm respectively. The factors of FB depths and kV settings did not have any statistically significant effects on FB visibility (p>0.05).

Conclusion: The performance of recent DR system in glass soft tissue FB visualisation appears more superior to CR. DR should be used to take orthogonal (AP and lateral) images for detecting any small FBs.


digital radiography, glass visibility, indirect flat panel detector, photostimulable phosphor plate receptor

Full Text:


The South African Radiographer | ISSN 0258 0241

This work is licensed under a Creative Commons Attribution-Noncommercial-No Derivative Works 2.5 South Africa License