Demonstration of correlations between clinical and physical image quality measures in chest and lumbar spine screen–film radiography

Sandborg, Michael; Tingberg, Anders; Dance, David R.; Lanhede, B; Almén, Anja; McVey, G.; Sund, Patrik; Kheddache, Susanne; Besjakov, Jack; Mattsson, Sören; Månsson, Lars Gunnar; Carlsson, Gudrun Alm

doi:10.1259/bjr.74.882.740520

Cited by 38 publications

(16 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Another situation when visual grading may be useful is when several types of post-processing are compared, as in the present study. Visual grading data have been shown to correlate with physical measures of image quality as well as with diagnostic performance evaluated with methods based on ROC 7,8,9 .…”

Section: Discussionmentioning

confidence: 99%

Visual grading regression with random effects

et al. 2012

View full text Add to dashboard Cite

To analyze visual grading experiments, ordinal logistic regression (here called visual grading regression, VGR) may be used in the statistical analysis. In addition to types of imaging or post-processing, the VGR model may include factors such as patient and observer identity, which should be treated as random effects. Standard software does not allow random factors in ordinal logistic regression, but using Generalized Linear Latent And Mixed Models (GLLAMM) this is possible. In a single-image study, 9 radiologists graded 24 cardiac Computed Tomography Angiography (CTA) images with reduced dose without and after post-processing with a 2D adaptive filter, using five image quality criteria. First, standard ordinal logistic regression was carried out, treating filtering, patient and observer identity as fixed effects. The same analysis was then repeated with GLLAMM, treating filtering as a fixed effect and patient and observer identity as random effects. With both approaches, a significant effect (p<0.01) of the filtering was found for all five criteria. No dramatic differences in parameter estimates or significance levels were found between the two approaches. It is concluded that random effects can be appropriately handled in VGR using GLLAMM, but no major differences in the results were found in a preliminary evaluation.

show abstract

Section: Discussionmentioning

confidence: 99%

Visual grading regression with random effects

et al. 2012

View full text Add to dashboard Cite

show abstract

“…Furthermore, we assumed that the visualisation of small structures in supine radiographs would be inferior to that of erect chest X-rays. Thus, according to Sandborg et al [12], the size threshold for rounded structures that needed to be visualised in the lung periphery was elevated from 0.7-2 mm to 2-6 mm and the threshold for linear and reticular structures remained at 0.3-2 mm.…”

Section: Image Evaluationmentioning

confidence: 98%

“…Until now, studies comparing quality of chest images with computed radiography (CR) and DR systems have analysed only erect images [10][11][12]. However, patients of both the groups described above often cannot stand, necessitating a supine radiograph.…”

Section: Introductionmentioning

confidence: 99%

Image quality of supine chest radiographs: intra-individual comparison of computed radiography and low-dose linear-slit digital radiography

Szücs-Farkas

Vock

2009

Eur Radiol

View full text Add to dashboard Cite

Image quality of supine chest radiographs: intra-individual comparison of computed radiography and low-dose linear-slit digital radiography Abstract The purpose of this retrospective study was to intraindividually compare the image quality of computed radiography (CR) and low-dose linear-slit digital radiography (LSDR) for supine chest radiographs. A total of 90 patients (28 female, 62 male; mean age, 55.1 years) imaged with CR and LSDR within a mean time interval of 2.8 days±3.0 were included in this study. Two independent readers evaluated the image quality of CR and LSDR based on modified European Guidelines for Quality Criteria for chest X-ray. The Wilcoxon test was used to analyse differences between the techniques. The overall image quality of LSDR was significantly better than the quality of CR (9.75 vs 8.16 of a maximum score of 10; p<0.001). LSDR performed significantly better than CR for delineation of anatomical structures in the mediastinum and the retrocardiac lung (p<0.001). CR was superior to LSDR for visually sharp delineation of the lung vessels and the thin linear structures in the lungs. We conclude that LSDR yields better image quality and may be more suitable for excluding significant pathological features of the chest in areas with high attenuation compared with CR.

show abstract

“…19 The analyses which follow often utilize methods developed for numerical data, such as the arithmetic mean. 20 This measure can be referred to as the mean opinion score (MOS) 7,[21][22][23] or the visual grading analysis score (VGAS). 7,10,24 Comparison of the summary values across different areas of interest are then used to answer the research question.…”

Section: Introductionmentioning

confidence: 99%

Methods for the analysis of ordinal response data in medical image quality assessment

Keeble

Baxter

Gislason‐Lee

et al. 2016

BJR

View full text Add to dashboard Cite

The assessment of image quality in medical imaging often requires observers to rate images for some metric or detectability task. These subjective results are used in optimization, radiation dose reduction or system comparison studies and may be compared to objective measures from a computer vision algorithm performing the same task. One popular scoring approach is to use a Likert scale, then assign consecutive numbers to the categories. The mean of these response values is then taken and used for comparison with the objective or second subjective response. Agreement is often assessed using correlation coefficients. We highlight a number of weaknesses in this common approach, including inappropriate analyses of ordinal data and the inability to properly account for correlations caused by repeated images or observers. We suggest alternative data collection and analysis techniques such as amendments to the scale and multilevel proportional odds models. We detail the suitability of each approach depending upon the data structure and demonstrate each method using a medical imaging example. Whilst others have raised some of these issues, we evaluated the entire study from data collection to analysis, suggested sources for software and further reading, and provided a checklist plus flowchart for use with any ordinal data. We hope that raised awareness of the limitations of the current approaches will encourage greater method consideration and the utilization of a more appropriate analysis. More accurate comparisons between measures in medical imaging will lead to a more robust contribution to the imaging literature and ultimately improved patient care.

show abstract

Demonstration of correlations between clinical and physical image quality measures in chest and lumbar spine screen–film radiography

Cited by 38 publications

References 12 publications

Visual grading regression with random effects

Visual grading regression with random effects

Image quality of supine chest radiographs: intra-individual comparison of computed radiography and low-dose linear-slit digital radiography

Methods for the analysis of ordinal response data in medical image quality assessment

Contact Info

Product

Resources

About