EVALUATION OF VGC ANALYZER BY COMPARISON WITH GOLD STANDARD ROC SOFTWARE AND ANALYSIS OF SIMULATED VISUAL GRADING DATA.

The purpose of the present work was to evaluate the use of resampling statistical methods for analysis of visual grading data-implemented in the software VGC Analyzer-by comparing the reanalyzed results from previously performed visual grading studies with the results calculated by gold standard receiver operating characteristic (ROC) methodology, Obuchowski-Rockette (OR)-Dorfman-Berbaum-Metz (DBM) multiple-readers and multiple-case (MRMC) and by analysis of simulated visual grading data where the true distribution was presumed to be known. The reanalysis was performed on two multiple-reader studies with non-paired data and paired data, respectively. The simulation study was performed by simulating a large number of visual grading characteristics (VGC) studies and by analyzing the statistical distribution of null hypothesis (H0) rejection rate. The comparison with OR-DBM MRMC showed good agreement when analyzing non-paired data for both fixed-reader and random-reader settings for the calculated area under the curve values and the confidence intervals (CIs). For paired data analysis, VGC Analyzer showed significantly lower CIs compared with the ROC software. This effect was also illustrated by the simulation study, where the VGC Analyzer, in general, showed good accuracy for simulated studies with stable statistical basis. For simulated studies with unstable statistics, the accuracy in the H0 rejection rate decreased. The present study has shown that resampling methodology can be used to accurately perform the statistical analysis of a VGC study, although the resampling technique used makes the method sensitive to small data sets.

THE VALIDITY OF USING ROC SOFTWARE FOR ANALYSING VISUAL GRADING CHARACTERISTICS DATA: AN INVESTIGATION BASED ON THE NOVEL SOFTWARE VGC ANALYZER.

The purpose of the present work was to investigate the validity of using single-reader-adapted receiver operating characteristics (ROC) software for analysis of visual grading characteristics (VGC) data. VGC data from four published VGC studies on optimisation of X-ray examinations, previously analysed using ROCFIT, were reanalysed using a recently developed software dedicated to VGC analysis (VGC Analyzer), and the outcomes [the mean and 95 % confidence interval (CI) of the area under the VGC curve (AUCVGC) and the p-value] were compared. The studies included both paired and non-paired data and were reanalysed both for the fixed-reader and the random-reader situations. The results showed good agreement between the softwares for the mean AUCVGC For non-paired data, wider CIs were obtained with VGC Analyzer than previously reported, whereas for paired data, the previously reported CIs were similar or even broader. Similar observations were made for the p-values. The results indicate that the use of single-reader-adapted ROC software such as ROCFIT for analysing non-paired VGC data may lead to an increased risk of committing Type I errors, especially in the random-reader situation. On the other hand, the use of ROC software for analysis of paired VGC data may lead to an increased risk of committing Type II errors, especially in the fixed-reader situation.

VGC ANALYZER: A SOFTWARE FOR STATISTICAL ANALYSIS OF FULLY CROSSED MULTIPLE-READER MULTIPLE-CASE VISUAL GRADING CHARACTERISTICS STUDIES.

Visual grading characteristics (VGC) analysis is a non-parametric rank-invariant method for analysis of visual grading data. In VGC analysis, image quality ratings for two different conditions are compared by producing a VGC curve, similar to how the ratings for normal and abnormal cases in receiver operating characteristic (ROC) analysis are used to create an ROC curve. The use of established ROC software for the analysis of VGC data has therefore previously been proposed. However, the ROC analysis is based on the assumption of independence between normal and abnormal cases. In VGC analysis, this independence cannot always be assumed, e.g. if the ratings are based on the same patients imaged under both conditions. A dedicated software intended for analysis of VGC studies, which takes possible dependencies between ratings into account in the statistical analysis of a VGC study, has therefore been developed. The software-VGC Analyzer-determines the area under the VGC curve and its uncertainty using non-parametric resampling techniques. This article gives an introduction to VGC Analyzer, describes the types of analyses that can be performed and instructs the user about the input and output data.

Estimating effective dose from 3D imaging with interventional fluoroscopy systems using limited exposure data.

BACKGROUND: Estimations of the effective dose from three-dimensional (3D) rotational imaging with interventional fluoroscopy systems are hampered by the fact that not all systems provide individual exposure values for each projection image included in the examination. PURPOSE: To investigate the error in resulting effective dose introduced by not using individual exposure values for each projection image in the dose calculations for 3D rotational imaging with interventional fluoroscopy systems. MATERIAL AND METHODS: An interventional fluoroscopy system was used to acquire images of two anthropomorphic phantoms. Calculations of the effective dose were performed using two different methods: 1, using individual exposure values for each projection image; and 2, using the mean tube voltage and the total dose-area product (DAP), evenly distributed over a selection of projection images. The second method was also tested in hypothetical examinations to investigate the effects of worst-case scenarios regarding the effect of exposure asymmetry on the error. RESULTS: The error in resulting effective dose obtained when simplifying the dose calculations by using Method 2 instead of Method 1 was within ±14%. The error increased slightly for the worst-case scenarios but was still smaller than ±20%, regardless of anatomical region, tube voltage variation, and patient size. CONCLUSION: Given the uncertainties associated with the effective dose concept as well as of reported DAP values, the present study indicates that dose calculations based on average exposure values distributed over a smaller selection of projection angles can provide reasonably accurate estimations of the radiation doses from 3D imaging using interventional fluoroscopy systems.

Comparison of three methods for determining CT dose profile: presenting the tritium method.

The purpose of the present work was to describe a method of using an imaging plate from a computed radiography system to determine the computed tomography (CT) dose profile (the tritium method) and to compare this method with point-dose measurements using a solid-state detector (CT Dose Profiler; RTI Electronics, Mölndal, Sweden) and the indirect method of comparing the air kerma-length product (P(KL)) at different beam collimations. The three methods were used to determine the full width at half maximum (FWHM) of the dose profile of a multi-slice CT at different nominal beam collimations. For all beam collimations, the obtained deviation between the tritium method and the CT Dose Profiler was smaller than 0.1 mm. The maximum relative error was 2 %. For the P(KL) method, the deviation from the CT Dose Profiler was between 0.2 and 0.4 mm, resulting in a relative error larger than 10 % for the smallest beam collimation even after normalisation to a known FWHM. In conclusion, the proposed method of using an imaging plate to determine the FWHM of the CT dose profile has a high accuracy and shows good agreement with the more advanced method of point-dose measurements using a solid-state detector.

The effect of radiation dose reduction on clinical image quality in chest radiography of premature neonates using a dual-side readout technique computed radiography system.

The purpose of the present study was to investigate if the exposure could be reduced from the clinical setting (resulting in an effective dose of 8 microSv for a neonate of weight 0.7 kg and height 25 cm at a tube voltage of 90 kV) without negatively influencing the image quality for a dual-side readout technique computed radiography (CR) system in chest radiography of premature neonates. Chest radiographs of premature neonates were acquired with the double-side readout technique CR system. The images underwent simulated dose reduction in steps of 20 % to represent five different radiation dose levels. Four image quality criteria, related to the visibility of important anatomical structures, were used in a visual grading study where five experienced radiologists rated how well the criteria were fulfilled for all images. When reducing the radiation dose, a decrease in image quality could be observed already at the 80 % dose level for all the structures. The results indicate that a decrease in exposure from the clinically used setting affects the image quality negatively for the CR system.

Optimisation of tube voltage for conventional urography using a Gd2O2S:Tb flat panel detector.

With the increasing use of computed tomography (CT) for urography examinations, the indications for 'conventional' projection urography have changed and are more focused on high-contrast details. The purpose of the present study was to optimise the beam quality for urography examinations performed with a Gd(2)O(2)S:Tb flat-panel detector for the new conditions. Images of an anthropomorphic phantom were collected at different tube voltages with a CXDI-40G detector (Canon Inc., Tokyo, Japan). The images were analysed by radiologists and residents in a visual grading characteristics (VGCs) study. The tube voltage resulting in the best image quality was 55 kV, which therefore was selected for a clinical study. Images from 62 patients exposed with either 55 or 73 kV (original tube voltage) at constant effective doses were included. The 55-kV images underwent simulated dose reduction to represent images collected at 80, 64, 50, 40 and 32 % of the original dose level. All images were included in a VGC study where the observers rated the visibility of important anatomical landmarks. For images collected at 55 kV, an effective dose of approximately 85 % resulted in the same image quality as for images collected at 73 kV at 100 % dose. In conclusion, a low tube voltage should be used for conventional urography focused on high-contrast details. The study indicates that using a tube voltage of 55 kV instead of 73 kV for a Gd(2)O(2)S:Tb flat-panel detector, the effective dose can be reduced by approximately 10-20 % for normal-sized patients while maintaining image quality.

A practical approach to prioritise among optimisation tasks in X-ray imaging: introducing the 4-bit concept.

According to European and national legislation, as well as international recommendations, X-ray examinations shall be optimised. However, with limited resources and hundreds of different types of X-ray examinations, it may be difficult to prioritise among the optimisation tasks at a radiology department. This work is focused on describing a method that can be used to determine the order of which the examinations should be optimised. In the Medical Exposure Directive from 1997, the European Commission prescribes the content of an optimisation process in relation to medical exposure. A reasonable interpretation of the directive is that the assurance of medical purpose for a justified examination is superior to the need of decreased radiation dose. This was used as a basis for developing a method for prioritisation among optimisation tasks. For each examination type, the following four yes/no questions are raised: (i) Is the present image quality unacceptable? (ii) Is the examination of particular importance? (iii) Is the radiation dose suspiciously high? (iv) Are there special dose level concerns, e.g. diagnostic reference levels? Arguing that a positive response to any of the four questions results in the examination being higher prioritised than otherwise and that the questions are labelled in order of decreasing relevance, it can be shown that the resulting flow chart, determining the order of which the examinations should be optimised, can be described by a 4-bit binary scale. In this way, each examination type is given a number from 0 to 15, a higher number corresponding to the examination being prioritised higher in the optimisation work. The method was applied to a general radiology department and resulted in a well-discriminated distribution of examinations prioritised for optimisation tasks. In conclusion, taking into account both medical outcome and potential risk, the proposed method can be used to determine the order in which examinations at a radiology department should be optimised.

Evaluation of subjective assessment of the low-contrast visibility in constancy control of computed tomography.

The purpose of the present work was to investigate the reliability of subjective assessments of the low-contrast visibility in constancy control of computed tomography (CT). Axial CT images of a low-contrast phantom were acquired on an 8-slice multi-detector CT scanner at nine tube current settings ranging from 75 to 440 mA. Five medical physicists assessed the visibility of the low-contrast details in two sessions. In the first session, containing 54 images, the visibility was rated on an absolute scale by determining the number of visible details in each contrast group in each image. In the second session, 180 image pairs were presented to the observers with the task of determining if the two images had been acquired under identical conditions or not. In the absolute session, both the intra- and inter-observer variabilities were high. In the relative session, the variability was smaller, but an exposure difference of 50 % was needed for all observers to correctly identify a change in all cases. In conclusion, the present study indicates that subjective assessments of the low-contrast visibility in constancy control of CT are not reliable.

A conceptual optimisation strategy for radiography in a digital environment.

Using a completely digital environment for the entire imaging process leads to new possibilities for optimisation of radiography since many restrictions of screen/film systems, such as the small dynamic range and the lack of possibilities for image processing, do not apply any longer. However, at the same time these new possibilities lead to a more complicated optimisation process, since more freedom is given to alter parameters. This paper focuses on describing an optimisation strategy that concentrates on taking advantage of the conceptual differences between digital systems and screen/film systems. The strategy can be summarised as: (a) always include the anatomical background during the optimisation, (b) perform all comparisons at a constant effective dose and (c) separate the image display stage from the image collection stage. A three-step process is proposed where the optimal setting of the technique parameters is determined at first, followed by an optimisation of the image processing. In the final step the optimal dose level-given the optimal settings of the image collection and image display stages-is determined.

An optimisation strategy in a digital environment applied to neonatal chest imaging.

The aim of this study was to find the optimum tube voltage for neonatal chest imaging in computed radiography. The study was designed to take full advantage of the benefits of digital imaging, for example, by comparing the tube voltages at constant effective dose. A phantom study using a living rabbit was first conducted. Images were collected at tube voltages ranging from 40 to 90 kV(p). The reproduction of four structures (central vessels, peripheral vessels, carina and thoracic vertebrae) was rated by 10 radiologists. The reproduction of both central and peripheral vessels was relatively independent of tube voltage. The carina was better reproduced at higher tube voltages whereas the opposite was true for the thoracic vertebrae. Based on the higher importance of the reproduction of the carina it was decided that 90 kV(p) was the optimal tube voltage. To validate the result from the phantom study, a follow-up study was conducted in which images of neonates collected at the tube voltage regularly used at Sahlgrenska University Hospital (70 kV(p)) were compared with images collected at the tube voltage proposed by the phantom study. The follow-up study confirmed the results from the phantom study that the reproduction of the carina was better at 90 than at 70 kV(p). In conclusion, for neonatal chest imaging-given the same effective dose-90 kVp gives better reproduction of important structures than the regularly used 70 kV(p).