Askell Löve

Six iterative reconstruction algorithms in brain CT- A phantom study on image quality at different radiation doses.

OBJECTIVEnTo evaluate the image quality produced by six different iterative reconstruction (IR) algorithms in four CT systems in the setting of brain CT, using different radiation dose levels and iterative image optimisation levels.nnnMETHODSnAn image quality phantom, supplied with a bone mimicking annulus, was examined using four CT systems from different vendors and four radiation dose levels. Acquisitions were reconstructed using conventional filtered back-projection (FBP), three levels of statistical IR and, when available, a model-based IR algorithm. The evaluated image quality parameters were CT numbers, uniformity, noise, noise-power spectra, low-contrast resolution and spatial resolution.nnnRESULTSnCompared with FBP, noise reduction was achieved by all six IR algorithms at all radiation dose levels, with further improvement seen at higher IR levels. Noise-power spectra revealed changes in noise distribution relative to the FBP for most statistical IR algorithms, especially the two model-based IR algorithms. Compared with FBP, variable degrees of improvements were seen in both objective and subjective low-contrast resolutions for all IR algorithms. Spatial resolution was improved with both model-based IR algorithms and one of the statistical IR algorithms.nnnCONCLUSIONnThe four statistical IR algorithms evaluated in the study all improved the general image quality compared with FBP, with improvement seen for most or all evaluated quality criteria. Further improvement was achieved with one of the model-based IR algorithms.nnnADVANCES IN KNOWLEDGEnThe six evaluated IR algorithms all improve the image quality in brain CT but show different strengths and weaknesses.

Hybrid Iterative Reconstruction Algorithm Improves Image Quality in Craniocervical CT Angiography

OBJECTIVEnThe purpose of this study was to evaluate the potential of a hybrid iterative reconstruction algorithm for improving image quality in craniocervical CT angiography (CTA) and to assess observer performance.nnnSUBJECTS AND METHODSnThirty patients (mean age, 58 years; range 16-80 years) underwent standard craniocervical CTA (volume CT dose index, 6.8 mGy, 2.8 mSv). Images were reconstructed using both filtered back projection (FBP) and a hybrid iterative reconstruction algorithm. Five neuroradiologists assessed general image quality and delineation of the vessel lumen in seven arterial segments using a 4-grade scale. Interobserver and intraobserver variability were determined. Mean attenuation and noise were measured and signal-to-noise and contrast-to-noise ratios calculated. Descriptive statistics are presented and data analyzed using linear mixed-effects models.nnnRESULTSnIn pooled data, image quality in iterative reconstruction was graded superior to FBP regarding all five quality criteria (p < 0.0001), with the greatest improvement observed in the vertebral arteries. Iterative reconstruction resulted in elimination of arterial segments graded poor. Interobserver percentage agreement was significantly better (p = 0.024) for iterative reconstruction (69%) than for FBP (66%) but worse than intraobserver percentage agreement (mean, 79%). Noise levels, signal-to-noise ratio, and contrast-to-noise ratio were significantly (p < 0.001) improved in iterative reconstruction at all measured levels.nnnCONCLUSIONnThe iterative reconstruction algorithm significantly improves image quality in craniocervical CT, especially at the thoracic inlet. Despite careful study design, considerable interobserver and intraobserver variability was noted.

Regression models for analyzing radiological visual grading studies – an empirical comparison

BackgroundFor optimizing and evaluating image quality in medical imaging, one can use visual grading experiments, where observers rate some aspect of image quality on an ordinal scale. To analyze the grading data, several regression methods are available, and this study aimed at empirically comparing such techniques, in particular when including random effects in the models, which is appropriate for observers and patients.MethodsData were taken from a previous study where 6 observers graded or ranked in 40 patients the image quality of four imaging protocols, differing in radiation dose and image reconstruction method. The models tested included linear regression, the proportional odds model for ordinal logistic regression, the partial proportional odds model, the stereotype logistic regression model and rank-order logistic regression (for ranking data). In the first two models, random effects as well as fixed effects could be included; in the remaining three, only fixed effects.ResultsIn general, the goodness of fit (AIC and McFadden’s Pseudo R2) showed small differences between the models with fixed effects only. For the mixed-effects models, higher AIC and lower Pseudo R2 was obtained, which may be related to the different number of parameters in these models. The estimated potential for dose reduction by new image reconstruction methods varied only slightly between models.ConclusionsThe authors suggest that the most suitable approach may be to use ordinal logistic regression, which can handle ordinal data and random effects appropriately.

Hybrid iterative reconstruction algorithm in brain CT: a radiation dose reduction and image quality assessment study.

Background Iterative reconstruction (IR) algorithms improve image quality and allow for radiation dose reduction in CT. Dose reduction is particularly challenging in brain CT where good low-contrast resolution is essential. Ideally, evaluation of image quality combines objective measurements and subjective assessment of clinically relevant quality criteria. Subjective assessment is associated with various pitfalls and biases. Purpose To evaluate the potential of the hybrid IR algorithm iDOSE4 to preserve image quality in phantom and clinical brain CT acquired with 30% reduced radiation dose, and to discuss the image quality assessment methods. Material and Methods Forty patients underwent two consecutive brain CTs with normal radiation dose (ND) and 30% reduced dose (RD). Both ND and RD were reconstructed with FBP. In addition the reduced dose CTs were reconstructed with two levels of IR (ID2, ID4). Three image quality criteria (grey-white-matter discrimination, basal ganglia delineation, general image quality) were graded and ranked by six neuroradiologists. Noise levels and contrast-to-noise ratios (CNR) were measured in clinical data. Noise, signal-to-noise ratio (SNR), spatial resolution, and noise-power spectrum (NPS) were also assessed in a phantom. Results Subjective image quality was considered adequate for clinical use for all reconstructions, graded good or excellent in 93% of cases for ND, 83% for ID4, 79% for ID2, and 67% for RD. For all quality parameters, ID4 and ID2 were graded better than RD (Pu2009<u20090.0055 and Pu2009<u20090.035), but worse than ND (Pu2009<u20090.001). In clinical images, objective measurements showed lower noise and significantly higher CNR in ID4 compared with ND and RD (Pu2009<u20090.001). CNR was similar for ID2 and ND. In the phantom, IR reduced noise while maintaining spatial resolution and NPS. Conclusion: The IR algorithm improves image quality of reduced dose CTs and consistently delivers sufficient image quality for clinical purposes. Pitfalls related to subjective assessment can be addressed with careful study design.

Comprehensive CT Evaluation in Acute Ischemic Stroke: Impact on Diagnosis and Treatment Decisions.

Background. With modern CT imaging a comprehensive overview of cerebral macro- and microcirculation can be obtained within minutes in acute ischemic stroke. This opens for patient stratification and individualized treatment. Methods. Four patients with acute ischemic stroke of different aetiologies and/or treatments were chosen for illustration of the comprehensive CT protocol and its value in subsequent treatment decisions. The patients were clinically evaluated according to the NIHSS-scale, examined with the comprehensive CT protocol including both CT angiography and CT perfusion, and followed up by MRI. Results. The comprehensive CT examination protocol increased the examination time but did not delay treatment initiation. In some cases CT angiography revealed the cause of stroke while CT perfusion located and graded the perfusion defect with reasonable accuracy, confirmed by follow-up MR-diffusion. In the presented cases findings of the comprehensive CT examination influenced the treatment strategy. Conclusions. The comprehensive CT examination is a fast and safe method allowing accurate diagnosis and making way for individualized treatment in acute ischemic stroke.

Radiation dose reduction in CT of the brain: can advanced noise filtering compensate for loss of image quality?

Background Computed tomography (CT) of the brain is performed with high local doses due to high demands on low contrast resolution. Advanced algorithms for noise reduction might be able to preserve critical image information when reducing radiation dose. Purpose To evaluate the effect of advanced noise filtering on image quality in brain CT acquired with reduced radiation dose. Material and Methods Thirty patients referred for non-enhanced CT of the brain were examined with two helical protocols: normal dose (ND, CTDIvol 57 mGy) and low dose (LD, CTDIvol 40 mGy) implying a 30% radiation dose reduction. Images from the LD examinations were also postprocessed with a noise reduction software with non-linear filters (SharpView CT), creating filtered low dose images (FLD) for each patient. The three image stacks for each patient were presented side by side in randomized order. Five radiologists, blinded for dose level and filtering, ranked these three axial image stacks (ND, LD, FLD) as best to poorest (1 to 3) regarding three image quality criteria. Measurements of mean Hounsfield units (HU) and standard deviation (SD) of the HU were calculated for large region of interest in the centrum semiovale as a measure for noise. Results Ranking results in pooled data showed that the advanced noise filtering significantly improved the image quality in FLD as compared to LD images for all tested criteria. No significant differences in image quality were found between ND examinations and FLD. However, there was a notable inter-reader spread of the ranking. SD values were 15% higher for LD as compared to ND and FLD. Conclusion The advanced noise filtering clearly improves image quality of CT examinations of the brain. This effect can be used to significantly lower radiation dose.

Clinically essential requirement for brain CT with iterative reconstruction: author reply.

To the Editor n nDuring the past decade, CT technique has developed rapidly, with the main goal of reducing radiation dose while maintaining image quality. Since large diagnostic accuracy studies are at great risk of being already outdated at the time of publication, simpler clinical feasibility phantom studies1 and diagnostic acceptability studies have become important tools for evaluation and comparison of the new techniques. n nThe study by Love et al1 is a phantom study and is therefore limited when it comes to evaluation of specific clinically relevant aspects such as potential influence of beam-hardening artefacts on diagnostic accuracy in subarachnoidal haemorrhage, as described by Awai et al.2 Such evaluation is clearly outside the scope of the study. In our study, the bone-mimicking ring was instead used to introduce noise in the phantom and thereby more closely resemble the clinical situation with respect to low-contrast resolution, which is clinically very important as Awai et al point out. n nIn the brain, the difference in CT numbers between grey matter and white matter is normally in the range 5–10u2009HU. Although visual differentiation of adjacent structures is dependent on standard deviation (SD) of noise in the images, the relationship is far more complex, which is why our study also evaluates noise-power spectra and subjective evaluation. It is to us unclear on which scientific basis Awai et al argue that “to detect such minute CT number differences between the grey matter and white matter, the image noise must be less than at least 5u2009HU.” Nevertheless, as seen in Table 3, noise levels (SD) are below or close to 5u2009HU when applying the advanced IR algorithms at 48u2009mGy and even in some cases at 12u2009mGy. n nDespite increasing use of clinical guidelines for CT neuroimaging in clinical situations, such as in trauma, stroke, headache and dementia, utilization of CT is increasing in all age groups—with only a small minority of patients having detectable pathology. Assuming that currently employed clinical CT protocols already are optimized according to the as low as reasonably achievable principle, introduction of a new technique that improves image quality should primarily be used to reduce radiation dose so that image quality remains unchanged. From a clinical and radiation perspective, CT image quality higher than that necessary for the clinical task should not be pursued. n nWe do not claim that our results can directly be used to change current clinical protocols, but our conclusions can be used as a base for clinical diagnostic acceptability studies aiming at further dose reduction while retaining image quality. n nThe authors