Advances in multiscale image processing and its effects on image quality in skeletal radiography.

Multi-frequency processing (MFP) leads to enhanced image quality (IQ) of radiographs. This study is to determine the effect of third generation MFP (M3) on IQ in comparison to standard second-generation MFP (M2). 20 cadavers were examined and post-processing of radiographs was performed with both M2 and M3. Three readers blinded to the MFP used for each image independently compared corresponding image pairs according to overall IQ and depiction of bony structures and soft tissue (+ 2: notably better > 0: equal > - 2: notably worse). A significant deviation of the median grade from grade 0 (equal) (p < 0.01) for each evaluator A, B and C speaks against an equal image quality of M2- and M3-images. M3-images were categorized with better grades (+ 1, + 2) in 87.7% for overall image quality, in 90.4% for soft tissue and 81.8% for bony structures. M3 images showed significant higher averaged SNR and CNR for all investigated lower extremities than that of M2 images (0.031 < p < 0.049). The newest generation of MFP leads to significantly better depiction of anatomical structures and overall image quality than in images processed with the preceding generation of MFP. This provides increased diagnostic accuracy and further decreased radiation exposure.

Image quality is resilient against tube voltage variations in post-mortem skeletal radiography with a digital flat-panel detector.

In recent phantom studies low-contrast detectability was shown to be independent from variations in tube voltage in digital radiography (DR) systems. To investigate the transferability to a clinical setting, the lower extremities of human cadavers were exposed at constant detector doses with different tube voltages in a certain range, as proposed in the phantom studies. Three radiologists independently graded different aspects of image quality (IQ) in a comparative analysis. The grades show no correlation between IQ and kV, which means that the readers were not able to recognize a significant IQ difference at different kV. Signal-to-noise and contrast-to-noise ratios showed no significant differences in IQ despite the kV-setting variations. These findings were observed from a limited kV range setting. Higher kV-settings resulted in lowest patient exposure at constant IQ. These results confirm the potential of DR-systems to contribute to standardization of examination protocols comparable to computed tomography. This may prevent the trend to overexpose. Further investigations in other body regions and other DR-systems are encouraged to determine transferability.

Evaluation of ankle ligaments with CT: A feasibility study.

PURPOSE: This study aims to systematically investigate if normal ankle ligaments can be assessed with CT imaging, using MRI as reference standard. METHODS: 35 patients (mean age: 47 ±â€¯15 years; female n = 19) with combined CT and MRI exams and without MR-morphologic signs of ankle ligamental injury were retrospectively identified. 3 readers independently evaluated the syndesmotic, the lateral and medial ankle ligaments in terms of visibility on a 4-point Likert scale (0-3 points) in multiplanar MDCT images in standard bone kernel reconstructions. In consensus CT-based ligament density and thickness were measured and the appearance was rated for each ligament. Results were compared and validated with corresponding MRI images. RESULTS: Almost all ankle ligaments identified in MRI images could be adequately depicted in standard multiplanar bone kernel CT images with a mean visual score of 2.7/3 (± 0.2). Difficulties in CT morphological delineation of ankle ligaments occurred in cases of filiform TNL and TCL and in cases of concurrent soft tissue edema. Interreader agreement for the CT-assessment of ankle ligaments was excellent, with Fleiss Kappa values >0.8. Mean density of evaluated medial and lateral ankle ligaments was 68 ±â€¯2.9 HU, with substantially inter- and intraindividual variations. Thickness measurements and assessment of appearance of ankle ligaments showed a good concordance between CT and MRI. CONCLUSIONS: Assessment of normal ankle ligaments via standard CT in bone kernel reconstructions is feasible, with some restrictions concerning the medial collateral ligaments and in the presence of soft tissue edema.

Image quality of iterative reconstruction in cranial CT imaging: comparison of model-based iterative reconstruction (MBIR) and adaptive statistical iterative reconstruction (ASiR).

OBJECTIVES: The purpose of this study was to compare cranial CT (CCT) image quality (IQ) of the MBIR algorithm with standard iterative reconstruction (ASiR). METHODS: In this institutional review board (IRB)-approved study, raw data sets of 100 unenhanced CCT examinations (120 kV, 50-260 mAs, 20 mm collimation, 0.984 pitch) were reconstructed with both ASiR and MBIR. Signal-to-noise (SNR) and contrast-to-noise (CNR) were calculated from attenuation values measured in caudate nucleus, frontal white matter, anterior ventricle horn, fourth ventricle, and pons. Two radiologists, who were blinded to the reconstruction algorithms, evaluated anonymized multiplanar reformations of 2.5 mm with respect to depiction of different parenchymal structures and impact of artefacts on IQ with a five-point scale (0: unacceptable, 1: less than average, 2: average, 3: above average, 4: excellent). RESULTS: MBIR decreased artefacts more effectively than ASiR (p < 0.01). The median depiction score for MBIR was 3, whereas the median value for ASiR was 2 (p < 0.01). SNR and CNR were significantly higher in MBIR than ASiR (p < 0.01). CONCLUSIONS: MBIR showed significant improvement of IQ parameters compared to ASiR. As CCT is an examination that is frequently required, the use of MBIR may allow for substantial reduction of radiation exposure caused by medical diagnostics. KEY POINTS: • Model-Based iterative reconstruction (MBIR) effectively decreased artefacts in cranial CT. • MBIR reconstructed images were rated with significantly higher scores for image quality. • Model-Based iterative reconstruction may allow reduced-dose diagnostic examination protocols.

Upgrade to iterative image reconstruction (IR) in MDCT imaging: a clinical study for detailed parameter optimization beyond vendor recommendations using the adaptive statistical iterative reconstruction environment (ASIR) Part2: The chest.

PURPOSE: To compare the image quality in dose-reduced 64-row CT of the chest at different levels of adaptive statistical iterative reconstruction (ASIR) to full-dose baseline examinations reconstructed solely with filtered back projection (FBP) in a realistic upgrade scenario. MATERIALS AND METHODS: A waiver of consent was granted by the institutional review board (IRB). The noise index (NI) relates to the standard deviation of Hounsfield units in a water phantom. Baseline exams of the chest (NI = 29; LightSpeed VCT XT, GE Healthcare) were intra-individually compared to follow-up studies on a CT with ASIR after system upgrade (NI = 45; Discovery HD750, GE Healthcare), n = 46. Images were calculated in slice and volume mode with ASIR levels of 0 - 100 % in the standard and lung kernel. Three radiologists independently compared the image quality to the corresponding full-dose baseline examinations (-2: diagnostically inferior, -1: inferior, 0: equal, + 1: superior, + 2: diagnostically superior). Statistical analysis used Wilcoxon's test, Mann-Whitney U test and the intraclass correlation coefficient (ICC). RESULTS: The mean CTDIvol decreased by 53 % from the FBP baseline to 8.0 ± 2.3 mGy for ASIR follow-ups; p < 0.001. The ICC was 0.70. Regarding the standard kernel, the image quality in dose-reduced studies was comparable to the baseline at ASIR 70 % in volume mode (-0.07 ± 0.29, p = 0.29). Concerning the lung kernel, every ASIR level outperformed the baseline image quality (p < 0.001), with ASIR 30 % rated best (slice: 0.70 ± 0.6, volume: 0.74 ± 0.61). CONCLUSION: Vendors' recommendation of 50 % ASIR is fair. In detail, the ASIR 70 % in volume mode for the standard kernel and ASIR 30 % for the lung kernel performed best, allowing for a dose reduction of approximately 50 %.

Evaluation of a dedicated MDCT protocol using iterative image reconstruction after cervical spine trauma.

AIM: To evaluate radiation exposure for 64-row computed tomography (CT) of the cervical spine comparing two optimized protocols using filtered back projection (FBP) and adaptive statistical iterative reconstruction (ASIR), respectively. MATERIALS AND METHODS: Sixty-seven studies using FBP (scanner 1) were retrospectively compared with 80 studies using ASIR (scanner 2). The key scanning parameters were identical (120 kV dose modulation, 64 × 0.625 mm collimation, pitch 0.531:1). In protocol 2, the noise index (NI) was increased from 5 to 25, and ASIR and the high-definition (HD) mode were used. The scan length, CT dose index (CTDI), and dose-length product (DLP) were recorded. The image quality was analysed subjectively by using a three-point scale (0; 1; 2), and objectively by using a region of interest (ROI) analysis. Mann-Whitney U and Wilcoxon's test were used. RESULTS: In the FBP group, the mean CTDI was 21.43 mGy, mean scan length 186.3 mm, and mean DLP 441.15 mGy cm. In the ASIR group, the mean CTDI was 9.57 mGy, mean scan length 195.21 mm, and mean DLP 204.23 mGy cm. The differences were significant for CTDI and DLP (p < 0.001) and scan length (p = 0.01). There was no significant difference in the subjective image quality (p > 0.05). The estimated mean effective dose decreased from 2.38 mSv (FBP) to 1.10 mSv (ASIR). CONCLUSION: The radiation dose of 64-row MDCT can be reduced to a level comparable to plain radiography without loss of subjective image quality by implementation of ASIR in a dedicated cervical spine trauma protocol. These results might contribute to an improved relative risk-to-benefit ratio and support the justification of CT as a first-line imaging tool to evaluate cervical spine trauma.

Upgrade to iterative image reconstruction (IR) in abdominal MDCT imaging: a clinical study for detailed parameter optimization beyond vendor recommendations using the adaptive statistical iterative reconstruction environment (ASIR).

PURPOSE: To compare the image quality of dose-reduced 64-row abdominal CT reconstructed at different levels of adaptive statistical iterative reconstruction (ASIR) to full-dose baseline examinations reconstructed with filtered back-projection (FBP) in a clinical setting and upgrade situation. MATERIALS AND METHODS: Abdominal baseline examinations (noise index NI = 29; LightSpeed VCT XT, GE) were intra-individually compared to follow-up studies on a CT with an ASIR option (NI = 43; Discovery HD750, GE), n = 42. Standard-kernel images were calculated with ASIR blendings of 0 - 100 % in slice and volume mode, respectively. Three experienced radiologists compared the image quality of these 567 sets to their corresponding full-dose baseline examination (- 2: diagnostically inferior, - 1: inferior, 0: equal, + 1: superior, + 2: diagnostically superior). Furthermore, a phantom was scanned. Statistical analysis used the Wilcoxon - the Mann-Whitney U-test and the intra-class correlation (ICC). RESULTS: The mean CTDIvol decreased from 19.7 ±â€Š5.5 to 12.2 ±â€Š4.7 mGy (p < 0.001). The ICC was 0.861. The total image quality of the dose-reduced ASIR studies was comparable to the baseline at ASIR 50 % in slice (p = 0.18) and ASIR 50 - 100 % in volume mode (p > 0.10). Volume mode performed 73 % slower than slice mode (p < 0.01). CONCLUSION: After the system upgrade, the vendor recommendation of ASIR 50 % in slice mode allowed for a dose reduction of 38 % in abdominal CT with comparable image quality and time expenditure. However, there is still further dose reduction potential for more complex reconstruction settings.