Effectiveness of flexible noise control image processing for digital portal images using computed radiography

Image processing setting adaptions according to image dose and radiologist preference can improve image quality in computed radiography of the equine distal limb: A cadaveric study

Image processing (IP) in digital radiography has been steadily refined to improve image quality. Adaptable settings enable users to adjust systems to their specific requirements. This prospective, analytical study aimed to investigate the influence of different IP settings and dose reductions on image quality. Included were 20 cadaveric equine limb specimens distal to the metacarpophalangeal and metatarsophalangeal joints. Images were processed with the Dynamic Visualization II system (Fujifilm) using five different IP settings including multiobjective frequency processing, flexible noise control (FNC), and virtual grid processing (VGP). Seven criteria were assessed by three veterinary radiology Diplomates and one veterinary radiology resident in a blinded study using a scoring system. Algorithm comparison was performed using an absolute visual grading analysis. The rating of bone structures was improved by VGP at full dose (P < .05; AUCVGC  = 0.45). Überschwinger artifact perception was enhanced by VGP (P < .001; AUCVGC  = 0.66), whereas image noise perception was suppressed by FNC (P < .001; AUCVGC  = 0.29). The ratings of bone structures were improved by FNC at 50% dose (P < .05; AUCVGC  = 0.44), and 25% dose (P < .001; AUCVGC  = 0.32), and clinically acceptable image quality was maintained at 50% dose (mean rating 2.16; 95.8% ratings sufficient or better). The favored IP setting varied among observers, with higher agreement at lower dose levels. These findings supported using individualized IP settings based on the radiologist's preferences and situational image requirements, rather than using default settings.

Impact of flexible noise control (FNC) image processing parameters on portable chest radiography

There is a lack of understanding in the performance of flexible noise control (FNC) processing, which is used in digital radiography on a scanner vendor and has four parameters each involving multiple options. The aim of this study was to investigate the impact of FNC on portable chest imaging. An anthropomorphic chest phantom was imaged using a clinical chest program with 85 kV and five radiation dose levels at 40″ source-to-image distance with software-based scatter reduction method. All images were processed without and with FNC. Noise analysis was performed in two regions of interest (ROI) on subtracted noise-only images, and line profiles were generated through a lung-rib interface. In addition, noise power spectra (NPS) analysis was performed in solid water phantoms of 10 and 20 cm thicknesses, using the same acquisition program and a range of dose levels. Last, feedback on retrospectively deidentified, reprocessed, and randomized clinical images from 20 portable chest exams was gathered from two thoracic radiologists. Noise reduction performances of FNC were demonstrated, with the level depending on specific FNC parameters, dose levels, ROI placement, and phantom sizes. Higher frequency textural patterns were revealed through the NPS analysis, which varied based on FNC parameters, dose levels, and phantom sizes. Overall, the vendor default parameter FGA0.5 yielded the highest noise reduction and textural artifacts. Radiologist feedback showed consistent preference of no FNC due to the presence of textural artifacts in the FNC-processed images. An algorithm improvement to avoid introducing artifacts would be desired.

Optimizing image quality using automatic exposure control based on the signal-difference-to-noise ratio: a phantom study

This study proposes to adjust the sensitivity of automatic exposure control (AEC) for achieving consistent image quality over a range of subject thicknesses in abdominal radiography simulations. The relation between image quality and subject thickness was investigated using a digital radiography system with 10-, 15-, 20-, and 25-cm-thick acrylic phantom. Simple pixel signal-to-noise ratio (SNR) was measured to check the default AEC accuracy for each thickness, and image quality was evaluated using the signal-difference-to-noise ratio (SDNR) with an additional acrylic plate and bone-equivalent material. Based on the figure of merit theory, dose ratios to obtain constant image quality regardless of the subject thickness were calculated from SDNR results. The AEC setup was manually modified using this dose ratio, and visibility was examined using a CDRAD 2.0 contrast-detail analysis phantom. Moreover, the entrance surface dose (ESD) was estimated as an index of exposure dose using exposure parameters. The default AEC setup provided a constant simple pixel SNR for each subject thickness with a high accuracy. SDNRs decreased with an increase in the subject thickness. The calculated dose ratios relative to the results for 20 cm thickness were 0.424, 0.647, and 1.43 for 10, 15 and 25 cm, respectively, and a > 25% decrease in ESD was observed for smaller patients. CDRAD analysis using the modified AEC setup showed almost identical visibility for each thickness. Adjusting the sensitivity of AEC according to subject thickness can contribute toward the optimization of the exposure condition.

Dependency of image quality on acquisition protocol and image processing in chest tomosynthesis-a visual grading study based on clinical data

OBJECTIVE

To compare the quality of images obtained with two different protocols with different acquisition time and the influence from image post processing in a chest digital tomosynthesis (DTS) system.

METHODS

20 patients with suspected lung cancer were imaged with a chest X-ray equipment with tomosynthesis option. Two examination protocols with different acquisition times (6.3 and 12 s) were performed on each patient. Both protocols were presented with two different image post-processing (standard DTS processing and more advanced processing optimised for chest radiography). Thus, 4 series from each patient, altogether 80 series, were presented anonymously and in a random order. Five observers rated the quality of the reconstructed section images according to predefined quality criteria in three different classes. Visual grading characteristics (VGC) was used to analyse the data and the area under the VGC curve (AUC_VGC) was used as figure-of-merit. The 12 s protocol and the standard DTS processing were used as references in the analyses.

RESULTS

The protocol with 6.3 s acquisition time had a statistically significant advantage over the vendor-recommended protocol with 12 s acquisition time for the classes of criteria, Demarcation (AUC_VGC = 0.56, p = 0.009) and Disturbance (AUC_VGC = 0.58, p < 0.001). A similar value of AUC_VGC was found also for the class Structure (definition of bone structures in the spine) (0.56) but it could not be statistically separated from 0.5 (p = 0.21). For the image processing, the VGC analysis showed a small but statistically significant advantage for the standard DTS processing over the more advanced processing for the classes of criteria Demarcation (AUC_VGC = 0.45, p = 0.017) and Disturbance (AUC_VGC = 0.43, p = 0.005). A similar value of AUC_VGC was found also for the class Structure (0.46), but it could not be statistically separated from 0.5 (p = 0.31).

CONCLUSION

The study indicates that the protocol with 6.3 s acquisition time yields slightly better image quality than the vender-recommended protocol with acquisition time 12 s for several anatomical structures. Furthermore, the standard gradation processing  (the vendor-recommended post-processing for DTS), yields to some extent advantage over the gradation processing/multiobjective frequency processing/flexible noise control processing in terms of image quality for all classes of criteria. Advances in knowledge: The study proves that the image quality may be strongly affected by the selection of DTS protocol and that the vendor-recommended protocol may not always be the optimal choice.

Evaluation of three presets for four-dimensional cone beam CT in lung radiotherapy verification by visual grading analysis

OBJECTIVE

To evaluate three image acquisition presets for four-dimensional cone beam CT (CBCT) to identify an optimal preset for lung tumour image quality while minimizing dose and acquisition time.

METHODS

Nine patients undergoing radical conventionally fractionated radiotherapy for lung cancer had verification CBCTs acquired using three presets: Preset 1 on Day 1 (11 mGy dose, 240 s acquisition time), Preset 2 on Day 2 (9 mGy dose, 133 s acquisition time) and Preset 3 on Day 3 (9 mGy dose, 67 s acquisition time). The clarity of the tumour and other thoracic structures, and the acceptability of the match, were retrospectively graded by visual grading analysis (VGA). Logistic regression was used to identify the most appropriate preset and any factors that might influence the result.

RESULTS

Presets 1 and 2 met a clinical requirement of 75% of structures to be rated "Clear" or above and 75% of matches to be rated "Acceptable" or above. Clarity is significantly affected by preset, patient, observer and structure. Match acceptability is significantly affected by preset.

CONCLUSION

The application of VGA in this initial study enabled a provisional selection of an optimal preset (Preset 2) to be made.

ADVANCES IN KNOWLEDGE

This was the first application of VGA to the investigation of presets for CBCT.

Study on improvement of signal detectability considering noise characteristics in medical X-ray images

The purpose of this study is to develop method for improvement of low-contrast signal detectability of general X-ray images used for observing overviews, included noise reduction technique and signal amplification technique, that are difficult to detect in the observing overview. The proposal method consists signal amplification part and noise reduction part. The noise reduction part first identifies and mathematizes the characteristics of the noise attributed to the X-ray imaging system before the subject is imaged. The noise information derived from mathematization is used to remove the noise representing the system noise from the subject image. After noise reduction, all the signals on the subject image with sufficiently reduced noise are amplified by multiplication with an arbitrary coefficient. The feasibility of the proposed method was indicated by three evaluations that were performed vision assessment using the image including the virtual low-contrast signals by four radiological technologists. That of the proposed method was confirmed from results of the evaluations. That of the proposed method was confirmed from results of the evaluations.

[Electronic portal imaging device (EPID) portal image filtering for simplifying registration on radiation therapy]

A simple method for improving the quality of electronic portal imaging device (EPID) portal images was proposed for the reduction of the burden on the registration between digital reconstruction radiography (DRR) and EPID portal images in radiation therapy. Conventional image filtering techniques in the spatial-frequency domain are applied to the proposed method. While a band-pass filter (BPF) is employed to extract spatial-frequency components included in the bone edge, a high-pass filter (HPF) is employed to obtain the effect corresponding to the general dynamic range compression. The band-pass filtered image is weighted by a parameter for adjusting the bone edge enhancement, and is added to the high-pass filtered image. This method was applied to the portal images in the neck region. In the image obtained by the proposed filtering, the bone edge was clearly observed. In addition, soft tissue structures were identified in the same display settings (window level/width; WL/WW) as the bone edge observation; that is, the adjustment of the display settings was not required for the observation of each object. These results suggested that both bone edge enhancement and dynamic range compression would be achieved successfully. It was estimated that the images obtained by the proposed method were more appropriate for the registration than conventional portal images, in 47 times registrations of 50 times in total (the registrations by five radiological technologists in ten patients). The proposed method was concluded to be useful for improving the quality of portal images, enabling the efficient registration.

Effectiveness of the single-shot dual-energy subtraction technique for portal images

The aim of the present study was to evaluate the clinical efficacy of the single‐shot dual‐energy subtraction technique for obtaining portal images. We prepared two storage phosphor plates for this study. A 1 mm thick tungsten sheet was placed between the two storage phosphor plates. A single use of the double‐exposure technique provides two portal images simultaneously (i.e., a standard image and a low‐contrast image), using the same patient position and with no additional radiation delivered to the patient. A bone‐enhanced image is created by image subtraction between these two images. For evaluation of clinical efficacy, three treatment sites — the brain, lung, and pelvis — were imaged. Ten sets of images were obtained for each site, and five landmarks were selected for each treatment site. The visibility of each landmark and the ease of overall verification for the selected treatment sites were assessed separately for the standard and bone‐enhanced images. Four observers consisting of one radiation oncologist and three radiation therapists participated in the present study. For most of the landmarks studied, the bone‐enhanced images were significantly superior to the standard images. Regarding the ease of overall verification, the bone‐enhanced images were significantly superior to the standard images at all sites. The p‐values of mean rating for the brain, lung, and pelvis were 0.002, 0.012, and 0.003, respectively. The bone‐enhanced images obtained using our technique increased the image quality in terms of bone visibility, and are considered useful for routine clinical practice.

PACS number: 87.56.Da

Detection of patient setup errors with a portal image - DRR registration software application

The purpose of this study was to evaluate a custom portal image — digitally reconstructed radiograph (DRR) registration software application. The software works by transforming the portal image into the coordinate space of the DRR image using three control points placed on each image by the user, and displaying the fused image. In order to test statistically that the software actually improves setup error estimation, an intra‐ and interobserver phantom study was performed. Portal images of anthropomorphic thoracic and pelvis phantoms with virtually placed irradiation fields at known setup errors were prepared. A group of five doctors was first asked to estimate the setup errors by examining the portal and DRR image side‐by‐side, not using the software. A second group of four technicians then estimated the same set of images using the registration software. These two groups of human subjects were then compared with an auto‐registration feature of the software, which is based on the mutual information between the portal and DRR images. For the thoracic case, the average distance between the actual setup error and the estimated error was 4.3±3.0 mm for doctors using the side‐by‐side method, 2.1±2.4 mm for technicians using the registration method, and 0.8±0.4 mm for the automatic algorithm. For the pelvis case, the average distance between the actual setup error and estimated error was 2.0±0.5 mm for the doctors using the side‐by‐side method, 2.5±0.4 mm for technicians using the registration method, and 2.0±1.0 mm for the automatic algorithm. The ability of humans to estimate offset values improved statistically using our software for the chest phantom that we tested. Setup error estimation was further improved using our automatic error estimation algorithm. Estimations were not statistically different for the pelvis case. Consistency improved using the software for both the chest and pelvis phantoms. We also tested the automatic algorithm with a database of over 5,000 clinical cases from our hospital. The algorithm performed well for head and breast but performed poorly for pelvis cases, probably due to lack of contrast in the megavoltage portal image. The software incorporates an original algorithm to fuse portal and DRR images, which we describe in detail. The offset optimization algorithm used in the automatic mode of operation is also unique, and may be useful if the contrast of the portal images can be improved.

PACS numbers: 87.55.Qr, 87.57.nj

A new population-enrichment strategy to improve efficiency of placebo-controlled clinical trials of antidepressant drugs

The rate-limiting factor in the discovery of novel antidepressants is the inefficient methodology of traditional multicenter randomized clinical trials (RCTs). We applied a model-based approach to a large clinical database (five RCTs in major depressive disorder (MDD), involving 1,837 patients from 124 recruitment centers) with two objectives: (i) to learn about the role of center-specific placebo response in RCT failure and (ii) to apply what is learned to improve the efficiency of RCTs by enhancing the detection of treatment effect (TE). Sensitivity analysis indicated that center-specific placebo response was the most relevant predictor of RCT failure. To reduce the statistical "noise" generated by centers with nonplausible, excessively high/low placebo responses, we developed an enrichment-window strategy. Clinical trial simulation was used to assess the enrichment strategy applied before the standard statistical analysis, resulting in an overall reduction in failure of RCTs from ~50 to ~10%.

Relationship between radiation dose reduction and image quality change in photostimulable phosphor luminescence X-ray imaging systems

OBJECTIVES

The aim of the study was to clarify the change in image quality upon X-ray dose reduction and to re-analyse the possibility of X-ray dose reduction in photostimulable phosphor luminescence (PSPL) X-ray imaging systems. In addition, the study attempted to verify the usefulness of multiobjective frequency processing (MFP) and flexible noise control (FNC) for X-ray dose reduction.

METHODS

Three PSPL X-ray imaging systems were used in this study. Modulation transfer function (MTF), noise equivalent number of quanta (NEQ) and detective quantum efficiency (DQE) were evaluated to compare the basic physical performance of each system. Subjective visual evaluation of diagnostic ability for normal anatomical structures was performed. The NEQ, DQE and diagnostic ability were evaluated at base X-ray dose, and 1/3, 1/10 and 1/20 of the base X-ray dose.

RESULTS

The MTF of the systems did not differ significantly. The NEQ and DQE did not necessarily depend on the pixel size of the system. The images from all three systems had a higher diagnostic utility compared with conventional film images at the base and 1/3 X-ray doses. The subjective image quality was better at the base X-ray dose than at 1/3 of the base dose in all systems. The MFP and FNC-processed images had a higher diagnostic utility than the images without MFP and FNC.

CONCLUSIONS

The use of PSPL imaging systems may allow a reduction in the X-ray dose to one-third of that required for conventional film. It is suggested that MFP and FNC are useful for radiation dose reduction.

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.

A pilot study of architectural distortion detection in mammograms based on characteristics of line shadows

OBJECTIVE

We present herein a novel algorithm for architectural distortion detection that utilizes the point convergence index with the likelihood of lines (e.g., spiculations) relating to architectural distortion.

MATERIALS AND METHODS

Validation was performed using 25 computed radiography (CR) mammograms, each of which has an architectural distortion with radiating spiculations. The proposed method comprises five steps. First, the lines were extracted on mammograms, such as spiculations of architectural distortion as well as lines in the mammary gland. Second, the likelihood of spiculation for each extracted line was calculated. In the third step, point convergence index weighted by this likelihood was evaluated at each pixel to enhance distortion only. Fourth, local maxima of the index were extracted as candidates for the distortion, then classified based on nine features in the last step.

RESULTS

Point convergence index without the proposed likelihood generated 84.48/image false-positives (FPs) on average. Conversely, the proposed index succeeded in decreasing this number to 12.48/image on average when sensitivity was 100%. After the classification step, number of FPs was reduced to 0.80/image with 80.0% sensitivity.

CONCLUSION

Combination of the likelihood of lines with point convergence index is effective in extracting architectural distortion with radiating spiculations.

Improvement of lung abnormality detection in computed radiography using multi-objective frequency processing: Evaluation by receiver operating characteristics (ROC) analysis

Computed radiography (CR) has been shown to have relatively low sensitivity for detection of pulmonary nodules. This poor sensitivity precludes its use as a screening modality despite the low cost, low dose and wide distribution of devices. The purpose of this study was to apply multi-objective frequency processing (MFP) to CR images and to evaluate its usefulness for diagnosing subtle lung abnormalities. Fifty CR images with simulated subtle lung abnormalities were obtained from 50 volunteers. Each image was processed with MFP. We cut chest images. The chest image was divided into two rights and left. A total of 200 half-chest images (100 MFP-processed images and 100 MFP-unprocessed images) were prepared. Five radiologists participated in this study. ROC analyses demonstrated that the detection rate of simulated subtle lung abnormalities on the CR images was significantly better with MFP (Az=0.8508) than without MFP (Az=0.7925). The CR images processed with MFP could be useful for diagnosing subtle lung abnormalities. In conclusion, MFP appears to be useful for increasing the sensitivity and specificity in the detection of pulmonary nodules, ground-glass opacity (GGO) and reticular shadow.