To investigate proper radiographic contrast and latitude based on density perception of the high-density areas of X-ray films, we studied the human characteristics of darkness perception for those areas. Magnitude estimation was used to provide a numerical estimate of the sensory magnitudes for contrast. Experimental figures that imitated the head were imaged on X-ray films. Magnitude estimation of darkness in the surrounding fields showed greater misperception or darker perception. When the density of the background field was high, the larger the density difference between the background and surrounding field, the darker the surrounding field was perceived. This implies that the lighter field is assimilated into the darker field. Since assimilation reduces the perceived density difference between soft tissue and the area around it, using high contrast film is less effective for attempting to increase the physical density difference between such areas. We concluded that psychological assimilation of darkness on contrast perception had a detrimental influence on the density perception of high-density areas on head radiographs.
The PROPELLER (periodically rotated overlapping parallel lines with enhanced reconstruction) MRI method is available with a technique for motion correction. We studied changes in correction according to differences in echo train length (ETL)and number of blades, and measured correction accuracy according to linear translation and rotation, using computer simulation. A T2-weighted axial image of the head taken by FSE was utilized as the basic MR image. We reconstructed other images with differences in quantity of motion, ETL, and number of blades, also using computer simulation. After motion correction was performed, we measured correction accuracy with cross-correlation to the basic image. The method of motion correction was performed by the conversion of k-space data for each blade step using 2D Fourier-transform. After motion correction of the obtained image had been carried out, the image was converted to k-space data using reverse Fourier-trans-form. For data of 30 pixels with horizontal translation, cross-correlation coefficients for the stationary image were 0.6 for FSE without motion correction, 0.74 for the PROPELLER (ETL 8, blades 32) image without motion correction, and 0.99 for the PROPELLER (ETL 8, blades 32) image with motion correction. For data of 24 degrees with rotation, cross-correlation coefficients were 0.38 for FSE without motion correction, 0.53 for the PROPELLER (ETL 8, blades 32) image without motion correction, and 0.93 for the PROPELLER (ETL 8, blades 32) image with motion correction. The cross-correlation coefficient of liner translation is higher than its rotation. Correction accuracy was better with larger numbers of ETL than without motion correction. The spatial resolution of the image was decreased in the corrected image more by rotation than linear translation. This study indicated that the PR method was able to inspect the imaging technique with little influence on movement.
In this study, we investigated the usefulness of the fluorescent glass dosimeter for measuring patient dose. The fluorescent glass dosimeter is constructed of a glass element and its holder. One type has a tin (Sn) filter and the other does not. The characteristics of these two types of fluorescent glass dosimeters were studied in the range of diagnostic X-ray energy. The result was excellent for each characteristic. Directional dependency, however, was recognized in the fluorescent glass dosimeter with tin (Sn) filter. Based on these evaluations, patient skin dose was measured for abdominal interventional radiology and diagnostic digital subtraction angiography using the holder without filter, which is less direction-dependent and eliminates obstructive shadows in radiography and fluoroscopy. The average skin dose of 30 patients for abdominal IVR was 1.17±0.44 Gy (0.51-1.94 Gy), while those for diagnostic DSA examination was 0.54±0.21 Gy (0.15-1.02 Gy). The fluorescent glass dosimeter provides high capability for skin dose measurement. The fluorescent glass dosimeter is also useful for controlling patient dose during IVR procedures.
Coronary artery visualization by multi-slice CT and the evaluation of coronary stenosis were examined. Multi-slice CT could not be used for evaluation after the coronary artery stent custody method. We examined various coronary artery stent custody phantoms and stent visualization of clinical examples by using multi-slice CT with a slice thickness of 0.5 mm. ECG-gated scanning was done by inputting ECG using 4DAS (Data Acquisition System) multi-slice CT with a slice thickness of 0.5 mm. We carried out image reconstruction, measured the CT number, and examined the inside stent. The CT number of the inside stent rose in accordance with stent diameter. The inside stent was visualized clearly at 0.5 mm in comparison with 1.0 mm in slice thickness. Visualization of the lumen was influenced in the Stent Custody Phantom by the quality of the stent material, Strut form, and the size of the diameter. The inside stent could be evaluated with the S670 4 mm0 stent. Form evaluation of the inside stent was possible, although the cavity in the stent was influenced by the difference in distance between peaks.
Recent reports have indicated that depiction of the thoracic duct is possible without administration of a contrast agent using fast advanced spin echo(FASE), ECG-triggered, half-Fourier fast spin echo(FSE), by depicting blood vessels. In this study, we attempted to depict the thoracic duct using FASE, which is generally used for MR-hydrography. By varying effective echo time (effective TE), the contrast-to-noise ratios (CNR)for saline and baby oil were measured with and without fat suppression. Without fat suppression, the effective TE of 500 msec yielded the highest CNR. With fat suppression, the effective TE of 250 msec provided the highest CNR. Next, examinations of the thoracic duct were performed in volunteers in order to obtain the highest CNR. Results indicated that the best depiction of the thoracic duct was obtained using the effective TE of 500 msec in 3D-FASE without fat suppression. Thoracic duct imaging using heavily T2-weighted parameters allows better control of signal intensities of background and surrounding tissues than can be obtained with fat suppression. Furthermore, the heavily T2-weighted parameter only depicts the long T2 components of the thoracic duct.
We studied the objective diagnosis of Alzheimer-type dementia based on changes in the corpus callosum. We examined midsagittal head MR images of 40 Alzheimer-type dementia patients(15 men and 25 women; mean age, 75.4±5.5 years)and 31 healthy elderly persons(10 men and 21 women; mean age, 73.4+7.5 years), 71 subjects altogether. First, the corpus callosum was automatically extracted from midsagittal head MR images. Next, Alzheimer-type dementia was compared with the healthy elderly individuals using the features of shape factor and six features of Co-occurrence Matrix from the corpus callosum. Automatic extraction of the corpus callosum succeeded in 64 of 71 individuals, for an extraction rate of 90.1%. A statistically significant difference was found in 7 of the 9 features between Alzheimer-type dementia patients and the healthy elderly adults. Discriminant analysis using the 7 features demonstrated a sensitivity rate of 82.4%, specificity of 89.3%, and overall accuracy of 85.5%. These results indicated the possibility of an objective diagnostic system for Alzheimer-type dementia using feature analysis based on change in the corpus callosum.