MR imaging is an important method for diagnosing abnormalities of the brain. This paper presents an automated method to segment the cerebellum and brainstem for brain MR images. MR images were obtained from 10 normal subjects (male 4, female 6; 22-75 years old, average 31.0 years) and 15 patients with brain atrophy (male 3, female 12; 62-85 years of age, average 76.0 years). The automated method consisted of the following four steps : (1) segmentation of the brain on original images, (2) detection of an upper plane of the cerebellum using the Hough transform, (3) correction of the plane using three-dimensional (3D) information, and (4) segmentation of the cerebellum and brainstem using the plane. The results indicated that the regions obtained by the automated method were visually similar to those obtained by a manual method. The average rates of coincidence between the automated method and manual method were 83.0±9.0% in normal subjects and 86.4±3.6% in patients.
The multi-sector reconstruction (MSR) algorithm and cardiac half-reconstruction (CHR) algorithm are the main algorithms used in cardiac reconstruction. Analysis of effective temporal resolution (TR) confirmed that optimal rotation speed depends on different heart rates when using MSR. During visualization (3D/MPR image) and quantitative (EF : ejection fraction) evaluations, it was found that image quality and measurement accuracy are well correlated with effective temporal resolution (TR) by the different algorithms. The CHR algorithm resulted in less desirable image quality at TR 250 ms than that from MSR at high heart rates (>75 bpm) in the phantom experiment. We determined that the combination of the MSR algorithm and the optimal selection of gantry rotation speed is important for obtaining high-quality cardiac imaging in the high heart rate region.
In abdominal X-ray computerized tomography (CT), exposure conditions can be established by real-time exposure control and manual procedures, depending on the body type. In the former, appropriate dose control can be attained. In the latter, dose cannot be controlled to a very precise level, since the dose is controlled manually, according to body type. In head X-ray CT, the conditions are in most cases established manually. We determined the causal relationship between the body mass index (BMI) showing the body type of a subject and the abdominal cross section/circumferential length of the subject. Then a water phantom that simulated the abdominal area was used to take pictures, with the current time integrator changed, relating the section area and circumferential length of the water phantom with the body mass index, thereby determining the relationship with the current time integrator. Further, a water phantom simulating the head was used to take pictures, with the current time integrator changed, thereby determining the relationship between the circumferential length of the water phantom with the current time integrator. We determined that pictures can be taken at a dose that is close to an appropriate dose, when these relationships are plotted to determine the standard deviation (SD) for establishing exposure conditions.
We have developed a system to assist in patient positioning for radiation therapy that includes an automated registration process between the reference image and target volume data before irradiation. The system uses the cross-correlation function between images for the automatic positioning processing algorithm. Evaluation of the system showed good results in providing highly accurate registration. Therefore, the system is suitable for use in clinical study. In addition, evaluation of clinical data of the prostate using the system revealed displacement of the internal organs.
The results of statistical image analyses, such as three-dimensional stereotactic surface projection (3D-SSP), are affected by the normal database (NDB) used. As a general rule, each institution must establish its own NDB for analysis. However, the numbers of institutions with the ability to establish their own NDB are limited. Therefore, it would be ideal if an NDB suitable for other institutions could be established from standard volunteer data gathered at a single institution. In this study, we examined the validity of establishing an NDB suitable for other institutions from the standard volunteer data gathered at our institution, using E.CAM (used as a reference) and IRIX. We established three NDBs. One was established from E.CAM, and the other two were from IRIX (in which overall FWHM was the same as that of E.CAM with Butterworth filter cutoff frequency and using an identical algorithm for image-reconstruction of E.CAM). We then compared the mean values and standard deviations of the NDBs, and also examined the effects on the Z-score map. We determined that the NDB that had been established was suitable for other institutions, using identical FWHM (resolution of SPECT imaging) and an algorithm for image-reconstruction.