We investigated the effect on image data resampling in an evaluation of the basic imaging properties for a digital radiographic system based on a flat panel detector (FPD). One of the latest digital radiographic systems was used in this study. This system was based on a direct-conversion FPD of amorphous selenium. The basic imaging properties of the system were evaluated by measuring characteristic curve, presampled modulation transfer function (MTF), and Wiener spectrum (WS) using DICOM image with a matrix size of 2048×2048. The evaluations were performed under two conditions because matrix size automatically changes according to the selection of imaging size. One of the conditions was a different matrix size between image data acquired on the FPD and the output image (DICOM image for which resampling was performed). The other condition was that these matrices be the same size (DICOM image with no resampling performed). Resampling did not affect the characteristic curves. However, MTF and the WS obtained from the resampled data were different from those of the one not resampled, which is considered to be the "inherent" basic imaging properties, and this phenomenon was remarkable, especially in terms of the MTFs. Our study indicates that the effect on resampling should not be disregarded in evaluating the basic imaging properties of digital radiographic systems. Therefore, it is mandatory to use DICOM images for which no resampling was performed in order to evaluate the inherent basic imaging properties for digital radiographic systems.
Objective: Image fusion has been recognized as a useful technique in diagnostic imaging. We have been evaluating the manual image fusion of PET and contrast-enhanced (CE) CT obtained separately. The CT images can be used for attenuation correction as well as for image fusion; however, the quantitative accuracy of CT-corrected PET images has yet to be assessed. The purpose of this study was to compare the radioactivity concentration between conventional 68Ge-corrected and CECT-corrected PET images. Methods: Twenty patients underwent a whole-body PET scan, followed by a CT scan with intravenous contrast material, after careful positioning using an individually molded vacuum cushion. Two different attenuation-corrected emission data sets were produced, i.e., 68Ge-corrected images and CECT-corrected images. Image registration was performed by maximizing mutual information-based cost function, between the CT and the combination of emission and transmission PET volumes. The CT pixel values in Hounsfield units were transformed into linear attenuation coefficients in cm−1, using a conversion formula for a lookup-table from phantom experiments. Measured activity concentrations from identical regions of interest in representative normal organs and in 25 pathologic foci of uptake were compared. In addition, the quality of the reconstructed images was assessed using the normal mean square error (NMSE). Results: Measured average radioactivity concentrations were 1.38-9.56% higher for CECT-corrected images than for 68Ge-corrected images. Overall, the NMSE value of CECT-corrected images compared with 68Ge-corrected images was 0.02±0.01. Conclusions: The difference in quantitative values between 68Ge-corrected and CECT-corrected PET images was comparable to that of an integrated PET/CT system. Diagnostic CT images with intravenous contrast performed separately before or after a PET scan could be used clinically not only for fusion but also for attenuation correction.
Intensity-modulated radiation therapy (IMRT) represents one of the most significant technical advances in radiation therapy. In the dynamic multileaf collimator (MLC) method of IMRT delivery, because of the relatively small gaps between opposed leaves and because most regions are shielded by leaves most of the time, the delivered dose is very sensitive to MLC leaf positional accuracy. A variation of ±0.2 mm in the gap width can result in a dose variation of ±3% for each clinical dynamic MLC field. Most often the effects of leaf motion are inferred from dose deviations on film or from variations in ionization measurements. These techniques provide dosimetric information but do not provide detailed information for diagnosing delivery problems. Therefore, a dynamic log file (Dynalog file) was used to verify dynamic MLC leaf positional accuracy. Measuring for narrow gaps using the thickness gauge could detect a log file accuracy of approximately 0.1 mm. The accuracy of dynamic MLC delivery depends on the accuracy with which the velocity of each leaf is controlled. We studied the relationship between leaf positional accuracy and leaf velocity. Leaf velocity of 0.7 cm/sec caused approximately 0.2 mm leaf positional variation. We then analyzed leaf positional accuracy for the clinical dynamic MLC field using Dynalog File Viewer (Varian Medical Systems, Inc., Palo Alto, CA), and developed a new program that can analyze more detailed leaf motions. Using this program, we can obtain more detailed information, and therefore can determine the source of dose uncertainties for the dynamic MLC field.
Image evaluation with Response Evaluation Criteria in Solid Tumors (RECIST) evaluates the change in a measurable lesion as determined by ruler or micrometer caliper. However, there is no definition of the conditions thought to influence the precision of measurement. We therefore examined the effects on measurement precision by changing image amplification, WW, WL, and time phase. Moreover, to determine response rate, one-dimensional evaluation with RECIST was compared with the two-dimensional evaluation of World Health Organization (WHO) for a hepatocellular carcinoma. The results of measuring the object lesion for measured value variation were as follows. Under image conditions of 1 time expansion/ WW 150/ WL 100 was (4.92±1.94)%. Under image conditions of 1 time expansion/ WW 350/ WL 75 was (4.42±1.70)%. Under image conditions of 4 times expansion/ WW 150/ WL 100 was (2.52±0.82)%. Under image conditions of 4 times expansion/ WW 350/ WL 75 was (2.83±1.10)%. When an image was enlarged to 4 times, precision doubled. There was no a difference in comparing RECIST to WHO in terms of response rate. Thus the best method was considered to be RECIST because of its convenience.