Takashi Yokoi

Osaka University, Ōsaka-shi, Osaka-fu, Japan

Are you Takashi Yokoi?

Claim your profile

Publications (12)17.49 Total impact

  • [Show abstract] [Hide abstract]
    ABSTRACT: The registration of images from positron emission tomography (PET) to those from magnetic resonance imaging (MRI) using mutual information is usually effective, but fails occasionally because of small region of overlap, low-activity defects in the PET image, difference in spatial resolution, etc. In this article, the authors propose the pixel-based individual entropy correlation coefficient (IECC) as a new, more accurate and more robust registration criterion. The authors compare it to the current criteria: Mutual information (MI), normalized mutual information (NMI), and the entropy correlation coefficient (ECC). The anatomical region to be registered was the head. A rigid-body registration was used; no deformation was employed. The authors established the effectiveness of IECC by both simulated data and clinical studies using brain fluorodeoxyglucose (FDG) PET and MRI. Both a normal-activity FDG model and a FDG model with a perfusion defect were used for the PET image. Reconstruction by both filtered backprojection and by ordered subset-expectation maximization was investigated. The mean errors and SDs of IECC were 1.17 +/- 0.85 mm for translation and 1.04 +/- 1.28 degrees for rotation in clinical PET. Those of MI, NMI, and ECC were 1.86 +/- 1.22, 1.86 +/- 0.96, and 1.68 +/- 4 1.05 mm for translations and 1.52 +/- 1.84 degrees, 1.74 +/- 1.68 degrees, and 1.70 +/- 1.90 degrees for rotations. The mean errors and SDs of IECC were 1.56 +/- 0.58 mm for translation and 1.46 +/- 1.66 degrees for rotation in clinical PET model with a 30% perfusion defect. Those of MI, NMI, and ECC were 2.81 +/- 1.41, 2.98 +/- 1.80, and 3.29 +/- 2.08 mm for translations and 3.34 +/- 3.800, 2.87 +/- 3.25 degrees, and 3.04 +/- 3.44 degrees for rotations. The mean errors and SDs of IECC were 1.79 +/- 1.04 mm for translation and 1.64 +/- 1.62 degrees for rotation in clinical PET model with a 50% perfusion defect. Those of MI, NMI, and ECC were 3.49 +/- 1.92, 3.57 +/- 2.22, and 3.49 +/- 1.89 mm for translations and 4.12 +/- 4.24 degrees, 3.62 +/- 4.87 degrees, and 3.44 +/- 3.80 degrees for rotations. The significant differences between IECC and each of the other three criteria were found for displacement misregistration in almost all parameters (p < 0.01). Accuracy of the IECC criterion was higher than that of the other criteria, usually in a statistically significant way.
    No preview · Article · Feb 2011 · Medical Physics
  • [Show abstract] [Hide abstract]
    ABSTRACT: Recently, whole-body positron emission tomography (PET) examination has greatly developed. To reduce the overall examination time, the transmission scan has been increasingly shortened. Many noise-reduction processes have been developed for count-limited transmission data. Segmented attenuation correction (SAC) is one method by which the pixel values of transmission image are transformed into several groups. The median root prior-ordered subset convex (MRP-OSC) algorithm is another method that is applicable to control the noise level on the basis that the change of the pixel value is locally monotonous. This article presents an alternative approach on the basis of the Bayesian iterative reconstruction technique incorporating a median prior and an anatomical prior from the segmented mu-map for count-limited transmission data. The proposed method is based on the Bayesian iterative reconstruction technique. The median prior and the anatomical prior are represented as two Gibbs distributions. The product of these distributions was used as a penalty function. In the thorax simulation study, the mean square error from the true transmission image of the presented method (5.74 x 10(-5)) was lower than MRP-OSC (6.72 x 10(-5)) and SAC (7.08 x 10(-5)). The results indicate that the noise of the image reconstructed from the proposed technique was decreased more than that of MRP-OSC without segmentation error such as that of an SAC image. In the thorax phantom study, the emission image that was corrected using the proposed technique displayed little noise and bias (27.42 +/- 0.96 kBq/ml, calculated from a region of interest drawn on the liver of the phantom); it was very similar to the true value (28.0 kBq/ml). The proposed method is effective for reducing propagation of noise from transmission data to emission data without loss of the quantitative accuracy of the PET image.
    No preview · Article · Jun 2008 · Annals of Nuclear Medicine
  • [Show abstract] [Hide abstract]
    ABSTRACT: In nuclear medicine, cerebral vascular reserve(CVR) is evaluated using technetium-99m ethyl cysteinate dimer [99mTc-ECD] and acetazolamide(ACZ). We developed a protocol involving the intravenous injection of 99mTc-ECD in three divided doses(TIE method), and have found that the cerebrovascular response to ACZ depended on time after ACZ administration. However, it was difficult to obtain high-precision quantitative SPECT images by the conventional method because of complicated image processing and image degradation accompanying image subtraction. We recently developed software known as the Automatic Quantitative CVR Estimation Tool(hereinafter referred to as Triple AQCEL), which, after the input of simple parameters, enables us to carry out automatic reconstruction of quantitative SPECT images without image degradation due to subtraction. Triple AQCEL was determined to reduce image degradation caused by subtraction and to provide valid quantitative data. Because Triple AQCEL does not require manual determination of ROI or image selection for the reconstruction of quantitative SPECT images, reproducibility of regional cerebral blood flow by 3DSRT is ensured. Since all analyses in evaluation by the TIE method are automated and the operator plays no part in them, with the resulting increase of throughput, this software will contribute to improved reproducibility of regional cerebral blood flow data, and will be useful in clinical pathophysiological assessment both preoperatively and during postoperative follow-up.
    No preview · Article · Jun 2007 · Nippon Hoshasen Gijutsu Gakkai zasshi
  • [Show abstract] [Hide abstract]
    ABSTRACT: The following process conventionally has been followed to develop quantitative images of cerebral blood flow: (1) mean cerebral blood flow (mCBF) is calculated by the Patlak plot method; (2) a SPECT slice that includes the basal ganglia is selected; and (3) based on the value of mCBF calculated by the Patlak plot method, the SPECT slice is corrected by the Lassen method and developed into a SPECT image of quantitative regional cerebral blood flow. However, this process is complicated, and the values of rCBF have been reported to fluctuate because selection of the SPECT slice and the ROI setting are in the hands of the operator. We have developed new software that automates this analysis. This software enables automatic processing simply by inputting the value of mCBF in the normal hemisphere. Since there is no need for manual operations such as setting the ROI, reproducibility is improved as well. Regional cerebral blood flow as determined by this software is quite similar to that calculated by the conventional method, so the existing clinical evaluation does not need to be changed. This software is considered to be useful.
    No preview · Article · Jun 2006 · Nippon Hoshasen Gijutsu Gakkai zasshi

  • No preview · Article · Mar 2005 · Annals of Nuclear Medicine
  • [Show abstract] [Hide abstract]
    ABSTRACT: Mean cerebral blood flow (mCBF) in the slice including the basal ganglia (reference slice) is necessary for the quantification of regional CBF using Patlak plot and BUR methods on 99mTc-ECD cerebral perfusion SPECT. The mCBF was calculated from the mean counts of this slice. A region of interest (ROI) has been manually set on the reference slice to obtain the mean counts (manual ROI method). However, there was large variability observed in the value of rCBF in this method. We developed a 3DSRT method for improving the accuracy of the mean counts in the reference slice and evaluated the difference between the value of rCBF on manual ROI method and that on 3DSRT method in consecutive 11 patients with cerebral vascular disease. Difference in the value of mean counts of the reference slice was distributed within the 2 standard deviations (SD) with Blant-Altman analysis in 9 of 11 patients. Significant difference in the value of mean counts between two methods was observed in 2 of 11 patients. 3DSRT method is superior accuracy to the manual ROI method in the evaluation of the counts in the ROI. Lower accuracy in manual ROI method, therefore, results in the difference of the value of mean counts. 3DSRT method provides high accuracy with the various quantitative methods for the evaluation of rCBF using 99mTc-ECD.
    No preview · Article · Mar 2005 · Kaku igaku. The Japanese journal of nuclear medicine
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: We implemented a 3D co-registration technique based on mutual information (MI) including 2D image matching as a coarse pre-registration. The 2D coarse pre-registration was performed in the transverse, sagittal and coronal planes sequentially, and all six parameters were then optimized as fine registration. Normalized mutual information (NMI) was also examined as another entropy-based measure that was invariant to the overlapped area of two images. In order to compare accuracy and precision of the present method with a conventional two-level multiresolution approach, simulation was performed by 100 trials with the random initial mismatch of +/-10 degrees and +/-17.92 mm (Type-I) and +/-20 degrees and +/-40.32 mm (Type-II). For Type-I, no significant differences were found between registration errors of the multiresolution approach and the present method with the MI criterion. No biases were observed (< or =0.13 degrees and < or =0.57 mm for the multiresolution approach; < or =0.12 degrees and < or =0.57 mm for the present method) and the SDs were very small (< or =0.18 degrees and < or =0.12 mm for the multiresolution approach; < or =0.11 degrees and < or =0.11 mm for the present method). For Type-II, SDs for the multiresolution approach (< or =1.8 degrees and < or =0.88 mm) were markedly larger than those for the present method (< or =0.64 degrees and < or =0.20 mm) with MI. Success rate for the present method was 99.9%, which was higher than 97.6% for the multiresolution approach. Simulation also revealed that MI and NMI performance were almost equivalent. The choice of optimization strategy more affected accuracy and reproducibility than the choice of the registration criterion (MI or NMI) in our simulation condition. The present method is sufficiently accurate and reproducible for MRI-SPECT registration in clinical use.
    Full-text · Article · Jan 2005 · Annals of Nuclear Medicine
  • Takashi Yokoi · Hiroyuki Shinohara · Akihiro Takaki
    [Show abstract] [Hide abstract]
    ABSTRACT: Split-dose injection using technetium-99m ethyl cysteinate dimer ((99m)Tc-ECD) and consecutive SPET measurements performed before and after acetazolamide (ACZ) loading was used to estimate the cerebral perfusion reserve. The disadvantage of the split-dose method is that the signal-to-noise ratio (S/N) of ACZ-loaded images is decreased by subtraction of the 1st SPET data (rest) from the 2nd SPET data (ACZ loaded). To improve the S/N of reconstructed images, we implemented an iterative reconstruction algorithm including the term of remaining radioactivity in the brain from the 1st injection. It was expected that this method (the "addition method") would improve the S/N of rest and ACZ images compared with the conventional subtraction method owing to exclusion of the subtraction process. To evaluate the effect of statistical noise, we estimated the percentage coefficient of variation (%COV) as a function of total photon counts (from 1.35 to 86.5 Mcounts/slice) by Monte Carlo simulation with equal-volume split-dose injection. The %COV of the 2nd SPET study was higher than that of the 1st (e.g. 50.3% for the 1st and 80.5% for the 2nd at a total count of 2.70 Mcounts/slice) when using the conventional subtraction method. By contrast, the %COV of the 1st and 2nd SPET studies was almost equivalent (e.g. 43.1% for the 1st and 41.4% for the 2nd at a total count of 2.70 Mcounts/slice) when using the addition method. We also determined the optimal injection dose ratio of the 2nd to the 1st SPET study, which provides the equivalent %COV value between the 1st and 2nd images. With the subtraction method, the optimal injection dose ratio of the 2nd to the 1st SPET study was approximately 2.0, while with the addition method it was approximately 1.0. The absolute value of %COV at the optimal injection dose was about 54% and 43% with the subtraction method and the addition method, respectively. The addition method gave a lower %COV value than the subtraction method even at the optimal injection dose ratio. In a clinical study, the addition method provided better quality images than the subtraction method. The ROI values of rest images estimated by the subtraction method were close to the results obtained with the addition method (ROI(sub)=1.01 ROI(add)-0.312, r=0.999). The ROI values of the ACZ images estimated by the subtraction method also agreed with the results obtained using the addition method, but the correlation was slightly worse (ROI(sub)=1.03 ROI(add)-2.23, r=0.995). Quantitative ROI values were quite similar between the methods. Our results demonstrated that the quality of reconstructed rest and ACZ-loaded images were significantly better with the addition method than with the conventional subtraction method. We conclude that the proposed method will be useful as a practical reconstruction algorithm to improve the S/N in an equal-volume split-dose injection protocol using (99m)Tc-ECD.
    No preview · Article · Sep 2003 · European journal of nuclear medicine and molecular imaging
  • Source
    Takashi Yokoi · Hiroyuki Shinohara · Hideo Onishi
    [Show abstract] [Hide abstract]
    ABSTRACT: Iterative reconstruction techniques such as an ordered subsets-expectation maximization (OSEM) algorithm can easily incorporated various physical models of attenuation or scatter. We implemented OSEM reconstruction algorithm incorporating compensation for distance-dependent blurring due to the collimator in SPECT. The algorithm was examined by computer simulation to estimate the accuracy for brain perfusion study. The detector response was assumed to be a two-dimensional Gauss function and the width of the function varied linearly with the source-to-detector distance. The attenuation compensation (AC) was also included. To investigate the properties of the algorithm, we performed computer simulations with the point source and digital brain phantoms. In the point source phantom, the uniformity of FWHM for the radial, tangential and longitudinal directions was evaluated on the reconstruction image. As for the brain phantom, quantitative accuracy was estimated by comparing the reconstructed images with the true image by the mean square error (MSE) and the ratio of gray and white matter counts (G/W). Both noise free and noisy simulations were examined. In the point source simulation, FWHM in radial, tangential and longitudinal directions were 14.7, 14.7 and 15.0 mm at the image center and were 15.9, 9.83 and 10.6 mm at a distance of 15 cm from the center by using FBP, respectively. On the other hand, they were 8.12, 8.12 and 7.83 mm at the image center, and were 7.45, 7.44 and 7.01 mm at 15 cm from the center by OSEM with distance-dependent resolution compensation (DRC). An isotropic and stationary resolution was obtained at any location by OSEM with DRC. The spatial resolution was also improved about 6.5 mm by OSEM with DRC at the image center. In the brain phantom simulation, the blurring at the edge of the brain structure was eliminated by using OSEM with both DRC and AC. The G/W was 2.95 and 2.68 for noise free and noisy cases, respectively, when no compensation was performed. But the values for G/W without and with noise became 3.45 and 3.21 with AC only and were improved to 3.75 and 3.71 with both AC and DRC. The G/W approached the true value (4.00) by using OSEM with both AC and DRC even when there was statistical noise. In conclusion, OSEM reconstruction including the distance-dependent resolution compensation algorithm was reasonably successful in achieving isotropic and stationary resolution and improving the quantitative accuracy for brain perfusion SPECT.
    Preview · Article · Mar 2002 · Annals of Nuclear Medicine
  • [Show abstract] [Hide abstract]
    ABSTRACT: Three differing exact methods of inverting the two-dimensional (2D) exponential Radon transform were implemented and evaluated quantitatively with a phantom study. The phantom had the shape of a pie-chart divided into six cavities, each 480 ml in volume and 10 cm in height, that were symmetrically positioned in a cylinder that was 20 cm in diameter and 10 cm in height. This phantom tests for linearity between true activity concentration and measured activity concentration, and it is denoted as a linearity phantom in the present study. Each cavity contained a different concentration of a homogeneous solution of 99mTc (74, 148, 222, 296, 370 and 444 kBq ml(-1)). Data acquisition was performed with two energy windows: a 20% photopeak energy window set symmetrically over the 140 keV of 99mTc and a secondary 5% energy window set over the 122 keV peak. We optimized a triple-energy window scatter correction method for a gamma camera-collimator system to obtain accurate scatter-corrected projections. A circular ROI 3 cm in diameter was identified over each cavity region, and count density (counts per pixel) was calculated. This value was converted to activity concentration (kBq ml(-1)) using a cross-calibration coefficient between SPECT counts and the gamma well counter. The relation between true activity (x) and measured activity concentration (y) was fitted to a line using the least-squares method. Regression lines were y = 0.63 + 1.0255x (R2 = 0.9987), y = -2.62 + 1.0278x (R2 = 0.9995), and y = 0.092 + 1.0241x (R2 = 0.9989) for the Bellini, Inouye and Metz-Pan methods respectively. In another phantom study using two different types of phantoms, contrast of a cold region in the two was 96% and 101% for all three methods. Combined optimized scatter correction and analytical attenuation correction methods achieve good accuracy in quantification of activity distribution with a uniform attenuating medium.
    No preview · Article · Nov 1999 · Physics in Medicine and Biology
  • [Show abstract] [Hide abstract]
    ABSTRACT: Because attenuation gradually decreases reconstructed counts from the periphery to the center portion of the object, the deep region far from the detector is not clearly observed if attenuation is not compensated. In brain perfusion SPECT, diagnosis is sometimes made using filtered back projection images without attenuation compensation (FBP). Brain perfusion SPECT has the unique characteristic that the radiopharmaceutical accumulates only within the brain and is never taken up by the surrounding skull. This study investigated the effect of skull on brain perfusion SPECT reconstructed with FBP. We theoretically derived the relation between the counts of brain and the linear attenuation coefficient of skull. It was found that the difference in reconstructed counts between the deep gray matter and peripheral gray matter decreased due to the existence of the skull. This result indicated that the deep gray matter was inclined to be visible if the FBP images were displayed according to relative counts normalized to the maximum count of each image. In order to confirm this, we made a numerical phantom with realistic human brain and skull contours on the basis of MR images from a normal volunteer. The linear attenuation coefficient of brain was assumed to be 0.15 cm-1, while that of skull was assumed to be 0.26 cm-1 (denoted as BONE+) or 0 cm-1 (BONE-). In accordance with the theoretical results, the deep gray matter of BONE+ images was more clearly observed than that of BONE- images, if these were displayed using the relative counts of each images. The physical phantom experiments also supported the theoretical and numerical phantom studies.
    No preview · Article · Jan 1999 · Nihon Igaku Hoshasen Gakkai zasshi. Nippon acta radiologica
  • [Show abstract] [Hide abstract]
    ABSTRACT: There are two possible ways to obtain scatter-corrected images with the ML-EM (maximum likelihood expectation maximization) algorithm: one is the subtraction of scatter estimate si from projection data pi, and then (pi-si) is used for scatter-corrected projection data (denoted as SC(T)); the other method is the addition of scatter estimate si to the projections calculated from the reconstructed image without performing data subtraction (SC(E)). This paper investigated these two ML-EM algorithms of combined scatter and attenuation correction on 201Tl myocardial perfusion SPECT imaging. Scatter windows were placed one full width at half maximum (FWHM) below and above the photopeak centerline. The scatter fraction in the primary peak was estimated using trapezoidal approximation by the triple energy window method. Phantom and clinical images were reconstructed using 6 iterations of ordered subsets EM algorithm (OS-EM). A cylindrical phantom with a cold-rod insert and a heart/thorax phantom with liver insert were used to evaluate scatter and the attenuation compensation technique. A cylindrical phantom filled with uniform 201Tl solution was used to evaluate statistical noise. The percent root-mean-square uncertainty (%RMSU) was used as a quantitative measure of noise amplification. %RMSU showed that the SC(E) method amplified noise less in comparison with the SC(T) method, however, no significant difference in image quality was observed between the two methods. In conclusion, both the SC(T) and SC(E) methods provided significant and similar improvement in the removal of scatter in 201Tl myocardial perfusion SPECT imaging.
    No preview · Article · Dec 1998 · Nihon Igaku Hoshasen Gakkai zasshi. Nippon acta radiologica