Nihon Hoshasen Gijutsu Gakkai zasshi最新文献

Purpose: The purpose of this study is to evaluate the detection accuracy of a high-definition optical surface imaging (OSI) system for non-coplanar radiotherapy (by rotating a phantom instead of a couch rotation).

Methods: The constancy, reproducibility, and accuracy of the positioning of the OSI system, Catalyst HD (CHD), for non-coplanar treatment were examined by rotating the head phantom around the isocenter. For all the tests, the phantom was rotated by ±30°, ±45°, ±60°, ±90° after correction of the phantom position within 0.0 mm±0.2 mm, and 0.0°±0.1° using Cone Beam CT (CBCT); the CBCT images were acquired again after rotation. We compared the phantom position derived from CHD, translational displacements of the isocenter (Dev.), and rotational displacements (Rot.) to the position derived from CBCT. The constancy of monitoring was evaluated by observing the variation in the isocenter position for 30 min. For evaluating reproducibility, the positions derived from CHD were compared with those from the planning data. The accuracy of positioning was evaluated by comparing CHD and CBCT findings after the couch rotation of ±0.5°.

Results: The constancy test revealed a maximum Rot. of 0.02±0.01° and Dev. of 0.20±0.08 mm, and the reproducibility test showed a maximum Rot. of 0.26±0.15° and Dev. of 0.93±0.26 mm. In the accuracy tests, when the phantom was further rotated by +0.5°, the maximum values were Rot. of 0.73±0.05° and Dev. of 0.35±0.15 mm; at -0.5°, the values were Rot. of -0.37±0.34° and Dev. of 0.43±0.24 mm.

Conclusion: A high-resolution OSI system is useful for position detection during treatment, even in non-coplanar irradiation.

Purpose: This study was to investigate the cases of ferromagnetic objects brought into the MRI room by using incident reports.

Methods: We included incident reports on ferromagnetic objects brought into the MRI room over the past 10 years from January 2012 to December 2021. We investigated the incidence rate of the ferromagnetic objects into the MRI room, the cause for bringing ferromagnetic objects, the number of years of MRI experience, the time of occurrence, and the names of ferromagnetic objects.

Results: There were 248 incident reports, including 26 cases related to the ferromagnetic objects. The frequency of occurrence shows that the highest number of cases occurred within the first year of experience, accounting for approximately half of the case related to MRI staff. For newcomers, there are more incidents in the second quarter than in other quarters.

Conclusion: The number of cases of bringing in ferromagnetic objects was higher in the second quarter as there was less experience with MRI.

Purpose: Cone beam computed tomography (CBCT) is the most commonly used technique for target localization in radiation therapy. Four-dimensional CBCT (4D CBCT) is valuable for localizing tumors in the lung and liver regions, where the localization accuracy is affected by respiratory motions. However, in image-guided radiation therapy for organs subject to respiratory motion, position verification is often performed using 3D cone beam CT or 2D X-ray images. While it is possible to collimate tumors at specific respiratory phases during breath-holding and to determine the tumor's motion range by taking inspiratory and expiratory breath-hold images, it remains difficult to track the tumor's trajectory at each respiratory phase. The aim of this study is to investigate the positional phases of targets that move with respiration using phantom experiments with 4D CT and 4D CBCT.

Methods: To simulate respiratory motion, we captured images of a moving phantom with a simulated tumor synchronized to simulated breathing using 4D CT and 4D CBCT. The simulated tumor was set to have respiratory cycles of 3, 4, 5, and 7.5 s, with displacements 20, 16, 10, 8, and 4 mm per breath. Under these conditions, 4D CT and 4D CBCT images were captured. Using the treatment planning system, regions of interest for the simulated tumors were set from the obtained images of each respiratory phase, identifying the tumor and setting the region as the target. Volume, positional error, and Dice coefficient of the target centroid in the corresponding phase images of 4D CT and 4D CBCT were measured with the treatment planning system.

Results: The positional error of the target centroid between 4D CT and 4D CBCT was generally within ±1 mm. The Dice coefficient for each respiratory phase under each condition of 4D CT and 4D CBCT was generally above 0.8.

Conclusion: It has been suggested that 4D CBCT has the same detection ability as 4D CT for targets with respiratory movement.

Purpose: In single-photon emission computed tomography (SPECT), the standardized uptake value requires a becquerel calibration factor (BCF). The changes over time in BCF due to different radionuclides and collimators was examined.

Methods: The BCF (cps/MBq) was monthly calculated from the radioactivity of syringe formulations and dispensed sources measured with a dose calibrator, and planar acquisition counts. In addition, relative errors with respect to the maximum BCF over 44 months were calculated.

Results: The average BCF was 46.0 for ¹²³I with a low-penetration high-resolution collimator (¹²³I-LPHR) and 124.1 for ¹³¹I with a medium-energy low-penetration collimator (¹³¹I-MELP). The standard deviation of the BCF for any radionuclides and collimators was less than 3.4 and the differences between detectors were small. Relative errors of the BCF were less than 10% for ^99mTc with a low-energy high-resolution collimator (^99mTc-LEHR), ¹²³I-MELP, ⁶⁷Ga-MELP, and ¹³¹I-MELP, and less than 5% for ¹²³I-MELP. Relative errors for ¹²³I-LEHR and ¹²³I-LPHR were initially slightly higher but decreased to less than 10% after 7 months.

Conclusion: The BCF measured by planar acquisition were stable and reproducible over time with a wide variety of nuclides and collimators.

Purpose: There are attempts to assess tumor heterogeneity by texture analysis. However, the ordered subsets-expectation maximization (OSEM) reconstruction method has problems depicting heterogeneities. The aim of this study was to identify image reconstruction parameters that improve the ability to depict internal tumor necrosis using a self-made phantom that simulates internal necrosis.

Methods: Self-made phantoms were prepared using polypropylene cylinders with inner diameters of 18.0 mm and 6.0 mm. The concentration ratios of the simulated tumor : tumor interior were 4 : 0 and 4 : 1. For each reconstruction method, the iteration for OSEM and OSEM+point spread function (PSF) were 1 to 25 and the subset was 12. The β values for block sequential regularized expectation maximization (BSREM) were set between 10 and 400. We evaluated the features of the profile curve, contrast-to-noise ratio, and grey-level co-occurrence matrix (GLCM).

Results: In the phantom study, OSEM and OSEM+PSF showed a better delineation of the differences between the inside and outside of the cylinder as iteration was increased and BSREM showed a better delineation as β was decreased. The highest value for each feature, both 4 : 0 and 4 : 1, was BSREM β 10 for angular second moment (ASM) and inverse differential moment (IDM), OSEM iteration 25 for contrast and entropy.

Conclusion: We have identified image reconstruction parameters that improve the ability to visualize internal tumor necrosis. The parameter was BRSEM β 10.

Purpose: This study developed a deep learning-based artificial intelligence (AI) observer to address the shortage of skilled human observers and evaluated the impact of substituting human observers with AI.

Methods: We used a CT system (Aquilion Prime SP; Canon Medical Systems, Tochigi) and modules CTP682 and CTP712 to scan the phantom (Catphan 700; Toyo Medic, Tokyo). The imaging conditions were set to a tube voltage of 120 kV and tube currents of 200, 160, 120, 80, 40, and 20 mA. Each condition was scanned twice, resulting in a total of 24 images. After the paired comparison experiment with 5 observers, deep learning models based on VGG19 and VGG16 were trained. We evaluated the variance, including both human and AI observers, and examined the impact of replacing humans with AI on the average degree of preference and statistical significance. These evaluations were conducted both when the training and assessments were from the same module and when they were from different modules.

Results: Variance ranged from 0.085 to 0.177 (mean: 0.124). Despite using different modules for training and evaluation, the variance remained consistent, indicating that the results are independent of the training data. The average degree of preference and image rankings were nearly identical. Between 200 mA and 160 mA, AI results differed from human results in terms of statistical significance, though the difference was minimal. The discrepancy arose from differences in observations between humans and AI, yet it fell within the expected range of variation typically observed among human observers.

Conclusion: Our results suggest that replacing human observers with AI has a minimal impact and may help alleviate observer shortages. The main limitation is the inability to modify evaluation criteria or stages with the trained models.

Purpose: The purpose of this study was to propose a method for measuring the maximum leaf velocity (V_max) of the multileaf collimator (MLC) in a dynamic MLC irradiation.

Methods: The irradiation was carried out with a plan in which the MLC leaves were constantly and gradually accelerated. Based on this plan, it was assumed that the velocity of each leaf v(t) (t is the elapsed time) would initially increase but plateau once it reached its maximum velocity. In the proposed method, v(t) was calculated from the log file data during irradiation, and fitted by a piecewise linear function consisting of 2 linear segments (constant acceleration and constant velocity segments); V_max was determined as the velocity in the constant velocity segments. The V_max values in each accelerator were obtained periodically for 7 months (20 measurements in total).

Results: In all measurements, the constant acceleration and constant velocity segments in v(t) were clearly distinguished by the piecewise linear approximation, and the V_max was determined. The mean V_max value of each leaf ranged from 3.63 to 4.32 cm/s with standard deviations (SD) less than 0.04 cm/s.

Conclusion: The proposed method made it possible to confirm the long-term stability of the V_max easily.

Purpose: StarGuide (GE HealthCare, Haifa, Israel) is a full-ring SPECT/CT system based on Cadmium Zinc Telluride (CZT) technology. In this study, we aimed to compare the image quality of this CZT-based SPECT/CT to a conventional Anger-type SPECT/CT system (NM/CT 870 DR, 870DR; GE HealthCare).

Methods: Tomographic sensitivity was calculated by recording the total number of counts detected during tomographic acquisition for a point source. We evaluated spatial resolution and image uniformity on each system using the full width half maximum (FWHM) of line sources and root mean square uniformity (%RMSU) of pool phantom, respectively. The voxel size of the StarGuide SPECT images was 2.46×2.46×2.46 mm³, compared to 4.42×4.42×4.42 mm³ on 870DR. These projection data were reconstructed using 3D-OSEM with a resolution recovery technique (RR). We compared 3 different algorithms: non-correction (NCRR), scatter correction (SCRR), and attenuation correction and scatter correction (ACSCRR).

Results: Tomographic sensitivity of StarGuide and 870DR were estimated at 200.0 counts・s^-1・MBq^-1 and 193.3 counts・s^-1・MBq^-1, respectively. Spatial resolution at the center of the FOV was estimated at 2.6 mm for StarGuide and 5.4 mm for 870DR with ACSCRR. Likewise, the %RMSU was 21.7 for StarGuide and 24.6 for 870DR.

Conclusion: The full-ring CZT SPECT/CT system has a superior spatial resolution and better image uniformity than the conventional Anger-type SPECT instrument, whereas tomographic sensitivity remains similar.