Acquisition time reduction in pediatric 99m Tc-DMSA planar imaging using deep learning.

PURPOSE: Given the potential risk of motion artifacts, acquisition time reduction is desirable in pediatric 99m Tc-dimercaptosuccinic acid (DMSA) scintigraphy. The aim of this study was to evaluate the performance of predicted full-acquisition-time images from short-acquisition-time pediatric 99m Tc-DMSA planar images with only 1/5th acquisition time using deep learning in terms of image quality and quantitative renal uptake measurement accuracy. METHODS: One hundred and fifty-five cases that underwent pediatric 99m Tc-DMSA planar imaging as dynamic data for 10 min were retrospectively collected for the development of three deep learning models (DnCNN, Win5RB, and ResUnet), and the generation of full-time images from short-time images. We used the normalized mean squared error (NMSE), peak signal-to-noise ratio (PSNR), and structural similarity index metrics (SSIM) to evaluate the accuracy of the predicted full-time images. In addition, the renal uptake of 99m Tc-DMSA was calculated, and the difference in renal uptake from the reference full-time images was assessed using scatter plots with Pearson correlation and Bland-Altman plots. RESULTS: The predicted full-time images from the deep learning models showed a significant improvement in image quality compared to the short-time images with respect to the reference full-time images. In particular, the predicted full-time images obtained by ResUnet showed the lowest NMSE (0.4 [0.4-0.5] %) and the highest PSNR (55.4 [54.7-56.1] dB) and SSIM (0.997 [0.995-0.997]). For renal uptake, an extremely high correlation was achieved in all short-time and three predicted full-time images (R2  > 0.999 for all). The Bland-Altman plots showed the lowest bias (-0.10) of renal uptake in ResUnet, while short-time images showed the lowest variance (95% confidence interval: -0.14, 0.45) of renal uptake. CONCLUSIONS: Our proposed method is capable of producing images that are comparable to the original full-acquisition-time images, allowing for a reduction of acquisition time/injected dose in pediatric 99m Tc-DMSA planar imaging.

Optimal choice of OSEM and SD reconstruction algorithms in CZT SPECT for hypertrophic cardiomyopathy patients.

BACKGROUND: The spectrum dynamics (SD) algorithm is a cardiac reconstruction algorithm of D-SPECT, which improves spatial resolution compared with the ordered-subsets expectation maximization (OSEM) algorithm. We evaluated the wall thickness and left ventricular (LV) volume in patients with hypertrophic cardiomyopathy (HCM) using the SD algorithm. METHODS: In a phantom study, the myocardial wall was scanned with varying wall thicknesses (10-40 mm). In the clinical study, 30 and 27 normal and HCM patients underwent myocardial perfusion imaging. RESULTS: In a phantom study, LV volume using the SD algorithm was increased by thickening the wall of the phantom. In the clinical study, the wall thickness and LV volume of OSEM and SD algorithms showed a difference between HCM and normal groups. The wall thickness using OSEM and SD algorithms were 19.4 ± 2.0 and 16.7 ± 1.5 mm in patients with normal, and 27.9 ± 4.9 and 21.8 ± 2.6 mm in patients with HCM. CONCLUSION: The SD algorithm in cases of HCM may not be able to correctly assess wall thickness and LV volume. Our study suggests that the OSEM is more suitable in cases of HCM than the SD algorithm.

[A Nationwide Survey on Additional Scan in Nuclear Medicine Imaging].

The aim of the present study was to clarify the routine protocols and the frequency of added or omitted imaging on nuclear medicine imaging in Japan. A nationwide survey on routine protocols and current state of added or omitted imaging in major nuclear medicine imaging were performed for Japanese nuclear medicine technologist. The survey showed that the routine protocols were almost 100% fixed, some of the routine protocols were found to be useful and percentage of imaging techniques such as single photon emission computed tomography/computed tomography that increased patient burden and reduced through put were low. Furthermore, the survey showed that additional or omission imaging were frequently performed on bone scintigraphy and positron emission tomography and added or omitted judgements were often depend upon the rule of thumb by nuclear medicine technologist. In this study, we have concluded that the quality of examination and the diagnosis might depend on the knowledge of nuclear medicine technologist, performed added or omitted imaging.

[Development of Body Phantom for Evaluation of Appropriate Administered Radioactivities and Image Quality on 99mTc-DMSA Scintigraphy in Pediatric Nuclear Medicine].

PURPOSE: We conducted a field survey about pediatric nuclear medicine. As a result, it was suggested that 99mTc-DMSA scintigraphy was performed at many institutions, whereas various examinations such as image acquisition and processing are not carried out using the renal phantom. Therefore, we developed the body phantom for the evaluation of appropriate administered radioactivities and image quality with renal scintigraphy in pediatric nuclear medicine. METHODS: We created three differently sized body phantoms (1-, 5-, and 20-year-old models). These pediatric body phantoms were filled with a 99mTc solution based on the consensus guideline of pediatric radiopharmaceutical administered radioactivity in Japan. The planar image was evaluated using acquisition count, uniformity and defect contrast. SPECT images were evaluated with a recovery coefficient (RC). RESULTS: The acquisition counts for pediatric body phantoms were relatively corresponded to the clinical study. The appropriate acquisition counts and the pixel size for the planar image were approximately 140 counts per pixel and 1.23-1.35 mm at 5 min acquisition times in 1- and 5-year-old pediatric body phantom studies, respectively. Although the uniformity and the cold contrast did not depend on pixel size and body size, the cold contrast was affected by body size. The RC for SPECT images depended on the performance of SPECT systems, the resolution recovery algorithm and body phantom size. CONCLUSION: The developed pediatric body phantom could allow us to establish optimal image acquisition and more evidence on renal scintigraphy in pediatric nuclear medicine.

Effect of Reconstruction Strategies for the Quantification and Diagnostic Accuracy of (123)I-FP-CIT SPECT.

PURPOSE: This study evaluates the effect of reconstruction strategies for the quantification and diagnostic accuracy of (123)I-FP-CIT SPECT. METHODS: We evaluated the quantification of (123)I-FP-CIT SPECT obtained by several combinations of reconstruction using the striatal phantom. The phantom images were reconstructed using FBP and OSEM with/without attenuation correction (AC) and scatter correction (SC). We calculated the specific binding ratio (SBR) using volume of interest (VOI) analysis on each reconstructed images. For the clinical study, 40 patients who underwent (123)I-FP-CIT SPECT were selected. We grouped the patients into the normal binding group and decreased binding group according to their clinical diagnosis. The clinical images were reconstructed under the same conditions as the phantom study. The SBRs were calculated, and a receiver operating characteristic (ROC) analysis was performed to evaluate the diagnostic accuracy. RESULTS: The SBRs with AC and SC significantly increased compared with no corrections. In the clinical study, although ROC analysis showed no significant difference in the all combinations of reconstruction, the area under the curve using SC and AC tended to be higher than that obtained by other reconstruction. CONCLUSIONS: Quantification of (123)I-FP-CIT SPECT was affected by reconstruction strategies. In addition, both the AC and SC improved the diagnostic accuracy of (123)I-FP-CIT SPECT. Our results suggest that both the AC and SC are recommended for the improving the quantification and diagnostic accuracy in (123)I-FP-CIT SPECT.

[Evaluation of Resolution Correction in Single Photon Emission Computed Tomography Reconstruction Method Using a Body Phantom: Study of Three Different Models].

PURPOSE: The aim of this study was to evaluate the resolution recovery techniques of Flash3D, Astonish, and Evolution in single photon emission computed tomography(SPECT) using a body phantom. METHODS: We scanned a National Electrical Manufactures Association body phantom filled with 99mTc. The body of the phantom with radioactive sphere and background was filled with either water or radioactive solution. We investigated image quality using profile curves, recovery coefficient, and image contrast. RESULTS: The profile curve at the edge of the hot sphere showed artifact due to Gibbs oscillation for all techniques, and also over estimation of recovery coefficient was seen in the hot sphere, as had been previously reported in a simulation study. These phenomena were more remarkable than Evolution in the Flash3D and Astonish techniques. For the contrast between hot sphere and background, the contrast recover was enough for the <17-mm hot spheres. These results showed that the effect of contrast correction was less as the radius of rotation diameter became large. CONCLUSION: In the present study using the body phantom, overestimated counts and edge artifacts due to Gibbs oscillation were shown. These phenomena were different by each resolution correction algorithms. Also, there were limitation regarding image quality improvement by resolution correction depending on sphere size and length of radius of rotation.

[Validation of noise reduction in iterative reconstruction: a simulation phantom study].

PURPOSE: In the iterative reconstruction method, image noise tends to increase in proportion to falling available photon count and increasing update number. Image filtering is an important factor in single photon emission computed tomography (SPECT) image reconstruction, but it is frequently treated in a subjective way. The aim of this study was to evaluate the effects of pre-reconstruction filtering and post-reconstruction filtering on the iterative reconstruction process. METHODS: Using simulation phantoms, projection data were reconstructed using ordered subsets expectation maximization (OSEM) with or without compensation for resolution recovery. Pre- and post-reconstruction filtering was performed using a Butterworth filter (BW) (range: 0.3-1.3 cycles/cm) and a Gaussian filter (GA) (range: 0.3-1.3 mm) with various parameters. We evaluated the variances of full width at half maximum (FWHM), coefficients of variation (CV), image contrast and normalized mean squared error (NMSE) values. RESULTS: The FWHM values for pre-reconstruction filtering tended to be lower than those observed for post-reconstruction filtering. These values were 5.1 mm (pre-reconstruction) and 6.7 mm (post-reconstruction). The CV on pre- and post-reconstruction filtering was 7.5% and 11.6%. Pre-reconstruction filtering reduced image noise more effectively than post-reconstruction filtering. The contrast for pre-reconstruction filtering was similar to that observed after post-reconstruction filtering. However, contrast after filtering with a GA slowly decreased as compared to the BW. NMSE values obtained by pre-reconstruction filtering tended to be lower than those observed for post-reconstruction filtering. CONCLUSIONS: Pre-reconstruction filtering provided SPECT image quality comparable to that from post-reconstruction filtering, especially when using the BW. Our results suggest that pre-reconstruction filtering is a beneficial method when applied to the iterative reconstruction method with or without compensation for resolution recovery.

[Fundamental evaluation of wavelet transform based noise reduction using soft threshold method in single photon emission computed tomography image].

PURPOSE: The wavelet transform is a newly developed signal-processing tool that decomposes a signal into various levels of resolution. The wavelet transform based noise reduction has the characteristics of optimally separating signal from noise, preserving the rapid rises and falls of a signal, and reconstructing a smooth signal from noise-imposed observations. The aim of this study was to evaluate the effects of applying a new noise reduction technique, the wavelet transform based noise reduction, to single photon emission computed tomography (SPECT) images. METHODS: Three experiments were performed using cylindrical phantom, line source, and hot-rod phantom, respectively. We acquired SPECT image datasets of each phantom, and reconstructed SPECT images using the wavelet transform based noise reduction with filter back projection (FBP). Images were de-noised by 3 parameters of wavelet transform based noise reduction: 1st wavelet weight (WW), 2nd WW, and 3rd WW, respectively. We evaluated the variances of full width at half maximum (FWHM), coefficients of variation (%CV), and frequency domains (radius direction distribution function in the power spectrum), respectively. RESULTS: In the cylindrical phantom test, %CV was reduced from 27.92% to 15.38% using the wavelet approach. On the other hand, FWHM values showed no significant change. However, the increases of wavelet weights caused artifacts on the reconstructed images in some cases. CONCLUSIONS: The wavelet based noise reduction had the significant potential to improve SPECT image. Therefore, the wavelet method should prove to be a robust approach to improve image quantification and fidelity.

Feasibility of using counts-per-volume approach with a new SPECT phantom to optimize the relationship between administered dose and acquisition time.

We developed a phantom for single-photon emission computed tomography (SPECT), with the objective of assessing image quality to optimize administered dose and acquisition time. We investigated whether the concept of counts-per-volume (CPV), which is used as a predictor of visual image quality in positron emission tomography, can be used to estimate the acquisition time required for each SPECT image. QIRE phantoms for the head (QIRE-h) and torso (QIRE-t) were developed to measure four physical indicators of image quality in a single scan: uniformity, contrast of both hot and defective lesions with respect to the background, and linearity between radioactivity concentration and count density. The target organ's CPV (TCPV), sharpness index (SI), and contrast-to-noise ratio (CNR) were measured for QIRE-h and QIRE-t phantoms, and for anthropomorphic brain and torso phantoms. The SPECT image quality of the four phantoms was visually assessed on a 5-point scale. The acquisition time and TCPV were correlated for all four phantoms. The SI and CNR values were nearly identical for the QIRE and anthropomorphic phantoms with comparable TCPV. The agreement between the visual scores of QIRE-h and brain phantoms, as well as QIRE-t and torso phantoms, was moderate and substantial, respectively. Comparison of SPECT image quality between QIRE and anthropomorphic phantoms revealed close agreement in terms of physical indicators and visual assessments. Therefore, the TCPV concept can also be applied to SPECT images of QIRE phantoms, and optimization of imaging parameters for nuclear medicine examinations may be possible using QIRE phantoms alone.

Feasibility of ultra-high-speed acquisition in xSPECT bone algorithm: a phantom study with advanced bone SPECT-specific phantom.

OBJECTIVE: Although xSPECT Bone (xB) provides quantitative single-photon emission computed tomography (SPECT) high-resolution images, patients' burden remains high due to long acquisition time; therefore, this study aimed to investigate the feasibility of shortening the xB acquisition time using a custom-designed phantom. METHODS: A custom-designed xSPECT bone-specific (xSB) phantom with simulated cortical and spongious bones was developed based on the thoracic bone phantom. Both standard- and ultra-high-speed (UHS) xB acquisitions were performed in a male patient with lung cancer. In this phantom study, SPECT was acquired for 3, 6, 9, 12, and 30 min. The clinical SPECT acquisition time per rotation was 9 and 3 min for standard and UHS, respectively. SPECT images were reconstructed using ordered subset expectation maximization with three-dimensional resolution recovery (Flash3D; F3D) and xB algorithms. Quantitative SPECT value (QSV) and coefficient of variation (CV) were measured using the volume of interests (VOIs) placed at the center of the vertebral body and hot sphere. A linear profile was plotted on the spinous process at the center of the xSB phantom; then, the full width at half maximum (FWHM) was measured. The standardized uptake value (SUV) and standard deviation from the first thoracic to the fifth lumbar vertebrae in clinical standard- and UHS-xB images were measured using a 1-cm3 VOI. RESULTS: The QSV of F3D images was underestimated even in large regions, whereas those of xB images were close to actual radioactivity concentration. The CV was similar or lower for xB images than that for F3D images but was not decreased with increasing acquisition time for both reconstruction images. The FWHM of xB images was lower than those of F3D images at all acquisition times. The mean SUV values from the first thoracic to fifth lumbar vertebrae for standard- and UHS-xB images were 6.73 ± 0.64 and 6.19 ± 0.87, respectively, showing a strong positive correlation. CONCLUSIONS: Results of this phantom study suggest that xB imaging can be obtained in only one-third of the acquisition time without compromising the image quality. The SUV of UHS-xB images can be similar to that of standard-xB images in terms of clinical interpretation.

[Restructuring Torso CT Protocol for Radiation Dose Management].

PURPOSE: There are problems with dose management in X-ray computed tomography (CT) because the protocol used for any examination is not always in the same scan range. The purpose of this study was to investigate the usefulness of setting the CT protocol based on the scan range. METHODS: We evaluated the examination data of patients who underwent plain CT based on a scan range of chest to pelvis and abdomen to pelvis. The previous protocol [Chest-Abdomen Routine] was changed to the current protocols [Chest_Abdomen] and [Chest_Pelvis], and the previous protocol of [Abdomen Routine] was changed to the current protocols [Abdomen] and [Abdomen_Pelvis]. Examination data of height, scan length, volume CT dose index (CTDIvol), and dose length product (DLP) were obtained from digital imaging and communications in medicine, and radiation dose structured report using Radimetrics. The relationship between patient height and scan range, and CTDIvol and DLP was indicated in a scatter plot. Standard deviation (SD) of scan length and DLP were compared between current and previous protocols. Outliers were defined as the data exceeding average ±2SD. RESULTS: The SD of scan length decreased by 77.1% on abdomen to pelvis, and the SD of DLP decreased by 65.2% on abdomen to pelvis. The causes of the outliers were CT scan range, scan parameter, arm position, metal implants, and body thickness of patients. CONCLUSION: Setting CT protocols based on the scan range reduced SD of scan length and DLP. It was helpful for reducing the number of scan range outliers and analyzing the cause of outliers.

The setting of heartbeat acceptance windows on gated myocardial perfusion single-photon emission computed tomography using CZT camera: effect of left ventricular functional parameters in patients with arrhythmia.

OBJECTIVES: Consistently variable with several peaks in heart rate histogram (e.g. bigeminy and trigeminy) is one of the arrhythmia types. We investigated the effects of gating error by consistently variable patients on left ventricular (LV) functional with cadmium zinc telluride (CZT) camera. The purpose of this study is to evaluate LV functional parameters by setting different heartbeat acceptance windows on gated myocardial perfusion single-photon emission computed tomography (SPECT) (MPS) in consistently variable patients, using echocardiography (echo) as a reference. MATERIALS AND METHODS: Sixteen consistently variable patients underwent the gated MPS using a D-SPECT. The MPS images were obtained by setting two different types of heartbeat acceptance windows. The heartbeat acceptance windows were set to include only one peak and two peaks of the maximum count peaks, respectively. RESULTS: Mean end-diastolic volume, end-systolic volume and left ventricular ejection fraction (LVEF) were 112.7 mL, 62.2 mL and 51.3% for one peak, 114.5 mL, 66.1 mL and 47.1% for two peak, and 113.0 mL, 54.2 mL and 54.1% for echo, respectively. The mean differences between two peaks and echo in LVEF were larger than those of between one peak and echo. CONCLUSION: Our study suggests that setting the heartbeat acceptance window of one peak was suitable for accurate measurement of LV function in consistently variable patients.

Current state of oncologic 18F-FDG PET/CT in Japan: A nationwide survey.

OBJECTIVES: Combined positron emission tomography/computed tomography (PET/CT) has gradually advanced with the introduction of newly developed techniques. However, the recent status of imaging techniques (e.g., scanning range, availability of correction methods, and decisions on performing delayed scan) in oncologic PET/CT with 18F-fluorodeoxyglucose (18F-FDG) in Japan is unclear. We conducted a nationwide cross-sectional survey to document 18F-FDG PET/CT protocols and clarify the recent status of imaging techniques for oncologic 18F-FDG PET/CT in Japan. METHODS: We conducted a web survey hosted by the Japanese Society of Radiological Technology between October and December 2017. The questionnaire included nine items on the demographics of the respondents, their scan protocols, and additional imaging to their routine protocols. RESULTS: We received responses from 119 Japanese technologists who performed 18F-FDG PET/CT in practice. Almost all the respondents stated that the scanning range was from the top of the head to the pelvis or mid-thigh region. Newly developed techniques were used by fewer than half of the respondents. Most respondents performed additional imaging in consultation with physicians, such as delayed imaging (83%) or an extended scanning range for early imaging (55%). CONCLUSIONS: Our survey helps in clarifying the recent state of oncologic 18F-FDG PET/CT imaging techniques in Japan. Given that 18F-FDG PET/CT practices most frequently performed additional imaging along with their routine scan protocol, the practice constitutes the most varied examination performed in Japanese nuclear medicine.

[Comparison of FBP and ML-EM reconstruction used by simulation data; effect of the projection number for the SPECT image].

PURPOSE: The influence of the numbers of projection of SPECT exerting on a re-constructed image cannot be strictly evaluated by phantom studies. Therefore, we compared re-constructed images of the FBP method and the ML-EM method by using simulation data. METHODS: Simulation data was entered in the image processing software, and the projection data that changed the numbers of projection was made. Afterwards, reconstructed images of the FBP and the ML-EM methods were compared with respect to contrast, %COV, and the NMSE value. RESULT: When the numbers of projection of the FBP and the ML-EM method were decreased, all of the contrast, %COV, and the NMSE value were more deteriorated than that of the ideal image. Therefore, the image quality of SPECT improves with both FBP and ML-EM methods when there are many numbers of projection. Moreover, the FBP method was excellent in a cold contrast, and the ML-EM method was uniformly excellent. Therefore, an understanding of features and their inspection are effective for the selection of each image reconstruction method.

Current state of bone scintigraphy protocols and practice in Japan.

OBJECTIVES: Nuclear medicine technologists in Japan often perform additional single-photon emission computed tomography (SPECT) with or without computed tomography (CT) after whole-body imaging for bone scintigraphy. In this study, we wanted to identify the bone scanning protocols used in Japan, together with the current clinical practices. METHODS: The study was conducted between October and December 2017. We created a web survey that was hosted by the Japanese Society of Radiological Technology. The questionnaire included 12 items regarding the demographics of the responders, their scan protocols, and the imaging added to, or omitted from, routine protocols. RESULTS: In total, 228 eligible responses were collected from participants with a mean of 11.6±8.4 years' experience in nuclear medicine examination. All responders reported using routine scan protocols that included whole-body imaging. However, only 2%, 4%, 20%, and 14% of the responders also acquired single-field SPECT, single-field SPECT/CT, multi-field SPECT, and multi-field SPECT/CT, respectively. CONCLUSION: Our survey results indicate that nuclear medicine practice in Japan is beginning to shift from planar whole-body imaging with additional spot planar images to additional SPECT or SPECT/CT. Further study is required to examine the optimal protocols for bone scintigraphy.

Imaging technology for myocardial perfusion single-photon emission computed tomography 2018 in Japan.

AIM: Recently, nuclear cardiology has dramatically advanced by a new technology development such as the device, short-term acquisition system, image reconstruction algorithm and image analysis. Although these innovations have been gradually employed in routine examinations, we did not investigate the current use of image acquisition, image reconstruction, and image analysis with myocardial perfusion single-photon emission computed tomography (MPS). We investigated the current status of MPS imaging technology in Japan. METHODS: We carried out a survey using a Web-based questionnaire system, the opening of which was announced via e-mail, and it was available on a website for 3 months. We collected data on the current use of MPS with 201Tl and/or 99mTc agents with respect to routine protocols, image acquisition, image reconstruction, and image analysis. RESULTS: We received responses to the Web-based questionnaire from 178 and 174 people for 99mTc and 201Tl MPS, respectively. The routine protocols of MPS of stress-rest and rest-stress MPS on 1-day protocols with 99mTc were 41.2% and 14.5%, respectively, and the rest-only scan response rate was 23.7%, whereas that of 201Tl MPS was 65.9% with stress-rest MPS, 19.0% with rest-only MPS, and 10.9% with stress-rest MPS adding a rest scan 24 h after injection. The filtered back projection (FBP) method is most commonly used image reconstruction method, yielding 70.5% for 99mTc MPS and 76.8% for 201Tl MPS, including combined FBP and ordered subset expectation maximization method. The results for no-correction (NC) images were 49.2% with 99mTc MPS and 55.2% with 201Tl MPS including the response of NC and combined attenuation correction (AC) and scatter correction (SC) (i.e., ACSC) images. The AC or ACSC images of 99mTc and 201Tl were provided by 30-40% of the institutions surveyed. CONCLUSIONS: We investigated the current status of MPS imaging technology in Japan, and found that although the use of various technical developments has been reported, some of these technologies have not been utilized effectively. Hence, we expect that nuclear medicine technology will be used more effectively to improve diagnosis.

[Evaluation of factor for one-point venous blood sampling method based on the causality model].

PURPOSE: One-point venous blood sampling method (Mimura, et al.) can evaluate the rCBF value with a high degree of accuracy. However, the method is accompanied by complexity of technique because it requires a venous blood Octanol value, and its accuracy is affected by factors of input function. Therefore, we evaluated the factors that are used for input function to determine the accuracy input function and simplify the technique. METHODS: The input function which uses the time-dependent brain count of 5 minutes, 15 minutes, and 25 minutes from administration, and the input function in which an objective variable is used as the artery octanol value to exclude the venous blood octanol value are created. Therefore, a correlation between these functions and rCBF value by the MS method is evaluated. RESULTS: Creation of a high-accuracy input function and simplification of technique are possible. The rCBF value obtained by the input function, the factor of which is a time-dependent brain count of 5 minutes from administration, and the objective variable is artery octanol value, had a high correlation with the MS method (y=0.899x+4.653, r=0.842).

Validation of the CT iterative reconstruction technique for low-dose CT attenuation correction for improving the quality of PET images in an obesity-simulating body phantom and clinical study.

OBJECTIVE: The aim of this study was to validate the efficacy of computed tomography (CT) iterative reconstruction (CT-IR) for low-dose CT attenuation correction in terms of the estimation of attenuation coefficient and quality of PET images. MATERIALS AND METHODS: We used normal and obesity-simulating body phantoms. PET images were reconstructed using two attenuation correction maps obtained using filtered back projection (CT-FBP) and CT-IR. The CT numbers, attenuation coefficients, contrast-to-noise ratio (CNR10 mm), and coefficient of variation were evaluated. Fifty-two consecutive patients who underwent F-FDG PET/CT with low-dose CT scans were selected for the clinical study. Clinical PET images were reconstructed using CT-FBP and CT-IR, and the effects of CT-IR were examined according to the maximum standardized uptake value (SUVmax), contrast-to-noise ratio in the tumor (CNRtumor), and signal-to-noise ratio in the liver (SNRliver). RESULTS: The CT number on the CT-IR was significantly lower than that of CT-FBP in the obesity-simulating body phantom. The decrease in attenuation coefficients obtained using CT-IR was smaller than that obtained using CT-FBP. The CNR10 mm and coefficient of variation obtained using CT-IR were superior to those obtained using CT-FBP. The SUVmax was not significantly different between the CT-FBP and CT-IR. Although the difference in the SNRliver between the CT-FBP and CT-IR was not significant, the CNRtumor of the CT-IR was significantly higher than that obtained using CT-FBP in obese patients. CONCLUSION: We demonstrated that CT-IR improved the estimation of the attenuation coefficient and provided significant improvement in the CNR of the clinical PET images.