Iodine contrast volume reduction in preoperative transcatheter aortic valve implantation computed tomography: Comparison with 64- and 256-multidetector row computed tomography.

INTRODUCTION: This study aimed to compare the vascular enhancement and radiation dose in preoperative transcatheter aortic valve implantation (TAVI) computed tomography (CT) with a reduced contrast medium (CM) using volume scans in 256-multidetector row CT (MDCT) with a standard CM using 64-MDCT. METHODS: This study included 78 patients with preoperative TAVI CT with either 64- or 256-MDCT. The CM was injected at 1.5 mL/kg in the 64-MDCT group and 1.0 mL/kg in the 256-MDCT group. We compared vascular enhancement of the aortic root and access routes, image quality (IQ) scores, and radiation dose in both groups. RESULTS: Despite the reduced CM (by 33 %) in the 256-MDCT group, the mean vascular enhancement of the right and left subclavian arteries was significantly higher than that in the 64-MDCT group [284 and 267 Hounsfield units (HU) vs. 376 and 359 HU; p < 0.05]; however, no significant differences in the mean vascular enhancement in the ascending aorta, abdominal aorta at the celiac level, and bilateral common femoral arteries were observed between the two groups (p > 0.05 for all). The median IQ scores at the aortic root were higher in the 256-MDCT group than in the 64-MDCT group (3 vs. 4; p < 0.05), and those at the femoral access routes were comparable (4 vs. 4; p = 0.33). The mean effective dose was significantly reduced by 30 % in the 256-MDCT group (23.6 vs. 16.3 mSv; p < 0.05). CONCLUSION: In preoperative TAVI CT, volume scans using 256-MDCT provide comparable or better vascular enhancement and IQ with a 30 % reduction in CM and radiation dose than those using 64-MDCT. IMPLICATIONS FOR PRACTICE: Volume scan using 256-MDCT for preoperative TAVI CT may reduce CM and radiation dose in TAVI patients with renal dysfunction.

Radiation dose reduction method combining the ECG-Edit function and high helical pitch in retrospectively-gated CT angiography.

INTRODUCTION: The purpose of this study was to demonstrate that dose reduction does not compromise image quality when combining high helical pitch (HP) and the ECG-Edit function during low HP retrospectively gated computed tomography angiography (CTA). METHODS: This study made use of a pulsating cardiac phantom (ALPHA 1 VTPC). The heart rate (HR) of the cardiac phantom was changed in five intervals, every 5 beats per minute (bpm), from 40 to 60 bpm. Evaluation of a range of HR was important because data loss might occur when combining a low HR and high HP. We performed retrospectively gated CTA scans five times using a low HP (0.16) and high HP (0.24), for each of the five HR intervals, using a 64-detector row CT scanner. The CT volume dose index (CTDIvol) was recorded from the CT console of each scan. For the images with data loss, data were repaired using the ECG-Edit function. We compared the CTDIvol, estimated cardiac phantom volume, and the visualization of the coronary ladder phantom between HP 0.16, with or without repaired HP 0.24, using the ECG-Edit function. RESULTS: Data loss occurred with a HR of 40 bpm and 45 bpm when using HP 0.24. The CTDIvol was reduced by approximately 33% with HP 0.24 when compared with HP 0.16. There were no significant differences in the mean cardiac motion phantom volume and visualization scores between HP 0.16 and with and without repaired HP 0.24 using the ECG-Edit function (p < 0.05). CONCLUSION: The ECG-Edit function is potential useful for repairing the lost data in patients with a low HR, and when combined with a high HP, it is possible to reduce the radiation dose by approximately 33%. IMPLICATIONS FOR PRACTICE: The ECG-Edit function and high HP may be a viable option in pediatric CTA studies.

Virtual magnetic resonance lumbar spine images generated from computed tomography images using conditional generative adversarial networks.

INTRODUCTION: The aim of this study was to generate virtual Magnetic resonance (MR) from computed tomography (CT) using conditional generative adversarial networks (cGAN). METHODS: We selected examinations from 22 adults who obtained their CT and MR lumbar spine examinations. Overall, 4 examinations were used as test data, and 18 examinations were used as training data. A cGAN was trained to generate virtual MR images from the CT images using the corresponding MR images as targets. After training, the generated virtual MR images from test data in epochs 1, 10, 50, 100, 500, and 1000 were compared with the original ones using the mean square error (MSE) and structural similarity index (SSIM). Additionally, two radiologists also performed qualitative assessments. RESULTS: The MSE of the virtual MR images decreased as the epoch of the cGANs increased from the original CT images: 8876.7 ± 1192.9 (original CT), 1567.5 ± 433.9 (Epoch 1), 1242.4 ± 442.0 (Epoch 10), 1065.8 ± 478.1 (Epoch 50), 1276.1 ± 718.9 (Epoch 100), 1046.7 ± 488.2 (Epoch 500), and 1031.7 ± 400.0 (Epoch 1000). No considerable differences were observed in the qualitative evaluation between the virtual MR images and the original ones, except in the structure of the spinal canal. CONCLUSION: Virtual MR lumbar spine images using cGANs could be a feasible technique to generate near-MR images from CT without MR examinations for evaluation of the vertebral body and intervertebral disc. IMPLICATIONS FOR PRACTICE: Virtual MR lumbar spine images using cGANs can offer virtual CT images with sufficient quality for attenuation correction for PET or dose planning in radiotherapy.

Deep learning with convolutional neural network for estimation of the characterisation of coronary plaques: Validation using IB-IVUS.

INTRODUCTION: Deep learning approaches have shown high diagnostic performance in image classifications, such as differentiation of malignant tumors and calcified coronary plaque. However, it is unknown whether deep learning is useful for characterizing coronary plaques without the presence of calcification using coronary computed tomography angiography (CCTA). The purpose of this study was to compare the diagnostic performance of deep learning with a convolutional neural network (CNN) with that of radiologists in the estimation of coronary plaques. METHODS: We retrospectively enrolled 178 patients (191 coronary plaques) who had undergone CCTA and integrated backscatter intravascular ultrasonography (IB-IVUS) studies. IB-IVUS diagnosed 81 fibrous and 110 fatty or fibro-fatty plaques. We manually captured vascular short-axis images of the coronary plaques as Portable Network Graphics (PNG) images (150 × 150 pixels). The display window level and width were 100 and 700 Hounsfield units (HU), respectively. The deep-learning system (CNN; GoogleNet Inception v3) was trained on 153 plaques; its performance was tested on 38 plaques. The area under the curve (AUC) obtained by receiver operating characteristic analysis of the deep learning system and by two board-certified radiologists was compared. RESULTS: With the CNN, the AUC and the 95% confidence interval were 0.83 and 0.69-0.96, respectively; for radiologist 1 they were 0.61 and 0.42-0.80; for radiologist 2 they were 0.68 and 0.51-0.86, respectively. The AUC for CNN was significantly higher than for radiologists 1 (p = 0.04); for radiologist 2 it was not significantly different (p = 0.22). CONCLUSION: DL-CNN performed comparably to radiologists for discrimination between fatty and fibro-fatty plaque on CCTA images. IMPLICATIONS FOR PRACTICE: The diagnostic performance of the CNN and of two radiologists in the assessment of 191 ROIs on CT images of coronary plaques whose type corresponded with their IB-IVUS characterization was comparable.

Machine learning to identify lymph node metastasis from thyroid cancer in patients undergoing contrast-enhanced CT studies.

INTRODUCTION: We compared the diagnostic performance of morphological methods such as the major axis, the minor axis, the volume and sphericity and of machine learning with texture analysis in the identification of lymph node metastasis in patients with thyroid cancer who had undergone contrast-enhanced CT studies. METHODS: We sampled 772 lymph nodes with histology defined tissue types (84 metastatic and 688 benign lymph nodes) that were visualised on CT images of 117 patients. A support vector machine (SVM), free programming software (Python), and the scikit-learn machine learning library were used to discriminate metastatic-from benign lymph nodes. We assessed 96 texture and 4 morphological features (major axis, minor axis, volume, sphericity) that were reported useful for the differentiation between metastatic and benign lymph nodes on CT images. The area under the curve (AUC) obtained by receiver operating characteristic analysis of univariate logistic regression and SVM classifiers were calculated for the training and testing datasets. RESULTS: The AUC for all classifiers in training and testing datasets was 0.96 and 0.86, at the SVM for machine learning. When we applied conventional methods to the training and testing datasets, the AUCs were 0.63 and 0.48 for the major axis, 0.70 and 0.44 for the minor axis, 0.66 and 0.43 for the volume, and 0.69 and 0.54 for sphericity, respectively. The SVM using texture features yielded significantly higher AUCs than univariate logistic regression models using morphological features (p = 0.001). CONCLUSION: For the identification of metastatic lymph nodes from thyroid cancer on contrast-enhanced CT images, machine learning combined with texture analysis was superior to conventional diagnostic methods with the morphological parameters. IMPLICATIONS FOR PRACTICE: Our findings suggest that in patients with thyroid cancer and suspected lymph node metastasis who undergo contrast-enhanced CT studies, machine learning using texture analysis is high diagnostic value for the identification of metastatic lymph nodes.

The combined application of the contrast-to-noise index and 80 kVp for cardiac CTA scanning before atrial fibrillation ablation reduces radiation dose exposure.

INTRODUCTION: To compare the radiation dose, diagnostic accuracy, and the resultant ablation procedures using 80 and 120-kVp cardiac computed tomography angiography (CCTA) protocols with the same contrast-to-noise ratio in patients scheduled for atrial fibrillation (AF) ablation. METHODS: This retrospective study was performed following institutional review board approval. We divided 140 consecutive patients who had undergone CCTA using a 64-MDCT scanner into two equal groups. Standard deviation (SD) of the CT number was set at 25 Hounsfield units (HU) for the 120-kVp protocol. To facilitate a reduction in radiation dose it was set at 40 HU for the 80 kVp protocol. We compared the two protocols with respect to the radiation dose, the diagnostic accuracy for detecting left atrial appendage (LAA) thrombi, matching for surface registration, and the resultant ablation procedures. RESULTS: At 120 kVp, the dose length product (DLP) was 2.2 times that at 80 kVp (1269.0 vs 559.0 mGy cm, p < 0.01). The diagnostic accuracy for thrombus detection was 100% using both protocols. There was no difference between the two protocols with respect to matching for surface registration. The protocols did not differ with respect to the subsequent time required for the ablation procedures and the ablation fluoroscopy time, and the radiation dose (p = 0.54, 0.33, and 0.32, respectively). CONCLUSION: For the same CNR, the DLP at 80 kVp (559.0 mGy cm) was 56% of that delivered at 120 kVp (1269.0 mGy cm). There was no reduction in diagnostic accuracy. IMPLICATIONS FOR PRACTICE: Maintaining CNR allows for a reduction in the radiation dose without reducing the image quality.

Dual-layer detector CT of chest, abdomen, and pelvis with a one-third iodine dose: image quality, radiation dose, and optimal monoenergetic settings.

AIM: To compare the image quality and radiation dose of reduced iodine dose dual-layer detector (DL) computed tomography (CT) with those of a conventional 120 kVp protocol for chest-abdomen-pelvis CT (CAP-CT). MATERIALS AND METHODS: Forty patients with renal dysfunction (estimated glomerular filtrating ratio <45 ml/min/1.73 m2) underwent reduced iodine dose CAP-CT (120 kVp, 200 mg iodine/kg) on DLCT. Virtual monochromatic images (VMI) at 40-70 keV (5 keV interval) were reconstructed retrospectively. Forty matched patients who underwent conventional CAP-CT (120 kVp, 600 mg iodine/kg, iterative reconstruction) were included as controls. The size-specific dose estimate (SSDE), image noise, CT attenuation, and contrast-to-noise ratio (CNR) were compared between the protocols. Two radiologists rated image contrast, image noise, streak artefact, and diagnostic confidence on a five-point scale. RESULTS: The SSDE of the DLCT group was approximately 20% lower than that of the 120 kVp group (15.4±1.9 versus 19.4±2.3 mGy, p<0.01). DLCT-VMI provided almost constant image noise throughout the range of energies (differences of ≤13%), with the noise being equivalent or lower than 120 kVp in the abdomen. CT attenuation and CNR gradually increased as the energy decreased, with values comparable to 120 kVp being attained at around 45-50 keV. Although streak artefact was accentuated at 40-50 keV (p<0.01), the highest scores for diagnostic confidence were assigned at 40 and 45 keV, both of which were equivalent to 120 kVp (p=1.0). CONCLUSION: For CAP-CT with a one-third iodine dose, DLCT-VMI at 40-45 keV allows for a 20% reduction in radiation dose, while preserving image quality comparable to that of conventional 120 kVp protocol.

Reduction in radiation and contrast medium dose via optimization of low-kilovoltage CT protocols using a hybrid iterative reconstruction algorithm at 256-slice body CT: phantom study and clinical correlation.

AIM: To optimize low-kilovoltage (kV) computed tomography (CT) protocols using a hybrid iterative reconstruction (HIR) algorithm at 256-detector-row body CT. MATERIALS AND METHODS: Based on preliminary phantom studies, three different tube voltage protocols with an equal contrast-to-noise ratio (CNR) were developed. They were a conventional 120 kV protocol with filtered back-projection (FBP), an 80 kV protocol with HIR (a 160% increase in the tube current-time product and a 40% reduction in the contrast medium dose), and a 100 kV protocol with HIR (a 20% reduction in the tube current-time product and the contrast medium dose). The clinical study included 70 patients (34 women, 36 men; mean age 70.5 ± 9.1 years, range 44-92 years) who had undergone CT at 120 kV a mean of 148 ± 137 days before undergoing low kV contrast-enhanced body CT (80 kV with HIR, n = 35; 100 kV with HIR, n = 35). The estimated effective radiation dose (ED), image noise, and CNR were calculated and the visual image quality was scored on a four-point scale. RESULTS: Mean ED was 12.3, 8.4, and 15.4 mSv for the 80, 100, and 120 kV protocol, respectively, and significantly lower using the low kV protocols. There was no significant difference in the image noise and CNR between the low kV protocols with HIR and the 120 kV protocol with FBP, or in the visual scores among the three protocols. CONCLUSION: Without ensuing image-quality degradation, the radiation and contrast medium dose can be reduced with optimal contrast-enhanced CT protocols using a low kV technique and an HIR algorithm.

Detection of small pulmonary nodules on chest radiographs: efficacy of dual-energy subtraction technique using flat-panel detector chest radiography.

AIM: To investigate the effect of a double-exposure dual-energy subtraction (DES) technique on the diagnostic performance of radiologists detecting small pulmonary nodules on flat-panel detector (FPD) chest radiographs. MATERIALS AND METHODS: Using FPD radiography 41 sets of chest radiographs were obtained from 26 patients with pulmonary nodules measuring