Diagnostic value of deep learning reconstruction for radiation dose reduction at abdominal ultra-high-resolution CT.

OBJECTIVES: We evaluated lower dose (LD) hepatic dynamic ultra-high-resolution computed tomography (U-HRCT) images reconstructed with deep learning reconstruction (DLR), hybrid iterative reconstruction (hybrid-IR), or model-based IR (MBIR) in comparison with standard-dose (SD) U-HRCT images reconstructed with hybrid-IR as the reference standard to identify the method that allowed for the greatest radiation dose reduction while preserving the diagnostic value. METHODS: Evaluated were 72 patients who had undergone hepatic dynamic U-HRCT; 36 were scanned with the standard radiation dose (SD group) and 36 with 70% of the SD (lower dose [LD] group). Hepatic arterial and equilibrium phase (HAP, EP) images were reconstructed with hybrid-IR in the SD group, and with hybrid-IR, MBIR, and DLR in the LD group. One radiologist recorded the standard deviation of attenuation in the paraspinal muscle as the image noise. The overall image quality was assessed by 3 other radiologists; they used a 5-point confidence scale ranging from 1 (unacceptable) to 5 (excellent). Superiority and equivalence with prespecified margins were assessed. RESULTS: With respect to the image noise, in the HAP and EP, LD DLR and LD MBIR images were superior to SD hybrid-IR images; LD hybrid-IR images were neither superior nor equivalent to SD hybrid-IR images. With respect to the quality scores, only LD DLR images were superior to SD hybrid-IR images. CONCLUSIONS: DLR preserved the quality of abdominal U-HRCT images even when scanned with a reduced radiation dose. KEY POINTS: • Lower dose DLR images were superior to the standard-dose hybrid-IR images quantitatively and qualitatively at abdominal U-HRCT. • Neither hybrid-IR nor MBIR may allow for a radiation dose reduction at abdominal U-HRCT without compromising the image quality. • Because DLR allows for a reduction in the radiation dose and maintains the image quality even at the thinnest slice section, DLR should be applied to abdominal U-HRCT scans.

Utility of Radial Scanning for the Identification of Arterial Hypervascularity of Hepatocellular Carcinoma on Gadoxetic Acid-Enhanced Magnetic Resonance Images.

OBJECTIVES: This study aimed to compare the accuracy of assessing the arterial hypervascularity of hepatocellular carcinoma (HCC) on dynamic computed tomography (CT) scans and gadoxetic acid (EOB)-enhanced magnetic resonance imaging (MRI) scans performed with radial sampling. METHODS: We studied the images of 40 patients with hypervascular HCC. A radiologist recorded the standard deviation of the attenuation (or the signal intensity [SI]) in subcutaneous fat tissue as the image noise (N) and calculated the contrast-to-noise ratio (CNR) as follows: (CNR) = (n-ROIT - n-ROIL)/N, where n-ROIT is the mean attenuation (or SI) of the tumor divided by the mean attenuation (or SI) of the aorta and n-ROIL is the mean attenuation (or SI) of the liver parenchyma divided by the mean attenuation (or SI) of the aorta. RESULTS: The CNR was significantly higher on EOB-enhanced MRI than on dynamic CT scans. CONCLUSIONS: For the assessment of HCC vascularity, EOB-enhanced MRI scans acquired with radial sampling were more accurate than dynamic CT images.

Advanced CT techniques for assessing hepatocellular carcinoma.

Hepatocellular carcinoma (HCC) is the sixth-most common cancer in the world, and hepatic dynamic CT studies are routinely performed for its evaluation. Ongoing studies are examining advanced imaging techniques that may yield better findings than are obtained with conventional hepatic dynamic CT scanning. Dual-energy CT-, perfusion CT-, and artificial intelligence-based methods can be used for the precise characterization of liver tumors, the quantification of treatment responses, and for predicting the overall survival rate of patients. In this review, the advantages and disadvantages of conventional hepatic dynamic CT imaging are reviewed and the general principles of dual-energy- and perfusion CT, and the clinical applications and limitations of these technologies are discussed with respect to HCC. Finally, we address the utility of artificial intelligence-based methods for diagnosing HCC.

Possibility of Deep Learning in Medical Imaging Focusing Improvement of Computed Tomography Image Quality.

Deep learning (DL), part of a broader family of machine learning methods, is based on learning data representations rather than task-specific algorithms. Deep learning can be used to improve the image quality of clinical scans with image noise reduction. We review the ability of DL to reduce the image noise, present the advantages and disadvantages of computed tomography image reconstruction, and examine the potential value of new DL-based computed tomography image reconstruction.

Deep learning reconstruction improves image quality of abdominal ultra-high-resolution CT.

OBJECTIVES: Deep learning reconstruction (DLR) is a new reconstruction method; it introduces deep convolutional neural networks into the reconstruction flow. This study was conducted in order to examine the clinical applicability of abdominal ultra-high-resolution CT (U-HRCT) exams reconstructed with a new DLR in comparison to hybrid and model-based iterative reconstruction (hybrid-IR, MBIR). METHODS: Our retrospective study included 46 patients seen between December 2017 and April 2018. A radiologist recorded the standard deviation of attenuation in the paraspinal muscle as the image noise and calculated the contrast-to-noise ratio (CNR) for the aorta, portal vein, and liver. The overall image quality was assessed by two other radiologists and graded on a 5-point confidence scale ranging from 1 (unacceptable) to 5 (excellent). The difference between CT images subjected to hybrid-IR, MBIR, and DLR was compared. RESULTS: The image noise was significantly lower and the CNR was significantly higher on DLR than hybrid-IR and MBIR images (p < 0.01). DLR images received the highest and MBIR images the lowest scores for overall image quality. CONCLUSIONS: DLR improved the quality of abdominal U-HRCT images. KEY POINTS: • The potential degradation due to increased noise may prevent implementation of ultra-high-resolution CT in the abdomen. • Image noise and overall image quality for hepatic ultra-high-resolution CT images improved with deep learning reconstruction as compared to hybrid- and model-based iterative reconstruction.

Correction to: Deep learning reconstruction improves image quality of abdominal ultra-high-resolution CT.

The original version of this article, published on 11 April 2019, unfortunately, contained a mistake. The following correction has therefore been made in the original: The image in Fig. 3c was wrong. The corrected figure is given below. The original article has been corrected.

Clinical staging of upper urinary tract urothelial carcinoma for T staging: Review and pictorial essay.

Upper urinary tract urothelial carcinoma is staged using the TNM classification of malignant tumors. Preoperative TNM is important for treatment planning. Computed tomography urography is now widely used for clinical survey of upper urinary tract carcinoma because of its diagnostic accuracy. Computed tomography urography is recommended as the first-line imaging procedure in several guidelines. Several reports stated that computed tomography urography is also useful for staging. However, no educational and practical reviews detailing the T staging of upper urinary tract urothelial carcinomas using imaging are available. We discuss the scanning protocol, T staging using computed tomography urography, limitations, magnetic resonance imaging, computed tomography comparison and pitfalls in imaging of upper urinary tract urothelial carcinoma. A recent study reported the high diagnostic accuracy of computed tomography urography with respect to T3 or higher stage tumors. To date, images that show a Tis-T2 stage have not been reported, but various studies are ongoing. Although magnetic resonance imaging has lower spatial resolution than computed tomography urography, magnetic resonance imaging can be carried out without radiation exposure or contrast agents. Magnetic resonance imaging also offers the unique ability of diffusion-weighted imaging without contrast agent use. Some researchers reported that diffusion-weighted imaging is useful not only for detecting lesions, but for predicting the T stage and tumor grade. We recommend the appropriate use of computed tomography and magnetic resonance while considering the limitations of each modality and the pitfalls in upper urinary tract urothelial carcinoma imaging.

Introduction to the Technical Aspects of Computed Diffusion-weighted Imaging for Radiologists.

Diffusion-weighted (DW) imaging is a magnetic resonance (MR) imaging method. It is an indispensable sequence for the diagnosis of acute cerebral infarction and is recognized as a standard tool in oncologic imaging. Computed DW imaging refers to the synthesizing of arbitrary b-value DW images from a set of measured b-value images by voxelwise fitting. Computed DW imaging is advantageous because it generates DW images with a higher diffusion effect than that achievable by using the MR imaging units in use today. Additionally, computed DW imaging can reduce imaging time while producing images characterized by a higher signal-to-noise ratio than what the acquired DW images would display at the corresponding b values. By fitting input images acquired at a lower b value and correspondingly a shorter echo time, the signal intensity of the resulting computed DW image is closer to the ideal case. Computed DW images are generated by employing mathematical models that use mono-, bi-, or triexponential equations. To generate accurate computed DW images, the appropriate model must be selected, and the image parameters for the input data must be chosen accordingly. In addition, to reduce artifacts on computed DW images, the misalignment of input data must be corrected with the aid of image registration techniques. ©RSNA, 2018.

Preliminary Results of High-Precision Computed Diffusion Weighted Imaging for the Diagnosis of Hepatocellular Carcinoma at 3 Tesla.

OBJECTIVE: To compare the utility of high-precision computed diffusion-weighted imaging (hc-DWI) and conventional computed DWI (cc-DWI) for the diagnosis of hepatocellular carcinoma (HCC) at 3 T. METHODS: We subjected 75 HCC patients to DWI (b-value 150 and 600 s/mm). To generate hc-DWI we applied non-rigid image registration to avoid the mis-registration of images obtained with different b-values. We defined c-DWI with a b-value of 1500 s/mm using DWI with b-value 150 and 600 s/mm as cc-DWI, and c-DWI with b-value 1500 s/mm using registered DWI with b-value 150 and 600 s/mm as hc-DWI. A radiologist recorded the contrast ratio (CR) between HCC and the surrounding hepatic parenchyma. RESULTS: The CR for HCC was significantly higher on hc- than cc-DWIs (median 2.0 vs. 1.8, P < 0.01). CONCLUSION: The CR of HCC can be improved with image registration, indicating that hc-DWI is more useful than cc-DWI for the diagnosis of HCC.

Deep learning reconstruction of equilibrium phase CT images in obese patients.

PURPOSE: To compare abdominal equilibrium phase (EP) CT images of obese and non-obese patients to identify the reconstruction method that preserves the diagnostic value of images obtained in obese patients. METHODS: We compared EP images of 50 obese patients whose body mass index (BMI) exceeded 25 (group 1) with EP images of 50 non-obese patients (BMI < 25, group 2). Group 1 images were subjected to deep learning reconstruction (DLR), hybrid iterative reconstruction (hybrid-IR), and model-based IR (MBIR), group 2 images to hybrid-IR; group 2 hybrid-IR images served as the reference standard. A radiologist recorded the standard deviation of attenuation in the paraspinal muscle as the image noise. The overall image quality was assessed by 3 other radiologists; they used a confidence scale ranging from 1 (unacceptable) to 5 (excellent). Non-inferiority and potential superiority were assessed. RESULTS: With respect to the image noise, group 1 DLR- were superior to group 2 hybrid-IR images; group 1 hybrid-IR- and MBIR images were neither superior nor non-inferior to group 2 hybrid-IR images. The quality scores of only DLR images in group 1 were superior to hybrid-IR images of group 2 while the quality scores of group 1 hybrid-IR- and MBIR images were neither superior nor non-inferior to group 2 hybrid-IR images. CONCLUSIONS: DLR preserved the quality of EP images obtained in obese patients.

Deep learning reconstruction of drip-infusion cholangiography acquired with ultra-high-resolution computed tomography.

PURPOSE: Deep learning reconstruction (DLR) introduces deep convolutional neural networks into the reconstruction flow. We examined the clinical applicability of drip-infusion cholangiography (DIC) acquired on an ultra-high-resolution CT (U-HRCT) scanner reconstructed with DLR in comparison to hybrid and model-based iterative reconstruction (hybrid-IR, MBIR). METHODS: This retrospective, single-institution study included 30 patients seen between January 2018 and November 2019. A radiologist recorded the standard deviation of attenuation in the paraspinal muscle as the image noise and calculated the contrast-to-noise ratio (CNR) in the common bile duct. The overall visual image quality of the bile duct on thick-slab maximum intensity projections was assessed by two other radiologists and graded on a 5-point confidence scale ranging from 1 (not delineated) to 5 (clearly delineated). The difference among hybrid-IR, MBIR, and DLR images was compared. RESULTS: The image noise was significantly lower on DLR than hybrid-IR and MBIR images and the CNR and the overall visual image quality of the bile duct were significantly higher on DLR than on hybrid-IR and MBIR images (all: p < 0.001). CONCLUSION: DLR resulted in significant quantitative and qualitative improvement of DIC acquired with U-HRCT.

Pseudo-random Trajectory Scanning Suppresses Motion Artifacts on Gadoxetic Acid-enhanced Hepatobiliary-phase Magnetic Resonance Images.

PURPOSE: Hepatobiliary-phase (HBP) MRI with gadoxetic acid facilitates the differentiation between lesions with and without functional hepatocytes. Thus, high-quality HBP images are required for the detection and evaluation of hepatic lesions. However, the long scan time may increase artifacts due to intestinal peristalsis, resulting in the loss of diagnostic information. Pseudo-random acquisition order disperses artifacts into the background. The aim of this study was to investigate the clinical applicability of pseudo-random trajectory scanning for the suppression of motion artifacts on T1-weighted images including HBP. METHODS: Our investigation included computer simulation, phantom experiments, and a clinical study. For computer simulation and phantom experiments a region of interest (ROI) was placed on the area with motion artifact and the standard deviation inside the ROI was measured as image noise. For clinical study we subjected 62 patients to gadoxetic acid-enhanced hepatobiliary-phase imaging with a circular- and a pseudo-random trajectory (c-HBP and p-HBP); two radiologists graded the motion artifacts, sharpness of the liver edge, visibility of intrahepatic vessels, and overall image quality using a five-point scale where 1 = unacceptable and 5 = excellent. Differences in the qualitative scores were determined using the two-sided Wilcoxon signed-rank test. RESULTS: The image noise was higher on the circular image compared with pseudo-random image (101.0 vs 60.9 on computer simulation image, 91.2 vs 67.7 on axial, 95.5 vs 86.9 on reformatted sagittal image for phantom experiments). For clinical study the score for motion artifacts was significantly higher with p-HBP than c-HBP imaging (left lobe: mean 3.4 vs 3.2, P < 0.01; right lobe: mean 3.6 vs 3.4, P < 0.01) as was the qualitative score for the overall image quality (mean 3.6 vs 3.3, P < 0.01). CONCLUSION: At gadoxetic acid-enhanced hepatobiliary-phase imaging, p-HBP scanning suppressed motion artifacts and yielded better image quality than c-HBP scanning.

Preoperative PET-CT is useful for predicting recurrent extrahepatic metastasis of hepatocellular carcinoma after resection.

PURPOSE: In recent years, it has been reported that use of 18F-FDG PET-CT can reveal the degree of hepatocellular carcinoma malignancy. We evaluate the ability of a preoperative 18F-FDG PET-CT to predict the recurrence of extrahepatic metastasis of HCC after surgery. METHODS: We retrospectively examined 67 patients who received 18F-FDG PET-CT prior to curative hepatic resection for HCC between April 2010 and March 2016. Multivariate Cox regression analysis was performed to identify the factors associated with recurrence of extrahepatic metastasis of HCC after surgery. We also evaluated the sensitivity, specifity, positive predictive value, negative predictive value and accuracy of diagnosis of 18F-FDG PET-CT for recurrent extrahepatic metastasis of HCC after surgery. RESULTS: The multivariate analysis identified a tumor-to-normal liver standardized uptake value ratio (TNR) ≥ 1.53 (hazard ratio [HR], 0.037; P = 0.003), multiple tumor nodules (HR, 0.121; P = 0.007), and presence of microvascular invasion (HR, 0.094; P = 0.003) as independent predictors of distant metastasis recurrence. A TNR ≥ 1.53 showed a sensitivity of 91.7 %, specificity of 76.4 %, positive predictive value of 45.8 %, negative predictive value of 97.7 %, and accuracy of 79.1 % for diagnosing distant metastasis recurrence of HCC. In a binomial logistic regression analysis of tumor factors associated with a TNR ≥ 1.53, poor tumor differentiation and large tumor size were significant factors. CONCLUSION: 18F-FDG PET-CT and microvascular invasion may be useful for predicting the recurrence of extrahepatic metastasis of HCC after surgery.

Tubulocystic renal cell carcinoma: a review of literature focused on radiological findings for differential diagnosis.

Tubulocystic renal cell carcinoma (TC-RCC) has been classified as an independent subtype according to the 2016 World Health Organization (WHO) classification. It is a rare subtype that predominantly affects men. Although few in number, radiological imaging reports have suggested that TC-RCC is characterized by multilocular cystic lesions, which are categorized as the Bosniak classification II-IV, with signature pathological characteristics comprising numerous small cysts or a tubular structure. The Bosniak classification system facilitates patient management; however, the differentiation of cystic tumors exhibiting similar imaging findings remains impossible; in fact, the differentiation of multilocular cystic RCC, adult cystic nephroma, and mixed epithelial and stromal tumor remains challenging. This review aims to discuss TC-RCC with a focus on implications of radiological findings in the differential diagnosis of TC-RCC.