Image Quality and Lesion Detection of Multiplanar Reconstruction Images Using Deep Learning: Comparison with Hybrid Iterative Reconstruction.

Background: We assessed and compared the image quality of normal and pathologic structures as well as the image noise in chest computed tomography images using "adaptive statistical iterative reconstruction-V" (ASiR-V) or deep learning reconstruction "TrueFidelity". Methods: Forty consecutive patients with suspected lung disease were evaluated. The 1.25-mm axial images and 2.0-mm coronal multiplanar images were reconstructed under the following three conditions: (i) ASiR-V, lung kernel with 60% of ASiR-V; (ii) TF-M, standard kernel, image filter (Lung) with TrueFidelity at medium strength; and (iii) TF-H, standard kernel, image filter (Lung) with TrueFidelity at high strength. Two radiologists (readers) independently evaluated the image quality of anatomic structures using a scale ranging from 1 (best) to 5 (worst). In addition, readers ranked their image preference. Objective image noise was measured using a circular region of interest in the lung parenchyma. Subjective image quality scores, total scores for normal and abnormal structures, and lesion detection were compared using Wilcoxon's signed-rank test. Objective image quality was compared using Student's paired t-test and Wilcoxon's signed-rank test. The Bonferroni correction was applied to the P value, and significance was assumed only for values of P < 0.016. Results: Both readers rated TF-M and TF-H images significantly better than ASiR-V images in terms of visualization of the centrilobular region in axial images. The preference score of TF-M and TF-H images for reader 1 were better than that of ASiR-V images, and the preference score of TF-H images for reader 2 were significantly better than that of ASiR-V and TF-M images. TF-M images showed significantly lower objective image noise than ASiR-V or TF-H images. Conclusion: TrueFidelity showed better image quality, especially in the centrilobular region, than ASiR-V in subjective and objective evaluations. In addition, the image texture preference for TrueFidelity was better than that for ASiR-V.

MR imaging findings of massive perivillous fibrin deposition of the placenta: A case report.

Massive perivillous fibrin deposition (MPFD) of the placenta is characterized by the obliteration of the villous trophoblast with extensive deposition of fibrinoid material in the intervillous space. Here, we describe the MRI findings of a case of MPFD. The placenta demonstrates linear and geographical hypointensity on T2-weighted imaging, which is suggested to mainly reflect fibrin deposition. This finding should be noted, particularly in patients with miscarriage in their past history.

MR imaging findings of primary ovarian carcinoid: A novel finding of T1 hyperintense solid tissue.

Ovarian carcinoid is a rare well-differentiated neuroendocrine tumor resembling those arising in the gastrointestinal tract. We present a case of ovarian carcinoid with magnetic resonance imaging (MRI) findings. A 50-year-old woman with genital bleeding and severe constipation was referred to our hospital. On MR imaging, a left ovarian tumor showed iso to high signal intensity on T1-weighted images (T1WI), relatively low signal intensity on T2WI, and slightly high signal intensity on diffusion-weighted images. Additionally, the tumor demonstrated early and delayed strong contrast enhancement on dynamic contrast-enhanced images. The tumor was pathologically diagnosed with ovarian strumal carcinoid. High signal intensity on T1WI should be recognized as the MRI findings in ovarian carcinoids.

Clinical Utility of Diffusion-Weighted Imaging in Gynecological Imaging: Revisited.

ABSTRACT: Diffusion-weighted imaging (DWI) is an increasingly valuable sequence in daily clinical practice, providing both functional and morphological information. The use of DWI can help quantify diffusion using the apparent diffusion coefficient, which reflects the physiological features of the tissue and tumor microcirculation. This knowledge is crucial for understanding and interpreting gynecological imaging. This article reviews the clinical utility of DWI for gynecological imaging, highlighting its ability to aid in the detection of endometrial and cervical cancers, as well as tumor extension and metastasis. In addition, DWI can easily detect the solid components of ovarian cancer (including dissemination), assist in the diagnosis of adnexal torsion, and potentially show bone marrow status. Apparent diffusion coefficient measurement is useful for differentiating between endometrial lesions, uterine leiomyomas, and sarcomas, and may provide important information for predicting the prognosis of gynecological cancers.

Amide proton transfer imaging in differentiation of type II and type I endometrial carcinoma: a pilot study.

PURPOSE: This study aimed at evaluating the efficacy of amide proton transfer (APT) imaging in differentiation of type II and type I uterine endometrial carcinoma. MATERIALS AND METHODS: Thirty-three patients diagnosed with uterine endometrial carcinoma, including 24 with type I and 9 with type II carcinomas, underwent APT imaging. Two readers evaluated the magnetization transfer ratio at 3.5 ppm [MTRasym (3.5 ppm)] in each type of carcinoma. The average MTRasym (APTmean) and the maximum MTRasym (APTmax) were analyzed. The receiver operating characteristic (ROC) curve analysis was performed. RESULTS: The APTmax was significantly higher in type II carcinomas than in type I carcinomas (reader1, p = 0.004; reader 2, p = 0.014; respectively). However, APTmean showed no significant difference between type I and II carcinomas. Based on the results reported by reader 1, the area under the curve (AUC) pertaining to the APTmax for distinguishing type I from type II carcinomas was 0.826, with a cut-off, sensitivity, and specificity of 9.90%, 66.7%, and 91.3%, respectively. Moreover, based on the results reported by reader 2, the AUC was 0.750, with a cut-off, sensitivity, and specificity of 9.80%, 62.5%, and 87.5%, respectively. CONCLUSION: APT imaging has the potential to determine the type of endometrial cancer.

Magnetic resonance imaging findings of endosalpingiosis: a case report.

Endosalpingiosis is characterized by the presence of glands lined by benign tubal-type epithelium outside the fallopian tube. It is usually an incidental finding and rarely occurs as a tumor-like mass lesion. Here, we describe the magnetic resonance imaging findings of endosalpingiosis that presented as a paraovarian multicystic lesion. It exhibited iso to low intensity on T1-weighted images and inhomogeneous high intensity on T2-weighted images. The septa presented relatively iso to slight high intensity on T2-weighted images and strong contrast enhancement on dynamic contrast-enhanced imaging. Endosalpingiosis should be considered as a differential diagnosis in cases of paraovarian multicystic lesions along the uterine serosa.

CT and MR imaging findings of bilateral ovarian metastasis from renal cell carcinoma: a case report.

Secondary ovarian involvement by renal cell carcinoma rarely occurs. Here, we describe the computed tomography and magnetic resonance imaging findings of bilateral ovarian metastases from renal cell carcinoma that demonstrated heterogeneous strong contrast enhancing tumors with flow voids around and within the tumors. In addition, the apparent diffusion coefficients of the malignant tumors were high. These findings were similar to those of renal cell carcinomas at primary and other metastatic sites.

MR imaging findings of unusual leiomyoma and malignant uterine myometrial tumors: what the radiologist should know.

Uterine sarcomas account for less than 1% of gynecological malignancies and 2-5% of all uterine malignancies. Such sarcomas mainly include leiomyosarcoma (LMS) and endometrial stromal sarcoma (ESS). Additionally, inflammatory myofibroblastic tumor (IMT) and endometrial carcinoma arising in adenomyosis can occur as uterine myometrial tumors. Their differentiation from leiomyoma (LM), particularly degenerated LM and the malignant tumors, is challenging, but preoperative diagnosis is very important for the patient's management. We demonstrate the useful and compulsory findings to differentiate between uterine myometrial malignant tumors and degenerated LM with an unusual appearance.

Effect of energy difference in the evaluation of calcification size and luminal diameter in calcified coronary artery plaque using spectral CT.

PURPOSE: This study evaluated the calcium blooming-reducing effect and the differences of luminal diameter among various-energy virtual monochromatic images (VMIs) using rapid kilovolt-switching dual-energy computed tomography (DECT). MATERIALS AND METHODS: Forty-five calcified segments in 31 patients were analyzed. For the analysis, 40- to 140-keV VMIs on both non-contrast CT and coronary CT angiography were generated at 10-keV steps, and calcification size and luminal diameter were measured using CT number profile curve and full-width at half-maximum method. We compared calcification size and luminal diameter on each keV VMIs with those on 70-keV VMI. RESULTS: There was no significant differences among the 40- to 140-keV VMIs regarding calcification size or luminal diameter. CONCLUSION: The 40- to 140-keV VMIs produced by single-source DECT had no effect on the calcification size or luminal diameter in the coronary artery.

Investigation of myocardial extracellular volume fraction in heart failure patients using iodine map with rapid-kV switching dual-energy CT: Segmental comparison with MRI T1 mapping.

PURPOSE: To measure myocardial extracellular volume fraction (ECV) for each region or segment using iodine density image (IDI) with single-source dual-energy computed tomography (DECT) and compare the results with an MRI T1 mapping approach. MATERIALS AND METHODS: For this prospective study, 79 consecutive heart failure patients referred for MRI were included and 23 patients (14 men, 63 ± 14 years) who underwent both MRI and late contrast enhancement DECT following coronary CT angiography were evaluated. CT-ECV was computed from IDI using late acquisition projection data. MR-ECV was computed from native and post-contrast T1 maps using non-rigid image registration for segments with evaluable image quality from 3.0-T MRI. Regional CT-ECV and MR-ECV were measured based on 16-segment models. CT-ECV and MR-ECV were compared using Pearson correlations. Agreement among methods was assessed using Bland-Altman comparisons. RESULTS: In the 368 segments, although all segments were evaluable on IDI, 37 segments were rated as non-evaluable on T1 maps. Overall, 331 segments were analyzed. Mean CT-ECV and MR-ECV were 31.6 ± 9.1 and 33.2 ± 9.1, respectively. Strong correlations were seen between CT-ECV and MR-ECV for each region, as follows: all segments, r = 0.837; septal, r = 0.871; mid-septal, r = 0.895; anterior, r = 0.869; inferior, r = 0.793; and lateral, 0.864 (all p < 0.001). Differences between CT-ECV and MR-ECV were as follows: all segments, 1.13 ± 4.98; septal, -1.51 ± 4.37; mid-septal, -1.85 ± 4.22; anterior, 2.54 ± 4.89; inferior, 1.2 ± 5.78; and lateral, 2.65 ± 3.98. CONCLUSION: ECV using DECT and from cardiac MRI showed a strong correlation on regional and segmental evaluations. DECT is useful for characterizing myocardial ECV changes as well as MRI.

Usefulness of Preoperative 18F-FDG PET/CT for Patients with Thymic Epithelial Tumors.

BACKGROUND: The purpose of this study was to investigate the relationship between preoperative FDG-PET parameters and the World Health Organization (WHO) classification or Masaoka staging system of thymic epithelial tumors. METHODS: We retrospectively reviewed 32 patients with histologically proven thymic epithelial tumors who underwent FDG-PET/CT before surgical resection. FDG-PET parameters, including the maximum standardized uptake value (SUVmax), metabolic tumor volume (MTV), and total lesion glycolytic activity (TLG), were measured. These PET parameters were compared in the Masaoka staging system and WHO classification. A receiver operating characteristics (ROC) analysis was performed to identify the cut-off values of PET parameters for the accurate differentiation of early and advanced stages in the Masaoka staging system. RESULTS: There were 17 low-risk thymomas (1 type A, 9 type AB, and 7 type B1), 8 high-risk thymomas (4 type B2 and 4 type B3), and 7 thymic carcinomas (7 squamous cell carcinoma). Their Masaoka stages were as follows: 24 in the early stage (stages I and II) and 8 in the advanced stage (stage III). Regarding the WHO classification, only SUVmax showed a significant difference (P < 0.05). In the Masaoka stage, all PET parameters were significantly higher in the advanced stage than in the early stage (P < 0.05). In the ROC analysis to predict the early and advanced stages in thymic epithelial tumors, the area under the curve was the highest for TLG among the PET parameters examined and the cut-off value of TLG for discriminating the early from advanced stage with maximal sensitivity and specificity was 30.735. CONCLUSION: Although volumetric PET parameters, such as MTV and TLG, did not correlate with the WHO classification, a significant correlation was observed between SUVmax and the WHO classification. In the Masaoka staging system, volumetric PET parameters may achieve more precise staging than SUVmax.

Detection and Classification of Myocardial Delayed Enhancement Patterns on MR Images with Deep Neural Networks: A Feasibility Study.

PURPOSE: To evaluate whether deep neural networks trained on a similar number of images to that required during physician training in the American College of Cardiology Core Cardiovascular Training Statement can acquire the capability to detect and classify myocardial delayed enhancement (MDE) patterns. MATERIALS AND METHODS: The authors retrospectively evaluated 1995 MDE images for training and validation of a deep neural network. Images were from 200 consecutive patients who underwent cardiovascular MRI and were obtained from the institutional database. Experienced cardiac MR image readers classified the images as showing the following MDE patterns: no pattern, epicardial enhancement, subendocardial enhancement, midwall enhancement, focal enhancement, transmural enhancement, and nondiagnostic. Data were divided into training and validation datasets by using a fourfold cross-validation method. Three untrained deep neural network architectures using the convolutional neural network (CNN) technique were trained with the training dataset images. The detection and classification accuracies of the trained CNNs were calculated with validation data. RESULTS: The 1995 MDE images were classified by human readers as follows: no pattern, 926; epicardial enhancement, 91; subendocardial enhancement, 458; midwall enhancement, 118; focal enhancement, 141; transmural enhancement, 190; and nondiagnostic, 71. GoogLeNet, AlexNet, and ResNet-152 CNNs demonstrated accuracies of 79.5% (1592 of 1995 images), 78.9% (1574 of 1995 images), and 82.1% (1637 of 1995 images), respectively. CONCLUSION: Deep learning with CNNs using a limited amount of training data, less than that required during physician training, achieved high diagnostic performance in the detection of MDE on MR images.© RSNA, 2019Supplemental material is available for this article.

Quantitative evaluation of non-ischemic dilated cardiomyopathy by late iodine enhancement using rapid kV switching dual-energy computed tomography: A feasibility study.

OBJECTIVES: The aim of this study was to evaluate the feasibility of myocardial iodine density and extracellular volume fraction (ECV) from delayed iodine density images using dual-energy computed tomography (DECT) for differentiation between non-ischemic dilated cardiomyopathy (NIDCM) patients and normal subjects. METHODS: Forty-six subjects were imaged, including 35 normal subjects and 11 patients with NIDCM. All subjects underwent myocardial delayed enhancement (MDE) imaging on rapid-kVp switching DECT. Global and segmental iodine density and ECV were calculated from MDE images. Histogram analysis was also performed. Receiver-operator characteristic (ROC) analysis was used to determine the cut-off value and diagnostic performances in differentiating NIDCM patients from normal subjects. RESULTS: Global iodine density and ECV were significantly higher in NIDCM compared with normal controls (iodine: 14.19 ±â€¯3.90 vs. 10.69 ±â€¯1.88 in 100 µg/cm3, p = 0.015; ECV: 31.35 ±â€¯2.53% vs. 26.62 ±â€¯2.69%, p < 0.001). In histogram analyses, kurtosis was higher in NIDCM than in controls (0.47 ±â€¯0.46 vs. 1.26 ±â€¯0.88, p < 0.001). On segmental analysis, ECV showed higher values in NIDCM than in controls for all segments. ECV could differentiate between normal myocardium and NIDCM with 91.0% sensitivity and 86.0% specificity at a cut-off of 28.82% (area under the curve of ROC, 0.906). Iodine density could differentiate between normal myocardium and NIDCM with 91% sensitivity and 60% specificity at a cut-off of 11.18 (area under the curve of ROC, 0.812). CONCLUSIONS: Iodine density and ECV values from DECT may provide indices offering high diagnostic accuracy for discriminating between NIDCM and normal myocardium.

Myocardial Delayed Enhancement CT for the Evaluation of Heart Failure: Comparison to MRI.

Purpose To assess the diagnostic performance of dual-energy CT with myocardial delayed enhancement (MDE) in the detection and classification of myocardial scar in patients with heart failure, with late gadolinium enhancement (LGE) MRI as the standard of reference. Materials and Methods MDE CT and LGE MRI were performed in 44 patients with heart failure (30 men; mean patient age, 66 years ± 14) between 2013 and 2016, and images were retrospectively analyzed. The presence and patterns of MDE on iodine-density and virtual monochromatic (VM) images were assessed by two independent readers. Contrast-to-noise ratio (CNR) and percentage signal intensity increase relative to normal myocardium were measured. Diagnostic performance and area under the receiver operating characteristic curve for MDE CT and kappa values for reader agreement were determined. Results Thirty-five of the 44 patients (80%) demonstrated a focal area of LGE, with a nonischemic pattern in 22 of the 44 patients (50%) and an ischemic pattern in 13 (30%). Iodine-density images demonstrated the highest CNR and percentage signal intensity increase on CT images (P < .05), resulting in the highest diagnostic performance in the detection of any MDE CT abnormality (92% sensitivity [195 of 213 segments] and 98% specificity [481 of 491 segments]). The areas under the receiver operating characteristic curve for iodine-density images and 40-keV VM images in the detection of MDE were 0.97 and 0.95, respectively (P < .001). Kappa values for reader agreement were 0.82 for iodine-density images and 0.72 for 40-keV VM images. Conclusion Myocardial delayed enhancement CT enables accurate detection and localization of scar in patients with heart failure when compared with late gadolinium enhancement MRI, the reference standard.

Diagnostic performance of calcification-suppressed coronary CT angiography using rapid kilovolt-switching dual-energy CT.

OBJECTIVES: Multi-detector-row computed tomography angiography (MDCTA) plays an important role in the assessment of patients with suspected coronary artery disease. However, MDCTA tends to overestimate stenosis in calcified coronary artery lesions. The aim of our study was to evaluate the diagnostic performance of calcification-suppressed material density (MD) images produced by using a single-detector single-source dual-energy computed tomography (ssDECT). METHODS: We enrolled 67 patients with suspected or known coronary artery disease who underwent ssDECT with rapid kilovolt-switching (80 and 140 kVp). Coronary artery stenosis was evaluated on the basis of MD images and virtual monochromatic (VM) images. The diagnostic performance of the two methods for detecting coronary artery disease was compared with that of invasive coronary angiography as a reference standard. RESULTS: We evaluated 239 calcified segments. In all the segments, the sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) and accuracy for detecting significant stenosis were respectively 88%, 88%, 75%, 95% and 88% for the MD images, 91%, 71%, 56%, 95% and 77% for the VM images. PPV was significantly higher on the MD images than on the VM images (P < 0.0001). CONCLUSIONS: Calcification-suppressed MD images improved PPV and diagnostic performance for calcified coronary artery lesions. KEY POINTS: • Computed tomography angiography tends to overestimate stenosis in calcified coronary artery. • Dual-energy CT enables us to suppress calcification of coronary artery lesions. • Calcification-suppressed material density imaging reduces false-positive diagnosis of calcified lesion.