Hybrid deep-learning-based denoising method for compressed sensing in pituitary MRI: comparison with the conventional wavelet-based denoising method.

OBJECTIVES: This study aimed to evaluate the efficacy of a combined wavelet and deep-learning reconstruction (DLR) method for under-sampled pituitary MRI. METHODS: This retrospective study included 28 consecutive patients who underwent under-sampled pituitary T2-weighted images (T2WI). Images were reconstructed using either the conventional wavelet denoising method (wavelet method) or the wavelet and DLR methods combined (hybrid DLR method) at five denoising levels. The signal-to-noise ratio (SNR) of the CSF, hypothalamic, and pituitary images and the contrast between structures were compared between the two image types. Noise quality, contrast, sharpness, artifacts, and overall image quality were evaluated by two board-certified radiologists. The quantitative and the qualitative analyses were performed with robust two-way repeated analyses of variance. RESULTS: Using the hybrid DLR method, the SNR of the CSF progressively increased as denoising levels increased. By contrast, with the wavelet method, the SNR of the CSF, hypothalamus, and pituitary did not increase at higher denoising levels. There was a significant main effect of denoising methods (p < 0.001) and denoising levels (p < 0.001), and an interaction between denoising methods and denoising levels (p < 0.001). For all five qualitative scores, there was a significant main effect of denoising methods (p < 0.001) and an interaction between denoising methods and denoising levels (p < 0.001). CONCLUSIONS: The hybrid DLR method can provide higher image quality for T2WI of the pituitary with compressed sensing (CS) than the wavelet method alone, especially at higher denoising levels. KEY POINTS: • The signal-to-noise ratios of cerebrospinal fluid progressively increased with the hybrid DLR method, with an increase in the denoising level for cerebrospinal fluid in pituitary T2WI with CS. • The signal-to-noise ratios of cerebrospinal fluid using the conventional wavelet method did not increase at higher denoising levels. • All qualitative scores of hybrid deep-learning reconstructions at all denoising levels were higher than those for the wavelet denoising method.

A preliminary study of deep learning-based reconstruction specialized for denoising in high-frequency domain: usefulness in high-resolution three-dimensional magnetic resonance cisternography of the cerebellopontine angle.

PURPOSE: Deep learning-based reconstruction (DLR) has been developed to reduce image noise and increase the signal-to-noise ratio (SNR). We aimed to evaluate the efficacy of DLR for high spatial resolution (HR)-MR cisternography. METHODS: This retrospective study included 35 patients who underwent HR-MR cisternography. The images were reconstructed with or without DLR. The SNRs of the CSF and pons, contrast of the CSF and pons, and sharpness of the normal-side trigeminal nerve using full width at half maximum (FWHM) were compared between the two image types. Noise quality, sharpness, artifacts, and overall image quality of these two types of images were qualitatively scored. RESULTS: The SNRs of the CSF and pons were significantly higher with DLR than without DLR (CSF 21.81 ± 7.60 vs. 15.33 ± 4.03, p < 0.001; pons 5.96 ± 1.38 vs. 3.99 ± 0.48, p < 0.001). There were no significant differences in the contrast of the CSF and pons (p = 0.225) and sharpness of the normal-side trigeminal nerve using FWHM (p = 0.185) without and with DLR, respectively. Noise quality and the overall image quality were significantly higher with DLR than without DLR (noise quality 3.95 ± 0.19 vs. 2.53 ± 0.44, p < 0.001; overall image quality 3.97 ± 0.17 vs. 2.97 ± 0.12, p < 0.001). There were no significant differences in sharpness (p = 0.371) and artifacts (p = 1) without and with DLR. CONCLUSION: DLR can improve the image quality of HR-MR cisternography by reducing image noise without sacrificing contrast or sharpness.

Deep Learning Based Noise Reduction for Brain MR Imaging: Tests on Phantoms and Healthy Volunteers.

PURPOSE: To test whether our proposed denoising approach with deep learning-based reconstruction (dDLR) can effectively denoise brain MR images. METHODS: In an initial experimental study, we obtained brain images from five volunteers and added different artificial noise levels. Denoising was applied to the modified images using a denoising convolutional neural network (DnCNN), a shrinkage convolutional neural network (SCNN), and dDLR. Using these brain MR images, we compared the structural similarity (SSIM) index and peak signal-to-noise ratio (PSNR) between the three denoising methods. Two neuroradiologists assessed the image quality of the three types of images. In the clinical study, we evaluated the denoising effect of dDLR in brain images with different levels of actual noise such as thermal noise. Specifically, we obtained 2D-T2-weighted image, 2D-fluid-attenuated inversion recovery (FLAIR) and 3D-magnetization-prepared rapid acquisition with gradient echo (MPRAGE) from 15 healthy volunteers at two different settings for the number of image acquisitions (NAQ): NAQ2 and NAQ5. We reconstructed dDLR-processed NAQ2 from NAQ2, then compared with SSIM and PSNR. Two neuroradiologists separately assessed the image quality of NAQ5, NAQ2 and dDLR-NAQ2. Statistical analysis was performed in the experimental and clinical study. In the clinical study, the inter-observer agreement was also assessed. RESULTS: In the experimental study, PSNR and SSIM for dDLR were statistically higher than those of DnCNN and SCNN (P < 0.001). The image quality of dDLR was also superior to DnCNN and SCNN. In the clinical study, dDLR-NAQ2 was significantly better than NAQ2 images for SSIM and PSNR in all three sequences (P < 0.05), except for PSNR in FLAIR. For all qualitative items, dDLR-NAQ2 had equivalent or better image quality than NAQ5, and superior quality to that of NAQ2 (P < 0.05), for all criteria except artifact. The inter-observer agreement ranged from substantial to near perfect. CONCLUSION: dDLR reduces image noise while preserving image quality on brain MR images.

Perfusion abnormality on three-dimensional arterial spin labeling in patients with acute encephalopathy with biphasic seizures and late reduced diffusion.

PURPOSE: Acute encephalopathy with biphasic seizures and late reduced diffusion (AESD) is the most common encephalopathy subtype in Japanese children. Few case reports have shown perfusion abnormality on arterial spin labeling (ASL) in patients with AESD. The present study aimed to review the chronological change of cerebral perfusion on three-dimensional (3D) ASL in patients with AESD. METHODS: Twenty consecutive patients with AESD were enrolled; the patients underwent MRI including 3D ASL. The clinical course of AESD was divided into four phases according to the time from occurrence of seizures to MRI. Two neuroradiologists independently assessed presence or absence, distribution, and severity of perfusion abnormality using ASL and qualitatively scored perfusion abnormality using a five-point grading system. The level of interobserver agreement in the evaluation was analyzed using weighted κ statistics. Additionally, the signal ratio of abnormal perfusion region and peri-central sulcus region on ASL was semi-quantitatively evaluated. Moreover, we qualitatively compared the distribution between perfusion abnormality on ASL and bright tree appearance (BTA) on diffusion-weighted image (DWI). RESULTS: ASL showed hypoperfusion from 8.5 to 22 h after early seizures (ESs) and hyperperfusion within 24 h after late seizures (LSs). Various perfusions were found >3 days after LSs. Interobserver agreement for qualitative scored perfusion abnormality was good (κ = 0.77). The distribution of abnormal perfusion was relatively consistent with BTA. CONCLUSION: In AESD, cerebral perfusion changes with time. ASL showed hypoperfusion from 8.5 to 22 h after ESs, hyperperfusion within 24 h after LSs in patients with AESD.

An initial experience of machine learning based on multi-sequence texture parameters in magnetic resonance imaging to differentiate glioblastoma from brain metastases.

PURPOSE: To evaluate the performance of a machine learning method based on texture parameters in conventional magnetic resonance imaging (MRI) in differentiating glioblastoma (GB) from brain metastases (METs). MATERIALS AND METHODS: In this retrospective study conducted between November 2008 and July 2017, we included 73 patients diagnosed with GB (n = 73) and METs (n = 53) who underwent contrast-enhanced 3 T brain MRI. Twelve histogram and texture parameters were assessed on T2-weighted images (T2WIs), apparent diffusion coefficient maps (ADCs), and contrast-enhanced T1-weighted images (CE-T1WIs). A prediction model was developed for a machine learning method, and the area under the receiver operating characteristic curve of this model was calculated through 5-fold cross-validation. Furthermore, machine learning method's performance was compared with three board-certified radiologists' judgments. RESULTS: Univariate logistic regression model showed that the area under the curve (AUC) was highest with the standard value of T2WIs (0.78), followed by the maximum value of T2WIs (0.764), minimum value of T2WIs (0.738), minimum values of CE-T1WIs and contrast of T2WIs (0.733), and mean value of T2WIs (0.724). AUC calculated using the support vector machine was comparable to that calculated by the three radiologists (0.92 vs. 0.72, p < .01; 0.92 vs. 0.73, p < .01; and 0.92 vs. 0.86, p = .096). CONCLUSION: In differentiating GB from METs on the basis of texture parameters in MRI, the performance of the machine learning method based on convention MRI was superior to that of the univariate method, and comparable to that of the radiologists.

Machine learning based on multi-parametric magnetic resonance imaging to differentiate glioblastoma multiforme from primary cerebral nervous system lymphoma.

PURPOSE: To evaluate the performance of a machine learning method based on texture features in multi-parametric magnetic resonance imaging (MRI) to differentiate a glioblastoma multiforme (GBM) from a primary cerebral nervous system lymphoma (PCNSL). MATERIALS AND METHODS: We included 70 patients who underwent contrast enhanced brain MRI at 3 T with brain tumors diagnosed as GBM (n = 45) and PCNSL (n = 25) in this retrospective study. Twelve histograms and texture parameters were assessed on T2-weighted images (T2WIs), apparent diffusion coefficient maps, relative cerebral blood volume (rCBV) map, and contrast-enhanced T1-weighted images (CE-T1WIs). A prediction model was developed using a machine learning method (univariate logistic regression and multivariate eXtreme gradient boosting-XGBoost) and the area under the receiver operating characteristic curve of this model was calculated via 10-fold cross validation. In addition, the performance of the machine learning method was compared with the judgments of two board certified radiologists. RESULTS: With the univariate logistic regression model, the standard deviation of rCBV offered the highest AUC (0.86), followed by mean value of rCBV (0.83), skewness of CE-T1WI (0.78), mean value of CET1 (0.78), and max value of rCBV (0.77). The AUC of the XGBoost was significantly higher than the two radiologists (0.98 vs. 0.84; p < 0.01 and 0.98 vs. 0.79; p < 0.01, respectively). CONCLUSION: The performance of machine learning based on histogram and texture features in multi-parametric MRI was superior to that of conventional cut-off method and the board certified radiologists to differentiate a GBM from a PCNSL.

Perfusion abnormality on three-dimensional arterial spin labeling with a 3T MR system in pediatric and adolescent patients with migraine.

PURPOSE: Few studies have assessed the prevalence of perfusion abnormality with migraine. This study aimed to determine the prevalence and topography of perfusion abnormality on three-dimensional (3D) arterial spin labeling (ASL) and assess the correlation between perfusion abnormality and clinical data in pediatric and adolescent patients with migraine. METHODS: Forty-nine consecutive pediatric and adolescent patients with migraine were enrolled, and they underwent 3 T MRI, including 3D ASL. Perfusion abnormality on 3D ASL was qualitatively evaluated using a five-point grading system and was compared with non-ASL MR findings. In patients with perfusion abnormality, relative cerebral perfusion signal intensity (rCPS) was measured. Moreover, we compared clinical data and 3D ASL findings between patients with and those without perfusion abnormality. RESULTS: Of the 49 patients, 11 (22%) exhibited perfusion abnormality, and the occipital lobe was the most frequently involved (73%). One patient showed mild hyperperfusion (rCPS =2.474), and 10 showed hypoperfusion (mean rCPS = 0.405 ±â€¯0.134). There was no abnormality on non-ASL MRI, except in one case. We found statistically significant differences in the presence of aura (P < .001), motor disabilities (P = .019), confusion (P = .004), hospitalization (P = .004), between patients with and those without perfusion abnormality. CONCLUSION: In pediatric and adolescent patients with migraine, 3D ASL shows a high prevalence of perfusion abnormality, especially in the occipital lobe. Patients with perfusion abnormality tend to show the specific clinical symptoms at disease onset and need hospitalization.

Application of 80-kVp scan and raw data-based iterative reconstruction for reduced iodine load abdominal-pelvic CT in patients at risk of contrast-induced nephropathy referred for oncological assessment: effects on radiation dose, image quality and renal function.

OBJECTIVE: To evaluate the image quality, radiation dose, and renal safety of contrast medium (CM)-reduced abdominal-pelvic CT combining 80-kVp and sinogram-affirmed iterative reconstruction (SAFIRE) in patients with renal dysfunction for oncological assessment. METHODS: We included 45 patients with renal dysfunction (estimated glomerular filtration rate  <45 ml per min per 1.73 m2) who underwent reduced-CM abdominal-pelvic CT (360 mgI kg-1, 80-kVp, SAFIRE) for oncological assessment. Another 45 patients without renal dysfunction (estimated glomerular filtration rate >60 ml per lmin per 1.73 m2) who underwent standard oncological abdominal-pelvic CT (600 mgI kg-1, 120-kVp, filtered-back projection) were included as controls. CT attenuation, image noise, and contrast-to-noise ratio (CNR) were compared. Two observers performed subjective image analysis on a 4-point scale. Size-specific dose estimate and renal function 1-3 months after CT were measured. RESULTS: The size-specific dose estimate and iodine load of 80-kVp protocol were 32 and 41%,, respectively, lower than of 120-kVp protocol (p < 0.01). CT attenuation and contrast-to-noise ratio of parenchymal organs and vessels in 80-kVp images were significantly better than those of 120-kVp images (p < 0.05). There were no significant differences in quantitative or qualitative image noise or subjective overall quality (p > 0.05). No significant kidney injury associated with CM administration was observed. CONCLUSION: 80-kVp abdominal-pelvic CT with SAFIRE yields diagnostic image quality in oncology patients with renal dysfunction under substantially reduced iodine and radiation dose without renal safety concerns. Advances in knowledge: Using 80-kVp and SAFIRE allows for 40% iodine load and 32% radiation dose reduction for abdominal-pelvic CT without compromising image quality and renal function in oncology patients at risk of contrast-induced nephropathy.

Differentiating between Alzheimer Disease Patients and Controls with Phase-difference-enhanced Imaging at 3T: A Feasibility Study.

PURPOSE: To test the feasibility of the phase difference enhanced (PADRE) imaging for differentiation between Alzheimer disease (AD) patients and control subjects on 3T MR imaging. MATERIALS AND METHODS: Fifteen patients with AD and 10 age-matched control subjects underwent two-dimensional fast field echo imaging to obtain PADRE images on a 3T MR scanner. A double Gaussian distribution model was used to determine the threshold phase value for differentiation between the physiologic and non-physiologic iron in the cerebral cortices, and PADRE images were processed with the threshold. Using a 4-point grading system, two readers independently assessed the signal of the four cerebral cortices on PADRE images: the cuneus, precuneus, superior frontal gyrus, and superior temporal gyrus. The difference in the signals in each cortex between the AD patients and age-matched control subjects was determined by using Mann-Whitney U test. Inter-rater reliability was determined by Kappa analysis. We also evaluated the correlation between Mini-Mental State Examination (MMSE) score and the hypointense grade, and between disease duration and the hypointense grade using the Spearman rank correlation test. RESULTS: The threshold phase value for differentiation between the physiologic and non-physiologic iron was -4.6% π (radian). The mean grades of the cuneus, precuneus, and superior temporal gyrus were significantly higher for the AD patients than for the control subjects (P = 0.002). Excellent inter-rater reliability was seen in the precuneus (kappa = 0.93), superior temporal gyrus (kappa = 0.94), and superior frontal gyrus (kappa = 0.93); good inter-rater reliability was observed in the cuneus (kappa = 0.75). We found a statistical correlation between MMSE score and the hypointense grade in superior temporal gyrus (STG) (P = 0.008), and no correlation between disease duration and the hypointense grade in any gyrus. CONCLUSION: Our results suggest the feasibility of PADRE imaging at 3T for differentiation between AD patients and control subjects.

Benefit of 3T Diffusion-weighted Imaging in Comparison to Contrast-enhanced MR Imaging for the Evaluation of Disseminated Lesions in Primary Malignant Brain Tumors.

PURPOSE: We aimed to determine whether 3T diffusion-weighted imaging (DWI) has an additive value relative to contrast-enhanced MR imaging for the detection of disseminated lesions in patients with primary malignant brain tumors. METHODS: We included consecutive 12 patients with nodular disseminated lesions of primary malignant brain tumors that were confirmed by surgery or follow-up MR imaging. All underwent conventional MR imaging, DWI at b = 1000 and 3000 s/mm2, post-contrast T1-weighted and 3D gradient-echo imaging at 3T. For the largest lesion per person, two radiologists independently evaluated the presence of additional information on DWI compared with postcontrast MR images using a 4-point scoring system. On DW images, one radiologist measured the lesion-to-brain contrast ratio (LBCR). RESULTS: Compared with postcontrast studies, radiologists 1 and 2, respectively, assigned more apparent lesion conspicuity in 2 (17%) and 1 (8%) DWI at b = 1000 s/mm2 and 4 (33%) and 5 (42%) DWI at b = 3000 s/mm2 studies. For one of them, the mean score was significantly higher for b = 3000 s/mm2 than b = 1000 s/mm2 (P < 0.05). Interobserver agreement for DWI at b = 1000 s/mm2 and b = 3000 s/mm2 was very good (κ = 0.85; 95% CI, 0.63-1.00) and excellent (κ = 0.93; 95% CI, 0.78-1.00), respectively. The mean LBCR was significantly higher for DWI at b = 3000 s/mm2 than DWI at b = 1000 s/mm2 (P < 0.01). CONCLUSION: In the detection of disseminated lesions in patients with primary malignant brain tumors, 3T DWI has an additive value relative to contrast-enhanced MR imaging. DWI at b = 3000 s/mm2 may be more useful than DWI at b = 1000 s/mm2.

Evaluation of Intracranial Vasculatures in Healthy Subjects with Arterial-Spin-Labeling-Based 4D-MR Angiography at 3T.

Contrast inherent inflow-enhanced multi-phase angiography combining multiple-phase flow-alternating inversion-recovery (CINEMA-FAIR) is an arterial-spin-labeling-based four-dimensional magnetic resonance angiography (4D-MRA) technique. Two neuroradiologists independently evaluated the depiction of the intracranial vasculatures in healthy subjects with 3T 4D-MRA using CINEMA-FAIR. Our results indicated that this technique can provide good visualization of the cerebral arteries with a high spatial and temporal resolution. It appears to have sufficient resolution for identifying flow difference in the anterior and posterior circulation in healthy subjects.

[Anesthesia and perioperative management in a patient with acquired angioedema].

We experienced anesthesia and perioperative management for hysterectomy in a patient with acquired angioedema. Angioedema due to C1 esterase inhibitor (C1-INH) deficiency (loss or dysfunction of C1-INH) is one of the rarest diseases, and is characterized by recurrent episodes of regional hard edema and ascites induced by mechanical stimuli or mental stress. Edema spreads to the subcutaneous and submucosal layer, and laryngeal edema may cause the upper airway obstruction. Tranexamic acid and C1-INH concentrates were administered perioperatively for prophylaxis of attacks, and combined spinal and epidural anesthesia was performed for hysterectomy. We could manage perioperative care without causing edema.