Simultaneous brain structure segmentation in magnetic resonance images using deep convolutional neural networks.

In brain magnetic resonance imaging (MRI) examinations, rapidly acquired two-dimensional (2D) T1-weighted sagittal slices are typically used to confirm brainstem atrophy and the presence of signals in the posterior pituitary gland. Image segmentation is essential for the automatic evaluation of chronological changes in the brainstem and pituitary gland. Thus, the purpose of our study was to use deep learning to automatically segment internal organs (brainstem, corpus callosum, pituitary, cerebrum, and cerebellum) in midsagittal slices of 2D T1-weighted images. Deep learning for the automatic segmentation of seven regions in the images was accomplished using two different methods: patch-based segmentation and semantic segmentation. The networks used for patch-based segmentation were AlexNet, GoogLeNet, and ResNet50, whereas semantic segmentation was accomplished using SegNet, VGG16-weighted SegNet, and U-Net. The precision and Jaccard index were calculated, and the extraction accuracy of the six convolutional network (DCNN) systems was evaluated. The highest precision (0.974) was obtained with the VGG16-weighted SegNet, and the lowest precision (0.506) was obtained with ResNet50. Based on the data, calculation times, and Jaccard indices obtained in this study, segmentation on a 2D image may be considered a viable and effective approach. We found that the optimal automatic segmentation of organs (brainstem, corpus callosum, pituitary, cerebrum, and cerebellum) on brain sagittal T1-weighted images could be achieved using SegNet with VGG16.

Exploration of an odorous aldehydes and ketones produced by Uroglena americana using high resolution mass spectrometry, GC-Olfactometry, and multivariate analysis.

Off-flavor events in tap water have been reported from various regions of Japan. Fishy smell is the second most common off-flavor in Japan and Uroglena americana (U. americana) is known to be a major contributor to the smell. However, the causative compound of the smell it produces still remains unrevealed to the best of our knowledge. In this study, an exploration of odorous aldehydes and ketones originating from U. americana was performed with a view to discovering a possible candidate substance of causative compounds. Environmental samples containing U. americana colony and cultured media with U. americana were analyzed with two high resolution mass spectrometers, one of them is coupled with liquid chromatography (LC-HRMS), and the other is with gas chromatography and a sniffing port (GC-O-HRMS). Multivariate analyses (MVA) were utilized to explore a compound that is likely to be odorous aldehydes or ketones with a reduced time of exploration. A combination of LC-HRMS and MVA resulted in the selection of one candidate substance and its formula was determined to be C13H20O3 on the basis of its accurate mass and natural isotopic pattern. The candidate substance underwent GC-O-HRMS analyses and milk-like smell was detected at around its retention time. Although the detected smell was different from fishy smell, it is expected that the fishy smell is caused by multiple compounds to which the candidate substance belongs. First generation product ion spectra of the candidate substance suggested that it contains a hydroxyl group, a cyclohexene ring, and a ketone moiety.

Evaluation of capillary blood volume in the lower limb muscles after exercise by intravoxel incoherent motion.

PURPOSE: The purpose of this study was to evaluate the change in capillary blood volume in the muscles of the lower limbs, before and after exercise, using intravoxel incoherent motion (IVIM). MATERIALS AND METHODS: Diffusion-weighted images were obtained (with 16 b values) from the thigh and leg muscles of 11 subjects. The imaging was obtained before and immediately exercise; additionally, imaging was performed at 3 h, 6 h, and 24 h after exercise, and an IVIM index was calculated. The exercise involved walking up and down a flight of stairs (140 steps) ten times. The IVIM of each time course-before and after the exercise-was compared. In addition, we examined the correlation of IVIM measurements with the degree of the muscle ache that occurred at 24 h following the exercise. RESULTS: The IVIM index significantly increased after exercise compared with that before exercise (P < 0.01). IVIM decreased at 3 h following exercise, but increased again at 24 h. A correlation was found between the IVIM index at 24 h after exercise and the degree of the muscle ache (r = 0.80) CONCLUSIONS: The capillary blood volume significantly increased after exercise when compared to before exercise. The capillary blood volume decreased after exercise at 3 h and 6 h following exercise, but it increased again at 24 h. There was a correlation between the degree of muscle ache and the amount of capillary blood volume measured from the femoral muscle at 24 h after exercise.

Evaluating medical images using deep convolutional neural networks: A simulated CT phantom image study.

BACKGROUND: Applied research on artificial intelligence, mainly in deep learning, is widely performed. If medical images can be evaluated using artificial intelligence, this could substantially improve examination efficiency. OBJECTIVE: We investigated an evaluation system for medical images with different noise characteristics using a deep convolutional neural network. METHODS: Simulated computed tomography images are the targets of the system. We used an AlexNet trained with natural images for the deep convolutional neural network and a support vector machine for classification. Synthetic computed tomography images with circular and rectangular signal bodies at different levels of contrast and added Gaussian noise were used for training and testing. RESULTS: Two transfer learning methods were tested: classification by a re-trained support vector machine using the AlexNet features, and a method that fine-tuned the deep convolutional neural network. Using the first method, all the test image noise levels could be classified correctly. The fine-tuning method achieved an accuracy rate of 92.6%. CONCLUSIONS: An image quality evaluation method using artificial intelligence will be useful for clinical images and different image quality indices in the future.

Development of a quantitative statistical analysis system for double inversion recovery (DIR) MRI: A preliminary clinical study.

BACKGROUND: Gray matter (GM) imaging is important in the investigation of many neurological diseases, including schizophrenia, multiple sclerosis, stroke, Alzheimer's disease, tuberous sclerosis, and epilepsy, which are all associated with changes in cortical GM. OBJECTIVE: The aim of this study was to develop a quantitative statistical analysis system for double inversion recovery (DIR) MRI and to evaluate the new system using preliminary clinical data. METHODS: The study population comprised of 10 healthy volunteers and six patients with or without brain degeneration. A quantitative statistical analysis system for DIR images was developed using the following steps: 1) brain spatial normalization, 2) mean and standard deviation (SD) map creation, and 3) Z-score map creation. To evaluate the new voxel-based morphometry system, Z-scores of lesions in patients with brain degeneration were measured and then compared with Z-scores of normal regions. RESULTS: All DIR images were adequately spatially normalized to Montreal Neurological Institute MNI coordinate. Lesions in each patient were indicated by high Z-score values on a Z-score map, which were significantly higher than Z-scores of normal regions (p< 0.05). CONCLUSIONS: In this study, we developed a quantitative statistical analysis system for DIR MRI. Using our system, clinicians might accurately diagnose early brain degeneration.

[Setting Method of the Optimal Slab of Brain MRA for Detective Unruptured Cerebral Aneurysm].

Generally, the imaging range of the brain magnetic resonance angiography (MRA) is determined in the subjectivity by the operator used by the plan image that a blood vessel is not depicted. However, a necessary blood vessel may not be often depicted by an error of the setting of the imaging area. Therefore, optimal slab angle, thickness, distance, and image contrast for depiction of the unruptured cerebral aneurysm were examined. The brain MRA of 14 subjects was imaged in a wide area parallel to an orbitomeatal line (OM) line. The line which linked the arteria vertebral (the first cervical vertebrae curved section) to anterior cerebral artery (A3) was determined with an optimal slab base line, and the angle with the OM line was evaluated. Moreover, slavic range including the unruptured aneurysm was calculated. In addition, the distance from the inferior margin of pons to the slavic bottom end was measured. Furthermore, the cerebrovascular contrast by the slave angle was compared. As a result, the slave setting of the range was recommended in brain-MRA as an angle was 34.3 degrees, and the thickness was 56.4 mm, and, as the distance from the inferior margin of pons was 27.6 mm. The cerebrovascular contrast of the optimal slab base line angle did not have a significant difference for an OM line.

[Verification of "IEC 62464-12007": The Evaluation Method of Spatial Resolution in MRI].

Recently, many methods are suggested to evaluate spatial resolution in MRI. However, those techniques are not simple and easy. The International Electrotechnical Commission (IEC) recommends a method to evaluate spatial resolution using a periodic pattern image as IEC 62464-1. IEC 62464-1 prescribes specifications and placement of phantom, and a method of analysis, but these details grounds are not clear. A purpose of this study is to examine the effect in each factor of IEC 62464-1 method and define the characteristics of this method. Nine phantoms with different plate thickness were made including prescribed specifications of IEC 62464-1. Imaging was conducted with changing the placement angle of these phantoms. Also, analysis was carried out in region of interest (ROI) of three different size. As a result, the placement angle of the phantom, measurement error was <1% on a condition prescribed by a method of IEC 62464-1. There was not the effect if the transverse diameter for the longitudinal diameter exceeded 100% fort the size of ROI. In specifications of the phantom, there was not the dependence for the thickness of the plate of the phantom in IEC 62464-1 prescribes.

[Improvement of Fat Suppressing Effect for Finger and Neck MRI by Using Small Glass Beads].

The defectiveness of the fat suppression becomes the factor of the decrease of the quality of the diagnosis of magnetic resonance imaging. It is reported that the use of magnetic field uniformity adjuvant pad is effective for reduce poor fat suppression. The ball bullets, polystyrene balls, and polished rice are used for pad packing material, in recently, it was reported that fat suppression effect was good by the use of the small glass beads. Therefore, we tested the utility of small glass beads pad in the neck and fingers in this study. Neck and the fingers of subjects were imaged with T1-weighted image with fat suppression and T1-high resolution isotropic volume excitation image. The fat suppression effect of each image was compared with the polished rice and glass beads as material of pad used by physical, observation, and contact evaluation. In the result, satisfactory results were obtained by using glass beads, and it is suggested that fat suppression effect is improved by using glass beads as a filling material of pad in clinical study as a conclusion.

Slow component apparent diffusion coefficient for prostate cancer: Comparison and correlation with pharmacokinetic evaluation from dynamic contrast-enhanced MR imaging.

PURPOSE: Dynamic contrast enhancement (DCE)-MRI has high diagnostic performance of prostate cancer. However, it is preferable to avoid the use of MRI contrast media. A study reported that the diagnosability of the wash-in index of DCE-MRI was equivalent to the intravoxel incoherent motion of the diffusion weighted image. The purpose of this study was to examine the correlation between the slow component apparent diffusion coefficient (ADC) and the wash-out index of the DCE. MATERIALS AND METHODS: Thirty-eight patients diagnosed with prostate cancer by biopsy were enrolled in this study. The fast and slow component ADCs of the DWI were calculated for 76 points of the tumor and the contralateral normal parts. Furthermore, the wash-in and wash-out indices of the DCE-MRI were calculated. The correlations for each calculated index were compared. RESULTS: There was a significant difference between the tumor and the contralateral normal parts for both fast (p = 0.03) and slow component (p < 0.01) ADCs. In addition, the slow component ADC was correlated with the wash-out index (r = 0.64). CONCLUSION: The slow component ADC was correlated with the wash-out index, and may, therefore, be a suitable substitute for DCE-MRI.

[Capillary Blood Flow of Muscle Before and After the Exercise Used by Intravoxel Incoherent Motion].

Capillaries are the most basic and important blood vessel of the circulatory systems. The evaluation of the blood flow may contribute to many studies in future. We evaluated the capillary blood flow change of lower limb muscle over time before and after the exercise used by magnetic resonance imaging-intravoxel incoherent motion (MRI-IVIM) obtained perfusion information. Furthermore, we examined an association between the muscle pain after the exercise and the diffusion weighted image (DWI) indexes. DWI was imaged using multi-b values for a thigh and calf muscles. MRI was performed just after an exercise test, 3, 6, and 24 hours later, and the IVIM index and diffusion index were calculated. Furthermore, we interviewed the degree of the muscle ache 24 hours later. As a result, pseudo diffusion coefficient (D*) and f value as IVIM index increased after-exercise as compared with pre-exercise and decreased in 3 hours later. A similar tendency was found in the apparent diffusion coefficient and the diffusion coefficient as diffusion index. Furthermore, all indexes increased in after exercise from before exercise and decreased with time passed and increased again 24 hours later. In conclusion, IVIM could obtain capillary blood flow information, and it was suggested to contribute for sports medicine in future.

[The Novel Evaluation Method of Geometric Distortion for 3D-MRI].

Recently, many imaging methods are developed in magnetic resonance imaging and in particular, the release of the 3D acquisition method is remarkable. The image distortion often becomes the problem by the gradient echo method and the echo planar imaging (EPI) -based method, but the image distortion of the 3D acquisition is not established. A purpose of this study is to examine the utility of the novel evaluation method of the image distortion for the 3D acquisition image. The spin echo image was used as a criteria image, and EPI was used as a target image for 3D acquisition imaging. Image J was used for the image processing. The value that divided the volume of the different shape of criteria image and the target image by the volume of the criteria image was defined as a distortion rate. Hence, we refer this method to the volume method. It is reported that the distortion rate of the EPI is proportional to a rectangle rate of rectangular field of view (RFOV). The distortion rate of the volume method for 50-100% of rectangle ratio was compared with the theoretical value. In addition, the dependence by the threshold of the binarization was considered. Furthermore, the comparison with the NEMA (National Electrical Manufacturers Association) method was carried out. As a result, the distortion rate decreased according to a rectangular rate of RFOV by the volume method, and the numerical value was equal with a theoretical value. In addition, the distortion rate did not have the effect by the thresholding of binarizing. The volume method had less error of measurement than the NEMA method.

Comparison of medical image classification accuracy among three machine learning methods.

BACKGROUND: Low-quality medical images may influence the accuracy of the machine learning process. OBJECTIVE: This study was undertaken to compare accuracy of medical image classification among machine learning methods, as classification is a basic aspect of clinical image inspection. METHODS: Three types of machine learning methods were used, which include Support Vector Machine (SVM), Artificial Neural Network (ANN), and Convolution Neural Network (CNN). To investigate changes in accuracy related to image quality, we constructed a single dataset using two different file formats of DICOM (Digital Imaging and Communications in Medicine) and JPEG (Joint Photographic Experts Group). RESULTS: The JPEG format contains less color information and data capacity than the DICOM format. CNN classification was accurate for both datasets, whereas SVM and ANN accuracy decreased with the loss of data from DICOM to JPEG formats. CONCLUSIONS: CNN is more accurate than conventional machine learning methods that utilize the manual feature extraction.

[Effect of Slice Thickness for Apparent Diffusion Coefficient Measurement of Mass].

Apparent diffusion coefficient (ADC) values calculated from diffusion-weighted magnetic resonance imaging (DW-MRI) can be used for differentiation of tumors. Clinically, ADC values are used for monitoring treatment response after chemotherapy or radiation. However, it is reported that the threshold of the ADC value differs among institutions. In addition, there are reports regarding the change factor of the ADC value. Slice thickness may induce error in the ADC value by the influence of the partial volume effect in thicker objects, and by the influence of signal-to-noise ratio (SNR) in thinner objects. Therefore, in this study, the effect of slice thickness was examined. The signal body of spherical high-diffusion coefficients of 6, 7.9, and 9.3 mm in diameter was fixed in the low-circumference material of the diffusion coefficient. These phantoms were imaged using DW imaging (DWI) of 1, 1.5, 2, 2.5, 3, 4, 5, 6, 7, 8, 9, 10, 15, and 20 mm slice thickness using the multi-b values. In addition, different SNR were imaged by changing field-of-view and the number of additions. ADC was calculated by DWI of the different b values. As a result, slice thickness showed a peak at 50-65% of the diameter of the signal body. Furthermore, ADC values fluctuated in the slice thickness in front of the peak with a change in SNR. In conclusion, the ADC value was most accurate at a setting of 50-65% of slice thickness for the object diameter.

Effectiveness of an e-Learning Platform for Image Interpretation Education of Medical Staff and Students.

Medical staff must be able to perform accurate initial interpretations of radiography to prevent diagnostic errors. Education in medical image interpretation is an ongoing need that is addressed by text-based and e-learning platforms. The effectiveness of these methods has been previously reported. Here, we describe the effectiveness of an e-learning platform used for medical image interpretation education. Ten third-year medical students without previous experience in chest radiography interpretation were provided with e-learning instructions. Accuracy of diagnosis using chest radiography was provided before and after e-learning education. We measured detection accuracy for two image groups: nodular shadow and ground-glass shadow. We also distributed the e-learning system to the two groups and analyzed the effectiveness of education for both types of image shadow. The mean correct answer rate after the 2-week e-learning period increased from 34.5 to 72.7%. Diagnosis of the ground glass shadow improved significantly more than that of the mass shadow. Education using the e-leaning platform is effective for interpretation of chest radiography results. E-learning is particularly effective for the interpretation of chest radiography images containing ground glass shadow.

[Comparison of Organ Dose Calculation Using Monte Carlo Simulation and In-phantom Dosimetry in CT Examination].

Direct measurement of each patient organ dose from computed tomography (CT) is not possible. Most methods to estimate patient organ dose is using Monte Carlo simulation with dedicated software. However, the method and the relative differences between organ dose simulation and measurement is unclear. The purpose of this study was to compare organ doses evaluated by Monte Carlo simulation with doses evaluated by in-phantom dosimetry. The simulation software Radimetrics (Bayer) was used for the calculation of organ dose. Measurement was performed with radio-photoluminescence glass dosimeter (RPLD) set at various organ positions within RANDO phantom. To evaluate difference of CT scanner, two different CT scanners were used in this study. Angular dependence of RPLD and measurement of effective energy were performed for each scanner. The comparison of simulation and measurement was evaluated by relative differences. In the results, angular dependence of RPLD at two scanners was 31.6±0.45 mGy for SOMATOM Definition Flash and 29.2±0.18 mGy for LightSpeed VCT. The organ dose was 42.2 mGy (range, 29.9-52.7 mGy) by measurements and 37.7 mGy (range, 27.9-48.1 mGy) by simulations. The relative differences of organ dose between measurement and simulation were 13%, excluding of breast's 42%. We found that organ dose by simulation was lower than by measurement. In conclusion, the results of relative differences will be useful for evaluating organ doses for individual patients by simulation software Radimetrics.

The Yellow Scale Is Superior to the Gray Scale for Detecting Acute Ischemic Stroke on a Monitor Display in Computed Tomography.

RATIONALE AND OBJECTIVES: The purpose of this study was to compare the efficacy of the color scale with regard to focal detection with computed tomography in acute ischemic stroke. MATERIALS AND METHODS: Computed tomography images of the brain of 19 patients diagnosed with acute stroke, based on magnetic resonance diffusion-weighted images obtained within an onset of 24 hours, and the images of five normal patients were displayed in each color look-up table on a monitor. The detection of acute stroke was compared among 15 radiologists. The images were compared in the gray, green, yellow, red, and blue scales of the look-up tables. The observers recorded acute ischemic stroke as "present" or "absent." They also located the position of the stroke lesion and described the degree of their conviction as to whether a lesion existed. Detection was evaluated by receiver operating characteristic analysis. The area under the receiver operating characteristic curves was compared. In addition, reduced fatigue and the ease in image observation were compared. RESULTS: Compared to the other scales, the yellow scale had a significantly higher area under the receiver operating characteristic curve, which indicated that this scale allowed better detection of acute ischemic stroke. The gray scale produced the least fatigue in image observation. CONCLUSIONS: The detection of acute ischemic stroke is improved by changing the display monitor from the gray scale to the yellow scale. From the perspective of color psychology, yellow is associated with higher arousal, cheerfulness, confidence, creativity, and excitement. Therefore, the yellow scale may be suitable for a medical imaging display.