Search

Purpose To develop an anomaly detection system in PET/CT with the tracer 18F-fluorodeoxyglucose (FDG) that requires only normal PET/CT images for training and can detect abnormal FDG uptake at any location in the chest region. Materials and methods We trained our model based on a Bayesian deep learning framework using 1878 PET/CT scans with no abnormal findings. Our model learns the distribution of standard uptake values in these normal training images and detects out-of-normal uptake regions. We evaluated this model using 34 scans showing focal abnormal FDG uptake in the chest region. This evaluation dataset includes 28 pulmonary and 17 extrapulmonary abnormal FDG uptake foci. We performed per-voxel and per-slice receiver operating characteristic (ROC) analyses and per-lesion free-response receiver operating characteristic analysis. Results Our model showed an area under the ROC curve of 0.992 on discriminating abnormal voxels and 0.852 on abnormal slices. Our model detected 41 of 45 (91.1%) of the abnormal FDG uptake foci with 12.8 false positives per scan (FPs/scan), which include 26 of 28 pulmonary and 15 of 17 extrapulmonary abnormalities. The sensitivity at 3.0 FPs/scan was 82.2% (37/45). Conclusion Our model trained only with normal PET/CT images successfully detected both pulmonary and extrapulmonary abnormal FDG uptake in the chest region.

The three-dimensional (3D) voxel labeling of lesions requires significant radiologists’ effort in the development of computer-aided detection software. To reduce the time required for the 3D voxel labeling, we aimed to develop a generalized semiautomatic segmentation method based on deep learning via a data augmentation-based domain generalization framework. In this study, we investigated whether a generalized semiautomatic segmentation model trained using two types of lesion can segment previously unseen types of lesion. We targeted lung nodules in chest CT images, liver lesions in hepatobiliary-phase images of Gd-EOB-DTPA-enhanced MR imaging, and brain metastases in contrast-enhanced MR images. For each lesion, the 32 × 32 × 32 isotropic volume of interest (VOI) around the center of gravity of the lesion was extracted. The VOI was input into a 3D U-Net model to define the label of the lesion. For each type of target lesion, we compared five types of data augmentation and two types of input data. For all considered target lesions, the highest dice coefficients among the training patterns were obtained when using a combination of the existing data augmentation-based domain generalization framework and random monochrome inversion and when using the resized VOI as the input image. The dice coefficients were 0.639 ± 0.124 for the lung nodules, 0.660 ± 0.137 for the liver lesions, and 0.727 ± 0.115 for the brain metastases. Our generalized semiautomatic segmentation model could label unseen three types of lesion with different contrasts from the surroundings. In addition, the resized VOI as the input image enables the adaptation to the various sizes of lesions even when the size distribution differed between the training set and the test set.

Objective This study aimed to determine a standardized cut-off value for abnormal 18F-fluorodeoxyglucose (FDG) accumulation in the thyroid gland. Methods Herein, 7013 FDG–PET/CT scans were included. An automatic thyroid segmentation method using two U-nets (2D- and 3D-U-net) was constructed; mean FDG standardized uptake value (SUV), CT value, and volume of the thyroid gland were obtained from each participant. The values were categorized by thyroid function into three groups based on serum thyroid-stimulating hormone levels. Thyroid function and mean SUV with increments of 1 were analyzed, and risk for thyroid dysfunction was calculated. Thyroid dysfunction detection ability was examined using a machine learning method (LightGBM, Microsoft) with age, sex, height, weight, CT value, volume, and mean SUV as explanatory variables. Results Mean SUV was significantly higher in females with hypothyroidism. Almost 98.9% of participants in the normal group had mean SUV < 2 and 93.8% participants with mean SUV < 2 had normal thyroid function. The hypothyroidism group had more cases with mean SUV ≥ 2. The relative risk of having abnormal thyroid function was 4.6 with mean SUV ≥ 2. The sensitivity and specificity for detecting thyroid dysfunction using LightGBM (Microsoft) were 14.5 and 99%, respectively. Conclusions Mean SUV ≥ 2 was strongly associated with abnormal thyroid function in this large cohort, indicating that mean SUV with FDG–PET/CT can be used as a criterion for thyroid evaluation. Preliminarily, this study shows the potential utility of detecting thyroid dysfunction based on imaging findings.

The performance of computer-aided detection (CAD) software depends on the quality and quantity of the dataset used for machine learning. If the data characteristics in development and practical use are different, the performance of CAD software degrades. In this study, we investigated changes in detection performance due to differences in training data for cerebral aneurysm detection software in head magnetic resonance angiography images. We utilized three types of CAD software for cerebral aneurysm detection in MRA images, which were based on 3D local intensity structure analysis, graph-based features, and convolutional neural network. For each type of CAD software, we compared three types of training pattern, which were two types of training using single-site data and one type of training using multisite data. We also carried out internal and external evaluations. In training using single-site data, the performance of CAD software largely and unpredictably fluctuated when the training dataset was changed. Training using multisite data did not show the lowest performance among the three training patterns for any CAD software and dataset. The training of cerebral aneurysm detection software using data collected from multiple sites is desirable to ensure the stable performance of the software.

To evaluate the spatial patterns of missed lung nodules in a real-life routine screening environment.In a screening institute, 4,822 consecutive adults underwent chest CT, and each image set was independently interpreted by two radiologists in three steps: (1) independently interpreted without computer-assisted detection (CAD) software, (2) independently referred to the CAD results, (3) determined by the consensus of the two radiologists. The locations of nodules and the detection performance data were semi-automatically collected using a CAD server integrated into the reporting system. Fisher's exact test was employed for evaluating findings in different lung divisions. Probability maps were drawn to illustrate the spatial distribution of radiologists' missed nodules.Radiologists significantly tended to miss lung nodules in the bilateral hilar divisions (p0.01). Some radiologists had their own spatial pattern of missed lung nodules.Radiologists tend to miss lung nodules present in the hilar regions significantly more often than in the rest of the lung.

The purposes of this study are to propose an unsupervised anomaly detection method based on a deep neural network (DNN) model, which requires only normal images for training, and to evaluate its performance with a large chest radiograph dataset. We used the auto-encoding generative adversarial network (α-GAN) framework, which is a combination of a GAN and a variational autoencoder, as a DNN model. A total of 29,684 frontal chest radiographs from the Radiological Society of North America Pneumonia Detection Challenge dataset were used for this study (16,880 male and 12,804 female patients; average age, 47.0 years). All these images were labeled as “Normal,” “No Opacity/Not Normal,” or “Opacity” by board-certified radiologists. About 70% (6,853/9,790) of the Normal images were randomly sampled as the training dataset, and the rest were randomly split into the validation and test datasets in a ratio of 1:2 (7,610 and 15,221). Our anomaly detection system could correctly visualize various lesions including a lung mass, cardiomegaly, pleural effusion, bilateral hilar lymphadenopathy, and even dextrocardia. Our system detected the abnormal images with an area under the receiver operating characteristic curve (AUROC) of 0.752. The AUROCs for the abnormal labels Opacity and No Opacity/Not Normal were 0.838 and 0.704, respectively. Our DNN-based unsupervised anomaly detection method could successfully detect various diseases or anomalies in chest radiographs by training with only the normal images.

Medical artificial intelligence (AI) systems need to learn to recognize synonyms or paraphrases describing the same anatomy, disease, treatment, etc. to better understand real-world clinical documents. Existing linguistic resources focus on variants at the word or sentence level. To handle linguistic variations on a broader scale, we proposed the Medical Text Radiology Report section Japanese version (MedTxt-RR-JA), the first clinical comparable corpus. MedTxt-RR-JA was built by recruiting nine radiologists to diagnose the same 15 lung cancer cases in Radiopaedia, an open-access radiological repository. The 135 radiology reports in MedTxt-RR-JA were shown to contain word-, sentence- and document-level variations maintaining similarity of contents. MedTxt-RR-JA is also the first publicly available Japanese radiology report corpus that would help to overcome poor data availability for Japanese medical AI systems. Moreover, our methodology can be applied widely to building clinical corpora without privacy concerns.

Purpose To build a novel, open-source, purely web-based platform system to address problems in the development and clinical use of computer-assisted detection/diagnosis (CAD) software. The new platform system will replace the existing system for the development and validation of CAD software, Clinical Infrastructure for Radiologic Computation of United Solutions (CIRCUS). Methods In our new system, the two top-level applications visible to users are the web-based image database (CIRCUS DB; database) and the Docker plug-in-based CAD execution platform (CIRCUS CS; clinical server). These applications are built on top of a shared application programming interface server, a three-dimensional image viewer component, and an image repository. Results We successfully installed our new system into a Linux server at two clinical sites. A total of 1954 cases were registered in CIRCUS DB. We have been utilizing CIRCUS CS with four Docker-based CAD plug-ins. Conclusions We have successfully built a new version of the CIRCUS system. Our platform was successfully implemented at two clinical sites, and we plan to publish it as an open-source software project.

To develop an anomaly detection system in PET/CT with the tracerWe trained our model based on a Bayesian deep learning framework using 1878 PET/CT scans with no abnormal findings. Our model learns the distribution of standard uptake values in these normal training images and detects out-of-normal uptake regions. We evaluated this model using 34 scans showing focal abnormal FDG uptake in the chest region. This evaluation dataset includes 28 pulmonary and 17 extrapulmonary abnormal FDG uptake foci. We performed per-voxel and per-slice receiver operating characteristic (ROC) analyses and per-lesion free-response receiver operating characteristic analysis.Our model showed an area under the ROC curve of 0.992 on discriminating abnormal voxels and 0.852 on abnormal slices. Our model detected 41 of 45 (91.1%) of the abnormal FDG uptake foci with 12.8 false positives per scan (FPs/scan), which include 26 of 28 pulmonary and 15 of 17 extrapulmonary abnormalities. The sensitivity at 3.0 FPs/scan was 82.2% (37/45).Our model trained only with normal PET/CT images successfully detected both pulmonary and extrapulmonary abnormal FDG uptake in the chest region.

Background It is essential for radiologists to communicate actionable findings to the referring clinicians reliably. Natural language processing (NLP) has been shown to help identify free-text radiology reports including actionable findings. However, the application of recent deep learning techniques to radiology reports, which can improve the detection performance, has not been thoroughly examined. Moreover, free-text that clinicians input in the ordering form (order information) has seldom been used to identify actionable reports. This study aims to evaluate the benefits of two new approaches: (1) bidirectional encoder representations from transformers (BERT), a recent deep learning architecture in NLP, and (2) using order information in addition to radiology reports. Methods We performed a binary classification to distinguish actionable reports (i.e., radiology reports tagged as actionable in actual radiological practice) from non-actionable ones (those without an actionable tag). 90,923 Japanese radiology reports in our hospital were used, of which 788 (0.87%) were actionable. We evaluated four methods, statistical machine learning with logistic regression (LR) and with gradient boosting decision tree (GBDT), and deep learning with a bidirectional long short-term memory (LSTM) model and a publicly available Japanese BERT model. Each method was used with two different inputs, radiology reports alone and pairs of order information and radiology reports. Thus, eight experiments were conducted to examine the performance. Results Without order information, BERT achieved the highest area under the precision-recall curve (AUPRC) of 0.5138, which showed a statistically significant improvement over LR, GBDT, and LSTM, and the highest area under the receiver operating characteristic curve (AUROC) of 0.9516. Simply coupling the order information with the radiology reports slightly increased the AUPRC of BERT but did not lead to a statistically significant improvement. This may be due to the complexity of clinical decisions made by radiologists. Conclusions BERT was assumed to be useful to detect actionable reports. More sophisticated methods are required to use order information effectively.

Recently, deep learning has been exploited in the field of medical image analysis. However, the training of deep learning models with medical images is time-consuming since most medical image data are three-dimensional volumes or high-resolution two-dimensional images. Moreover, the optimization of numerous hyperparameters strongly affects the performance of deep learning. If a framework for training deep learning with hyperparameter optimization on a supercomputer system can be realized, it is expected to accelerate the training of deep learning with medical images. In this study, we described our novel environment for training deep learning with medical images on the supercomputer system in our institute (Reedbush-H supercomputer system) based on asynchronous parallel Bayesian optimization. We trained two types of automated lesion detection application in a constructed environment. The constructed environment enabled us to train deep learning with hyperparameter tuning in a short time.

Gadolinium-ethoxybenzyl-diethylenetriamine pentaacetic acid (Gd-EOB-DTPA)-enhanced magnetic resonance imaging (MRI) tends to show higher diagnostic accuracy than other modalities. There is a demand for computer-assisted detection (CAD) software for Gd-EOB-DTPA-enhanced MRI. Segmentation with high accuracy is important for CAD software. We propose a liver segmentation method for Gd-EOB-DTPA-enhanced MRI that is based on a four-dimensional (4D) fully convolutional residual network (FC-ResNet). The aims of this study are to determine the best combination of an input image and output image in our proposed method and to compare our proposed method with the previous rule-based segmentation method. We prepared a five-phase image set and a hepatobiliary phase image set as the input image sets to determine the best input image set. We also prepared a labeled liver image and labeled liver and labeled body trunk images as the output image sets to determine the best output image set. In addition, we optimized the hyperparameters of our proposed model. We used 30 cases to train our model, 10 cases to determine the hyperparameters of our model, and 20 cases to evaluate our model. Our network with the five-phase image set and the output image set of labeled liver and labeled body trunk images showed the highest accuracy. Our proposed method showed higher accuracy than the previous rule-based segmentation method. The Dice coefficient of the liver region was 0.944 ± 0.018. Our proposed 4D FC-ResNet showed satisfactory performance for liver segmentation as preprocessing in CAD software.

A novel image feature set named histogram of triangular paths in graph (HoTPiG) is presented. The purpose of this study is to evaluate the feasibility of the proposed HoTPiG feature set through two clinical computer-aided detection tasks: nodule detection in lung CT images and aneurysm detection in head MR angiography images. The HoTPiG feature set is calculated from an undirected graph structure derived from a binarized volume. The features are derived from a 3-D histogram in which each bin represents a triplet of shortest path distances between the target node and all possible node pairs near the target node. First, the vessel structure is extracted from CT/MR volumes. Then, a graph structure is extracted using an 18-neighbor rule. Using this graph, a HoTPiG feature vector is calculated at every foreground voxel. After explicit feature mapping with an exponential-χ2 kernel, each voxel is judged by a linear support vector machine classifier. The proposed method was evaluated using 300 CT and 300 MR datasets. The proposed method successfully detected lung nodules and cerebral aneurysms. The sensitivity was about 80% when the number of false positives was three per case for both applications. The HoTPiG image feature set was presented, and its high general versatility was shown through two medical lesion detection applications.

To assess the clinical effectiveness of temporal subtraction computed tomography (TS CT) using deep learning to improve vertebral bone metastasis detection.This retrospective study used TS CT comprising bony landmark detection, bone segmentation with a multi-atlas-based method, and spatial registration of two images by a log-domain diffeomorphic Demons algorithm. Paired current and past CT images of 50 patients without vertebral metastasis, recorded during June 2011-September 2016, were included as training data. A deep learning-based method estimated registration errors and suppressed false positives. Thereafter, paired CT images of 40 cancer patients with newly developed vertebral metastases and 40 control patients without vertebral metastases were evaluated. Six board-certified radiologists and five radiology residents independently interpreted 80 paired CT images with and without TS CT.Records of 40 patients in the metastasis group (median age: 64.5 years; 20 males) and 40 patients in the control group (median age: 64.0 years; 20 males) were evaluated. With TS CT, the overall figure of merit (FOM) of the board-certified radiologist and resident groups improved from 0.848 to 0.876 (p = 0.01) and from 0.752 to 0.799 (p = 0.02), respectively. The sub-analysis focusing on attenuation changes in lesions revealed that the FOM of osteoblastic lesions significantly improved in both the board-certified radiologist and resident groups using TS CT. The sub-analysis focusing on lesion location showed that the FOM of the resident group significantly improved in the vertebral arch (p = 0.04).TS CT was effective in detecting bone metastasis by both board-certified radiologists and radiology residents.

Radiologists interpret many medical images and clinical practice demands timely interpretation, resulting in a heavy workload. To reduce the workload, here we formulate and validate a method that can handle different types of medical image and can detect virtually all types of lesion in a medical image. For the first time, we show that two flow-based deep generative (FDG) models can predict the logarithm posterior probability in a semi-supervised approach. We adopt two FDG models in conjunction with Bayes’ theorem to predict the logarithm posterior probability that a medical image is normal. We trained one of the FDG models with normal images and the other FDG model with normal and non-normal images. We validated the method using two types of medical image: chest X-ray images (CXRs) and brain computed tomography images (BCTs). The area under the receiver operating characteristic curve for pneumonia-like opacities in CXRs was 0.839 on average, and for infarction in BCTs was 0.904. We formulated a method of predicting the logarithm posterior probability using two FDG models. We validated that the method can detect abnormal findings in CXRs and BCTs with both an acceptable performance for testing and a comparatively light workload for training.

The clinical usefulness of computer-aided detection of cerebral aneurysms has been investigated using different methods to present lesion candidates, but suboptimal methods may have limited its usefulness. We compared three presentation methods to determine which can benefit radiologists the most by enabling them to detect more aneurysms. We conducted a multireader multicase observer performance study involving six radiologists and using 470 lesion candidates output by a computer-aided detection program, and compared the following three different presentation methods using the receiver operating characteristic analysis: (1) a lesion candidate is encircled on axial slices, (2) a lesion candidate is overlaid on a volume-rendered image, and (3) combination of (1) and (2). The response time was also compared. As compared with axial slices, radiologists showed significantly better detection performance when presented with volume-rendered images. There was no significant difference in response time between the two methods. The combined method was associated with a significantly longer response time, but had no added merit in terms of diagnostic accuracy. Even with the aid of computer-aided detection, radiologists overlook many aneurysms if the presentation method is not optimal. Overlaying colored lesion candidates on volume-rendered images can help them detect more aneurysms.

Muography is a novel method of visualizing the internal structures of active volcanoes by using high-energy near-horizontally arriving cosmic muons. The purpose of this study is to show the feasibility of muography to forecast the eruption event with the aid of the convolutional neural network (CNN). In this study, seven daily consecutive muographic images were fed into the CNN to compute the probability of eruptions on the eighth day, and our CNN model was trained by hyperparameter tuning with the Bayesian optimization algorithm. By using the data acquired in Sakurajima volcano, Japan, as an example, the forecasting performance achieved a value of 0.726 for the area under the receiver operating characteristic curve, showing the reasonable correlation between the muographic images and eruption events. Our result suggests that muography has the potential for eruption forecasting of volcanoes.

Evaluate the effect of vaginal delivery on pelvic organ positions and vaginal cross-sectional areas.MRI of 119 premenopausal women were grouped according to the number of deliveries. The distances from the three 3-reference points (bladder, uterus, and rectum) to two 2-lines (pubococcygeal-line (PCL) and midpubic-line (MPL)), length of H- and M-lines and vaginal cross-sectional area were compared between the groups.With increasing parity, distance from the rectum to PCL tended to increase (nullipara vs. bipara; p0.01). Vaginal cross-sectional area was larger in bipara and tripara than in nullipara (p0.01).Rectal position is more caudally located and vaginal cross-sectional area is larger in bipara than in nullipara.

To compare the quality of helical computed tomography (CT) images of the pelvis in patients with metal hip prostheses reconstructed using adaptive iterative dose reduction (AIDR) and AIDR with single-energy metal artifact reduction (SEMAR-A). This retrospective study included 28 patients (mean age, 64.6 ± 11.4 years; 6 men and 22 women). CT images were reconstructed using AIDR and SEMAR-A. Two radiologists evaluated the extent of metal artifacts and the depiction of structures in the pelvic region and looked for mass lesions. A radiologist placed a region of interest within the bladder and recorded CT attenuation. The metal artifacts were significantly reduced in SEMAR-A as compared to AIDR (p

To compare the image quality of high-resolution computed tomography (HRCT) for evaluating lung nodules reconstructed with the new version of model-based iterative reconstruction and spatial resolution preference algorithm (MBIRn) vs. conventional model-based iterative reconstruction (MBIRc) and adaptive statistical iterative reconstruction (ASIR).This retrospective clinical study was approved by our institutional review board and included 70 lung nodules in 58 patients (mean age, 71.2±10.9years; 34 men and 24 women). HRCT of lung nodules were reconstructed using MBIRn, MBIRc and ASIR. Objective image noise was measured by placing the regions of interest on lung parenchyma. Two blinded radiologists performed subjective image analyses.Significant improvements in the following points were observed in MBIRn compared with ASIR (p0.005): objective image noise (24.4±8.0 vs. 37.7±10.4), subjective image noise, streak artifacts, and adequateness for evaluating internal characteristics and borders of nodules. The sharpness of small vessels and bronchi and diagnostic acceptability with MBIRn were significantly better than with MBIRc and ASIR (p0.008).HRCT reconstructed with MBIRn provides diagnostically more acceptable images for the detailed analyses of lung nodules compared with MBIRc and ASIR.

BACKGROUND AND PURPOSE: Experiences with computer-assisted detection of cerebral aneurysms in diagnosis by radiologists in real-life clinical environments have not been reported. The purpose of this study was to evaluate the usefulness of computer-assisted detection in a routine reading environment. MATERIALS AND METHODS: During 39 months in a routine clinical practice environment, 2701 MR angiograms were each read by 2 radiologists by using a computer-assisted detection system. Initial interpretation was independently made without using the detection system, followed by a possible alteration of diagnosis after referring to the lesion candidate output from the system. We used the final consensus of the 2 radiologists as the reference standard. The sensitivity and specificity of radiologists before and after seeing the lesion candidates were evaluated by aneurysm- and patient-based analyses. RESULTS: The use of the computer-assisted detection system increased the number of detected aneurysms by 9.3% (from 258 to 282). Aneurysm-based analysis revealed that the apparent sensitivity of the radiologists9 diagnoses made without and with the detection system was 64% and 69%, respectively. The detection system presented 82% of the aneurysms. The detection system more frequently benefited radiologists than being detrimental. CONCLUSIONS: Routine integration of computer-assisted detection with MR angiography for cerebral aneurysms is feasible, and radiologists can detect a number of additional cerebral aneurysms by using the detection system without a substantial decrease in their specificity. The low confidence of radiologists in the system may limit its usefulness.

Purpose The purpose of this study was to determine whether temporal subtraction (TS) computed tomography (CT) contributes to the detection of vertebral bone metastases. Method The calculation of TS CT was composed of bony landmark detection, bone segmentation with a multiatlas-based method, and spatial registration. Temporal increase and decrease of the CT values were visualized in blue and red, respectively. Paired CT images of 20 patients with cancer and newly-developed vertebral metastases were analyzed. Control CT examinations of 20 different patients were also included. The presence of vertebral metastases on the TS CT was evaluated by two board-certified radiologists. Five additional board-certified radiologists and five radiology residents independently interpreted the 40 paired CT images with and without TS CT. Results In the lesion conspicuity evaluation, 96% of vertebral metastases were scored as excellent or good. In the image interpretation examination, according to free-response receiver operating characteristics analysis, the overall figure of merit (FOM) of the board-certified radiologist group was 0.892 and 0.898 with and without TS CT, respectively. The FOM of the resident group improved from 0.849 to 0.902 with viewing TS CT. In the sub-analysis focusing on the location of the lesion, the FOM of the resident group significantly improved from 0.75 to 0.92 in vertebral arch lesions (p = 0.001). Conclusions The TS CT may be useful to detect vertebral metastases because almost all the vertebral metastases were shown to be favorable visualization. The TS CT was proven to be especially helpful for radiology residents in detecting vertebral arch metastases.

The reading workload for radiologists is increasing because the numbers of examinations and images per examination are increasing due to the technical progress on imaging modalities such as computed tomography and magnetic resonance imaging. A computer-assisted detection (CAD) system based on machine learning is expected to assist radiologists. The preliminary results of a multi-institutional study indicate that the performance of the CAD system for each institution improved using training data of other institutions. This indicates that transfer learning may be useful for developing the CAD systems among multiple institutions. In this paper, we focus on transfer learning without sharing training data due to the need to protect personal information in each institution. Moreover, we raise a problem of negative transfer in CAD system and propose an algorithm for inhibiting negative transfer. Our algorithm provides a theoretical guarantee for managing CAD software in terms of transfer learning and exhibits experimentally better performance compared to that of the current algorithm in cerebral aneurysm detection.

For the development of computer-assisted detection (CAD) software using voxel-based classification, gold standards defined by pixel-by-pixel painting, called painted gold standards, are desirable. However, for radiologists who define gold standards, a simplified method of definition is desirable. One of the simplest methods of defining gold standards is a spherical region, called a spherical gold standard. In this study, we investigated whether spherical gold standards can be used as an alternative to painted gold standards for computerized detection using voxel-based classification. The spherical gold standards were determined by the center of gravity and the maximum diameter. We compared two types of gold standard, painted gold standards and spherical gold standards, by two types of CAD software using voxel-based classification. The time required to paint the area of one lesion was 4.7–6.5 times longer than the time required to define a spherical gold standard. For the same performance of the CAD software, the number of training cases required for the spherical gold standard was 1.6–7.6 times that for the painted gold standards. Spherical gold standards can be used as an alternative to painted gold standards for the computerized detection of lesions with simple shapes.

Purpose: The target disorders of emergency head CT are wide-ranging. Therefore, people working in an emergency department desire a computer-aided detection system for general disorders. In this study, we proposed an unsupervised anomaly detection method in emergency head CT using an autoencoder and evaluated the anomaly detection performance of our method in emergency head CT. Methods: We used a 3D convolutional autoencoder (3D-CAE), which contains 11 layers in the convolution block and 6 layers in the deconvolution block. In the training phase, we trained the 3D-CAE using 10,000 3D patches extracted from 50 normal cases. In the test phase, we calculated abnormalities of each voxel in 38 emergency head CT volumes (22 abnormal cases and 16 normal cases) for evaluation and evaluated the likelihood of lesion existence. Results: Our method achieved a sensitivity of 68% and a specificity of 88%, with an area under the curve of the receiver operating characteristic curve of 0.87. It shows that this method has a moderate accuracy to distinguish normal CT cases to abnormal ones. Conclusion: Our method has potentialities for anomaly detection in emergency head CT.

Background The usefulness of computer-assisted detection (CAD) for detecting cerebral aneurysms has been reported; therefore, the improved performance of CAD will help to detect cerebral aneurysms. Purpose To develop a CAD system for intracranial aneurysms on unenhanced magnetic resonance angiography (MRA) images based on a deep convolutional neural network (CNN) and a maximum intensity projection (MIP) algorithm, and to demonstrate the usefulness of the system by training and evaluating it using a large dataset. Study Type Retrospective study. Subjects There were 450 cases with intracranial aneurysms. The diagnoses of brain aneurysms were made on the basis of MRA, which was performed as part of a brain screening program. Field Strength/Sequence Noncontrast-enhanced 3D time-of-flight (TOF) MRA on 3T MR scanners. Assessment In our CAD, we used a CNN classifier that predicts whether each voxel is inside or outside aneurysms by inputting MIP images generated from a volume of interest (VOI) around the voxel. The CNN was trained in advance using manually inputted labels. We evaluated our method using 450 cases with intracranial aneurysms, 300 of which were used for training, 50 for parameter tuning, and 100 for the final evaluation. Statistical Tests Free-response receiver operating characteristic (FROC) analysis. Results Our CAD system detected 94.2% (98/104) of aneurysms with 2.9 false positives per case (FPs/case). At a sensitivity of 70%, the number of FPs/case was 0.26. Data Conclusion We showed that the combination of a CNN and an MIP algorithm is useful for the detection of intracranial aneurysms. Level of Evidence: 4 Technical Efficacy Stage 1 J. Magn. Reson. Imaging 2017.

We propose a generalized framework for developing computer-aided detection (CADe) systems whose characteristics depend only on those of the training dataset. The purpose of this study is to show the feasibility of the framework. Two different CADe systems were experimentally developed by a prototype of the framework, but with different training datasets. The CADe systems include four components; preprocessing, candidate area extraction, candidate detection, and candidate classification. Four pretrained algorithms with dedicated optimization/setting methods corresponding to the respective components were prepared in advance. The pretrained algorithms were sequentially trained in the order of processing of the components. In this study, two different datasets, brain MRA with cerebral aneurysms and chest CT with lung nodules, were collected to develop two different types of CADe systems in the framework. The performances of the developed CADe systems were evaluated by threefold cross-validation. The CADe systems for detecting cerebral aneurysms in brain MRAs and for detecting lung nodules in chest CTs were successfully developed using the respective datasets. The framework was shown to be feasible by the successful development of the two different types of CADe systems. The feasibility of this framework shows promise for a new paradigm in the development of CADe systems: development of CADe systems without any lesion specific algorithm designing.

In this study, we propose a novel method of lung lesion detection in FDG-PET/CT volumes without labeling lesions. In our method, the probability distribution over normal standardized uptake values (SUVs) is estimated from the features extracted from the corresponding volume of interest (VOI) in the CT volume, which include gradient-based and texture-based features. To estimate the distribution, we use Gaussian process regression with an automatic relevance determination kernel, which provides the relevance of feature values to estimation. Our model was trained using FDG-PET/CT volumes of 121 normal cases. In the lesion detection phase, the actual SUV is judged as normal or abnormal by comparison with the estimated SUV distribution. According to the validation using 28 FDG-PET/CT volumes with 34 lung lesions, the sensitivity of the proposed method at 5.0 false positives per case was 81.9%.

Objective. To develop automatic visceral fat volume calculation software for computed tomography (CT) volume data and to evaluate its feasibility.Methods. A total of 24 sets of whole-body CT volume data and anthropometric measurements were obtained, with three sets for each of four BMI categories (under 20, 20 to 25, 25 to 30, and over 30) in both sexes. True visceral fat volumes were defined on the basis of manual segmentation of the whole-body CT volume data by an experienced radiologist. Software to automatically calculate visceral fat volumes was developed using a region segmentation technique based on morphological analysis with CT value threshold. Automatically calculated visceral fat volumes were evaluated in terms of the correlation coefficient with the true volumes and the error relative to the true volume.Results. Automatic visceral fat volume calculation results of all 24 data sets were obtained successfully and the average calculation time was 252.7 seconds/case. The correlation coefficients between the true visceral fat volume and the automatically calculated visceral fat volume were over 0.999.Conclusions. The newly developed software is feasible for calculating visceral fat volumes in a reasonable time and was proved to have high accuracy.

This chapter presents examples of medical image understanding algorithms using computational anatomy models explained in Chap. 2. After the introductory in Sect. 3.1, Sect. 3.2 shows segmentation algorithms for vertebrae, ribs, and hip joints. Segmentation algorithms for skeletal muscle and detection algorithms for lymph nodes are explained in Sects. 3.3 and 3.4, respectively. Section 3.5 deals with algorithms for understanding organs/tissues in the head and neck regions and starts with computational neuroanatomy, followed by analysis and segmentation algorithms for white matter, brain CT, oral regions, fundus oculi, and retinal optical coherence tomography (OCT). Algorithms useful in the thorax, specifically for the lungs, tracheobronchial tree, vessels, and interlobar fissures from a thoracic CT volume, are presented in Sect. 3.6. Section 3.7 provides algorithms for breast ultrasound imaging, i.e., mammography and breast MRI. Cardiac imaging algorithms in an echocardiographic image sequence and MR images as well as coronary arteries in a CT volume are explained in Sect. 3.8. Section 3.9 deals with segmentation algorithms of abdominal organs, including the liver, pancreas, spleen, kidneys, gastrointestinal tract, and abdominal blood vessels, followed by anatomical labeling of segmented vessels.

A fully automatic multiatlas-based method for segmentation of the spine and pelvis in a torso CT volume is proposed. A novel landmark-guided diffeomorphic demons algorithm is used to register a given CT image to multiple atlas volumes. This algorithm can utilize both grayscale image information and given landmark coordinate information optimally. The segmentation has four steps. Firstly, 170 bony landmarks are detected in the given volume. Using these landmark positions, an atlas selection procedure is performed to reduce the computational cost of the following registration. Then the chosen atlas volumes are registered to the given CT image. Finally, voxelwise label voting is performed to determine the final segmentation result. The proposed method was evaluated using 50 torso CT datasets as well as the public SpineWeb dataset. As a result, a mean distance error of $$0.59\pm 0.14\hbox { mm}$$ and a mean Dice coefficient of $$0.90\pm 0.02$$ were achieved for the whole spine and the pelvic bones, which are competitive with other state-of-the-art methods. From the experimental results, the usefulness of the proposed segmentation method was validated.

The purpose of this study is to evaluate the feasibility of a novel feature generation, which is based on multiple deep neural networks (DNNs) with boosting, for computer-assisted detection (CADe). It is hard and time-consuming to optimize the hyperparameters for DNNs such as stacked denoising autoencoder (SdA). The proposed method allows using SdA based features without the burden of the hyperparameter setting. The proposed method was evaluated by an application for detecting cerebral aneurysms on magnetic resonance angiogram (MRA). A baseline CADe process included four components; scaling, candidate area limitation, candidate detection, and candidate classification. Proposed feature generation method was applied to extract the optimal features for candidate classification. Proposed method only required setting range of the hyperparameters for SdA. The optimal feature set was selected from a large quantity of SdA based features by multiple SdAs, each of which was trained using different hyperparameter set. The feature selection was operated through ada-boost ensemble learning method. Training of the baseline CADe process and proposed feature generation were operated with 200 MRA cases, and the evaluation was performed with 100 MRA cases. Proposed method successfully provided SdA based features just setting the range of some hyperparameters for SdA. The CADe process by using both previous voxel features and SdA based features had the best performance with 0.838 of an area under ROC curve and 0.312 of ANODE score. The results showed that proposed method was effective in the application for detecting cerebral aneurysms on MRA.

Major depressive disorder (MDD) is a globally prevalent psychiatric disorder that results from disruption of multiple neural circuits involved in emotional regulation. Although previous studies using diffusion tensor imaging (DTI) found smaller values of fractional anisotropy (FA) in the white matter, predominantly in the frontal lobe, of patients with MDD, studies using diffusion kurtosis imaging (DKI) are scarce. Here, we used DKI whole‐brain analysis with tract‐based spatial statistics (TBSS) to investigate the brain microstructural abnormalities in MDD. Twenty‐six patients with MDD and 42 age‐ and sex‐matched control subjects were enrolled. To investigate the microstructural pathology underlying the observations in DKI, a compartment model analysis was conducted focusing on the corpus callosum. In TBSS, the patients with MDD showed significantly smaller values of FA in the genu and frontal portion of the body of the corpus callosum. The patients also had smaller values of mean kurtosis (MK) and radial kurtosis (RK), but MK and RK abnormalities were distributed more widely compared with FA, predominantly in the frontal lobe but also in the parietal, occipital, and temporal lobes. Within the callosum, the regions with smaller MK and RK were located more posteriorly than the region with smaller FA. Model analysis suggested significantly smaller values of intra‐neurite signal fraction in the body of the callosum and greater fiber dispersion in the genu, which were compatible with the existing literature of white matter pathology in MDD. Our results show that DKI is capable of demonstrating microstructural alterations in the brains of patients with MDD that cannot be fully depicted by conventional DTI. Though the issues of model validation and parameter estimation still remain, it is suggested that diffusion MRI combined with a biophysical model is a promising approach for investigation of the pathophysiology of MDD.

Segmentation of vertebral bones in computed tomographic data is important as a first stage of image-based radiological tasks. However, it is a challenging problem to segment an affected spine correctly. In this study, we propose a new method of segmentation of thoracic and lumbar vertebral bodies from thin-slice computed tomographic images. Especially, we focus on a deformable model-based segmentation scheme to confirm the feasibility in clinical data sets with various bone diseases, such as bone metastases and scoliosis. As an application of this algorithm, virtual straightening of the thoracolumbar spine is also performed. Results on a database of 16 patients indicate the applicability of our method to spines affected by scoliosis and multiple bone metastases.

A novel feature set for medical image analysis, named HoTPiG (Histogram of Triangular Paths in Graph), is presented. The feature set is designed to detect morphologically abnormal lesions in branching tree-like structures such as vessels. Given a graph structure extracted from a binarized volume, the proposed feature extraction algorithm can effectively encode both the morphological characteristics and the local branching pattern of the structure around each graph node (e.g., each voxel in the vessel). The features are derived from a 3-D histogram whose bins represent a triplet of shortest path distances between the target node and all possible node pairs near the target node. The extracted feature set is a vector with a fixed length and is readily applicable to state-of-the-art machine learning methods. Furthermore, since our method can handle vessel-like structures without thinning or centerline extraction processes, it is free from the “short-hair” problem and local features of vessels such as caliper changes and bumps are also encoded as a whole. Using the proposed feature set, a cerebral aneurysm detection application for clinical magnetic resonance angiography (MRA) images was implemented. In an evaluation with 300 datasets, the sensitivities of aneurysm detection were 81.8% and 89.2% when the numbers of false positives were 3 and 10 per case, respectively, thus validating the effectiveness of the proposed feature set.

Purpose The purpose of this study is to introduce the new concept of a four-dimensional (4D) cone-beam computed tomography (CBCT) reconstruction approach for non-periodic organ motion in cooperation with the time-ordered chain graph model (TCGM) and to compare it with previously developed methods such as total variation-based compressed sensing (TVCS) and prior-image constrained compressed sensing (PICCS). Materials and Methods Our proposed reconstruction is based on a model including the constraint originating from the images of neighboring time phases. Namely, the reconstructed time-series images depend on each other in this TCGM scheme, and the time-ordered images are concurrently reconstructed in the iterative reconstruction approach. In this study, iterative reconstruction with the TCGM was carried out with 90° projection ranges. The images reconstructed by the TCGM were compared with the images reconstructed by TVCS (200° projection ranges) and PICCS (90° projection ranges). Two kinds of projection data sets–an elliptic-cylindrical digital phantom and two clinical patients’ data–were used. For the digital phantom, an air sphere was contained and virtually moved along the longitudinal axis by 3 cm/30 s and 3 cm/60 s; the temporal resolution was evaluated by measuring the penumbral width of the air sphere. The clinical feasibility of the non-periodic time-ordered 4D CBCT image reconstruction was examined with the patient data in the pelvic region. Results In the evaluation of the digital-phantom reconstruction, the penumbral widths of the TCGM yielded the narrowest result; the results obtained by PICCS and TCGM using 90° projection ranges were 2.8% and 18.2% for 3 cm/30 s, and 5.0% and 23.1% for 3 cm/60 s narrower than that of TVCS using 200° projection ranges. This suggests that the TCGM has a better temporal resolution, whereas PICCS seems similar to TVCS. These reconstruction methods were also compared using patients’ projection data sets. Although all three reconstruction results showed motion related to rectal gas or stool, the result obtained by the TCGM was visibly clearer with less blurring. Conclusion The TCGM is a feasible approach to visualize non-periodic organ motion. The digital-phantom results demonstrated that the proposed method provides 4D image series with a better temporal resolution compared to TVCS and PICCS. The clinical patients’ results also showed that the present method enables us to visualize motion related to rectal gas and flatus in the rectum.

Magnetic resonance imaging (MRI) enables the evaluation of organ structure and function. Oxygen-enhanced MRI (O2-enhanced MRI) is a method for evaluating the pulmonary ventilation function using oxygen as a contrast agent. We created the Cine View of Relative Enhancement Ratio Map (Cine RER map) in O2-enhanced MRI to easily observe the contrast effect for clinical use. Relative enhancement ratio (RER) was determined as the pixel values of the Cine RER map. Moreover, six healthy volunteers underwent O2-enhanced MRI to determine the appropriate scale width of the Cine RER map. We calculated each RER and set 0 to 1.27 as the scale width of the Cine RER map based on the results. The Cine RER map made it possible to observe the contrast effect over time and thus is a convenient tool for evaluating the pulmonary ventilation function in O2-enhanced MRI.

Hepatic transarterial interventional therapies such as chemoembolization and radiation embolization are important treatment options for hepatocellular carcinoma. Understanding the anatomy of individual arterial branches and hepatic segments is critical for selecting the correct embolization technique for treatment and to avoid complications. The authors describe the morphologic characteristics of hepatic arterial branches (and their mimickers) and hepatic segments on conventional angiograms. These vessels and segments include the celiac artery, the common and proper hepatic arteries, the left and right hepatic arteries and branches, the caudate lobe, and the portal vein and branches. Mimickers of hepatic arteries include the cystic, accessory left gastric, and right gastric arteries, as well as branches of the left gastric artery that resemble segmental branches of the replaced left hepatic artery. The authors describe how each segmental branch of the hepatic artery and the area it supplies correlates at computed tomography (CT) and angiography. Finally, the authors demonstrate how the vascular anatomy changes with the respiratory cycle by creating a virtual movie from calculations with dynamic CT data, in which the arterial and venous phases are acquired at end expiration and inspiration, respectively. Each segmental branch of the hepatic artery has morphologic characteristics that help distinguish it from mimickers. The location of each hepatic segment can be estimated if the artery supplying the segment can be correctly identified on angiograms. Notably, morphologic differences in the hepatic artery system caused by respiration should be recognized.

Introduction: The performance of computer-assisted detection (CAD) software depends on the quality and quantity of the dataset used for supervised learning. To realize the continuous clinical use and performance improvement of CAD software, it is necessary to continuously collect data for supervised learning in practical use and to improve CAD software by retraining with the collected data. In this study, we investigated the performance improvement of cerebral aneurysm detection software based on retraining the classifier through a simulation-based study. Methods: We collected data for retraining during the practical use of our cerebral aneurysm detection software and retrained the classifier for false positive (FP) reduction using the collected data. The effect on improving the performance was compared by changing the number of training cases and the training algorithms. Results: The performance was improved significantly ( p < .05) by retraining using additional training cases. In contrast, there were no statistical differences in the performance upon retraining among the four training algorithms for boosting. The sensitivity at 3 FPs/case was improved from 81.5% to 89.5% by retraining with additional training cases. Conclusions: The performance of the software was effectively improved by adding training cases rather than by changing the training algorithm.