Search

Purpose:The purpose of this study is to compare two libraries dedicated to Markov chain Monte Carlo method: pystan and numpyro.Materials and methods:Bayesian item response theory (IRT), 1PL-IRT and 2PL-IRT, were implemented with pystan and numpyro. Then, the Bayesian 1PL-IRT and 2PL-IRT were applied to two types of medical data obtained from a published paper. The same prior distributions of latent parameters were used in both pystan and numpyro. Estimation results of latent parameters of 1PL-IRT and 2PL-IRT were compared between pystan and numpyro. Additionally, the computational cost of Markov chain Monte Carlo method was compared between the two libraries. To evaluate the computational cost of IRT models, simulation data were generated from the medical data and numpyro.Results:For all the combinations of IRT types (1PL-IRT or 2PL-IRT) and medical data types, the mean and standard deviation of the estimated latent parameters were in good agreement between pystan and numpyro. In most cases, the sampling time using Markov chain Monte Carlo method was shorter in numpyro than that in pystan. When the large-sized simulation data were used, numpyro with a graphics processing unit was useful for reducing the sampling time.Conclusion:Numpyro and pystan were useful for applying the Bayesian 1PL-IRT and 2PL-IRT.

Rationale and objectives: Several studies have evaluated the usefulness of deep learning for lung segmentation using chest x-ray (CXR) images with small- or medium-sized abnormal findings. Here, we built a database including both CXR images with severe abnormalities and experts' lung segmentation results, and aimed to evaluate our network's efficacy in lung segmentation from these images. Materials and Methods: For lung segmentation, CXR images from the Japanese Society of Radiological Technology (JSRT, N = 247) and Montgomery databases (N = 138), were included, and 65 additional images depicting severe abnormalities from a public database were evaluated and annotated by a radiologist, thereby adding lung segmentation results to these images. Baseline U-net was used to segment the lungs in images from the three databases. Subsequently, the U-net network architecture was automatically optimized for lung segmentation from CXR images using Bayesian optimization. Dice similarity coefficient (DSC) was calculated to confirm segmentation. Results: Our results demonstrated that using baseline U-net yielded poorer lung segmentation results in our database than those in the JSRT and Montgomery databases, implying that robust segmentation of lungs may be difficult because of severe abnormalities. The DSC values with baseline U-net for the JSRT, Montgomery and our databases were 0.979, 0.941, and 0.889, respectively, and with optimized U-net, 0.976, 0.973, and 0.932, respectively. Conclusion: For robust lung segmentation, the U-net architecture was optimized via Bayesian optimization, and our results demonstrate that the optimized U-net was more robust than baseline U-net in lung segmentation from CXR images with large-sized abnormalities.

To evaluate the utility of a convolutional neural network (CNN) with an increased number of contracting and expanding paths of U-net for sparse-view CT reconstruction.This study used 60 anonymized chest CT cases from a public database called "The Cancer Imaging Archive". Eight thousand images from 40 cases were used for training. Eight hundred and 80 images from another 20 cases were used for quantitative and qualitative evaluation, respectively. Sparse-view CT images subsampled by a factor of 20 were simulated, and two CNNs were trained to create denoised images from the sparse-view CT. A CNN based on U-net with residual learning with four contracting and expanding paths (the preceding CNN) was compared with another CNN with eight contracting and expanding paths (the proposed CNN) both quantitatively (peak signal to noise ratio, structural similarity index), and qualitatively (the scores given by two radiologists for anatomical visibility, artifact and noise, and overall image quality) using the Wilcoxon signed-rank test. Nodule and emphysema appearance were also evaluated qualitatively.The proposed CNN was significantly better than the preceding CNN both quantitatively and qualitatively (overall image quality interquartile range, 3.0-3.5 versus 1.0-1.0 reported from the preceding CNN; p0.001). However, only 2 of 22 cases used for emphysematous evaluation (2 CNNs for every 11 cases with emphysema) had an average score of ≥ 2 (on a 3 point scale).Increasing contracting and expanding paths may be useful for sparse-view CT reconstruction with CNN. However, poor reproducibility of emphysema appearance should also be noted.

To develop and evaluate a deep learning-based algorithm (DLA) for automatic detection of bone metastases on CT.This retrospective study included CT scans acquired at a single institution between 2009 and 2019. Positive scans with bone metastases and negative scans without bone metastasis were collected to train the DLA. Another 50 positive and 50 negative scans were collected separately from the training dataset and were divided into validation and test datasets at a 2:3 ratio. The clinical efficacy of the DLA was evaluated in an observer study with board-certified radiologists. Jackknife alternative free-response receiver operating characteristic analysis was used to evaluate observer performance.A total of 269 positive scans including 1375 bone metastases and 463 negative scans were collected for the training dataset. The number of lesions identified in the validation and test datasets was 49 and 75, respectively. The DLA achieved a sensitivity of 89.8% (44 of 49) with 0.775 false positives per case for the validation dataset and 82.7% (62 of 75) with 0.617 false positives per case for the test dataset. With the DLA, the overall performance of nine radiologists with reference to the weighted alternative free-response receiver operating characteristic figure of merit improved from 0.746 to 0.899 (p.001). Furthermore, the mean interpretation time per case decreased from 168 to 85 s (p = .004).With the aid of the algorithm, the overall performance of radiologists in bone metastases detection improved, and the interpretation time decreased at the same time.• A deep learning-based algorithm for automatic detection of bone metastases on CT was developed. • In the observer study, overall performance of radiologists in bone metastases detection improved significantly with the aid of the algorithm. • Radiologists' interpretation time decreased at the same time.

Endometrial cancer (EC) is the most common gynecological tumor in developed countries, and preoperative risk stratification is essential for personalized medicine. There have been several radiomics studies for noninvasive risk stratification of EC using MRI. Although tumor segmentation is usually necessary for these studies, manual segmentation is not only labor-intensive but may also be subjective. Therefore, our study aimed to perform the automatic segmentation of EC on MRI with a convolutional neural network. The effect of the input image sequence and batch size on the segmentation performance was also investigated. Of 200 patients with EC, 180 patients were used for training the modified U-net model; 20 patients for testing the segmentation performance and the robustness of automatically extracted radiomics features. Using multi-sequence images and larger batch size was effective for improving segmentation accuracy. The mean Dice similarity coefficient, sensitivity, and positive predictive value of our model for the test set were 0.806, 0.816, and 0.834, respectively. The robustness of automatically extracted first-order and shape-based features was high (median ICC = 0.86 and 0.96, respectively). Other high-order features presented moderate-high robustness (median ICC = 0.57–0.93). Our model could automatically segment EC on MRI and extract radiomics features with high reliability.

To develop and evaluate a radiomics approach for classifying histological subtypes and epidermal growth factor receptor (EGFR) mutation status in lung cancer on PET/CT images. PET/CT images of lung cancer patients were obtained from public databases and used to establish two datasets, respectively to classify histological subtypes (156 adenocarcinomas and 32 squamous cell carcinomas) and EGFR mutation status (38 mutant and 100 wild-type samples). Seven types of imaging features were obtained from PET/CT images of lung cancer. Two types of machine learning algorithms were used to predict histological subtypes and EGFR mutation status: random forest (RF) and gradient tree boosting (XGB). The classifiers used either a single type or multiple types of imaging features. In the latter case, the optimal combination of the seven types of imaging features was selected by Bayesian optimization. Receiver operating characteristic analysis, area under the curve (AUC), and tenfold cross validation were used to assess the performance of the approach. In the classification of histological subtypes, the AUC values of the various classifiers were as follows: RF, single type: 0.759; XGB, single type: 0.760; RF, multiple types: 0.720; XGB, multiple types: 0.843. In the classification of EGFR mutation status, the AUC values were: RF, single type: 0.625; XGB, single type: 0.617; RF, multiple types: 0.577; XGB, multiple types: 0.659. The radiomics approach to PET/CT images, together with XGB and Bayesian optimization, is useful for classifying histological subtypes and EGFR mutation status in lung cancer.

To compare observer performance of detecting bone metastases between bone scintigraphy, including planar scan and single-photon emission computed tomography, and computed tomography (CT) temporal subtraction (TS). Data on 60 patients with cancer who had undergone CT (previous and current) and bone scintigraphy were collected. Previous CT images were registered to the current ones by large deformation diffeomorphic metric mapping; the registered previous images were subtracted from the current ones to produce TS. Definitive diagnosis of bone metastases was determined by consensus between two radiologists. Twelve readers independently interpreted the following pairs of examinations: NM-pair, previous and current CTs and bone scintigraphy, and TS-pair, previous and current CTs and TS. The readers assigned likelihood levels to suspected bone metastases for diagnosis. Sensitivity, number of false positives per patient (FPP), and reading time for each pair of examinations were analysed for evaluating observer performance by performing the Wilcoxon signed-rank test. Figure-of-merit (FOM) was calculated using jackknife alternative free-response receiver operating characteristic analysis. The sensitivity of TS was significantly higher than that of bone scintigraphy (54.3% vs. 41.3%, p = 0.006). FPP with TS was significantly higher than that with bone scintigraphy (0.189 vs. 0.0722, p = 0.003). FOM of TS tended to be better than that of bone scintigraphy (0.742 vs. 0.691, p = 0.070). Sensitivity of TS in detecting bone metastasis was significantly higher than that of bone scintigraphy, but still limited to 54%. TS might be superior to bone scintigraphy for early detection of bone metastasis. • Computed tomography temporal subtraction was helpful in early detection of bone metastases. • Sensitivity for bone metastasis was higher for computed tomography temporal subtraction than for bone scintigraphy. • Figure-of-merit of computed tomography temporal subtraction was better than that of bone scintigraphy.

Purpose To evaluate radiomic machine learning (ML) classifiers based on multiparametric magnetic resonance images (MRI) in pretreatment assessment of endometrial cancer (EC) risk factors and to examine effects on radiologists' interpretation of deep myometrial invasion (dMI). Methods This retrospective study examined 200 consecutive patients with EC during January 2004 –March 2017, divided randomly to Discovery (n = 150) and Test (n = 50) datasets. Radiomic features of tumors were extracted from T2-weighted images, apparent diffusion coefficient map, and contrast enhanced T1-weighed images. Using the Discovery dataset, feature selection and hyperparameter tuning for XGBoost were performed. Ten classifiers were built to predict dMI, histological grade, lymphovascular invasion (LVI), and pelvic/paraaortic lymph node metastasis (PLNM/PALNM), respectively. Using the Test dataset, the diagnostic performances of ten classifiers were assessed by the area under the receiver operator characteristic curve (AUC). Next, four radiologists assessed dMI independently using MRI with a Likert scale before and after referring to inference of the ML classifier for the Test dataset. Then, AUCs obtained before and after reference were compared. Results In the Test dataset, mean AUC of ML classifiers for dMI, histological grade, LVI, PLNM, and PALNM were 0.83, 0.77, 0.81, 0.72, and 0.82. AUCs of all radiologists for dMI (0.83–0.88) were better than or equal to mean AUC of the ML classifier, which showed no statistically significant difference before and after the reference. Conclusion Radiomic classifiers showed promise for pretreatment assessment of EC risk factors. Radiologists' inferences outperformed the ML classifier for dMI and showed no improvement by review.

Purpose To develop and evaluate a three-dimensional (3D) generative model of computed tomography (CT) images of lung nodules using a generative adversarial network (GAN). To guide the GAN, lung nodule size was used. Materials and methods A public CT dataset of lung nodules was used, from where 1182 lung nodules were obtained. Our proposed GAN model used masked 3D CT images and nodule size information to generate images. To evaluate the generated CT images, two radiologists visually evaluated whether the CT images with lung nodule were true or generated, and the diagnostic ability was evaluated using receiver-operating characteristic analysis and area under the curves (AUC). Then, two models for classifying nodule size into five categories were trained, one using the true and the other using the generated CT images of lung nodules. Using true CT images, the classification accuracy of the sizes of the true lung nodules was calculated for the two classification models. Results The sensitivity, specificity, and AUC of the two radiologists were respectively as follows: radiologist 1: 81.3%, 37.7%, and 0.592; radiologist 2: 77.1%, 30.2%, and 0.597. For categorization of nodule size, the mean accuracy of the classification model constructed with true CT images was 85% (range 83.2–86.1%), and that with generated CT images was 85% (range 82.2–88.1%). Conclusions Our results show that it was possible to generate 3D CT images of lung nodules that could be used to construct a classification model of lung nodule size without true CT images.

Convolutional neural network (CNN), a class of artificial neural networks that has become dominant in various computer vision tasks, is attracting interest across a variety of domains, including radiology. CNN is designed to automatically and adaptively learn spatial hierarchies of features through backpropagation by using multiple building blocks, such as convolution layers, pooling layers, and fully connected layers. This review article offers a perspective on the basic concepts of CNN and its application to various radiological tasks, and discusses its challenges and future directions in the field of radiology. Two challenges in applying CNN to radiological tasks, small dataset and overfitting, will also be covered in this article, as well as techniques to minimize them. Being familiar with the concepts and advantages, as well as limitations, of CNN is essential to leverage its potential in diagnostic radiology, with the goal of augmenting the performance of radiologists and improving patient care. • Convolutional neural network is a class of deep learning methods which has become dominant in various computer vision tasks and is attracting interest across a variety of domains, including radiology. • Convolutional neural network is composed of multiple building blocks, such as convolution layers, pooling layers, and fully connected layers, and is designed to automatically and adaptively learn spatial hierarchies of features through a backpropagation algorithm. • Familiarity with the concepts and advantages, as well as limitations, of convolutional neural network is essential to leverage its potential to improve radiologist performance and, eventually, patient care.

Many recent studies on medical image processing have involved the use of machine learning (ML) and deep learning (DL) [...]

Rationale and Objectives To develop a computer-aided diagnosis system to differentiate between malignant and benign nodules. Materials and Methods Seventy-three lung nodules revealed on 60 sets of computed tomography (CT) images were analyzed. Contrast-enhanced CT was performed in 46 CT examinations. The images were provided by the LUNGx Challenge, and the ground truth of the lung nodules was unavailable; a surrogate ground truth was, therefore, constructed by radiological evaluation. Our proposed method involved novel patch-based feature extraction using principal component analysis, image convolution, and pooling operations. This method was compared to three other systems for the extraction of nodule features: histogram of CT density, local binary pattern on three orthogonal planes, and three-dimensional random local binary pattern. The probabilistic outputs of the systems and surrogate ground truth were analyzed using receiver operating characteristic analysis and area under the curve. The LUNGx Challenge team also calculated the area under the curve of our proposed method based on the actual ground truth of their dataset. Results Based on the surrogate ground truth, the areas under the curve were as follows: histogram of CT density, 0.640; local binary pattern on three orthogonal planes, 0.688; three-dimensional random local binary pattern, 0.725; and the proposed method, 0.837. Based on the actual ground truth, the area under the curve of the proposed method was 0.81. Conclusions The proposed method could capture discriminative characteristics of lung nodules and was useful for the differentiation between malignant and benign nodules.

4 TECHNICAL EFFICACY STAGE: 4 J. Magn. Reson. Imaging 2020;52:1263-1264.

To develop and evaluate a radiomics approach for classifying histological subtypes and epidermal growth factor receptor (EGFR) mutation status in lung cancer on PET/CT images.PET/CT images of lung cancer patients were obtained from public databases and used to establish two datasets, respectively to classify histological subtypes (156 adenocarcinomas and 32 squamous cell carcinomas) and EGFR mutation status (38 mutant and 100 wild-type samples). Seven types of imaging features were obtained from PET/CT images of lung cancer. Two types of machine learning algorithms were used to predict histological subtypes and EGFR mutation status: random forest (RF) and gradient tree boosting (XGB). The classifiers used either a single type or multiple types of imaging features. In the latter case, the optimal combination of the seven types of imaging features was selected by Bayesian optimization. Receiver operating characteristic analysis, area under the curve (AUC), and tenfold cross validation were used to assess the performance of the approach.In the classification of histological subtypes, the AUC values of the various classifiers were as follows: RF, single type: 0.759; XGB, single type: 0.760; RF, multiple types: 0.720; XGB, multiple types: 0.843. In the classification of EGFR mutation status, the AUC values were: RF, single type: 0.625; XGB, single type: 0.617; RF, multiple types: 0.577; XGB, multiple types: 0.659.The radiomics approach to PET/CT images, together with XGB and Bayesian optimization, is useful for classifying histological subtypes and EGFR mutation status in lung cancer.

Background This study was performed to evaluate the clinical feasibility of a U-net for fully automatic uterine segmentation on MRI by using images of major uterine disorders. Methods This study included 122 female patients (14 with uterine endometrial cancer, 15 with uterine cervical cancer, and 55 with uterine leiomyoma). U-net architecture optimized for our research was used for automatic segmentation. Three-fold cross-validation was performed for validation. The results of manual segmentation of the uterus by a radiologist on T2-weighted sagittal images were used as the gold standard. Dice similarity coefficient (DSC) and mean absolute distance (MAD) were used for quantitative evaluation of the automatic segmentation. Visual evaluation using a 4-point scale was performed by two radiologists. DSC, MAD, and the score of the visual evaluation were compared between uteruses with and without uterine disorders. Results The mean DSC of our model for all patients was 0.82. The mean DSCs for patients with and without uterine disorders were 0.84 and 0.78, respectively (p = 0.19). The mean MADs for patients with and without uterine disorders were 18.5 and 21.4 [pixels], respectively (p = 0.39). The scores of the visual evaluation were not significantly different between uteruses with and without uterine disorders. Conclusions Fully automatic uterine segmentation with our modified U-net was clinically feasible. The performance of the segmentation of our model was not influenced by the presence of uterine disorders.

Temporal subtraction of CT (TS) images improves detection of newly developed bone metastases (BM). We sought to determine whether TS improves detection of BM by radiology residents as well. We performed an observer study using a previously reported dataset, consisting of 60 oncology patients, each with previous and current CT images. TS images were calculated using in-house software. Four residents independently interpreted twice the 60 sets of CT images, without and with TS. They identified BM by marking suspicious lesions likely to be BM. Lesion-based sensitivity and number of false positives per patient were calculated. Figure-of-merit (FOM) was calculated. Detectability of BM, with and without TS, was compared between radiology residents and board-certified radiologists, as published previously. FOM of residents significantly improved by implementing TS (p value

We have developed a novel medical device software program that is intended to improve radiologists’ performance with regard to the detection of bone metastases on CT images. To establish the protocol of the clinical performance evaluation for regulatory approval, the feasibility of nonclinical study (NCS) and clinical study (CS) conforming to the principles of GCP was compared. There were two main issues with establishing the protocol in Japan. Firstly, retrospective case collection was not accepted for CS due to concerns regarding selection bias. However, it is difficult to prospectively collect cases for use in an observer study due to the wide variety of bone metastases cases required. Secondly, when setting the reference standard (RS) for bone metastases, the utilization of any patient information was not accepted for NCS due to concerns regarding an inappropriate RS setting. However, it was predicted that it would be difficult to set accurate RSs only using CT images. Consequently, following discussion with a Japanese regulatory body, the protocol was established as an NCS using retrospectively collected CT images. Furthermore, we agreed to use anonymized clinical images to set the RS and to rescue the test results by reviewing the RS after the observer study, hence ensuring reliability. An NCS was a feasible method of evaluating the novel medical device software while also complying with quality management systems. By using real-world clinical image databases, this clinical performance evaluation method may be both effective and efficient as a path to regulatory approval of certain medical device software.

The purpose of this study was to evaluate agreement between standard-dose CT (SDCT) and ultralow-dose CT (ULDCT) findings with respect to emphysema quantification. ULDCT images were reconstructed with and without iterative reconstruction (IR). Adaptive iterative dose reduction with 3D processing was used for IR.Fifty patients who underwent SDCT and ULDCT were included. The tube current for SDCT was 250 mA, and that for ULDCT was 10 mA. SDCT, ULDCT without IR, and ULDCT with IR were used for emphysema quantification. The low-attenuation volume percentage (LAV%) in the lungs at four thresholds (-970, -950, -930, and -910 HU), mean lung attenuation, and total lung volume were computed. Concordance correlation coefficients (CCC) were used to assess the agreement of emphysema quantification between SDCT and ULDCT.The LAV% CCC values were 0.310-0.789 between SDCT and ULDCT without IR and 0.934-0.966 between SDCT and ULDCT with IR. The agreement of LAV% improved when IR was used for ULDCT. The mean lung attenuation CCC value between SDCT and ULDCT without IR was substantial (0.957), whereas that between SDCT and ULDCT with IR was poor (0.890). The total lung volume CCC values were substantial (0.982 with IR, 0.983 without IR).ULDCT with and without IR can substitute for SDCT in emphysema quantification.

Background and objective Currently, it is challenging to detect acute ischemic stroke (AIS)-related changes on computed tomography (CT) images. Therefore, we aimed to develop and evaluate an automatic AIS detection system involving a two-stage deep learning model. Methods We included 238 cases from two different institutions. AIS-related findings were annotated on each of the 238 sets of head CT images by referring to head magnetic resonance imaging (MRI) images in which an MRI examination was performed within 24 h following the CT scan. These 238 annotated cases were divided into a training set including 189 cases and test set including 49 cases. Subsequently, a two-stage deep learning detection model was constructed from the training set using the You Only Look Once v3 model and Visual Geometry Group 16 classification model. Then, the two-stage model performed the AIS detection process in the test set. To assess the detection model's results, a board-certified radiologist also evaluated the test set head CT images with and without the aid of the detection model. The sensitivity of AIS detection and number of false positives were calculated for the evaluation of the test set detection results. The sensitivity of the radiologist with and without the software detection results was compared using the McNemar test. A p-value of less than 0.05 was considered statistically significant. Results For the two-stage model and radiologist without and with the use of the software results, the sensitivity was 37.3%, 33.3%, and 41.3%, respectively, and the number of false positives per one case was 1.265, 0.327, and 0.388, respectively. On using the two-stage detection model's results, the board-certified radiologist's detection sensitivity significantly improved (p-value = 0.0313). Conclusions Our detection system involving the two-stage deep learning model significantly improved the radiologist's sensitivity in AIS detection.

Background The purpose of this study was to develop and evaluate an algorithm for bone segmentation on whole-body CT using a convolutional neural network (CNN). Methods Bone segmentation was performed using a network based on U-Net architecture. To evaluate its performance and robustness, we prepared three different datasets: (1) an in-house dataset comprising 16,218 slices of CT images from 32 scans in 16 patients; (2) a secondary dataset comprising 12,529 slices of CT images from 20 scans in 20 patients, which were collected from The Cancer Imaging Archive; and (3) a publicly available labelled dataset comprising 270 slices of CT images from 27 scans in 20 patients. To improve the network's performance and robustness, we evaluated the efficacy of three types of data augmentation technique: conventional method, mixup, and random image cropping and patching (RICAP). Results The network trained on the in-house dataset achieved a mean Dice coefficient of 0.983 ± 0.005 on cross validation with the in-house dataset, and 0.943 ± 0.007 with the secondary dataset. The network trained on the public dataset achieved a mean Dice coefficient of 0.947 ± 0.013 on 10 randomly generated 15-3-9 splits of the public dataset. These results outperform those reported previously. Regarding augmentation technique, the conventional method, RICAP, and a combination of these were effective. Conclusions The CNN-based model achieved accurate bone segmentation on whole-body CT, with generalizability to various scan conditions. Data augmentation techniques enabled construction of an accurate and robust model even with a small dataset.

Training of a convolutional neural network (CNN) generally requires a large dataset. However, it is not easy to collect a large medical image dataset. The purpose of this study is to investigate the utility of synthetic images in training CNNs and to demonstrate the applicability of unrelated images by domain transformation. Mammograms showing 202 benign and 212 malignant masses were used for evaluation. To create synthetic data, a cycle generative adversarial network was trained with 599 lung nodules in computed tomography (CT) and 1430 breast masses on digitized mammograms (DDSM). A CNN was trained for classification between benign and malignant masses. The classification performance was compared between the networks trained with the original data, augmented data, synthetic data, DDSM images, and natural images (ImageNet dataset). The results were evaluated in terms of the classification accuracy and the area under the receiver operating characteristic curves (AUC). The classification accuracy improved from 65.7% to 67.1% with data augmentation. The use of an ImageNet pretrained model was useful (79.2%). Performance was slightly improved when synthetic images or the DDSM images only were used for pretraining (67.6 and 72.5%, respectively). When the ImageNet pretrained model was trained with the synthetic images, the classification performance slightly improved (81.4%), although the difference in AUCs was not statistically significant. The use of the synthetic images had an effect similar to the DDSM images. The results of the proposed study indicated that the synthetic data generated from unrelated lesions by domain transformation could be used to increase the training samples.

BACKGROUND AND PURPOSE: Pretreatment diagnosis for the location of shunts and arterial feeders of spinal arteriovenous fistulas is crucial. This study aimed to evaluate the utility of subtracted CT angiography imaging by using nonrigid registration (R-CTA) in patients with spinal arteriovenous fistulas compared with conventional CTA imaging. MATERIALS AND METHODS: The records of 15 consecutive subjects (mean age, 65 years; 2 women) who had undergone CTA and digital subtraction angiography for clinically suspected spinal arteriovenous fistula were reviewed. From CTA images obtained at the arterial and late arterial phases, warped images of the late arterial phase were obtained by using nonrigid registration that was adjusted to the arterial phase images. R-CTA images were then obtained by subtracting the warped images from the arterial phase images. The accuracies of using nonrigid registration and conventional spinal CTA and the time required for detecting arterial feeders in spinal arteriovenous fistulas were analyzed for each patient with DSA results as a standard reference. The difference between R-CTA and conventional spinal CTA was assessed by the Welch test and the McNemar χ2 test. RESULTS: R-CTA had a higher accuracy compared with conventional spinal CTA (80% versus 47%, P = .025). The time for interpretation was reduced in R-CTA compared with conventional spinal CTA (45.1 versus 97.1 seconds, P = .002). CONCLUSIONS: Our subtracted CTA imaging by using nonrigid registration detects feeders of spinal arteriovenous fistulas more accurately and quickly than conventional CTA.

The purpose of this study was to develop methodology to segment tumors on 18F-fluorodeoxyg- lucose (FDG) positron emission tomography (PET) images. Sixty-four metastatic bone tumors were included. Graph cut was used for tumor segmentation, with segmentation energy divided into unary and pairwise terms. Locally connected conditional random fields (LCRF) were proposed for the pairwise term. In LCRF, three-dimensional cubic window with length L was set for each voxel, and voxels within the window were considered for the pairwise term. Three other types of segmentation were applied: region-growing based on 35%, 40%, and 45% of the tumor maximum standardized uptake value (RG35, RG40, and RG45, respectively), SLIC superpixels (SS), and region-based active contour models (AC). To validate the tumor segmentation accuracy, dice similarity coefficients (DSC) were calculated between the result of each technique and manual segmentation. Differences in DSC were tested using the Wilcoxon signed-rank test. Mean DSCs for LCRF at L = 3, 5, 7, and 9 were 0.784, 0.801, 0.809, and 0.812, respectively. Mean DSCs for the other techniques were: RG35, 0.633; RG40, 0.675; RG45, 0.689; SS, 0.709; and AC, 0.758. The DSC differences between LCRF and other techniques were statistically significant (p < 0.05). Tumor segmentation was reliably performed with LCRF.

The purpose of this study was to assess usefulness of local binary patterns (LBP) and related texture features, namely completed local binary patterns (CLBP) and local ternary patterns (LTP), for the classification of emphysema subtypes on low-dose CT images. Fifty patients (34 men and 16 women; age, 67.5 ± 10.1 years) who underwent low-dose CT (60 mAs) were included. They were comprised of 17 never smokers, 13 smokers without COPD, and 20 smokers with COPD. By consensus reading of low-dose CT images from these patients, two radiologists selected 3681 nonoverlapping regions of interest (ROIs) and annotated them as one of the following three classes: normal tissue, centrilobular emphysema, and paraseptal emphysema. From these ROIs, histogram of CT densities, LBP, CLBP, and LTP were calculated, and the 3 types of texture histograms were concatenated with the CT density histogram. These 3 types of histograms (referred to as combined LBP, combined CLBP, and combined LTP) were used to classify ROI using linear support vector machine. For each type of the combined histogram, the accuracy of classification was determined by patient-based 10-fold cross validation. The best accuracy of combined LBP, combined CLBP, and combined LTP were 81.36%, 82.99%, and 83.29%, respectively. Compared to the classification accuracies obtained with combined LBP, those with combined LTP or combined CLBP were consistently improved. In conclusion, the results of this study suggest that, on low-dose CT, LTP and CLBP were more useful for the classification of emphysema subtypes than LBP.

To compare the use of dynamic oxygen-enhanced magnetic resonance (MR) imaging with the use of quantitatively assessed computed tomography (CT) for assessment of clinical stage and evaluation of pulmonary functional change due to treatment in patients with asthma.The institutional review board of Kobe University Hospital approved this study, and written informed consent was obtained from each subject. Thirty consecutive patients with asthma (17 men and 13 women; age range, 27-78 years) underwent dynamic oxygen-enhanced MR imaging, multidetector CT, and assessment of forced expiratory volume in 1 second. All patients were classified as having one of four stages of asthma according to the guidelines of the National Asthma Education and Prevention Program. Relative enhancement ratio ( RER relative enhancement ratio ) and wash-in time maps were generated by means of pixel-by-pixel analyses. Regions of interest were placed on images of the lung in all sections, and all measurements were averaged to determine mean RER relative enhancement ratio and mean wash-in time for each subject. Percentage of airway wall area and mean lung density were determined at quantitative CT. For comparison of the modalities for assessment of clinical stage, indexes of subjects at all clinical stages were compared by means of the Tukey honestly significant difference test. Evaluation of pulmonary functional improvement was assessed by correlating improvement of each index with that of forced expiratory volume.Mean wash-in time was significantly different among patients with asthma of different clinical stages (P.05), but significant differences between mean RER relative enhancement ratio and percentage of airway wall area were observed for a limited number of clinical stages (P.05). Improvement of mean RER relative enhancement ratio (r = 0.63, P = .0002) and mean wash-in time (r = -0.75, P.0001) was significantly correlated with forced expiratory volume.Dynamic oxygen-enhanced MR imaging has potential as a tool for assessment of clinical stage and evaluation of pulmonary functional change due to treatment in patients with asthma.

Rationale and Objectives The purpose of this study was to analyze the relationship between airflow limitation and two types of computed tomography (CT) measurements: expiratory/inspiratory (E/I) ratio and E/I difference of percentage of low-attenuation lung regions (LAA%). Materials and Methods Thirty patients who underwent inspiratory and expiratory CT scans were included in this study. The CT data were used to calculate the LAA% E/I ratio and E/I difference. Other types of CT measurements were also obtained, including the E/I ratio and E/I difference of lung volume, mean lung density, standard deviation, skewness, and kurtosis. LAA% was calculated at 20 thresholds (−990 to −800 HU). Pearson's correlation between the measurements and forced expiratory flow in 1 second was used to determine the efficacy of LAA% E/I ratio and E/I difference. P values of Results The LAA% E/I ratio and E/I difference significantly correlated with forced expiratory flow in 1 second. The best correlation coefficient for the LAA% E/I ratio was −0.699 (P = 1.75 × 10−5) and for the LAA% E/I difference was −0.723 (P = 6.53 × 10−6). The best correlation coefficient for the LAA% E/I difference was stronger than that for the other types of CT measurements. Conclusions The LAA% E/I ratio and E/I difference significantly correlated with airflow limitation in chronic obstructive pulmonary disease.

The purpose of this study was to investigate efficacy of two types of emphysema quantification: percentage of low-attenuation lung regions (%LA); and size distribution of these regions. On a log-log plot, cumulative frequency-size distribution of low-attenuation lung regions can be fitted by a straight line whose slope (D) has been reported to reflect diffusing capacity. In this study, %LA and D were compared with pulmonary function test (PFT) parameters, especially with ratio of diffusing capacity of carbon monoxide to effective alveolar ventilation (i.e., DLCO/VA).Thin-section unenhanced CT images were acquired from 30 patients (25 men, five women; mean [SD] age, 70.1 ± 12.1 years), of whom 25 had received diagnosis of COPD, and %LA and D were calculated at 20 thresholds, ranging from -995 to -900 HU. To determine utility of %LA and D, we used Pearson's correlation for emphysema quantification and PFT. Significance of the coefficients was determined with Bonferroni correction (p0.0025). Finally, the relationships between emphysema quantification and DLCO/VA were examined by linear models and Akaike information criterion (AIC).The correlation coefficients for %LA and DLCO/VA were statistically significant at all the thresholds (optimal coefficient, -0.761). The correlation coefficients for D and DLCO/VA were statistically significant at the thresholds from -945 to -900 HU (optimal coefficient, -0.646). AIC values showed that the most accurate prediction of DLCO/VA was obtained by the model incorporating both %LA and D.Both %LA and D showed significant correlation with DLCO/VA. Combining %LA and D resulted in more accurate evaluation of DLCO/VA than did using %LA or D alone.

To evaluate and compare the capability of diffusion-weighted MR imaging (DWI) and CT for assessment of management needs for anterior mediastinal solitary tumors.Thirty-five patients with pathologically confirmed anterior mediastinal tumors were enrolled. The tumors were divided into two groups according to need for management: tumors not needing further intervention or treatment (group A; thymoma type A, AB and B1) and tumors needing further intervention and treatment (group B; other thymoma types and malignancies). The apparent diffusion coefficient (ADC) of each tumor was measured, and probabilities of malignancy and need for further intervention and treatment were visually assessed on CT. The differences in ADCs between group A and B and between malignancies and thymomas in group B were evaluated with the Mann-Whitney's U-test. Feasible threshold values for differentiation of group B from group A and distinguishing malignancies from thymomas assessed as group B were determined by the ROC-based positive test, and McNemar's test was used for comparing diagnostic capabilities of DWI with those of CT.ADCs for the two groups were significantly different (p0.001). Application of the threshold value for differentiation of group B from A showed no significant difference (p0.05). Application of the feasible threshold value for distinguishing malignant from thymomas assessed as group B showed that specificity (76.9%) and accuracy (85.2%) of DWI were significantly better than those of visual score (p0.05).DWI has useful potential for the assessment of management needs for anterior mediastinum solitary tumors as well as CT.

To prospectively and directly compare oxygen-enhanced (O2-enhanced) MRI with thin-section CT for pulmonary functional loss and disease severity assessment in connective tissue disease (CTD) patients with interstitial lung disease (ILD).Thin-section CT, O2-enhanced MRI, pulmonary function test and serum KL-6 were administered to 36 CTD patients with ILD (23 men, 13 women; mean age: 63.9 years) and nine CTD patients without ILD (six men, and three women; mean age: 62.0 years). A relative-enhancement ratio (RER) map was generated from O2-enhanced MRI and mean relative enhancement ratio (MRER) for each subject was calculated from all ROI measurements. CT-assessed disease severity was evaluated with a visual scoring system from each of the thin-section CT data. MRER and CT-assessed disease severities of CTD patients with and without ILD were then statistically compared. To assess capability for pulmonary functional loss and disease severity assessment in CTD patients, correlations of MRER and CT-assessed disease severity with pulmonary functional parameters and serum KL-6 in all subjects were statistically determined.MRER and CT-assessed disease severity showed significant differences between CTD patients with (MRER: 0.15 ± 0.08, CT-assessed disease severity: 13.0 ± 7.4%) and without ILD (MRER: 0.25 ± 0.06, p=0.0011; CT-assessed disease severity: 1.6 ± 1.6%, p0.0001). MRER and CT-assessed disease severity correlated significantly with pulmonary functional parameters and serum KL-6 in all subjects (0.61 ≤ r ≤ 0.79, p0.05).O2-enhanced MRI was found to be as useful as thin-section CT for pulmonary functional loss and disease severity assessment of CTD patients with ILD.

Purpose The purpose of this study was to compare diagnostic capabilities for assessment of recurrence in non-small cell lung cancer (NSCLC) patients by contrast-enhanced whole-body MRI (CE-WB-MRI) with and without CE-Quick 3D and double RF fat suppression technique (DFS), FDG-PET/CT and conventional radiological examinations. Materials and methods A total of 134 pathologically proven and completely resected NSCLC patients (78 males, 56 females; mean age: 72 years) underwent FDG-PET/CT, CE-WB-MRI with and without Quick 3D and DFS at 3 T as well as conventional radiological examinations. The probability of recurrence was assessed with a 5-point scoring system on a per-patient basis, and final diagnosis was made by consensus between two readers. The capability for overall recurrence assessment by all the methods was compared by means of ROC analysis and their sensitivity, specificity and accuracy by means of McNemar's test. Results Although areas under the curve did not show any significant differences, specificity (100%) and accuracy (95.5%) of CE-WB-MRI with CE-Quick 3D and DFS were significantly higher than those of FDG-PET/CT (specificity: 93.6%, p = 0.02; accuracy: 89.6%, p = 0.01) and conventional radiological examinations (specificity: 92.7%, p = 0.01; accuracy: 91.0%, p = 0.03). In addition, specificity of CE-WB-MRI without CE-Quick 3D and DFS (100%) was significantly higher than that of FDG-PET/CT (p = 0.02) and conventional radiological examinations (p = 0.01). Conclusion Specificity and accuracy of CE-WB-MRI with CE-Quick 3D and DFS for assessment of recurrence in NSCLC patients are at least as high as, or higher than those of others.

Purpose To assess the influence of ultrashort TE (UTE) intervals on pulmonary magnetic resonance imaging (MRI) with UTEs (UTE-MRI) for pulmonary functional loss assessment and clinical stage classification of smokers. Materials and Methods A total 60 consecutive smokers (43 men and 17 women; mean age 70 years) with and without COPD underwent thin-section multidetector row computed tomography (MDCT), UTE-MRI, and pulmonary functional measurements. For each smoker, UTE-MRI was performed with three different UTE intervals (UTE-MRI A: 0.5 msec, UTE-MRI B: 1.0 msec, UTE-MRI C: 1.5 msec). By using the GOLD guidelines, the subjects were classified as: “smokers without COPD,” “mild COPD,” “moderate COPD,” and “severe or very severe COPD.” Then the mean T2* value from each UTE-MRI and CT-based functional lung volume (FLV) were correlated with pulmonary function test. Finally, Fisher's PLSD test was used to evaluate differences in each index among the four clinical stages. Results Each index correlated significantly with pulmonary function test results (P

Purpose To evaluate the utility of pulmonary magnetic resonance (MR) imaging with ultra-short echo times (UTEs) at a 3.0 T MR system for pulmonary functional loss and disease severity assessments of connective tissue disease (CTD) patients with interstitial lung disease (ILD). Materials and methods This prospective study was approved by the institutional review board, and written informed consent was obtained from 18 CTD patients (eight men and ten women) and eight normal subjects with suspected chest disease (three men and five women). All subjects underwent thin-section MDCT, pulmonary MR imaging with UTEs, pulmonary function test and serum KL-6. Regional T2* maps were generated from each MR data set, and mean T2* values were determined from ROI measurements. From each thin-section MDCT data set, CT-based disease severity was evaluated with a visual scoring system. Mean T2* values for normal and CTD subjects were statistically compared by using Student's t-test. To assess capability for pulmonary functional loss and disease severity assessments, mean T2* values were statistically correlated with pulmonary functional parameters, serum KL-6 and CT-based disease severity. Results Mean T2* values for normal and CTD subjects were significantly different (p = 0.0019) and showed significant correlations with %VC, %DLCO, serum KL-6 and CT-based disease severity of CTD patients (p Conclusion Pulmonary MR imaging with UTEs is useful for pulmonary functional loss and disease severity assessments of CTD patients with ILD.

The objective of our study was to prospectively compare the capability of dynamic area-detector CT analyzed with different mathematic methods and PET/CT in the management of pulmonary nodules.Fifty-two consecutive patients with 96 pulmonary nodules underwent dynamic area-detector CT, PET/CT, and microbacterial or pathologic examinations. All nodules were classified into the following groups: malignant nodules (n = 57), benign nodules with low biologic activity (n = 15), and benign nodules with high biologic activity (n = 24). On dynamic area-detector CT, the total, pulmonary arterial, and systemic arterial perfusions were calculated using the dual-input maximum slope method; perfusion was calculated using the single-input maximum slope method; and extraction fraction and blood volume (BV) were calculated using the Patlak plot method. All indexes were statistically compared among the three nodule groups. Then, receiver operating characteristic analyses were used to compare the diagnostic capabilities of the maximum standardized uptake value (SUVmax) and each perfusion parameter having a significant difference between malignant and benign nodules. Finally, the diagnostic performances of the indexes were compared by means of the McNemar test.No adverse effects were observed in this study. All indexes except extraction fraction and BV, both of which were calculated using the Patlak plot method, showed significant differences among the three groups (p0.05). Areas under the curve of total perfusion calculated using the dual-input method, pulmonary arterial perfusion calculated using the dual-input method, and perfusion calculated using the single-input method were significantly larger than that of SUVmax (p0.05). The accuracy of total perfusion (83.3%) was significantly greater than the accuracy of the other indexes: pulmonary arterial perfusion (72.9%, p0.05), systemic arterial perfusion calculated using the dual-input method (69.8%, p0.05), perfusion (66.7%, p0.05), and SUVmax (60.4%, p0.05).Dynamic area-detector CT analyzed using the dual-input maximum slope method has better potential for the diagnosis of pulmonary nodules than dynamic area-detector CT analyzed using other methods and than PET/CT.

Since the publication of the Radiologic Diagnostic Oncology Group Report in 1991, the clinical application of pulmonary magnetic resonance imaging (MRI) in patients with lung cancer has been limited. In contrast, MRI for lung cancer has undergone continuous development, and several promising techniques have been introduced to overcome the previously suggested limitations. In addition, comparative studies involving multidetector-row computed tomography and positron emission tomography or positron emission tomography/computed tomography with 2-deoxy-2-[F]fluoro-D-glucose have shown useful new clinical applications for MRI in lung cancer. Moreover, MRI can provide not only morphologic information based on various parameters such as T1 and T2 relaxation times, tissue diffusion, perfusion, etc. but also functional information; it also has a significant role in nuclear medicine studies. In this review article, we describe recent advances made in MRI with respect to lung cancer, focusing on (1) detection of solid pulmonary nodules; (2) characterization of solid pulmonary nodules; (3) TNM staging assessment using chest and whole-body MRI examinations; (4) prediction of postsurgical lung function; and (5) prediction of tumor treatment response. We believe that further basic studies, as well as studies on clinical applications of new MRI techniques, are important for improving the management of lung cancer patients.

The purpose of this study was to compare three-dimensional airway lumen volumetry with bronchial wall area and parenchymal densitometry in the assessment of airway obstruction in pulmonary emphysema.56 patients, who were smokers, underwent CT examination and pulmonary function tests (PFTs). For quantitative assessments, the following parameters were computationally calculated: (1) percentage of voxels -950, -960 and -970 HU in the lung (%LAA(-950), %LAA(-960) and %LAA(-970), respectively); (2) percentage of partial bronchi luminal volumes per total luminal volumes (LV(main), main and distal bronchial volume/total luminal volume; LV(lobe), lobar and distal bronchial volume/total luminal volume); and (3) mean wall area percentages of segmental bronchi of the right apical and left apicoposterior segment (WA%(seg)) and of subsegmental bronchi (WA%(sub)) in the upper lobes. These parameters were correlated with PFTs and statistically compared between a chronic obstructive pulmonary disease (COPD) group [forced expiratory volume in 1 s (FEV1)/forced vital capacity (FVC)70] and a non-COPD group.FEV1, maximum mid-expiratory flow rate and forced expiratory flow at 25% vital capacity had significant correlation with LV(main) (r0.53, p0.0001), LV(lobe) (r0.52, p0.0001), WA%(seg) (|r|0.29, p0.05) and WA%(sub) (|r|0.31, p0.05). FEV1/FVC had significant correlation with all parameters (0.27|r|0.52, p0.05). LV(main) and LV(lobe) and WA%(seg) and WA%(sub) were significantly different between the two groups (LV(main) and LV(lobe); p0.0001, WA%(seg) and WA%(sub); p0.05).Bronchial luminal volumetric assessment better reflected the airflow limitation parameters.Bronchial luminal volumetric assessment can potentially be used to gauge airflow limitation in pulmonary emphysema.

The purpose of our study was to prospectively and directly compare capability of O2-enhanced MRI, MDCT, and perfusion SPECT/CT to clinical outcome measurements in candidates for lung volume reduction surgery (LVRS).Twenty-five consecutive candidates for LVRS (20 men and five women; age range, 45-76 years) underwent MDCT, O2-enhanced MRI, and perfusion SPECT/CT before and after LVRS. Clinical outcomes for each candidate were evaluated in terms of differences between pre- and postoperative percentage forced expiratory volume in 1 second (%FEV1), Pao2, and 6-minute walking distance. Quantitatively assessed upper/lower lung ratios on O2-enhanced MRI, MDCT, and SPECT/CT were calculated from regional relative enhancement ratios, functional lung volumes, and radioisotope uptakes between upper and lower lungs. Qualitatively assessed upper/lower lung ratios on O2-enhanced MRI, MDCT, and SPECT/CT were estimated using visual scoring systems. To evaluate the correlation for individual upper/lower lung ratios and clinical outcomes, all upper/lower lung ratios were correlated with clinical outcomes. Improvements in mean relative enhancement ratio were directly correlated with clinical outcomes to assess the capability of O2-enhanced MRI to assess therapeutic effect.All quantitatively (-0.63≤r≤-0.47, p0.05) and qualitatively (0.41≤r≤0.57, p0.05) assessed upper/lower lung ratios showed moderate and statistically significant correlation with clinical outcomes, and improvement in mean relative enhancement ratio showed moderate or good correlation, both statistically significant (-0.44≤r≤0.71, p0.05).O2-enhanced MRI shows potential for more accurate evaluation of postoperative clinical outcome for LVRS candidates than SPECT/CT and can be considered at least as reliable as MDCT.

Purpose: To compare therapeutic effect assessment capability of multidetector-row computed tomography (MDCT), magnetic resonance angiography (MRA), and dynamic perfusion MRI for chronic thromboembolic pulmonary hypertension (CTEPH) patients. Materials and Methods: Twenty-four consecutive CTEPH patients treated with conventional therapy underwent pre- and posttherapeutic MDCT, MRA, dynamic perfusion MRI, 6-minute walk distance (6-MWD), cardiac ultrasound (US), and right heart catheterization. According to therapeutic results, all patients were divided into response (n = 13) and nonresponse (n = 11) groups. CTEPH indexes for MDCT (CTEPHCT) and MRA (CTEPHMRA) were calculated on the basis of embolic burden. Pulmonary perfusion parameter maps were generated from all perfusion MR data, followed by determination of improvements in mean perfusion parameter at regions of interest (ROIs) for each patient. Receiver operating characteristic (ROC)-based positive tests were performed to determine the feasible threshold values for distinguishing two groups. Finally, diagnostic capabilities were compared by means of McNemar's test. Results: When feasible threshold values adapted, specificity (90.9 〈10/11〉%, P < 0.05) and accuracy (95.8 〈23/24〉%, P < 0.05) for improvement in pulmonary blood flow were significantly higher than those for CTEPHCTA (specificity: 36.4 〈4/11〉%, accuracy: 70.8 〈17/24〉%). Conclusion: Dynamic perfusion MRI has better capability for assessment of therapeutic effect on CTEPH patients than does MDCT. J. Magn. Reson. Imaging 2012;36:612–623. © 2012 Wiley Periodicals, Inc.

Pulmonary magnetic resonance (MR) imaging has been put forward as a new research and diagnostic tool mainly to overcome the limitations of computed tomography and nuclear medicine studies. However, pulmonary MR imaging has been difficult to use because of inherently low proton density, a multitude of air-tissue interfaces, which create significant magnetic field distortions and are commonly referred to as susceptibility artifacts; diminishing signal in the lung; and respiratory and/or cardiac motion artifacts. To overcome these drawbacks of pulmonary MR imaging, technical advances made during the last decade in sequencing, scanner and coil, adaptation of parallel imaging techniques, and utilization of contrast media have been reported as being useful for functional and morphologic assessment of various pulmonary diseases including airway diseases. This review article covers (1) pulmonary MR techniques for morphologic and functional assessment of airway diseases, and (2) pulmonary MR imaging for cystic fibrosis, asthma, and chronic obstructive pulmonary disease. Pulmonary MR imaging provides not only morphology-related but also pulmonary function-related information. It has the potential to replace nuclear medicine studies for the identification of regional pulmonary function and may perform a complementary role in airway disease assessment instead of nuclear medicine study. We believe that the findings of further basic studies as well as clinical applications of this new technique will validate the real significance of pulmonary MR imaging for the future of airway disease assessment and its usefulness for diagnostic radiology and pulmonary medicine.

We developed a computer-aided diagnosis (CADx) method for classification between benign nodule, primary lung cancer, and metastatic lung cancer and evaluated the following: (i) the usefulness of the deep convolutional neural network (DCNN) for CADx of the ternary classification, compared with a conventional method (hand-crafted imaging feature plus machine learning), (ii) the effectiveness of transfer learning, and (iii) the effect of image size as the DCNN input. Among 1240 patients of previously-built database, computed tomography images and clinical information of 1236 patients were included. For the conventional method, CADx was performed by using rotation-invariant uniform-pattern local binary pattern on three orthogonal planes with a support vector machine. For the DCNN method, CADx was evaluated using the VGG-16 convolutional neural network with and without transfer learning, and hyperparameter optimization of the DCNN method was performed by random search. The best averaged validation accuracies of CADx were 55.9%, 68.0%, and 62.4% for the conventional method, the DCNN method with transfer learning, and the DCNN method without transfer learning, respectively. For image size of 56, 112, and 224, the best averaged validation accuracy for the DCNN with transfer learning were 60.7%, 64.7%, and 68.0%, respectively. DCNN was better than the conventional method for CADx, and the accuracy of DCNN improved when using transfer learning. Also, we found that larger image sizes as inputs to DCNN improved the accuracy of lung nodule classification.

Purpose: The role of early stage functional assessment of muscle blood flow response (MFR) by dynamic muscle blood oxygen level-dependent (BOLD) magnetic resonance (MR) imaging is unknown. We investigated the effect of smoking on vascular function according to MFR derived from dynamic muscle BOLD MR imaging during postocclusive reactive hyperemia in young smokers and nonsmokers. Methods: Sixteen healthy male volunteers (8 smokers, 8 nonsmokers; mean age, 30.4 « 4.6 years) underwent BOLD MR imaging of the left calf muscle. During reactive hyperemia provoked by a cuff-compression technique, we measured muscle BOLD (mB) using a 3-tesla single-shot multi-echo gradient-echo echo-planar imaging sequence. The 2 key mB variables in the reactive hyperemic phase that we studied were times to half hyperemic peak (T1/2peak) and peak (TTP), each measured from cuff deflation. We used the Welch test to assess differences in these between smokers and nonsmokers. Results: T1/2peak and TTP were significantly longer in smokers (P < 0.05) in reactive hyperemia. T1/2peak was 13.8 « 5.4s in smokers and 7.6 « 1.5s in nonsmokers, and TTP was 67.5 « 18.8s in smokers and 45.4« 7.1s in nonsmokers. Conclusion: Dynamic BOLD MR imaging of calf muscle during postocclusive reactive hyperemia demonstrated statistically significant differences in T1/2peak and TTP between young smokers and nonsmokers, indicating th ep resence of early stage smoking-related deterioration in MFR.

This study aimed to develop software for tumor segmentation on 18 F-fluorodeoxyglucose (FDG) positron emission tomography (PET). To segment the tumor from the background, we used graph cut, whose segmentation energy was generally divided into two terms: the unary and pairwise terms. Locally connected conditional random fields (LCRF) was proposed for the pairwise term. In LCRF, a three-dimensional cubic window with length L was set for each voxel, and voxels within the window were considered for the pairwise term. To evaluate our method, 64 clinically suspected metastatic bone tumors were tested, which were revealed by FDG-PET. To obtain ground truth, the tumors were manually delineated via consensus of two board-certified radiologists. To compare the LCRF accuracy, other types of segmentation were also applied such as region-growing based on 35%, 40%, and 45% of the tumor maximum standardized uptake value (RG35, RG40, and RG45, respectively), SLIC superpixels (SS), and region-based active contour models (AC). To validate the tumor segmentation accuracy, a dice similarity coefficient (DSC) was calculated between manual segmentation and result of each technique. The DSC difference was tested using the Wilcoxon signed rank test. The mean DSCs of LCRF at L = 3, 5, 7, and 9 were 0.784, 0.801, 0.809, and 0.812, respectively. The mean DSCs of other techniques were RG35, 0.633; RG40, 0.675; RG45, 0.689; SS, 0.709; and AC, 0.758. The DSC differences between LCRF and other techniques were statistically significant (p