[en] Weather is an important factor for air quality. While there have been increasing attentions to long-term (monthly and seasonal) air pollution such as regional hazes from land-clearing fires during El Niño, the weather-air quality relationships are much less understood at long-term than short-term (daily and weekly) scales. This study is aimed to fill this gap through analyzing correlations between meteorological variables and air quality at various timescales. A regional correlation scale was defined to measure the longest time with significant correlations at a substantial large number of sites. The air quality index (API) and five meteorological variables during 2001–2012 at 40 eastern China sites were used. The results indicate that the API is correlated to precipitation negatively and air temperature positively across eastern China, and to wind, relative humidity and air pressure with spatially varied signs. The major areas with significant correlations vary with meteorological variables. The correlations are significant not only at short-term but also at long-term scales, and the important variables are different between the two types of scales. The concurrent regional correlation scales reach seasonal at p < 0.05 and monthly at p < 0.001 for wind speed and monthly at p < 0.01 for air temperature and relative humidity. Precipitation, which was found to be the most important variable for short-term air quality conditions, and air pressure are not important for long-term air quality. The lagged correlations are much smaller in magnitude than the concurrent correlations and their regional correction scales are at long term only for wind speed and relative humidity. It is concluded that wind speed should be considered as a primary predictor for statistical prediction of long-term air quality in a large region over eastern China. Relative humidity and temperature are also useful predictors but at less significant levels.

[en] X-ray luminescence tomography (XLT) is an imaging technology based on X-ray-excitable materials. The main purpose of this paper is to obtain quantitative luminescence concentration using the structural information of the X-ray computed tomography (XCT) in the hybrid cone beam XLT/XCT system. A multi-wavelength luminescence cone beam XLT method with the structural a priori information is presented to relieve the severe ill-posedness problem in the cone beam XLT. The nanophosphors and phantom experiments were undertaken to access the linear relationship of the system response. Then, an in vivo mouse experiment was conducted. The in vivo experimental results show that the recovered concentration error as low as 6.67% with the location error of 0.85 mm can be achieved. The results demonstrate that the proposed method can accurately recover the nanophosphor inclusion and realize the quantitative imaging

[en] To evaluate the effectiveness of bag-of-features (BOF)-based radiomics for differentiating ocular adnexal lymphoma (OAL) and idiopathic orbital inflammation (IOI) from contrast-enhanced MRI (CE-MRI). Fifty-six patients with pathologically confirmed IOI (28 patients) and OAL (28 patients) were randomly divided into training (n = 42) and testing (n = 14) groups. One hundred sixty texture features extracted from the CE-MR image were encoded into the BOF representation with fewer features. The support vector machine (SVM) with a linear kernel was used as the classifier. Data augmented was performed by cropping orbital lesions in different directions to alleviate the over-fitting problem. Student’s t test and the Holm-Bonferroni method were employed to compare the performance of different analysis methods. The chi-square test was used to compare the analysis with MRI and human radiological diagnosis. In the independent testing group, the differentiation by the BOF features with augmentation achieved an area under the curve (AUC) of 0.803 (95% CI: 0.725–0.880), which was significantly higher than that of the BOF features without augmentation and that of the texture features (p < 0.05). In addition, the same radiomic analysis with pre-contrast MRI obtained an AUC of 0.618 (95% CI: 0.560–0.677), which was significantly lower than that with CE-MRI. The diagnostic performance of the analysis with CE-MRI was significantly better than the radiology resident (p < 0.05) but had no significant difference with the experienced radiologist, even though there was less consistency between the radiomic analysis and the human visual diagnosis. The BOF-based radiomics may be helpful for the differentiation between OAL and IOI.

[en] To evaluate the value of deep learning (DL) combining multimodal radiomics and clinical and imaging features for differentiating ocular adnexal lymphoma (OAL) from idiopathic orbital inflammation (IOI). Eighty-nine patients with histopathologically confirmed OAL (n = 39) and IOI (n = 50) were divided into training and validation groups. Convolutional neural networks and multimodal fusion layers were used to extract multimodal radiomics features from the T1-weighted image (T1WI), T2-weighted image, and contrast-enhanced T1WI. These multimodal radiomics features were then combined with clinical and imaging features and used together to differentiate between OAL and IOI. The area under the curve (AUC) was used to evaluate DL models with different features under five-fold cross-validation. The Student t-test, chi-squared, or Fisher exact test was used for comparison of different groups. In the validation group, the diagnostic AUC of the DL model using combined features was 0.953 (95% CI, 0.895-1.000), higher than that of the DL model using multimodal radiomics features (0.843, 95% CI, 0.786-0.898, p < 0.01) or clinical and imaging features only (0.882, 95% CI, 0.782-0.982, p = 0.13). The DL model built on multimodal radiomics features outperformed those built on most bimodalities and unimodalities (p < 0.05). In addition, the DL-based analysis with the orbital cone area (covering both the orbital mass and surrounding tissues) was superior to that with the region of interest (ROI) covering only the mass area, although the difference was not significant (p = 0.33). DL-based analysis that combines multimodal radiomics features with clinical and imaging features may help to differentiate between OAL and IOI. It is difficult to differentiate OAL from IOI due to the overlap in clinical and imaging manifestations. Radiomics has shown potential for noninvasive diagnosis of different orbital lymphoproliferative disorders. DL-based analysis combining radiomics and imaging and clinical features may help the differentiation between OAL and IOI.

[en] To solve the multicollinearity issue and unequal contribution of vascular parameters for the quantification of angiogenesis, we developed a quantification evaluation method of vascular parameters for angiogenesis based on in vivo micro-CT imaging of hindlimb ischemic model mice. Taking vascular volume as the ground truth parameter, nine vascular parameters were first assembled into sparse principal components (PCs) to reduce the multicolinearity issue. Aggregated boosted trees (ABTs) were then employed to analyze the importance of vascular parameters for the quantification of angiogenesis via the loadings of sparse PCs. The results demonstrated that vascular volume was mainly characterized by vascular area, vascular junction, connectivity density, segment number and vascular length, which indicated they were the key vascular parameters for the quantification of angiogenesis. The proposed quantitative evaluation method was compared with both the ABTs directly using the nine vascular parameters and Pearson correlation, which were consistent. In contrast to the ABTs directly using the vascular parameters, the proposed method can select all the key vascular parameters simultaneously, because all the key vascular parameters were assembled into the sparse PCs with the highest relative importance. (paper)

[en] Highlights: • A core-shell structural 420-nano WC-Co composite powder by high energy ball milling. • The composite powder has high circularity and high fluidity, which is suitable for laser cladding. • There are no holes and cracks in the cladding layer. Nano WC-Co is uniformly distributed in the cladding layer. -- Abstract: WC is considered as a good choice for the reinforced phase of many composite alloys because of its excellent hardness and good wear resistance. In the past, micro WC particles physically mixed with micro metal particles were also widely used in laser cladding, however, micron WC particles often sink down to the bottom and may cause uneven element distribution, holes, cracks, and other defects in the cladding layer due to its high density and high laser absorption rate. Here, we prepare a core-shell structural 420-nano WC-Co composite powder by high energy ball milling. The composite powder has high circularity and high fluidity, which is suitable for laser cladding. The cladding layer with no holes or cracks can be attained using the core-shell structural composite powders, moreover, the nano WC-Co particles are evenly distributed in the cladding layer and the microhardness of the cladding layer has enhanced significantly.

[en] Detection of different classes of atherosclerotic plaques is important for early intervention of coronary artery diseases. However, previous methods focused either on the detection of a specific class of coronary plaques or on the distinction between plaques and normal arteries, neglecting the classification of different classes of plaques. Therefore, we proposed an automatic multi-class coronary atherosclerosis plaque detection and classification framework. Firstly, we retrieved the transverse cross sections along centerlines from the computed tomography angiography. Secondly, we extracted the region of interests based on coarse segmentation. Thirdly, we extracted a random radius symmetry (RRS) feature vector, which incorporates multiple descriptions into a random strategy and greatly augments the training data. Finally, we fed the RRS feature vector into the multi-class coronary plaque classifier. In experiments, we compared our proposed framework with other methods on the cross sections of Rotterdam Coronary Datasets, including 729 non-calcified plaques, 511 calcified plaques, and 546 mixed plaques. Our RRS with support vector machine outperforms the intensity feature vector and the random forest classifier, with the average precision of 92.6 ± 1.9% and average recall of 94.3 ± 2.1%. The proposed framework provides a computer-aided diagnostic method for multi-class plaque detection and classification.

.