MSCDA: Multi-level semantic-guided contrast improves unsupervised domain adaptation for breast MRI segmentation in small datasets.

Deep learning (DL) applied to breast tissue segmentation in magnetic resonance imaging (MRI) has received increased attention in the last decade, however, the domain shift which arises from different vendors, acquisition protocols, and biological heterogeneity, remains an important but challenging obstacle on the path towards clinical implementation. In this paper, we propose a novel Multi-level Semantic-guided Contrastive Domain Adaptation (MSCDA) framework to address this issue in an unsupervised manner. Our approach incorporates self-training with contrastive learning to align feature representations between domains. In particular, we extend the contrastive loss by incorporating pixel-to-pixel, pixel-to-centroid, and centroid-to-centroid contrasts to better exploit the underlying semantic information of the image at different levels. To resolve the data imbalance problem, we utilize a category-wise cross-domain sampling strategy to sample anchors from target images and build a hybrid memory bank to store samples from source images. We have validated MSCDA with a challenging task of cross-domain breast MRI segmentation between datasets of healthy volunteers and invasive breast cancer patients. Extensive experiments show that MSCDA effectively improves the model's feature alignment capabilities between domains, outperforming state-of-the-art methods. Furthermore, the framework is shown to be label-efficient, achieving good performance with a smaller source dataset. The code is publicly available at https://github.com/ShengKuangCN/MSCDA.

The effects of the interaction of genetic predisposition with lifestyle factors on bladder cancer risk.

OBJECTIVES: To investigate the association of polygenic risk score (PRS) and bladder cancer (BC) risk and whether this PRS can be offset by a healthy lifestyle. METHODS: Individuals with BC (n = 563) and non-BC controls (n = 483 957) were identified in the UK Biobank, and adjusted Cox regression models were used. A PRS was constructed based on 34 genetic variants associated with BC development, while a healthy lifestyle score (HLS) was constructed based on three lifestyle factors (i.e., smoking, physical activity, and diet). RESULTS: Overall, a negative interaction was observed between the PRS and the HLS (P = 0.02). A 7% higher and 28% lower BC risk per 1-standard deviation (SD) increment in PRS and HLS were observed, respectively. A simultaneous increment of 1 SD in both HLS and PRS was associated with a 6% lower BC risk. In addition, individuals with a high genetic risk and an unfavourable lifestyle showed an increased BC risk compared to individuals with low genetic risk and a favourable lifestyle (hazard ratio 1.55, 95% confidence interval 1.16-1.91; P for trend <0.001). Furthermore, population-attributable fraction (PAF) analysis showed that 12%-15% of the BC cases might have been prevented if individuals had adhered to a healthy lifestyle. CONCLUSION: This large-scale cohort study shows that a genetic predisposition combined with unhealthy behaviours have a joint negative effect on the risk of developing BC. Behavioural lifestyle changes should be encouraged for people through comprehensive, multifactorial approaches, although high-risk individuals may be selected based on genetic risk.

Plasma metabolomic profiling of dietary patterns associated with glucose metabolism status: The Maastricht Study.

BACKGROUND: Glucose metabolism has been reported to be affected by dietary patterns, while the underlying mechanisms involved remain unclear. This study aimed to investigate the potential mediation role of circulating metabolites in relation to dietary patterns for prediabetes and type 2 diabetes. METHODS: Data was derived from The Maastricht Study that comprised of 3441 participants (mean age of 60 years) with 28% type 2 diabetes patients by design. Dietary patterns were assessed using a validated food frequency questionnaire (FFQ), and the glucose metabolism status (GMS) was defined according to WHO guidelines. Both cross-sectional and prospective analyses were performed for the circulating metabolome to investigate their associations and mediations with responses to dietary patterns and GMS. RESULTS: Among 226 eligible metabolite measures obtained from targeted metabolomics, 14 were identified to be associated and mediated with three dietary patterns (i.e. Mediterranean Diet (MED), Dietary Approaches to Stop Hypertension Diet (DASH), and Dutch Healthy Diet (DHD)) and overall GMS. Of these, the mediation effects of 5 metabolite measures were consistent for all three dietary patterns and GMS. Based on a 7-year follow-up, a decreased risk for apolipoprotein A1 (APOA1) and docosahexaenoic acid (DHA) (RR 0.60, 95% CI 0.55, 0.65; RR 0.89, 95% CI 0.83, 0.97, respectively) but an increased risk for ratio of ω-6 to ω-3 fatty acids (RR 1.29, 95% CI 1.05, 1.43) of type 2 diabetes were observed from prediabetes, while APOA1 showed a decreased risk of type 2 diabetes from normal glucose metabolism (NGM; RR 0.82, 95% CI 0.75, 0.89). CONCLUSIONS: In summary, this study suggests that adherence to a healthy dietary pattern (i.e. MED, DASH, or DHD) could affect the GMS through circulating metabolites, which provides novel insights into understanding the biological mechanisms of diet on glucose metabolism and leads to facilitating prevention strategy for type 2 diabetes.

Broad-UNet: Multi-scale feature learning for nowcasting tasks.

Weather nowcasting consists of predicting meteorological components in the short term at high spatial resolutions. Due to its influence in many human activities, accurate nowcasting has recently gained plenty of attention. In this paper, we treat the nowcasting problem as an image-to-image translation problem using satellite imagery. We introduce Broad-UNet, a novel architecture based on the core UNet model, to efficiently address this problem. In particular, the proposed Broad-UNet is equipped with asymmetric parallel convolutions as well as Atrous Spatial Pyramid Pooling (ASPP) module. In this way, the Broad-UNet model learns more complex patterns by combining multi-scale features while using fewer parameters than the core UNet model. The proposed model is applied on two different nowcasting tasks, i.e. precipitation maps and cloud cover nowcasting. The obtained numerical results show that the introduced Broad-UNet model performs more accurate predictions compared to the other examined architectures.

Vegetable intake and the risk of bladder cancer in the BLadder Cancer Epidemiology and Nutritional Determinants (BLEND) international study.

BACKGROUND: Although a potential inverse association between vegetable intake and bladder cancer risk has been reported, epidemiological evidence is inconsistent. This research aimed to elucidate the association between vegetable intake and bladder cancer risk by conducting a pooled analysis of data from prospective cohort studies. METHODS: Vegetable intake in relation to bladder cancer risk was examined by pooling individual-level data from 13 cohort studies, comprising 3203 cases among a total of 555,685 participants. Pooled multivariate hazard ratios (HRs), with corresponding 95% confidence intervals (CIs), were estimated using Cox proportional hazards regression models stratified by cohort for intakes of total vegetable, vegetable subtypes (i.e. non-starchy, starchy, green leafy and cruciferous vegetables) and individual vegetable types. In addition, a diet diversity score was used to assess the association of the varied types of vegetable intake on bladder cancer risk. RESULTS: The association between vegetable intake and bladder cancer risk differed by sex (P-interaction = 0.011) and smoking status (P-interaction = 0.038); therefore, analyses were stratified by sex and smoking status. With adjustment of age, sex, smoking, energy intake, ethnicity and other potential dietary factors, we found that higher intake of total and non-starchy vegetables were inversely associated with the risk of bladder cancer among women (comparing the highest with lowest intake tertile: HR = 0.79, 95% CI = 0.64-0.98, P = 0.037 for trend, HR per 1 SD increment = 0.89, 95% CI = 0.81-0.99; HR = 0.78, 95% CI = 0.63-0.97, P = 0.034 for trend, HR per 1 SD increment = 0.88, 95% CI = 0.79-0.98, respectively). However, no evidence of association was observed among men, and the intake of vegetable was not found to be associated with bladder cancer when stratified by smoking status. Moreover, we found no evidence of association for diet diversity with bladder cancer risk. CONCLUSION: Higher intakes of total and non-starchy vegetable are associated with reduced risk of bladder cancer for women. Further studies are needed to clarify whether these results reflect causal processes and potential underlying mechanisms.

Grain and dietary fiber intake and bladder cancer risk: a pooled analysis of prospective cohort studies.

BACKGROUND: Higher intakes of whole grains and dietary fiber have been associated with lower risk of insulin resistance, hyperinsulinemia, and inflammation, which are known predisposing factors for cancer. OBJECTIVES: Because the evidence of association with bladder cancer (BC) is limited, we aimed to assess associations with BC risk for intakes of whole grains, refined grains, and dietary fiber. METHODS: We pooled individual data from 574,726 participants in 13 cohort studies, 3214 of whom developed incident BC. HRs, with corresponding 95% CIs, were estimated using Cox regression models stratified on cohort. Dose-response relations were examined using fractional polynomial regression models. RESULTS: We found that higher intake of total whole grain was associated with lower risk of BC (comparing highest with lowest intake tertile: HR: 0.87; 95% CI: 0.77, 0.98; HR per 1-SD increment: 0.95; 95% CI: 0.91, 0.99; P for trend: 0.023). No association was observed for intake of total refined grain. Intake of total dietary fiber was also inversely associated with BC risk (comparing highest with lowest intake tertile: HR: 0.86; 95% CI: 0.76, 0.98; HR per 1-SD increment: 0.91; 95% CI: 0.82, 0.98; P for trend: 0.021). In addition, dose-response analyses gave estimated HRs of 0.97 (95% CI: 0.95, 0.99) for intake of total whole grain and 0.96 (95% CI: 0.94, 0.98) for intake of total dietary fiber per 5-g daily increment. When considered jointly, highest intake of whole grains with the highest intake of dietary fiber showed 28% reduced risk (95% CI: 0.54, 0.93; P for trend: 0.031) of BC compared with the lowest intakes, suggesting potential synergism. CONCLUSIONS: Higher intakes of total whole grain and total dietary fiber are associated with reduced risk of BC individually and jointly. Further studies are needed to clarify the underlying mechanisms for these findings.

Deep neural-kernel blocks.

This paper introduces novel deep architectures using the hybrid neural-kernel core model as the first building block. The proposed models follow a combination of a neural networks based architecture and a kernel based model enriched with pooling layers. In particular, in this context three kernel blocks with average, maxout and convolutional pooling layers are introduced and examined. We start with a simple merging layer which averages the output of the previous representation layers. The maxout layer on the other hand triggers competition among different representations of the input. Thanks to this pooling layer, not only the dimensionality of the output of multi-scale representations is reduced but also multiple sub-networks are formed within the same model. In the same context, the pointwise convolutional layer is also employed with the aim of projecting the multi-scale representations onto a new space. Experimental results show an improvement over the core deep hybrid model as well as kernel based models on several real-life datasets.

Regularized Semipaired Kernel CCA for Domain Adaptation.

Domain adaptation learning is one of the fundamental research topics in pattern recognition and machine learning. This paper introduces a regularized semipaired kernel canonical correlation analysis formulation for learning a latent space for the domain adaptation problem. The optimization problem is formulated in the primal-dual least squares support vector machine setting where side information can be readily incorporated through regularization terms. The proposed model learns a joint representation of the data set across different domains by solving a generalized eigenvalue problem or linear system of equations in the dual. The approach is naturally equipped with out-of-sample extension property, which plays an important role for model selection. Furthermore, the Nyström approximation technique is used to make the computational issues due to the large size of the matrices involved in the eigendecomposition feasible. The learned latent space of the source domain is fed to a multiclass semisupervised kernel spectral clustering model that can learn from both labeled and unlabeled data points of the source domain in order to classify the data instances of the target domain. Experimental results are given to illustrate the effectiveness of the proposed approaches on synthetic and real-life data sets.

Robust Support Vector Machines for Classification with Nonconvex and Smooth Losses.

This letter addresses the robustness problem when learning a large margin classifier in the presence of label noise. In our study, we achieve this purpose by proposing robustified large margin support vector machines. The robustness of the proposed robust support vector classifiers (RSVC), which is interpreted from a weighted viewpoint in this work, is due to the use of nonconvex classification losses. Besides the robustness, we also show that the proposed RSCV is simultaneously smooth, which again benefits from using smooth classification losses. The idea of proposing RSVC comes from M-estimation in statistics since the proposed robust and smooth classification losses can be taken as one-sided cost functions in robust statistics. Its Fisher consistency property and generalization ability are also investigated. Besides the robustness and smoothness, another nice property of RSVC lies in the fact that its solution can be obtained by solving weighted squared hinge loss-based support vector machine problems iteratively. We further show that in each iteration, it is a quadratic programming problem in its dual space and can be solved by using state-of-the-art methods. We thus propose an iteratively reweighted type algorithm and provide a constructive proof of its convergence to a stationary point. Effectiveness of the proposed classifiers is verified on both artificial and real data sets.

Incremental multi-class semi-supervised clustering regularized by Kalman filtering.

This paper introduces an on-line semi-supervised learning algorithm formulated as a regularized kernel spectral clustering (KSC) approach. We consider the case where new data arrive sequentially but only a small fraction of it is labeled. The available labeled data act as prototypes and help to improve the performance of the algorithm to estimate the labels of the unlabeled data points. We adopt a recently proposed multi-class semi-supervised KSC based algorithm (MSS-KSC) and make it applicable for on-line data clustering. Given a few user-labeled data points the initial model is learned and then the class membership of the remaining data points in the current and subsequent time instants are estimated and propagated in an on-line fashion. The update of the memberships is carried out mainly using the out-of-sample extension property of the model. Initially the algorithm is tested on computer-generated data sets, then we show that video segmentation can be cast as a semi-supervised learning problem. Furthermore we show how the tracking capabilities of the Kalman filter can be used to provide the labels of objects in motion and thus regularizing the solution obtained by the MSS-KSC algorithm. In the experiments, we demonstrate the performance of the proposed method on synthetic data sets and real-life videos where the clusters evolve in a smooth fashion over time.

Multiclass semisupervised learning based upon kernel spectral clustering.

This paper proposes a multiclass semisupervised learning algorithm by using kernel spectral clustering (KSC) as a core model. A regularized KSC is formulated to estimate the class memberships of data points in a semisupervised setting using the one-versus-all strategy while both labeled and unlabeled data points are present in the learning process. The propagation of the labels to a large amount of unlabeled data points is achieved by adding the regularization terms to the cost function of the KSC formulation. In other words, imposing the regularization term enforces certain desired memberships. The model is then obtained by solving a linear system in the dual. Furthermore, the optimal embedding dimension is designed for semisupervised clustering. This plays a key role when one deals with a large number of clusters.

Approximate solutions to ordinary differential equations using least squares support vector machines.

In this paper, a new approach based on least squares support vector machines (LS-SVMs) is proposed for solving linear and nonlinear ordinary differential equations (ODEs). The approximate solution is presented in closed form by means of LS-SVMs, whose parameters are adjusted to minimize an appropriate error function. For the linear and nonlinear cases, these parameters are obtained by solving a system of linear and nonlinear equations, respectively. The method is well suited to solving mildly stiff, nonstiff, and singular ODEs with initial and boundary conditions. Numerical results demonstrate the efficiency of the proposed method over existing methods.