Rtr1 is required for Rpb1-Rpb2 assembly of RNAPII and prevents their cytoplasmic clump formation.

RNA polymerase II (RNAPII) is an essential machinery for catalyzing mRNA synthesis and controlling cell fate in eukaryotes. Although the structure and function of RNAPII have been relatively defined, the molecular mechanism of its assembly process is not clear. The identification and functional analysis of assembly factors will provide new understanding to transcription regulation. In this study, we identify that RTR1, a known transcription regulator, is a new multicopy genetic suppressor of mutants of assembly factors Gpn3, Gpn2, and Rba50. We demonstrate that Rtr1 is directly required to assemble the two largest subunits of RNAPII by coordinating with Gpn3 and Npa3. Deletion of RTR1 leads to cytoplasmic clumping of RNAPII subunit and multiple copies of RTR1 can inhibit the formation of cytoplasmic clump of RNAPII subunit in gpn3-9 mutant, indicating a new layer function of Rtr1 in checking proper assembly of RNAPII. In addition, we find that disrupted activity of Rtr1 phosphatase does not trigger the formation of cytoplasmic clump of RNAPII subunit in a catalytically inactive mutant of RTR1. Based on these results, we conclude that Rtr1 cooperates with Gpn3 and Npa3 to assemble RNAPII core.

HC-1119, a deuterated Enzalutamide, inhibits Migration, Invasion and Metastasis of the AR-positive triple-negative breast Cancer cells.

Triple-negative breast cancers (TNBCs) are aggressive, and they develop metastasis at earlier stages, relapse more frequently, and exhibits poorer prognosis than other subtypes of breast cancer. Due to the lack of estrogen receptor for endocrine therapy and HER2 for targeted therapy, new targeted therapies for TNBCs are urgently needed. Enzalutamide is a second-generation androgen receptor (AR) inhibitor, and HC-1119 is a new synthetic deuterated enzalutamide. Owing to the isotope effect, HC-1119 has many advantages over enzalutamide, including slow metabolism, high plasma concentration and low brain exposure. However, the efficacy of HC-1119 in inhibition of AR function in triple-negative breast cancer (TNBC) has not been studied. In this study, we found high-level AR expression in both Hs578T and SUM159PT TNBC cell lines. Activation of AR by dihydrotestosterone (DHT) in both cell lines increased AR protein, induced AR-nuclear localization, enhanced cell migration and invasion in culture, and promoted liver metastasis in mice. Importantly, cotreatment with HC-1119 of these cells efficiently abolished all of these effects of DHT on both Hs578T and SUM159PT cells. These results indicate that HC-1119 is a very effective new second-generation AR antagonist that can inhibit the migration, invasion and metastasis of the AR-positive TNBC cells.

Three-dimensional change in temperature sensitivity of northern vegetation phenology.

Understanding how the temperature sensitivity of phenology changes with three spatial dimensions (altitude, latitude, and longitude) is critical for the prediction of future phenological synchronization. Here we investigate the spatial pattern of temperature sensitivity of spring and autumn phenology with altitude, latitude, and longitude during 1982-2016 across mid- and high-latitude Northern Hemisphere (north of 30°N). We find distinct spatial patterns of temperature sensitivity of spring phenology (hereafter "spring ST ") among altitudinal, latitudinal, and longitudinal gradient. Spring ST decreased with altitude mostly over eastern Europe, whereas the opposite occurs in eastern North America and the north China plain. Spring ST decreased with latitude mainly in the boreal regions of North America, temperate Eurasia, and the arid/semi-arid regions of Central Asia. This distribution may be related to the increased temperature variance, decreased precipitation, and radiation with latitude. Compared to spring ST , the spatial pattern of temperature sensitivity of autumn phenology (hereafter "autumn ST ") is more heterogeneous, only showing a clear spatial pattern of autumn ST along the latitudinal gradient. Our results highlight the three-dimensional view to understand the phenological response to climate change and provide new metrics for evaluating phenological models. Accordingly, establishing a dense, high-quality three-dimensional observation system of phenology data is necessary for enhancing our ability to both predict phenological changes under changing climatic conditions and to facilitate sustainable management of ecosystems.

Modeling leaf senescence of deciduous tree species in Europe.

Autumnal leaf senescence signals the end of photosynthetic activities in temperate deciduous trees and consequently exerts a strong control on various ecological processes. Predicting leaf senescence dates (LSD) with high accuracy is thus a prerequisite for better understanding the climate-ecosystem interactions. However, modeling LSD at large spatial and temporal scales is challenging. In this study, first, we used 19972 site-year records (848 sites and four deciduous tree species) from the PAN European Phenology network to calibrate and evaluate six leaf senescence models during the period 1980-2013. Second, we extended the spatial analysis by repeating the procedure across Europe using satellite-derived end of growing season and a forest map. Overall, we found that models that considered photoperiod and temperature interactions outperformed models using simple temperature or photoperiod thresholds for Betula pendula, Fagus sylvatica and Quercus robur. On the contrary, no model displayed reasonable predictions for Aesculus hippocastanum. This inter-model comparison indicates that, contrary to expectation, photoperiod does not significantly modulate the accumulation of cooling degree days (CDD). On the other hand, considering the carryover effect of leaf unfolding date could promote the models' predictability. The CDD models generally matched the observed LSD at species level and its interannual variation, but were limited in explaining the inter-site variations, indicating that other environmental cues need to be considered in future model development. The discrepancies remaining between model simulations and observations highlight the need of manipulation studies to elucidate the mechanisms behind the leaf senescence process and to make current models more realistic.

An Assisted Diagnosis System for Detection of Early Pulmonary Nodule in Computed Tomography Images.

Lung cancer is still the most concerned disease around the world. Lung nodule generates in the pulmonary parenchyma which indicates the latent risk of lung cancer. Computer-aided pulmonary nodules detection system is necessary, which can reduce diagnosis time and decrease mortality of patients. In this study, we have proposed a new computer aided diagnosis (CAD) system for detection of early pulmonary nodule, which can help radiologists quickly locate suspected nodules and make judgments. This system consists of four main sections: pulmonary parenchyma segmentation, nodule candidate detection, features extraction (total 22 features) and nodule classification. The publicly available data set created by the Lung Image Database Consortium (LIDC) is used for training and testing. This study selects 6400 slices from 80 CT scans containing totally 978 nodules, which is labeled by four radiologists. Through a fast segmentation method proposed in this paper, pulmonary nodules including 888 true nodules and 11,379 false positive nodules are segmented. By means of an ensemble classifier, Random Forest (RF), this study acquires 93.2, 92.4, 94.8, 97.6% of accuracy, sensitivity, specificity, area under the curve (AUC), respectively. Compared with support vector machine (SVM) classifier, RF can reduce more false positive nodules and acquire larger AUC. With the help of this CAD system, radiologist can be provided with a great reference for pulmonary nodule diagnosis timely.

Suppressing label noise in medical image classification using mixup attention and self-supervised learning.

Deep neural networks (DNNs) have been widely applied in medical image classification and achieve remarkable classification performance. These achievements heavily depend on large-scale accurately annotated training data. However, label noise is inevitably introduced in the medical image annotation, as the labeling process heavily relies on the expertise and experience of annotators. Meanwhile, DNNs suffer from overfitting noisy labels, degrading the performance of models. Therefore, in this work, we innovatively devise a noise-robust training approach to mitigate the adverse effects of noisy labels in medical image classification. Specifically, we incorporate contrastive learning and intra-group mixup attention strategies into vanilla supervised learning. The contrastive learning for feature extractor helps to enhance visual representation of DNNs. The intra-group mixup attention module constructs groups and assigns self-attention weights for group-wise samples, and subsequently interpolates massive noisy-suppressed samples through weighted mixup operation. We conduct comparative experiments on both synthetic and real-world noisy medical datasets under various noise levels. Rigorous experiments validate that our noise-robust method with contrastive learning and mixup attention can effectively handle with label noise, and is superior to state-of-the-art methods. An ablation study also shows that both components contribute to boost model performance. The proposed method demonstrates its capability of curb label noise and has certain potential toward real-world clinic applications.

GlanceSeg: Real-time microaneurysm lesion segmentation with gaze-map-guided foundation model for early detection of diabetic retinopathy.

Early-stage diabetic retinopathy (DR) presents challenges in clinical diagnosis due to inconspicuous and minute microaneurysms (MAs), resulting in limited research in this area. Additionally, the potential of emerging foundation models, such as the segment anything model (SAM), in medical scenarios remains rarely explored. In this work, we propose a human-in-the-loop, label-free early DR diagnosis framework called GlanceSeg, based on SAM. GlanceSeg enables real-time segmentation of MA lesions as ophthalmologists review fundus images. Our human-in-the-loop framework integrates the ophthalmologist's gaze maps, allowing for rough localization of minute lesions in fundus images. Subsequently, a saliency map is generated based on the located region of interest, which provides prompt points to assist the foundation model in efficiently segmenting MAs. Finally, a domain knowledge filtering (DKF) module refines the segmentation of minute lesions. We conducted experiments on two newly-built public datasets, i.e., IDRiD and Retinal-Lesions, and validated the feasibility and superiority of GlanceSeg through visualized illustrations and quantitative measures. Additionally, we demonstrated that GlanceSeg improves annotation efficiency for clinicians and further enhances segmentation performance through fine-tuning using annotations. The clinician-friendly GlanceSeg is able to segment small lesions in real-time, showing potential for clinical applications.

Multi-Physics Modeling of Melting-Solidification Characteristics in Laser Powder Bed Fusion Process of 316L Stainless Steel.

In the laser powder bed fusion process, the melting-solidification characteristics of 316L stainless steel have a great effect on the workpiece quality. In this paper, a multi-physics model was constructed using the finite volume method (FVM) to simulate the melting-solidification process of a 316L powder bed via laser powder bed fusion. In this physical model, the phase change process, the influence of temperature gradient on surface tension of molten pool, and the influence of recoil pressure caused by the metal vapor on molten pool surface were considered. Using this model, the effects of laser scanning speed, hatch space, and laser power on temperature distribution, keyhole depth, and workpiece quality were studied. This study can be used to guide the optimization of process parameters, which is beneficial to the improvement of workpiece quality.

Multiobjective optimization of roadheader shovel-plate parameters based on improved particle swarm optimization and grey decision.

The development of tunneling equipment still lags behind, limiting rapid and accurate tunneling and restricting efficient production in coal mines. Thus, improving the reliability and design of roadheaders becomes essential. As the shovel plate is an essential part of a roadheader, improving its parameters can increase the roadheader performance. The parameter optimization of roadheader shovel plate is multi-objective optimization. Because of conventional multiobjective optimization requires strong prior knowledge, often provides low-quality results, and presents vulnerability to initialization and other shortcomings when used in practice. We propose an improved particle swarm optimization (PSO) algorithm that takes the minimum Euclidean distance from a base value as the evaluation criterion for global and individual extreme values. The improved algorithm enables multiobjective parallel optimization by providing a non-inferior solution set. Then, the optimal solution is searched in this set using grey decision to obtain the optimal solution. To validate the proposed method, the multiobjective optimization problem of the shovel-plate parameters is formulated for its solution. Before optimization shovel-plate most important parameters l is the width of the shovel plate l = 3.2 m, ß is the inclination angle of the shovel plate and ß = ,19°. When doing optimization, set accelerated factor c1=c2=2, population size N = 20, and maximum number of iterations Tmax = 100. In addition, speed V was restricted by V=Vimax-Vimin, and inertia factor W was dynamic and linearly decreasing, w(t)=wmin+wmax-wminN(N-t), with wmax=0.9 and wmin=0.4. In addition, r1 and r2 were set randomly in [0, 1], while optimization degree η was set to 30%. And then we executed the improved PSO, obtaining 2000 non-inferior solutions. Apply grey decision to find the optimal solution. The optimal roadheader shovel-plate parameters are l = 3.144 m and ß = 16.88°. Comparative analysis is made before and after optimization, the optimized parameters were substituted into the model and simulated. Obtained that the optimized parameters of shovel-plate can reduce the mass of the shovel plate decreases by 14.3%, while the propulsive resistance decreases by 6.62%, and the load capacity increases by 3.68%. Thus jointly achieving the optimization goals of reducing the propulsive resistance while increasing the load capacity. The feasibility of the proposed multiobjective optimization method with improved particle swarm optimization and grey decision is verified, and the method can provide convenient multiobjective optimization in engineering practice.

Emergy-based sustainability measurement and evaluation of industrial production systems.

The sustainability of industrial production systems is considered to be the key to promoting green transformation and upgrading of manufacturing industry. This study proposes an emergy-based method for the measurement and evaluation of the sustainability of industrial production systems. The method uses emergy as a measure, expresses all types of inputs and outputs of production systems in terms of solar emergy, and constructs indicators for the sustainable evaluation of industrial production systems that account for economic and environmental benefits from the perspectives of system functions, ecology, and sustainability. This method was applied to a disc cam machining system at a machinery company, where the user was able to quantify the sustainability of the production system and, through feedback, optimise the production process to reduce energy consumption and environmental pollution and improve the sustainability of the production process. The results show that after optimization, the emergy yield ratio of the disc cam machining system is increased from 1.45 to 4.32, the environmental load ratio is reduced by 16%, and the emergy-based sustainability index is increased by 85%. The system has long-term sustainable development capability in the future. This study provides a new theoretical perspective on sustainability assessments of industrial production systems, and our findings provide a scientific basis for guiding the sound operation and sustainable development of industrial production systems.

Multi-Learner Based Deep Meta-Learning for Few-Shot Medical Image Classification.

Few-shot learning (FSL) is promising in the field of medical image analysis due to high cost of establishing high-quality medical datasets. Many FSL approaches have been proposed in natural image scenes. However, present FSL methods are rarely evaluated on medical images and the FSL technology applicable to medical scenarios need to be further developed. Meta-learning has supplied an optional framework to address the challenging FSL setting. In this paper, we propose a novel multi-learner based FSL method for multiple medical image classification tasks, combining meta-learning with transfer-learning and metric-learning. Our designed model is composed of three learners, including auto-encoder, metric-learner and task-learner. In transfer-learning, all the learners are trained on the base classes. In the ensuing meta-learning, we leverage multiple novel tasks to fine-tune the metric-learner and task-learner in order to fast adapt to unseen tasks. Moreover, to further boost the learning efficiency of our model, we devised real-time data augmentation and dynamic Gaussian disturbance soft label (GDSL) scheme as effective generalization strategies of few-shot classification tasks. We have conducted experiments for three-class few-shot classification tasks on three newly-built challenging medical benchmarks, BLOOD, PATH and CHEST. Extensive comparisons to related works validated that our method achieved top performance both on homogeneous medical datasets and cross-domain datasets.

Discriminative ensemble meta-learning with co-regularization for rare fundus diseases diagnosis.

Deep neural networks (DNNs) have been widely applied in the medical image community, contributing to automatic ophthalmic screening systems for some common diseases. However, the incidence of fundus diseases patterns exhibits a typical long-tailed distribution. In clinic, a small number of common fundus diseases have sufficient observed cases for large-scale analysis while most of the fundus diseases are infrequent. For these rare diseases with extremely low-data regimes, it is challenging to train DNNs to realize automatic diagnosis. In this work, we develop an automatic diagnosis system for rare fundus diseases, based on the meta-learning framework. The system incorporates a co-regularization loss and the ensemble-learning strategy into the meta-learning framework, fully leveraging the advantage of multi-scale hierarchical feature embedding. We initially conduct comparative experiments on our newly-constructed lightweight multi-disease fundus images dataset for the few-shot recognition task (namely, FundusData-FS). Moreover, we verify the cross-domain transferability from miniImageNet to FundusData-FS, and further confirm our method's good repeatability. Rigorous experiments demonstrate that our method can detect rare fundus diseases, and is superior to the state-of-the-art methods. These investigations demonstrate that the potential of our method for the real clinical practice is promising.

Identification and characterization of a novel aminoglycoside O-nucleotidyltransferase ANT(6)-If from Paenibacillus thiaminolyticus PATH554.

Background: Paenibacillus thiaminolyticus, a species of genus Paenibacillus of the family Paenibacillaceae, exists widely in environments and habitats in various plants and worms, and occasionally causes human infections. This work aimed to characterize the function of a novel aminoglycoside O-nucleotidyltransferase resistance gene, designated ant(6)-If, from a P. thiaminolyticus strain PATH554. Methods: Molecular cloning, antimicrobial susceptibility testing, enzyme expression and purification, and kinetic analysis were used to validate the function of the novel gene. Whole-genome sequencing and comparative genomic analysis were performed to investigate the phylogenetic relationship of ANT(6)-If and other aminoglycoside O-nucleotidyltransferases, and the synteny of ant(6)-If related sequences. Results: The recombinant with the cloned ant(6)-If gene (pMD19-ant(6)-If/DH5α) demonstrated a 128-fold increase of minimum inhibitory concentration level against streptomycin, compared with the control strains (DH5α and pMD19/DH5α). The kinetic parameter kcat/Km of ANT(6)-If for streptomycin was 9.01 × 103 M-1·s-1. Among the function-characterized resistance genes, ANT(6)-If shared the highest amino acid sequence identity of 75.35% with AadK. The ant(6)-If gene was located within a relatively conserved genomic region in the chromosome. Conclusion: ant(6)-If conferred resistance to streptomycin. The study of a novel resistance gene in an unusual environmental bacterium in this work contributed to elucidating the resistance mechanisms in the microorganisms.

Predicting ischemic stroke risk from atrial fibrillation based on multi-spectral fundus images using deep learning.

Background: Ischemic stroke (IS) is one of the most common serious secondary diseases of atrial fibrillation (AF) within 1 year after its occurrence, both of which have manifestations of ischemia and hypoxia of the small vessels in the early phase of the condition. The fundus is a collection of capillaries, while the retina responds differently to light of different wavelengths. Predicting the risk of IS occurring secondary to AF, based on subtle differences in fundus images of different wavelengths, is yet to be explored. This study was conducted to predict the risk of IS occurring secondary to AF based on multi-spectrum fundus images using deep learning. Methods: A total of 150 AF participants without suffering from IS within 1 year after discharge and 100 IS participants with persistent arrhythmia symptoms or a history of AF diagnosis in the last year (defined as patients who would develop IS within 1 year after AF, based on fundus pathological manifestations generally prior to symptoms of the brain) were recruited. Fundus images at 548, 605, and 810 nm wavelengths were collected. Three classical deep neural network (DNN) models (Inception V3, ResNet50, SE50) were trained. Sociodemographic and selected routine clinical data were obtained. Results: The accuracy of all DNNs with the single-spectral or multi-spectral combination images at the three wavelengths as input reached above 78%. The IS detection performance of DNNs with 605 nm spectral images as input was relatively more stable than with the other wavelengths. The multi-spectral combination models acquired a higher area under the curve (AUC) scores than the single-spectral models. Conclusions: The probability of IS secondary to AF could be predicted based on multi-spectrum fundus images using deep learning, and combinations of multi-spectrum images improved the performance of DNNs. Acquiring different spectral fundus images is advantageous for the early prevention of cardiovascular and cerebrovascular diseases. The method in this study is a beneficial preliminary and initiative exploration for diseases that are difficult to predict the onset time such as IS.

TNFAIP3 mediates FGFR1 activation-induced breast cancer angiogenesis by promoting VEGFA expression and secretion.

PURPOSE: To investigate the role and mechanism of TNF-inducible protein 3(TNFAIP3) in breast cancer angiogenesis induced by fibroblast growth factor receptor1 (FGFR1) activation. METHODS: The immunohistochemical assay was used to detect the expression of vascular endothelial cell marker CD31 and CD105 in mice DCIS.COM-iFGFR1 transplanted tumor (previously established by our group). The effects of TNFAIP3 knockout/knockdown breast cancer cell lines on angiogenesis, migration, and invasion of Human Umbilical Vein Endothelial Cells (HUVEC) were detected by the tubulogenesis and Trewells assay. RNA-seq analysis of TNFAIP3 downstreams differential genes after TNFAIP3 knockdown. The expression and secretion of VEGFA after FGFR1 activation in breast cancer cells were detected by qPCR, Western blot, and ELISA. RESULTS: Immunohistochemistry showed that TNFAIP3 knockout inhibited the expression of CD31 and CD105 in DCIS grafted tumors promoted by FGFR1 activation. Tubulogenesis and Trewells experiments showed that TNFAIP3 gene knockout/knockdown inhibited the angiogenesis, migration, and invasion of HUVEC cells promoted by FGFR1 activation. qPCR assay showed that VEGFA mRNA level in the TNFAIP3 knockdown cell line was significantly down-regulated (p < 0.05). qPCR, Western blot and ELISA results showed that TNFAIP3 gene knockout/knockdown could inhibit the expression and secretion of VEGFA in breast cancer cells induced by FGFR1 activation. CONCLUSION: TNFAIP3 promotes breast cancer angiogenesis induced by FGFR1 activation through the expression and secretion of VEGFA.

TNFAIP3 promotes ALDH-positive breast cancer stem cells through FGFR1/MEK/ERK pathway.

Breast cancer stem cells (BCSCs) are a tiny population of self-renewing cells that may contribute to cancer initiation, progression, and resistance to therapy in patients. In our prior study, we found that tumor necrosis factor alpha-induced protein 3 (TNFAIP3) is necessary for fibroblast growth factors receptor 1 (FGFR1) signaling-promoted tumor growth and progression in breast cancer (BC). This study aims to investigate the involvement of TNFAIP3 in regulating BCSCs. In this work, we showed that ALDH-positive BCSCs were increased by activating the FGFR1-MEK-ERK pathway, meanwhile utilizing FGFR1 inhibitor, MEK inhibitor, or ERK inhibitor reversed the phenomenon in BC cells. Moreover, ALDH-positive BCSCs were decreased in TNFAIP3-knockout or TNFAIP3-depressing cells. In vivo analysis displayed that TNFAIP3-silenced MDA-MB-231 xenografts developed smaller tumors and ALDH immunostaining levels were significantly lower in TNFAIP3-depressing or TNFAIP3-knockout tumor tissues. Besides, our results also revealed that TNFAIP3 influences the transcription stemness factors gene expression. Taken together, TNFAIP3 could be a potential regulator in FGFR1-MEK-ERK-promoted ALDH-positive BCSCs.

Bayesian statistics-guided label refurbishment mechanism: Mitigating label noise in medical image classification.

PURPOSE: Deep neural networks (DNNs) have been widely applied in medical image classification, benefiting from its powerful mapping capability among medical images. However, these existing deep learning-based methods depend on an enormous amount of carefully labeled images. Meanwhile, noise is inevitably introduced in the labeling process, degrading the performance of models. Hence, it is significant to devise robust training strategies to mitigate label noise in the medical image classification tasks. METHODS: In this work, we propose a novel Bayesian statistics-guided label refurbishment mechanism (BLRM) for DNNs to prevent overfitting noisy images. BLRM utilizes maximum a posteriori probability in the Bayesian statistics and the exponentially time-weighted technique to selectively correct the labels of noisy images. The training images are purified gradually with the training epochs when BLRM is activated, further improving classification performance. RESULTS: Comprehensive experiments on both synthetic noisy images (public OCT & Messidor datasets) and real-world noisy images (ANIMAL-10N) demonstrate that BLRM refurbishes the noisy labels selectively, curbing the adverse effects of noisy data. Also, the anti-noise BLRMs integrated with DNNs are effective at different noise ratio and are independent of backbone DNN architectures. In addition, BLRM is superior to state-of-the-art comparative methods of anti-noise. CONCLUSIONS: These investigations indicate that the proposed BLRM is well capable of mitigating label noise in medical image classification tasks.

Npa3-Gpn3 cooperate to assemble RNA polymerase II and prevent clump of its subunits in the cytoplasm.

RNA polymerase II (RNAPII) is an essential machinery in eukaryotes that catalyzes mRNA synthesis and controls cell fate. Although the structure and function of RNAPII are relatively well defined, the molecular mechanism of its assembly process is poorly understood. Three members of GPN-loop GTPase family Npa3/Gpn1, Gpn2, and Gpn3 participate in the biogenesis of RNAPII with non-redundant roles. In this study, we demonstrate that Gpn3 and Npa3 directly participate in the assembly of the two largest subunits during biogenesis of RNAPII. When Gpn3 is defective, assembly of RNAPII is disrupted, leading to cytoplasmic foci of RNAPII subunits. Long-term assembly factor defects will lead to the accumulation of different kind of newly synthesized RNAPII subunits in the cytoplasm to form foci, and this can be prevented by recovery of the defective assembly factor. Cytoplasmic foci of RNAPII subunits in mutants of these assembly factors reveals a new cellular rescue response named the 'RNAPII assembly stress response'.

Triplet Cross-Fusion Learning for Unpaired Image Denoising in Optical Coherence Tomography.

Optical coherence tomography (OCT) is a widely-used modality in clinical imaging, which suffers from the speckle noise inevitably. Deep learning has proven its superior capability in OCT image denoising, while the difficulty of acquiring a large number of well-registered OCT image pairs limits the developments of paired learning methods. To solve this problem, some unpaired learning methods have been proposed, where the denoising networks can be trained with unpaired OCT data. However, majority of them are modified from the cycleGAN framework. These cycleGAN-based methods train at least two generators and two discriminators, while only one generator is needed for the inference. The dual-generator and dual-discriminator structures of cycleGAN-based methods demand a large amount of computing resource, which may be redundant for OCT denoising tasks. In this work, we propose a novel triplet cross-fusion learning (TCFL) strategy for unpaired OCT image denoising. The model complexity of our strategy is much lower than those of the cycleGAN-based methods. During training, the clean components and the noise components from the triplet of three unpaired images are cross-fused, helping the network extract more speckle noise information to improve the denoising accuracy. Furthermore, the TCFL-based network which is trained with triplets can deal with limited training data scenarios. The results demonstrate that the TCFL strategy outperforms state-of-the-art unpaired methods both qualitatively and quantitatively, and even achieves denoising performance comparable with paired methods. Code is available at: https://github.com/gengmufeng/TCFL-OCT.

Retinal oxygen kinetics imaging and analysis (ROKIA) based on the integration and fusion of structural-functional imaging.

The retina is one of the most metabolically active tissues in the body. The dysfunction of oxygen kinetics in the retina is closely related to the disease and has important clinical value. Dynamic imaging and comprehensive analyses of oxygen kinetics in the retina depend on the fusion of structural and functional imaging and high spatiotemporal resolution. But it's currently not clinically available, particularly via a single imaging device. Therefore, this work aims to develop a retinal oxygen kinetics imaging and analysis (ROKIA) technology by integrating dual-wavelength imaging with laser speckle contrast imaging modalities, which achieves structural and functional analysis with high spatial resolution and dynamic measurement, taking both external and lumen vessel diameters into account. The ROKIA systematically evaluated eight vascular metrics, four blood flow metrics, and fifteen oxygenation metrics. The single device scheme overcomes the incompatibility of optical design, harmonizes the field of view and resolution of different modalities, and reduces the difficulty of registration and image processing algorithms. More importantly, many of the metrics (such as oxygen delivery, oxygen metabolism, vessel wall thickness, etc.) derived from the fusion of structural and functional information, are unique to ROKIA. The oxygen kinetic analysis technology proposed in this paper, to our knowledge, is the first demonstration of the vascular metrics, blood flow metrics, and oxygenation metrics via a single system, which will potentially become a powerful tool for disease diagnosis and clinical research.