Search

Federated learning (FL) enables multiple data owners (a.k.a. FL clients) to collaboratively train machine learning models without disclosing sensitive private data. Existing FL research mostly focuses on the monopoly scenario in which a single FL server selects a subset of FL clients to update their local models in each round of training. In practice, there can be multiple FL servers simultaneously trying to select clients from the same pool. In this paper, we propose a first-of-its-kind Fairness-aware Federated Job Scheduling (FairFedJS) approach to bridge this gap. Based on Lyapunov optimization, it ensures fair allocation of high-demand FL client datasets to FL jobs in need of them, by jointly considering the current demand and the job payment bids, in order to prevent prolonged waiting. Extensive experiments comparing FairFedJS against four state-of-the-art approaches on two datasets demonstrate its significant advantages. It outperforms the best baseline by 31.9% and 1.0% on average in terms of scheduling fairness and convergence time, respectively, while achieving comparable test accuracy., Comment: accepted by ICASSP 2024

Accurate staging of liver fibrosis from magnetic resonance imaging (MRI) is crucial in clinical practice. While conventional methods often focus on a specific sub-region, multi-view learning captures more information by analyzing multiple patches simultaneously. However, previous multi-view approaches could not typically calculate uncertainty by nature, and they generally integrate features from different views in a black-box fashion, hence compromising reliability as well as interpretability of the resulting models. In this work, we propose a new multi-view method based on evidential learning, referred to as MERIT, which tackles the two challenges in a unified framework. MERIT enables uncertainty quantification of the predictions to enhance reliability, and employs a logic-based combination rule to improve interpretability. Specifically, MERIT models the prediction from each sub-view as an opinion with quantified uncertainty under the guidance of the subjective logic theory. Furthermore, a distribution-aware base rate is introduced to enhance performance, particularly in scenarios involving class distribution shifts. Finally, MERIT adopts a feature-specific combination rule to explicitly fuse multi-view predictions, thereby enhancing interpretability. Results have showcased the effectiveness of the proposed MERIT, highlighting the reliability and offering both ad-hoc and post-hoc interpretability. They also illustrate that MERIT can elucidate the significance of each view in the decision-making process for liver fibrosis staging., Comment: Submitted to Medical Image Analysis

Yan Jia,1,* Ya Liu,1,* Yilin Huang,1,* Jie Wang,1 Hanyue Wang,1 Shu Tan,2 Yuxin Shi,1 Qingxia Wang,3 Jie Peng1 1Department of Gastroenterology, Xiangya Hospital, Central South University, Changsha, Hunan, People’s Republic of China; 2Department of Emergency, Loudi Central Hospital, Loudi, Hunan, People’s Republic of China; 3Gastrointestinal Endoscopy Center, Xiangya Hospital, Central South University, Changsha, Hunan, People’s Republic of China*These authors contributed equally to this workCorrespondence: Jie Peng, Department of Gastroenterology, Xiangya Hospital, Central South University, 87 Xiangya Road, Changsha, Hunan, People’s Republic of China, Tel/Fax +86-731-89753722, Email pengjie2014@csu.edu.cn Qingxia Wang, Gastrointestinal Endoscopy Center, Xiangya Hospital, Central South University, 87 Xiangya Road, Changsha, Hunan, People’s Republic of China, Tel/Fax +86-15874189639, Email 1772509440@qq.comObjective: Infection is a common complication of acute pancreatitis (AP). Klebsiella pneumoniae (KP) is one of the most common pathogens associated with nosocomial infections. Our study focuses on investigating the clinical characteristics and risk factors for death of Klebsiella pneumoniae infections in AP patients, further to quantify the prognosis of the patients, and provide evidence for guiding antibiotic use and improving prognosis.Methods: The data of epidemiology, clinical manifestations and drug resistance rate with K. pneumoniae infections in AP patients from January 1, 2012 to August 30, 2022 were retrospectively collected. Logistic regression model and Cox regression model were, respectively, used to determine the risk factors for carbapenem-resistant Klebsiella pneumoniae (CRKP) acquisition and death. The nomogram prediction model was built by RMS software package to predict the 90-day survival rate.Results: One hundred and twenty-six AP patients combined with K. pneumoniae infections, with a mortality rate of 34

Junli Ding,1,* Hao Wang,1,* Rui Hou,1,* Yuxin Shi,1 Honghong Fan,1 Yuting Li,1 Jie Mei,1 Qinglin Zhang,2 Tingyan Ruan,1 Junying Xu1 1Department of Oncology, The Affiliated Wuxi People’s Hospital of Nanjing Medical University, Wuxi, 214023, People’s Republic of China; 2Department of Gastroenterology, The Affiliated Wuxi People’s Hospital of Nanjing Medical University, Wuxi, 214023, People’s Republic of China*These authors contributed equally to this workCorrespondence: Junying Xu; Tingyan Ruan, Department of Oncology, The Affiliated Wuxi People’s Hospital of Nanjing Medical University, No. 299 Qingyang Road, Wuxi, Jiangsu, 214023, People’s Republic of China, Tel/Fax +86-510-82700775, Email doctorxjy123@163.com; ruantingyan165@163.comBackground: Immune checkpoint inhibitors have achieved limited clinical effectiveness in colon cancer. Stem memory T cells (TSCMs) and in-situ cytotoxic T cells are dominant contributors to host immunity. Currently, data on the correlation between TSCM and T cell abundance and clinicopathological characteristics in colon cancer are largely unavailable.Methods: In-situ cytotoxic T cells are identified based on the quantification of CD3+ and CD8+ markers using immunohistochemistry (IHC) in the core of the tumor and the invasive margin of the tumor. The expression of representative markers of TSCMs, CD27 and CD95, was assayed using IHC in colon cancer tissues. Correlations between the levels of each marker and the clinicopathological characteristics as well as prognosis were evaluated.Results: High densities of CD3+ and CD8+ T cells correlated with stage I-II tumors, whereas a lower infiltration of cytotoxic T cells correlated with advanced-stage tumors. CD27 and CD95 were both expressed in the membrane of T cells present in the tumor stroma and their levels showed a negative correlation with the TNM stage. CD3, CD8, and CD27 were expressed at the same locations simultaneously, indicating their coo

Wei Huang,1,* Zhengtao Qian,2,* Yuxin Shi,1,3,* Zheming Zhang,3 Rui Hou,1,3 Jie Mei,1,3 Junying Xu,1 Junli Ding1 1Department of Oncology, the Affiliated Wuxi People’s Hospital of Nanjing Medical University, Wuxi, 214023, People’s Republic of China; 2Department of Clinical Laboratory, Changshu Medicine Examination Institute, Changshu, 215500, People’s Republic of China; 3Wuxi School of Clinical Medicine, Nanjing Medical University, Wuxi, 214023, People’s Republic of China*These authors contributed equally to this workCorrespondence: Junli Ding; Junying Xu, Email dingjunliletters@163.com; doctorxjy123@163.com; xujunying@wuxiph.comBackground: Proteasome 26S subunit ATPase 2 (PSMC2) is a part of the 19S regulatory complex, which catalyzes the unfolding and transport of substrates into the 20S proteasome. Our previous research demonstrated that PSMC2 participates in the tumorigenesis and progression of pancreatic cancer (PC). However, no systematic analysis has been conducted to conclude its expression pattern and correlation with tumor immunity.Aim: To investigate the expression level of PSMC2 in PC, its prognostic value and its relationship with tumor immunity.Methods: In numerous public and internal cohorts, the expression, prognostic significance, and immunological connections of PSMC2 in PC were investigated. Additionally, using data from The Cancer Genome Atlas (TCGA), a pan-cancer analysis was carried out to examine PSMC2’s immunological assocaition, and the predictive power of PSMC2 for immunotherapy was also evaluated in numerous public cohorts.Results: PSMC2 was overexpressed in tumor tissues and linked to unfavorable prognosis in PC. PSMC2 was not only positively correlated with TIICs, also positively correlated with immune checkpoints in PC. In addition to PC, PSMC2 was expected to be an indicator of high immunogenicity in most cancer types. Importantly, PSMC2 could predict the immunotherapeutic responses in various c

Honghong Fan,* Yuxin Shi,* Huiyu Wang, Yuting Li, Jie Mei, Junying Xu, Chaoying Liu Department of Oncology, The Affiliated Wuxi People’s Hospital of Nanjing Medical University, Wuxi, Jiangsu, 214023, People’s Republic of China*These authors contributed equally to this workCorrespondence: Chaoying Liu; Junying Xu, Department of Oncology, The Affiliated Wuxi People’s Hospital of Nanjing Medical University, No. 299 Qingyang Road, Wuxi, Jiangsu, 214023, People’s Republic of China, Tel/Fax +86-510-82700775, Email liuchaoying666@163.com; doctorxjy123@163.comBackground: Immunotherapy drugs, immune checkpoint inhibitors (ICIs), have been approved for first- and second-line treatment of non-small cell lung cancer (NSCLC), but only a portion of patients respond to ICIs. It is crucial to screen the beneficiaries of immunotherapy through biomarkers accurately.Methods: Several datasets were used to explore the predictive value for immunotherapy and immune relevance of guanylate binding protein 5 (GBP5) in NSCLC, including the GSE126044 dataset, The Cancer Genome Atlas (TCGA) dataset, Clinical Proteomic Tumor Analysis Consortium (CPTAC) dataset, the Kaplan–Meier plotter dataset, the HLuA150CS02 cohort, and the HLugS120CS01 cohort.Results: GBP5 was upregulated in tumor tissues but associated with a good prognosis in NSCLC. Moreover, our findings demonstrated that GBP5 was strongly correlated with the expression of many immune-related genes, TIIC levels, and PD-L1 expression based on RNA-seq data onto online databases and validation of the NSCLC tissue microarray using IHC staining. Moreover, pan-cancer analysis has shown that GBP5 was a factor in identifying immuno-hot tumors, except for a few tumor types.Conclusion: In summary, our current research suggests that GBP5 expression is a potential biomarker for predicting the outcome of NSCLC patients treated with ICIs. More research with large-scale samples is needed to determine their value as bioma

Federated learning (FL) has enabled multiple data owners (a.k.a. FL clients) to train machine learning models collaboratively without revealing private data. Since the FL server can only engage a limited number of clients in each training round, FL client selection has become an important research problem. Existing approaches generally focus on either enhancing FL model performance or enhancing the fair treatment of FL clients. The problem of balancing performance and fairness considerations when selecting FL clients remains open. To address this problem, we propose the Fairness-aware Federated Client Selection (FairFedCS) approach. Based on Lyapunov optimization, it dynamically adjusts FL clients' selection probabilities by jointly considering their reputations, times of participation in FL tasks and contributions to the resulting model performance. By not using threshold-based reputation filtering, it provides FL clients with opportunities to redeem their reputations after a perceived poor performance, thereby further enhancing fair client treatment. Extensive experiments based on real-world multimedia datasets show that FairFedCS achieves 19.6% higher fairness and 0.73% higher test accuracy on average than the best-performing state-of-the-art approach., Comment: Accepted by ICME 2023

Staging of liver fibrosis is important in the diagnosis and treatment planning of patients suffering from liver diseases. Current deep learning-based methods using abdominal magnetic resonance imaging (MRI) usually take a sub-region of the liver as an input, which nevertheless could miss critical information. To explore richer representations, we formulate this task as a multi-view learning problem and employ multiple sub-regions of the liver. Previously, features or predictions are usually combined in an implicit manner, and uncertainty-aware methods have been proposed. However, these methods could be challenged to capture cross-view representations, which can be important in the accurate prediction of staging. Therefore, we propose a reliable multi-view learning method with interpretable combination rules, which can model global representations to improve the accuracy of predictions. Specifically, the proposed method estimates uncertainties based on subjective logic to improve reliability, and an explicit combination rule is applied based on Dempster-Shafer's evidence theory with good power of interpretability. Moreover, a data-efficient transformer is introduced to capture representations in the global view. Results evaluated on enhanced MRI data show that our method delivers superior performance over existing multi-view learning methods., Comment: Early accepted by MICCAI 2023

The development of effective antifungal agents remains a big challenge in view of the close evolutionary relationship between mammalian cells and fungi. Moreover, rapid mutations of fungal receptors at the molecular level result in the emergence of drug resistance. Here, with low tendency to develop drug-resistance, the subcellular organelle mitochondrion is exploited as an alternative target for efficient fungal killing by photodynamic therapy (PDT) of mitochondrial-targeting luminogens with aggregation-induced emission characteristics (AIEgens). With cationic isoquinolinium (IQ) moiety and proper hydrophobicity, three AIEgens, namely, IQ-TPE-2O, IQ-Cm, and IQ-TPA, can preferentially accumulate at the mitochondria of fungi over the mammalian cells. Upon white light irradiation, these AIEgens efficiently generate reactive 1O2, which causes irreversible damage to fungal mitochondria and further triggers the fungal death. Among them, IQ-TPA shows the highest PDT efficiency against fungi and negligible toxicity to mammalian cells, achieving the selective and highly efficient killing of fungi. Furthermore, we tested the clinical utility of this PDT strategy by treating fungal keratitis on a fungus-infected rabbit model. It was demonstrated that IQ-TPA presents obviously better therapeutic effects as compared with the clinically used rose bengal, suggesting the success of this PDT strategy and its great potential for clinical treatment of fungal infections. © 2021 American Chemical Society.

The development of effective antifungal agents remains a big challenge in view of the close evolutionary relationship between mammalian cells and fungi. Moreover, rapid mutations of fungal receptors at the molecular level result in the emergence of drug resistance. Here, with low tendency to develop drug-resistance, the subcellular organelle mitochondrion is exploited as an alternative target for efficient fungal killing by photodynamic therapy (PDT) of mitochondrial-targeting luminogens with aggregation-induced emission characteristics (AIEgens). With cationic isoquinolinium (IQ) moiety and proper hydrophobicity, three AIEgens, namely, IQ-TPE-2O, IQ-Cm, and IQ-TPA, can preferentially accumulate at the mitochondria of fungi over the mammalian cells. Upon white light irradiation, these AIEgens efficiently generate reactive 1O2, which causes irreversible damage to fungal mitochondria and further triggers the fungal death. Among them, IQ-TPA shows the highest PDT efficiency against fungi and negligible toxicity to mammalian cells, achieving the selective and highly efficient killing of fungi. Furthermore, we tested the clinical utility of this PDT strategy by treating fungal keratitis on a fungus-infected rabbit model. It was demonstrated that IQ-TPA presents obviously better therapeutic effects as compared with the clinically used rose bengal, suggesting the success of this PDT strategy and its great potential for clinical treatment of fungal infections. © 2021 American Chemical Society.

Recent advances in Federated Learning (FL) have brought large-scale collaborative machine learning opportunities for massively distributed clients with performance and data privacy guarantees. However, most current works focus on the interest of the central controller in FL,and overlook the interests of the FL clients. This may result in unfair treatment of clients that discourages them from actively participating in the learning process and damages the sustainability of the FL ecosystem. Therefore, the topic of ensuring fairness in FL is attracting a great deal of research interest. In recent years, diverse Fairness-Aware FL (FAFL) approaches have been proposed in an effort to achieve fairness in FL from different perspectives. However, there is no comprehensive survey that helps readers gain insight into this interdisciplinary field. This paper aims to provide such a survey. By examining the fundamental and simplifying assumptions, as well as the notions of fairness adopted by existing literature in this field, we propose a taxonomy of FAFL approaches covering major steps in FL, including client selection, optimization, contribution evaluation and incentive distribution. In addition, we discuss the main metrics for experimentally evaluating the performance of FAFL approaches, and suggest promising future research directions towards FAFL., Comment: Accepted by IEEE Transactions on Neural Networks and Learning Systems

The development of effective antifungal agents remains a big challenge in view of the close evolutionary relationship between mammalian cells and fungi. Moreover, rapid mutations of fungal receptors at the molecular level result in the emergence of drug resistance. Here, with low tendency to develop drug-resistance, the subcellular organelle mitochondrion is exploited as an alternative target for efficient fungal killing by photodynamic therapy (PDT) of mitochondrial-targeting luminogens with aggregation-induced emission characteristics (AIEgens). With cationic isoquinolinium (IQ) moiety and proper hydrophobicity, three AIEgens, namely, IQ-TPE-2O, IQ-Cm, and IQ-TPA, can preferentially accumulate at the mitochondria of fungi over the mammalian cells. Upon white light irradiation, these AIEgens efficiently generate reactive 1O2, which causes irreversible damage to fungal mitochondria and further triggers the fungal death. Among them, IQ-TPA shows the highest PDT efficiency against fungi and negligible toxicity to mammalian cells, achieving the selective and highly efficient killing of fungi. Furthermore, we tested the clinical utility of this PDT strategy by treating fungal keratitis on a fungus-infected rabbit model. It was demonstrated that IQ-TPA presents obviously better therapeutic effects as compared with the clinically used rose bengal, suggesting the success of this PDT strategy and its great potential for clinical treatment of fungal infections. © 2021 American Chemical Society.

Chemotherapy is the dominant treatment approach to many cancers. For decades, Gemcitabine (GEM) has been used as the first-line therapy for pancreatic ductal adenocarcinoma (PDAC), one of the most fatal solid tumor, but its competence is disappointingly constrained byintrinsic or adaptive resistance. The mechanisms beneath such resistance have been intensively studied with great efforts and its correlations with many genes and pathways have been found. However, our knowledge has not been unified to provide abundant information for major advances, since the whole landscape in which drug adaptation and gene expressions are associated is not yet clear. Moreover, little had we known about the initiation of adaptation, which makes it more difficult to define better therapies and achieve better clinical outcomes.Recently, epigenetic alterations have become potential prognostic biomarkers and offered vast options for PDAC detection. Previously, we had observed different histone 3 lysine 9 trimethylation (H3K9me3) patterns among GEM sensitive and resistant PDAC cells during the GEM treatment, suggesting the role of histone modification in GEM resistance initiation and exhibition. In this paper, we would provide our analysis workflow to identify the candidate genes that might be modulated by H3K9me3 upon GEM treatment, and predict how they dynamically affect GEM sensitivity. Our results would be considered as a general framework to construct the regulatory network from epigenetic drug response to transcriptomic plasticity, and discover new opportunities for therapeutic development.

Chemotherapy is the dominant treatment approach to many cancers. For decades, Gemcitabine (GEM) has been used as the first-line therapy for pancreatic ductal adenocarcinoma (PDAC), one of the most fatal solid tumor, but its competence is disappointingly constrained byintrinsic or adaptive resistance. The mechanisms beneath such resistance have been intensively studied with great efforts and its correlations with many genes and pathways have been found. However, our knowledge has not been unified to provide abundant information for major advances, since the whole landscape in which drug adaptation and gene expressions are associated is not yet clear. Moreover, little had we known about the initiation of adaptation, which makes it more difficult to define better therapies and achieve better clinical outcomes.Recently, epigenetic alterations have become potential prognostic biomarkers and offered vast options for PDAC detection. Previously, we had observed different histone 3 lysine 9 trimethylation (H3K9me3) patterns among GEM sensitive and resistant PDAC cells during the GEM treatment, suggesting the role of histone modification in GEM resistance initiation and exhibition. In this paper, we would provide our analysis workflow to identify the candidate genes that might be modulated by H3K9me3 upon GEM treatment, and predict how they dynamically affect GEM sensitivity. Our results would be considered as a general framework to construct the regulatory network from epigenetic drug response to transcriptomic plasticity, and discover new opportunities for therapeutic development.

CT imaging is crucial for diagnosis, assessment and staging COVID-19 infection. Follow-up scans every 3-5 days are often recommended for disease progression. It has been reported that bilateral and peripheral ground glass opacification (GGO) with or without consolidation are predominant CT findings in COVID-19 patients. However, due to lack of computerized quantification tools, only qualitative impression and rough description of infected areas are currently used in radiological reports. In this paper, a deep learning (DL)-based segmentation system is developed to automatically quantify infection regions of interest (ROIs) and their volumetric ratios w.r.t. the lung. The performance of the system was evaluated by comparing the automatically segmented infection regions with the manually-delineated ones on 300 chest CT scans of 300 COVID-19 patients. For fast manual delineation of training samples and possible manual intervention of automatic results, a human-in-the-loop (HITL) strategy has been adopted to assist radiologists for infection region segmentation, which dramatically reduced the total segmentation time to 4 minutes after 3 iterations of model updating. The average Dice simiarility coefficient showed 91.6% agreement between automatic and manual infaction segmentations, and the mean estimation error of percentage of infection (POI) was 0.3% for the whole lung. Finally, possible applications, including but not limited to analysis of follow-up CT scans and infection distributions in the lobes and segments correlated with clinical findings, were discussed., Comment: 23 pages, 6 figures

The dramatic reorganization of chromatin during mitosis is perhaps one of the most fundamental of all cell processes. It remains unclear how epigenetic histone modifications, despite their crucial roles in regulating chromatin architectures, are dynamically coordinated with chromatin reorganization in controlling this process. We have developed and characterized biosensors with high sensitivity and specificity based on fluorescence resonance energy transfer (FRET). These biosensors were incorporated into nucleosomes to visualize histone H3 Lys-9 trimethylation (H3K9me3) and histone H3 Ser-10 phosphorylation (H3S10p) simultaneously in the same live cell. We observed an anticorrelated coupling in time between H3K9me3 and H3S10p in a single live cell during mitosis. A transient increase of H3S10p during mitosis is accompanied by a decrease of H3K9me3 that recovers before the restoration of H3S10p upon mitotic exit. We further showed that H3S10p is causatively critical for the decrease of H3K9me3 and the consequent reduction of heterochromatin structure, leading to the subsequent global chromatin reorganization and nuclear envelope dissolution as a cell enters mitosis. These results suggest a tight coupling of H3S10p and H3K9me3 dynamics in the regulation of heterochromatin dissolution before a global chromatin reorganization during mitosis.

The dramatic reorganization of chromatin during mitosis is perhaps one of the most fundamental of all cell processes. It remains unclear how epigenetic histone modifications, despite their crucial roles in regulating chromatin architectures, are dynamically coordinated with chromatin reorganization in controlling this process. We have developed and characterized biosensors with high sensitivity and specificity based on fluorescence resonance energy transfer (FRET). These biosensors were incorporated into nucleosomes to visualize histone H3 Lys-9 trimethylation (H3K9me3) and histone H3 Ser-10 phosphorylation (H3S10p) simultaneously in the same live cell. We observed an anticorrelated coupling in time between H3K9me3 and H3S10p in a single live cell during mitosis. A transient increase of H3S10p during mitosis is accompanied by a decrease of H3K9me3 that recovers before the restoration of H3S10p upon mitotic exit. We further showed that H3S10p is causatively critical for the decrease of H3K9me3 and the consequent reduction of heterochromatin structure, leading to the subsequent global chromatin reorganization and nuclear envelope dissolution as a cell enters mitosis. These results suggest a tight coupling of H3S10p and H3K9me3 dynamics in the regulation of heterochromatin dissolution before a global chromatin reorganization during mitosis.