A checkpointing scheme for relevant distributed real-time tasks which can be scheduled as a DAG is proposed. A typical algorithm, OSA, is selected for DAG scheduling. A new methods based a new structure, Scheduled Clu...A checkpointing scheme for relevant distributed real-time tasks which can be scheduled as a DAG is proposed. A typical algorithm, OSA, is selected for DAG scheduling. A new methods based a new structure, Scheduled Cluster Tree, is presented to calculate the slack time of each task in the task cluster. In the checkpointing scheme, the optimal checkpoint intervals which minimize the approximated failure probability are derived formally and validated experimentally. The complexity of approximated failure probability is quite small compared with that of the exact probability. Meanwhile, the consistency of the checkpointing is discussed also.展开更多
Training large-scale deep neural networks(DNNs)is prone to software and hardware failures,with critical failures often requiring full-machine reboots that substantially prolong training.Existing checkpoint-recovery so...Training large-scale deep neural networks(DNNs)is prone to software and hardware failures,with critical failures often requiring full-machine reboots that substantially prolong training.Existing checkpoint-recovery solutions either cannot tolerate such critical failures or suffer from slow checkpointing and recovery due to constrained input/output bandwidth.In this paper,we propose FastCheck,a checkpoint-recovery framework that accelerates checkpointing and recovery through parallel transmission and tailored compression.First,FastCheck partitions checkpoints into shards and leverages multiple nodes for parallel checkpointing and recovery.Second,it further reduces checkpoint size and overhead with delta compression for weights and index compression for momentum.Third,FastCheck employs lightweight and consistent health status maintenance that accurately tracks node health,preventing checkpoint transmission to failed nodes.We implement FastCheck in PyTorch and evaluate it on multiple DNN models against two baselines.Experimental results show that FastCheck reduces the checkpointing time by up to 78.42%and the recovery time by up to 77.41%,while consistently improving efficiency across different training stages.展开更多
The reliability of high-performance computing(HPC)is essential for program execution stability.However,as the hardware fault rate constantly increases,fault-tolerance techniques such as Checkpoint/Restart(C/R)introduc...The reliability of high-performance computing(HPC)is essential for program execution stability.However,as the hardware fault rate constantly increases,fault-tolerance techniques such as Checkpoint/Restart(C/R)introduce significant system overhead.This paper proposes Program Error Resilience-Aware Checkpointing Mechanism(ResCheckpointer)to mitigate the overhead of the C/R mechanism.The primary motivation of ResCheckpointer is that we observe that crash proneness(i.e.,the probability of the program crashing after fault occurrence)varies significantly among inter-and intra-HPC programs,which prompts us to flexibly adjust checkpoint intervals for further C/R overhead optimization.Specifically,we first construct the graph neural network(GNN)based learning paradigms to excavate the complex error propagation and effect mechanisms hidden within the HPC program’s execution flow,and propose Crash-Predictor for efficiently predicting programs’crash proneness.Based on this,we build ResCheckpointer,which equips an intelligent checkpoint interval setting strategy for HPC programs,i.e.,denser for the crash proneness stage while sparser for the error resilience stage.Experimental results show that ResCheckpointer can achieve up to 55.37%C/R cost reduction compared with the baseline C/R mechanism.展开更多
Network of workstations (NOW) now becomes one of the main trends of parallel computing. But for long-running scientific programs, it needs effective fault tolerance for its changing property. Checkpointing and rollbac...Network of workstations (NOW) now becomes one of the main trends of parallel computing. But for long-running scientific programs, it needs effective fault tolerance for its changing property. Checkpointing and rollback recovery is a solution to this problem. First the main problems upon rollback recovery are discussed, the different checkpointing techniques for NOW are analyzed, and then the design and implementation of ChaRM (checkpoint-based rollback recovery and process migration) system are described. The comparison of three coordinated checkpointing systems is given.展开更多
A consistent checkpointing algorithm with short freezing time (SFT) is presented in this paper. It supports fault-tolerance in distributed systems. The algorithm has shorter freezing time, lower overhead, and simplici...A consistent checkpointing algorithm with short freezing time (SFT) is presented in this paper. It supports fault-tolerance in distributed systems. The algorithm has shorter freezing time, lower overhead, and simplicity of recovery. To make checkpoint time shorter, a special control message (Munblock) is used to ensure that a process can respond the checkpoint event quickly at any given time. Moreover, main memory algorithm is used to improve the concurrency of checkpointing. By using SFT, the freezing time resulted by checkpointing is less than 0.03s. Furthermore, the control message number of SFT is only O(n).展开更多
In this papert the hard problem of the thorough garbage collection in uncoordinated Checkpointing algorithms is studied. After introduction of the traditional garbage collecting scheme, with which only obsolete checkp...In this papert the hard problem of the thorough garbage collection in uncoordinated Checkpointing algorithms is studied. After introduction of the traditional garbage collecting scheme, with which only obsolete checkpoints can be discarded, it is shown that this kind of traditional method may fail to discard any checkpoint in some special cases, and it is necessary and urgent to find a thorough garbage collecting method, with which all the checkpoints useless for any future rollback-recovery including the obsolete ones can be discarded. Then, the Thorough Garbage Collection Theorem is proposed and proved, which ensures the feasibility of the thorough garbage collection, and gives the method to calculate the set of the useful checkpoints as well.展开更多
The landscape of breast cancer treatment has undergone a transformative shift with the integration of immunotherapy.Historically considered a“cold”tumor with limited immunogenicity,breast cancer management was domin...The landscape of breast cancer treatment has undergone a transformative shift with the integration of immunotherapy.Historically considered a“cold”tumor with limited immunogenicity,breast cancer management was dominated by surgery,chemotherapy,radiotherapy,and targeted therapies1.However,the advent of immune checkpoint inhibitors(ICIs)has challenged this paradigm,opening a new frontier.The initial breakthrough in triple-negative breast cancer(TNBC)demonstrated that a subset of patients could derive profound and durable clinical benefit from pembrolizumab and atezolizumab2,3.Today,precision immunotherapy aims to identify the patients most likely to respond,to convert immunologically silent tumors into responsive tumors,and to strategically combine immunotherapies with other modalities to overcome resistance.This evolution from empirical application to biomarker-driven strategies marks the critical juncture at which we stand,transitioning promising clinical trial data into refined,effective,and accessible clinical practice4.Recent key clinical studies on breast cancer immunotherapy are summarized in Table 1.展开更多
Background:Exercise exerts tumor-suppressive effects across multiple malignancies,partly through exerkines—exercise-induced secreted factors with immunomodulatory and metabolic functions.However,the prognostic releva...Background:Exercise exerts tumor-suppressive effects across multiple malignancies,partly through exerkines—exercise-induced secreted factors with immunomodulatory and metabolic functions.However,the prognostic relevance of exerkines across cancer types remains unclear,and the molecular determinants of exercise responsiveness are poorly defined.Methods:We systematically profiled 183 curated exerkine-related genes across 33 cancer types from The Cancer Genome Atlas(TCGA)using non-negative matrix factorization(NMF)to define molecular subtypes.Prognostic significance was evaluated via Kaplan-Meier analysis.For five cancers with consistent survival divergence(LGG,KIRC,LUAD,PAAD,ACC),we developed an Exerkine Prognostic Index(EPI)using LASSO Cox regression and validated its predictive performance through time-dependent ROC analysis.Immune cell infiltration(CIBERSORT),stromal/immune scores(ESTIMATE),and immune checkpoint expression were assessed to characterize immune landscape differences between EPI subgroups.Results:Exerkine-based NMF clustering identified prognostically distinct subtypes in 25 cancers.The EPI robustly stratified patients into high-and low-risk groups with significant differences in overall survival(p<0.001).High-EPI subgroups were associated with elevated infiltration of immunosuppressive cells(e.g.,Tregs,M0 macrophages),altered immune/stromal scores,and differential expression of immune checkpoints such as PD-L1 and CTLA4 in a cancer-type-specific manner.Discussion:Our findings reveal that exerkine expression patterns capture biologically and clinically relevant heterogeneity across cancers.The EPI provides a robust molecular tool to stratify patients by prognosis and immune contexture,offering insights into differential exercise responsiveness.Conclusions:Exerkines represent promising biomarkers for risk stratification and precision-guided exercise interventions in oncology.展开更多
Objectives:Immunotherapy based on immune checkpoint blockade(ICB)has become a key treatment for melanoma.However,the increasing number of cases of melanoma resistant to immunotherapy highlights the need to develop met...Objectives:Immunotherapy based on immune checkpoint blockade(ICB)has become a key treatment for melanoma.However,the increasing number of cases of melanoma resistant to immunotherapy highlights the need to develop methods to overcome this resistance.This study aims to collect the most recent information on melanoma immunotherapy,discuss potential strategies to overcome resistance to immunotherapy,and identify areas that require further analysis.Methods:To achieve this goal,scientific publications from 2021-2024 available in PubMed and Google Scholar databases were analyzed.The databases were searched using the following terms:“melanoma”,“immunotherapy”,“Immune Checkpoint Blockade”,and“immunoresistance”.Results:The results of preclinical and early-stage clinical research indicate the potential application of tank-binding kinase 1(TBK-1),fecal microbiota transplant(FMT),Toll-like Receptor 9(TLR9),lipid nanoparticles(LNPs)containing a stimulator of an interferon gene agonist(STING),BRAF inhibitors,Lymphocyte Activation Gene(LAG-3),T-Cell Immunoglobulin and ITIM Domain(TIGIT),and oncolytic viruses(OVs)as potential methods to enhance melanoma sensitivity to ICB.Discussion:To optimize immunotherapy,further research is needed to determine the detailed mechanisms of action,safety profiles,tolerability,and precise patient selection criteria for methods capable of overcoming melanoma’s immunoresistance.展开更多
BACKGROUND Chemotherapy with an immune checkpoint inhibitor is one of the standard regimens for treating advanced gastric cancer(AGC).Ascites and peritoneal dissemination are common complications and poor prognostic f...BACKGROUND Chemotherapy with an immune checkpoint inhibitor is one of the standard regimens for treating advanced gastric cancer(AGC).Ascites and peritoneal dissemination are common complications and poor prognostic factors of AGC;however,reports regarding its efficacy and safety in patients with AGC and massive ascites are limited.AIM To evaluate the safety and efficacy of nivolumab combined with chemotherapy in patients with AGC and ascites.METHODS We retrospectively collected clinical data from 124 patients with AGC who received chemotherapy plus nivolumab as first-line treatment from July 2017 to December 2024.Based on computed tomography scans,massive or moderate ascites were classified as high ascites burden(HAB),whereas mild or no ascites were classified as low ascites burden.RESULTS Ascites was detected in 47 patients(38%);26(21%)were classified into the HAB group.Patients in the HAB group exhibited a significantly poorer performance status,a higher prevalence of diffuse-type histology,and lower programmed cell death ligand 1(PD-L1)expression.Combination therapy with FOLFOX and neutropenia was significantly more common in the HAB group.Progression-free survival(PFS)(4.4 months vs 9.3 months,P=0.0012)and overall survival(OS)(7.3 months vs 21.2 months,P<0.0001)were significantly poorer in the HAB group.However,an improvement in ascites was observed in 61.5%of patients in the HAB group.PD-L1 expression did not correlate with either PFS or OS in the HAB group.CONCLUSION Nivolumab plus chemotherapy demonstrated modest efficacy and acceptable toxicity in patients with AGC and HAB.展开更多
Highly potent chemotherapy provides rapid therapeutic efficacy in melanoma, but is often limited by drug resistance, off-target toxicity, and systemic toxicity. Combination therapy with chemotherapy and immunotherapy ...Highly potent chemotherapy provides rapid therapeutic efficacy in melanoma, but is often limited by drug resistance, off-target toxicity, and systemic toxicity. Combination therapy with chemotherapy and immunotherapy has attracted much attention but still faces challenges such as inconsistent immune responses and systemic toxicity. To address these limitations, we developed cathepsin B-activatable doxorubicin(DOX) prodrug nanoparticles(Cat B-NPs) for inducing tumor-specific immunogenic cell death(ICD), while minimizing off-target toxicity in normal tissues with low cathepsin B expression. The cathepsin Bactivatable DOX prodrug was synthesized by conjugating the cathepsin B-cleavable peptide(FRRL) to DOX, yielding FRRL-DOX. The amphiphilic FRRL-DOX formed stable nanoparticles(163.6 ± 13.5 nm) through intermolecular hydrophobic interaction and π-π stacking. In melanoma cells overexpressing cathepsin B, Cat B-NPs effectively induced cancer cellspecific ICD, while sparing normal cells and immune cells. When Cat B-NPs-treated B16F10cells were co-cultured with immune cells, Cat B-NPs enhanced the phagocytic activity of macrophages and induced the maturation of dendritic cells(DCs). In melanoma models,Cat B-NPs passively accumulated at tumor tissues through the enhanced permeability and retention effect and were selectively activated by intratumoral cathepsin B, enabling highdose treatment that induced robust ICD. Importantly, combination therapy with Cat B-NPs and anti-PD-L1 antibody enhanced ICD, DC maturation and T-cell activation, resulting in complete tumor regression in 50% of treated mice by converting the immunosuppressive tumor environment into an immune-responsive state. In a lung metastasis model, highdose Cat B-NPs with anti-PD-L1 also suppressed metastatic burden without systemic toxicity, supporting their potential as a safe and effective chemo-immunotherapy for melanoma.展开更多
Objectives:The combination of atezolizumab plus bevacizumab(A+B)represents one of the standards first-line treatments for unresectable hepatocellular carcinoma(HCC).Metformin has garnered attention for its potential a...Objectives:The combination of atezolizumab plus bevacizumab(A+B)represents one of the standards first-line treatments for unresectable hepatocellular carcinoma(HCC).Metformin has garnered attention for its potential antitumour and immunomodulatory properties beyond glycaemic control.This study aimed to assess metformin’s impact in patients with type 2 diabetes mellitus(T2DM)receiving A+B therapy.Methods:This retrospective analysis of a prospectively-maintained multicentre database included 523 patients with HCC treated with A+B from the ARTE(Atezolizumab-bevacizumab Real-life Experience for Treatment of Hepatocellular Carcinoma)dataset across 18 Italian centres(May 2020-January 2024).We evaluated objective response rate(ORR),disease control rate(DCR),progression-free survival(PFS),overall survival(OS),and time to progression(TTP)using Cox regression analysis and Inverse Probability of Treatment Weighting(IPTW)to address confounding.Results:Among 523 patients,341(65.2%)did not have diabetes and 182(34.8%)had T2DM.In the overall population,metformin showed no significant benefit for PFS(HR=1.15,95%CI[0.88-1.50],p=0.316)or OS(HR=1.28,95%CI[0.94-1.74],p=0.124).In the subgroup with T2DM(N=180),metformin showed no significant benefit for PFS(HR=1.41,95%CI[0.97-2.05],p=0.069),OS(HR=1.23,95%CI[0.81-1.86],p=0.333),or TTP(HR=0.82,95%CI[0.53-1.26],p=0.363).IPTW analysis confirmed these negative findings.Conclusion:This study found no evidence of improved outcomes with metformin use in patients with HCC in particular with T2DM receiving A+B therapy.Routine metformin use should not be expected to enhance A+B efficacy based on current evidence.展开更多
Dear Editor,Lymphocyte activation gene 3(LAG3),the third established target for immune checkpoint blockade therapy,suppresses T cell function by binding to major histocompatibility complex classⅡ(MHCⅡ).Despite its s...Dear Editor,Lymphocyte activation gene 3(LAG3),the third established target for immune checkpoint blockade therapy,suppresses T cell function by binding to major histocompatibility complex classⅡ(MHCⅡ).Despite its significant therapeutic potential in cancer immunotherapy and the substantial attention it has received from academia and industry,the molecular mechanisms of LAG3-mediated immunosuppression remain poorly understood,primarily because of its unique ligand-binding characteristics and intracellular domains[1].展开更多
Background:Gastric cancer(GC)continues to pose a significant global health challenge due to its high rates of incidence and mortality,with the majority of cases identified at advanced stages.Immunotherapy,particularly...Background:Gastric cancer(GC)continues to pose a significant global health challenge due to its high rates of incidence and mortality,with the majority of cases identified at advanced stages.Immunotherapy,particularly immune checkpoint blockades(ICBs),has demonstrated considerable therapeutic potential;however,many patients do not exhibit a favorable response.As a result,constructing a predictive model to assess ICBs'responsiveness is essential for enhancing treatment outcomes.Methods:Using consensus clustering based on anoikis-related gene expression,GC patients were stratified into two subclusters.Differences in tumor immune microenvironment,ICB resistance,genomic alterations,methylation profiles,and transcriptional networks were analyzed.A machine learning-based strategy was employed to develop a consensus anoikis-related gene signature(ARGS).Potential therapeutic targets were identified through single-cell RNA sequencing(scRNA-seq),and validation was conducted using multiplex immunofluorescence and immunohistochemistry in an in-house cohort(n=28),including 14 ICB responders and 14 nonresponders.Results:The anoikis-resistant cluster(Cluster A)was associated with poorer survival,immunosuppressive infiltration,lower tumor mutation burden,and ICB resistance.ScRNA-seq revealed high fibroblast and endothelial infiltration,with GLI3+cancer-associated fibroblasts suggesting Hedgehog pathway involvement.The ARGS model effectively stratified patients,with elevated scores associ-ated with immunotherapy resistance,enhanced AR characteristics,and poorer clinical outcomes.展开更多
Cutaneous squamous cell carcinoma(CSCC)is the second most common type of skin cancer and typically involves the head and neck.Systemic therapy is often required for patients with advanced CSCC to achieve optimal disea...Cutaneous squamous cell carcinoma(CSCC)is the second most common type of skin cancer and typically involves the head and neck.Systemic therapy is often required for patients with advanced CSCC to achieve optimal disease control.Immune checkpoint inhibitors(ICIs)are now the standard of care for these patients,with a 50%-60%response rate and sustainable remission for at least 30%of patients.Given the activity of ICIs in advanced head and neck CSCC,ICIs are being studied in early-stage disease or neoadjuvant situations.The purpose of this review is to provide an overview of the innovative perioperative strategies in resectable disease and discuss novel immunotherapeutic strategies designed to overcome treatment resistance.In conclusion,neoadjuvant ICIs have demonstrated impressive response rates with unclear survival benefit.The effectiveness of adjuvant ICIs is currently being explored.Emerging biomarkers,such as circulating tumor DNA(ctDNA),will be crucial for optimizing patient selection and improving treatment outcomes.展开更多
Microglia are the resident macrophages of the central nervous system.They act as the first line of defense against pathogens and play essential roles in neuroinflammation and tissue repair after brain insult or in neu...Microglia are the resident macrophages of the central nervous system.They act as the first line of defense against pathogens and play essential roles in neuroinflammation and tissue repair after brain insult or in neurodegenerative and demyelinating diseases(Borst et al.,2021).Together with infiltrating monocyte-derived macrophages,microglia also play a critical role for brain tumor development,since immunosuppressive interactions between tumor cells and tumor-associated microglia and macrophages(TAM)are linked to malignant progression.This mechanism is of particular relevance in glioblastoma(GB),the deadliest form of brain cancer with a median overall survival of less than 15 months(Khan et al.,2023).Therefore,targeting microglia and macrophage activation is a promising strategy for therapeutic interference in brain disease.展开更多
Background:While the treatment of metastatic renal cell carcinoma(mRCC)is evolving due to immune checkpoint inhibitors(ICIs),optimal strategies for later lines of therapy have yet to be defined.The combination of lenv...Background:While the treatment of metastatic renal cell carcinoma(mRCC)is evolving due to immune checkpoint inhibitors(ICIs),optimal strategies for later lines of therapy have yet to be defined.The combination of lenvatinib and everolimus represents a viable option,and the present review aimed to summarize its activity,effectiveness,and safety.Methods:A systematic review of the literature was conducted using PubMed,targeting studies published between 2018 and 2025.Eligible studies included English-language prospective and retrospective trials reporting survival outcomes in mRCC patients treated with lenvatinib and everolimus after at least one ICI-containing regimen.Results:Nine studies met the inclusion criteria,encompassing a total of 441 patients.The lenvatinib and everolimus combination was primarily used in the third and subsequent lines of therapy.Median overall survival ranged from 7.5 to 24.5 months,while median progression-free survival was more consistent,between 6.1 and 6.7 months,except for one study reporting 12.9 months.Objective response rates varied widely(14.0%–55.7%).Adverse events of grade≥3 did not exceed the expected rate,with diarrhoea and proteinuria as the most reported events.Dose reductions and treatment discontinuations due to toxicity occurred but were generally lower than in prior pivotal trials.Conclusions:Real-world evidence suggests that lenvatinib and everolimus represent an effective and safe option after ICI failure in mRCC patients.Nevertheless,the lack of randomized phase III trials and the heterogeneity of existing studies highlight the need for more robust prospective research to guide post-ICI therapeutic strategies.展开更多
Anti-programmed cell death protein 1(PD-1)or its ligand(PD-L1)are immune checkpoint inhibitors(ICIs)that have revolutionized cancer therapy.However,the efficacy of anti-PD-1 and anti-PD-L1 is limited by resistance and...Anti-programmed cell death protein 1(PD-1)or its ligand(PD-L1)are immune checkpoint inhibitors(ICIs)that have revolutionized cancer therapy.However,the efficacy of anti-PD-1 and anti-PD-L1 is limited by resistance and inter-individual variability.In recent years increasing evidence has highlighted the pivotal role of the gut microbiota in modulating the response to PD-1/PD-L1 immunotherapy.Extensive preclinical studies have demonstrated that commensal microbes can increase the efficacy of PD-1/PD-L1 blockade through multiple mechanisms,including the production of metabolites,such as short-chain fatty acids(SCFAs),tryptophan derivatives,and extracellular polysaccharides that remodel the tumor microenvironment,as well as the activation of immune pathways involving dendritic cells,CD8+T cells,and M1 macrophages to increase antitumor immunity.Moreover,clinical studies have shown that fecal microbiota transplantation(FMT)and targeted probiotic interventions show promise for improving the response to PD-1/PD-L1 therapy,while reducing the risk of immune-related adverse events(irAEs).This review systematically explores the multifaceted regulatory roles of the commensal microbiota in PD-1/PD-L1 therapy and examines the preclinical prospects of microbiota-based personalized immunotherapeutic strategies.The integration of multiomics technologies,synthetic biology,and precise microbiota interventions may further optimize PD-1/PD-L1 immunotherapy and offer novel insights into antitumor immune modulation.展开更多
Glioblastoma(GBM),the most aggressive and lethal primary brain tumor in adults,continues to resist conventional therapeutic approaches,withmedian survival remaining dismally low.Immune checkpoint inhibitors(ICIs),whic...Glioblastoma(GBM),the most aggressive and lethal primary brain tumor in adults,continues to resist conventional therapeutic approaches,withmedian survival remaining dismally low.Immune checkpoint inhibitors(ICIs),which have revolutionized the treatment of several solid tumors,have shown limited efficacy inGBMowing to the highly immunosuppressive and heterogeneousmicroenvironment of the tumor.The unique immune landscape of the central nervous system(CNS),characterized by low immunogenicity,restricted T-cell infiltration,and an abundance of regulatory and myeloid-derived suppressor cells,poses considerable barriers to effective immune reactivation.This review provides a comprehensive synthesis of the mechanistic barriers undermining ICI efficacy in GBM,including the blood-brain barrier,low tumor mutational burden,adaptive immune resistance,and iatrogenic immunosuppression.It also explores emerging predictive and prognostic biomarkers,such as programmed death-ligand 1(PD-L1)expression,immune gene signatures,tumor-infiltrating lymphocyte profiles,and circulating markers in cerebrospinal fluid and plasma,which hold promise for guiding patient selection and therapeutic monitoring.Importantly,recent breakthroughs in combinatorial immunotherapy strategies are highlighted,including the integration of ICIs with radiotherapy,anti-angiogenic agents,oncolytic viruses,personalized neoantigen vaccines,and tumor microenvironment reprogramming approaches.Innovative delivery platforms,such as nanoparticles,focused ultrasound,and convection-enhanced delivery,are also discussed for their potential to improve drug bioavailability and local immune activation in the CNS.This review hypothesizes that the therapeutic efficacy of ICIs in GBM can be considerably enhanced by disrupting immune exclusion and reversing immunosuppression through integrated,multimodal strategies guided by dynamic biomarker profiling and spatially resolved immunemapping.This hypothesisdriven approach aims to bridge translational gaps and inform next-generation clinical trial designs that may unlock the potential of immunotherapy for GBM.展开更多
文摘A checkpointing scheme for relevant distributed real-time tasks which can be scheduled as a DAG is proposed. A typical algorithm, OSA, is selected for DAG scheduling. A new methods based a new structure, Scheduled Cluster Tree, is presented to calculate the slack time of each task in the task cluster. In the checkpointing scheme, the optimal checkpoint intervals which minimize the approximated failure probability are derived formally and validated experimentally. The complexity of approximated failure probability is quite small compared with that of the exact probability. Meanwhile, the consistency of the checkpointing is discussed also.
基金supported by the National Natural Science Foundation of China(Nos.62025203 and 62372419).
文摘Training large-scale deep neural networks(DNNs)is prone to software and hardware failures,with critical failures often requiring full-machine reboots that substantially prolong training.Existing checkpoint-recovery solutions either cannot tolerate such critical failures or suffer from slow checkpointing and recovery due to constrained input/output bandwidth.In this paper,we propose FastCheck,a checkpoint-recovery framework that accelerates checkpointing and recovery through parallel transmission and tailored compression.First,FastCheck partitions checkpoints into shards and leverages multiple nodes for parallel checkpointing and recovery.Second,it further reduces checkpoint size and overhead with delta compression for weights and index compression for momentum.Third,FastCheck employs lightweight and consistent health status maintenance that accurately tracks node health,preventing checkpoint transmission to failed nodes.We implement FastCheck in PyTorch and evaluate it on multiple DNN models against two baselines.Experimental results show that FastCheck reduces the checkpointing time by up to 78.42%and the recovery time by up to 77.41%,while consistently improving efficiency across different training stages.
基金supported by the National Key Research and Development Program of China under Grant No.2023YFB4502304the National Natural Science Foundation of China under Grant Nos.62272190 and 62302190.
文摘The reliability of high-performance computing(HPC)is essential for program execution stability.However,as the hardware fault rate constantly increases,fault-tolerance techniques such as Checkpoint/Restart(C/R)introduce significant system overhead.This paper proposes Program Error Resilience-Aware Checkpointing Mechanism(ResCheckpointer)to mitigate the overhead of the C/R mechanism.The primary motivation of ResCheckpointer is that we observe that crash proneness(i.e.,the probability of the program crashing after fault occurrence)varies significantly among inter-and intra-HPC programs,which prompts us to flexibly adjust checkpoint intervals for further C/R overhead optimization.Specifically,we first construct the graph neural network(GNN)based learning paradigms to excavate the complex error propagation and effect mechanisms hidden within the HPC program’s execution flow,and propose Crash-Predictor for efficiently predicting programs’crash proneness.Based on this,we build ResCheckpointer,which equips an intelligent checkpoint interval setting strategy for HPC programs,i.e.,denser for the crash proneness stage while sparser for the error resilience stage.Experimental results show that ResCheckpointer can achieve up to 55.37%C/R cost reduction compared with the baseline C/R mechanism.
基金Project supported by the National "863" High-tech Program of China.
文摘Network of workstations (NOW) now becomes one of the main trends of parallel computing. But for long-running scientific programs, it needs effective fault tolerance for its changing property. Checkpointing and rollback recovery is a solution to this problem. First the main problems upon rollback recovery are discussed, the different checkpointing techniques for NOW are analyzed, and then the design and implementation of ChaRM (checkpoint-based rollback recovery and process migration) system are described. The comparison of three coordinated checkpointing systems is given.
基金the National Natural Science Foundation of China !69673012
文摘A consistent checkpointing algorithm with short freezing time (SFT) is presented in this paper. It supports fault-tolerance in distributed systems. The algorithm has shorter freezing time, lower overhead, and simplicity of recovery. To make checkpoint time shorter, a special control message (Munblock) is used to ensure that a process can respond the checkpoint event quickly at any given time. Moreover, main memory algorithm is used to improve the concurrency of checkpointing. By using SFT, the freezing time resulted by checkpointing is less than 0.03s. Furthermore, the control message number of SFT is only O(n).
文摘In this papert the hard problem of the thorough garbage collection in uncoordinated Checkpointing algorithms is studied. After introduction of the traditional garbage collecting scheme, with which only obsolete checkpoints can be discarded, it is shown that this kind of traditional method may fail to discard any checkpoint in some special cases, and it is necessary and urgent to find a thorough garbage collecting method, with which all the checkpoints useless for any future rollback-recovery including the obsolete ones can be discarded. Then, the Thorough Garbage Collection Theorem is proposed and proved, which ensures the feasibility of the thorough garbage collection, and gives the method to calculate the set of the useful checkpoints as well.
基金supported by the Non-communicable Chronic Diseases National Science and Technology Major Project(Grant No.2025ZD0544003).
文摘The landscape of breast cancer treatment has undergone a transformative shift with the integration of immunotherapy.Historically considered a“cold”tumor with limited immunogenicity,breast cancer management was dominated by surgery,chemotherapy,radiotherapy,and targeted therapies1.However,the advent of immune checkpoint inhibitors(ICIs)has challenged this paradigm,opening a new frontier.The initial breakthrough in triple-negative breast cancer(TNBC)demonstrated that a subset of patients could derive profound and durable clinical benefit from pembrolizumab and atezolizumab2,3.Today,precision immunotherapy aims to identify the patients most likely to respond,to convert immunologically silent tumors into responsive tumors,and to strategically combine immunotherapies with other modalities to overcome resistance.This evolution from empirical application to biomarker-driven strategies marks the critical juncture at which we stand,transitioning promising clinical trial data into refined,effective,and accessible clinical practice4.Recent key clinical studies on breast cancer immunotherapy are summarized in Table 1.
基金supported by Beijing Sport University Graduate Innovation Programme(2024013).
文摘Background:Exercise exerts tumor-suppressive effects across multiple malignancies,partly through exerkines—exercise-induced secreted factors with immunomodulatory and metabolic functions.However,the prognostic relevance of exerkines across cancer types remains unclear,and the molecular determinants of exercise responsiveness are poorly defined.Methods:We systematically profiled 183 curated exerkine-related genes across 33 cancer types from The Cancer Genome Atlas(TCGA)using non-negative matrix factorization(NMF)to define molecular subtypes.Prognostic significance was evaluated via Kaplan-Meier analysis.For five cancers with consistent survival divergence(LGG,KIRC,LUAD,PAAD,ACC),we developed an Exerkine Prognostic Index(EPI)using LASSO Cox regression and validated its predictive performance through time-dependent ROC analysis.Immune cell infiltration(CIBERSORT),stromal/immune scores(ESTIMATE),and immune checkpoint expression were assessed to characterize immune landscape differences between EPI subgroups.Results:Exerkine-based NMF clustering identified prognostically distinct subtypes in 25 cancers.The EPI robustly stratified patients into high-and low-risk groups with significant differences in overall survival(p<0.001).High-EPI subgroups were associated with elevated infiltration of immunosuppressive cells(e.g.,Tregs,M0 macrophages),altered immune/stromal scores,and differential expression of immune checkpoints such as PD-L1 and CTLA4 in a cancer-type-specific manner.Discussion:Our findings reveal that exerkine expression patterns capture biologically and clinically relevant heterogeneity across cancers.The EPI provides a robust molecular tool to stratify patients by prognosis and immune contexture,offering insights into differential exercise responsiveness.Conclusions:Exerkines represent promising biomarkers for risk stratification and precision-guided exercise interventions in oncology.
文摘Objectives:Immunotherapy based on immune checkpoint blockade(ICB)has become a key treatment for melanoma.However,the increasing number of cases of melanoma resistant to immunotherapy highlights the need to develop methods to overcome this resistance.This study aims to collect the most recent information on melanoma immunotherapy,discuss potential strategies to overcome resistance to immunotherapy,and identify areas that require further analysis.Methods:To achieve this goal,scientific publications from 2021-2024 available in PubMed and Google Scholar databases were analyzed.The databases were searched using the following terms:“melanoma”,“immunotherapy”,“Immune Checkpoint Blockade”,and“immunoresistance”.Results:The results of preclinical and early-stage clinical research indicate the potential application of tank-binding kinase 1(TBK-1),fecal microbiota transplant(FMT),Toll-like Receptor 9(TLR9),lipid nanoparticles(LNPs)containing a stimulator of an interferon gene agonist(STING),BRAF inhibitors,Lymphocyte Activation Gene(LAG-3),T-Cell Immunoglobulin and ITIM Domain(TIGIT),and oncolytic viruses(OVs)as potential methods to enhance melanoma sensitivity to ICB.Discussion:To optimize immunotherapy,further research is needed to determine the detailed mechanisms of action,safety profiles,tolerability,and precise patient selection criteria for methods capable of overcoming melanoma’s immunoresistance.
文摘BACKGROUND Chemotherapy with an immune checkpoint inhibitor is one of the standard regimens for treating advanced gastric cancer(AGC).Ascites and peritoneal dissemination are common complications and poor prognostic factors of AGC;however,reports regarding its efficacy and safety in patients with AGC and massive ascites are limited.AIM To evaluate the safety and efficacy of nivolumab combined with chemotherapy in patients with AGC and ascites.METHODS We retrospectively collected clinical data from 124 patients with AGC who received chemotherapy plus nivolumab as first-line treatment from July 2017 to December 2024.Based on computed tomography scans,massive or moderate ascites were classified as high ascites burden(HAB),whereas mild or no ascites were classified as low ascites burden.RESULTS Ascites was detected in 47 patients(38%);26(21%)were classified into the HAB group.Patients in the HAB group exhibited a significantly poorer performance status,a higher prevalence of diffuse-type histology,and lower programmed cell death ligand 1(PD-L1)expression.Combination therapy with FOLFOX and neutropenia was significantly more common in the HAB group.Progression-free survival(PFS)(4.4 months vs 9.3 months,P=0.0012)and overall survival(OS)(7.3 months vs 21.2 months,P<0.0001)were significantly poorer in the HAB group.However,an improvement in ascites was observed in 61.5%of patients in the HAB group.PD-L1 expression did not correlate with either PFS or OS in the HAB group.CONCLUSION Nivolumab plus chemotherapy demonstrated modest efficacy and acceptable toxicity in patients with AGC and HAB.
基金supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIT)(RS-2024-00351185, RS-2025-02219039)by the Korea Health Industry Development Institute (KHIDI)grant funded by the Korea government (MOHW)(RS-2023-KH135133)by Korea Drug Development Fund (KDDF)funded by Ministry of Science and ICT,Ministry of Trade,Industry,and Energy,and Ministry of Health and Welfare(RS-2025-02223093)。
文摘Highly potent chemotherapy provides rapid therapeutic efficacy in melanoma, but is often limited by drug resistance, off-target toxicity, and systemic toxicity. Combination therapy with chemotherapy and immunotherapy has attracted much attention but still faces challenges such as inconsistent immune responses and systemic toxicity. To address these limitations, we developed cathepsin B-activatable doxorubicin(DOX) prodrug nanoparticles(Cat B-NPs) for inducing tumor-specific immunogenic cell death(ICD), while minimizing off-target toxicity in normal tissues with low cathepsin B expression. The cathepsin Bactivatable DOX prodrug was synthesized by conjugating the cathepsin B-cleavable peptide(FRRL) to DOX, yielding FRRL-DOX. The amphiphilic FRRL-DOX formed stable nanoparticles(163.6 ± 13.5 nm) through intermolecular hydrophobic interaction and π-π stacking. In melanoma cells overexpressing cathepsin B, Cat B-NPs effectively induced cancer cellspecific ICD, while sparing normal cells and immune cells. When Cat B-NPs-treated B16F10cells were co-cultured with immune cells, Cat B-NPs enhanced the phagocytic activity of macrophages and induced the maturation of dendritic cells(DCs). In melanoma models,Cat B-NPs passively accumulated at tumor tissues through the enhanced permeability and retention effect and were selectively activated by intratumoral cathepsin B, enabling highdose treatment that induced robust ICD. Importantly, combination therapy with Cat B-NPs and anti-PD-L1 antibody enhanced ICD, DC maturation and T-cell activation, resulting in complete tumor regression in 50% of treated mice by converting the immunosuppressive tumor environment into an immune-responsive state. In a lung metastasis model, highdose Cat B-NPs with anti-PD-L1 also suppressed metastatic burden without systemic toxicity, supporting their potential as a safe and effective chemo-immunotherapy for melanoma.
文摘Objectives:The combination of atezolizumab plus bevacizumab(A+B)represents one of the standards first-line treatments for unresectable hepatocellular carcinoma(HCC).Metformin has garnered attention for its potential antitumour and immunomodulatory properties beyond glycaemic control.This study aimed to assess metformin’s impact in patients with type 2 diabetes mellitus(T2DM)receiving A+B therapy.Methods:This retrospective analysis of a prospectively-maintained multicentre database included 523 patients with HCC treated with A+B from the ARTE(Atezolizumab-bevacizumab Real-life Experience for Treatment of Hepatocellular Carcinoma)dataset across 18 Italian centres(May 2020-January 2024).We evaluated objective response rate(ORR),disease control rate(DCR),progression-free survival(PFS),overall survival(OS),and time to progression(TTP)using Cox regression analysis and Inverse Probability of Treatment Weighting(IPTW)to address confounding.Results:Among 523 patients,341(65.2%)did not have diabetes and 182(34.8%)had T2DM.In the overall population,metformin showed no significant benefit for PFS(HR=1.15,95%CI[0.88-1.50],p=0.316)or OS(HR=1.28,95%CI[0.94-1.74],p=0.124).In the subgroup with T2DM(N=180),metformin showed no significant benefit for PFS(HR=1.41,95%CI[0.97-2.05],p=0.069),OS(HR=1.23,95%CI[0.81-1.86],p=0.333),or TTP(HR=0.82,95%CI[0.53-1.26],p=0.363).IPTW analysis confirmed these negative findings.Conclusion:This study found no evidence of improved outcomes with metformin use in patients with HCC in particular with T2DM receiving A+B therapy.Routine metformin use should not be expected to enhance A+B efficacy based on current evidence.
文摘Dear Editor,Lymphocyte activation gene 3(LAG3),the third established target for immune checkpoint blockade therapy,suppresses T cell function by binding to major histocompatibility complex classⅡ(MHCⅡ).Despite its significant therapeutic potential in cancer immunotherapy and the substantial attention it has received from academia and industry,the molecular mechanisms of LAG3-mediated immunosuppression remain poorly understood,primarily because of its unique ligand-binding characteristics and intracellular domains[1].
基金National Science and Technology Major Project,Grant/Award Number:2024ZD0533300Excellent Doctor Program of Zhongnan Hospital of Wuhan University,Grant/Award Number:ZNYB2021009。
文摘Background:Gastric cancer(GC)continues to pose a significant global health challenge due to its high rates of incidence and mortality,with the majority of cases identified at advanced stages.Immunotherapy,particularly immune checkpoint blockades(ICBs),has demonstrated considerable therapeutic potential;however,many patients do not exhibit a favorable response.As a result,constructing a predictive model to assess ICBs'responsiveness is essential for enhancing treatment outcomes.Methods:Using consensus clustering based on anoikis-related gene expression,GC patients were stratified into two subclusters.Differences in tumor immune microenvironment,ICB resistance,genomic alterations,methylation profiles,and transcriptional networks were analyzed.A machine learning-based strategy was employed to develop a consensus anoikis-related gene signature(ARGS).Potential therapeutic targets were identified through single-cell RNA sequencing(scRNA-seq),and validation was conducted using multiplex immunofluorescence and immunohistochemistry in an in-house cohort(n=28),including 14 ICB responders and 14 nonresponders.Results:The anoikis-resistant cluster(Cluster A)was associated with poorer survival,immunosuppressive infiltration,lower tumor mutation burden,and ICB resistance.ScRNA-seq revealed high fibroblast and endothelial infiltration,with GLI3+cancer-associated fibroblasts suggesting Hedgehog pathway involvement.The ARGS model effectively stratified patients,with elevated scores associ-ated with immunotherapy resistance,enhanced AR characteristics,and poorer clinical outcomes.
文摘Cutaneous squamous cell carcinoma(CSCC)is the second most common type of skin cancer and typically involves the head and neck.Systemic therapy is often required for patients with advanced CSCC to achieve optimal disease control.Immune checkpoint inhibitors(ICIs)are now the standard of care for these patients,with a 50%-60%response rate and sustainable remission for at least 30%of patients.Given the activity of ICIs in advanced head and neck CSCC,ICIs are being studied in early-stage disease or neoadjuvant situations.The purpose of this review is to provide an overview of the innovative perioperative strategies in resectable disease and discuss novel immunotherapeutic strategies designed to overcome treatment resistance.In conclusion,neoadjuvant ICIs have demonstrated impressive response rates with unclear survival benefit.The effectiveness of adjuvant ICIs is currently being explored.Emerging biomarkers,such as circulating tumor DNA(ctDNA),will be crucial for optimizing patient selection and improving treatment outcomes.
基金Deutsche Forschungsgemeinschaft(DFG,German Research Foundation),project numbers 324633948 and 409784463(DFG grants Hi 678/9-3 and Hi 678/10-2,FOR2953)to HHBundesministerium für Bildung und Forschung-BMBF,project number 16LW0463K to HT.
文摘Microglia are the resident macrophages of the central nervous system.They act as the first line of defense against pathogens and play essential roles in neuroinflammation and tissue repair after brain insult or in neurodegenerative and demyelinating diseases(Borst et al.,2021).Together with infiltrating monocyte-derived macrophages,microglia also play a critical role for brain tumor development,since immunosuppressive interactions between tumor cells and tumor-associated microglia and macrophages(TAM)are linked to malignant progression.This mechanism is of particular relevance in glioblastoma(GB),the deadliest form of brain cancer with a median overall survival of less than 15 months(Khan et al.,2023).Therefore,targeting microglia and macrophage activation is a promising strategy for therapeutic interference in brain disease.
文摘Background:While the treatment of metastatic renal cell carcinoma(mRCC)is evolving due to immune checkpoint inhibitors(ICIs),optimal strategies for later lines of therapy have yet to be defined.The combination of lenvatinib and everolimus represents a viable option,and the present review aimed to summarize its activity,effectiveness,and safety.Methods:A systematic review of the literature was conducted using PubMed,targeting studies published between 2018 and 2025.Eligible studies included English-language prospective and retrospective trials reporting survival outcomes in mRCC patients treated with lenvatinib and everolimus after at least one ICI-containing regimen.Results:Nine studies met the inclusion criteria,encompassing a total of 441 patients.The lenvatinib and everolimus combination was primarily used in the third and subsequent lines of therapy.Median overall survival ranged from 7.5 to 24.5 months,while median progression-free survival was more consistent,between 6.1 and 6.7 months,except for one study reporting 12.9 months.Objective response rates varied widely(14.0%–55.7%).Adverse events of grade≥3 did not exceed the expected rate,with diarrhoea and proteinuria as the most reported events.Dose reductions and treatment discontinuations due to toxicity occurred but were generally lower than in prior pivotal trials.Conclusions:Real-world evidence suggests that lenvatinib and everolimus represent an effective and safe option after ICI failure in mRCC patients.Nevertheless,the lack of randomized phase III trials and the heterogeneity of existing studies highlight the need for more robust prospective research to guide post-ICI therapeutic strategies.
基金supported by grants from the National Natural Science Foundation of China(Grant Nos.82222058,82425046,and 82273142).
文摘Anti-programmed cell death protein 1(PD-1)or its ligand(PD-L1)are immune checkpoint inhibitors(ICIs)that have revolutionized cancer therapy.However,the efficacy of anti-PD-1 and anti-PD-L1 is limited by resistance and inter-individual variability.In recent years increasing evidence has highlighted the pivotal role of the gut microbiota in modulating the response to PD-1/PD-L1 immunotherapy.Extensive preclinical studies have demonstrated that commensal microbes can increase the efficacy of PD-1/PD-L1 blockade through multiple mechanisms,including the production of metabolites,such as short-chain fatty acids(SCFAs),tryptophan derivatives,and extracellular polysaccharides that remodel the tumor microenvironment,as well as the activation of immune pathways involving dendritic cells,CD8+T cells,and M1 macrophages to increase antitumor immunity.Moreover,clinical studies have shown that fecal microbiota transplantation(FMT)and targeted probiotic interventions show promise for improving the response to PD-1/PD-L1 therapy,while reducing the risk of immune-related adverse events(irAEs).This review systematically explores the multifaceted regulatory roles of the commensal microbiota in PD-1/PD-L1 therapy and examines the preclinical prospects of microbiota-based personalized immunotherapeutic strategies.The integration of multiomics technologies,synthetic biology,and precise microbiota interventions may further optimize PD-1/PD-L1 immunotherapy and offer novel insights into antitumor immune modulation.
文摘Glioblastoma(GBM),the most aggressive and lethal primary brain tumor in adults,continues to resist conventional therapeutic approaches,withmedian survival remaining dismally low.Immune checkpoint inhibitors(ICIs),which have revolutionized the treatment of several solid tumors,have shown limited efficacy inGBMowing to the highly immunosuppressive and heterogeneousmicroenvironment of the tumor.The unique immune landscape of the central nervous system(CNS),characterized by low immunogenicity,restricted T-cell infiltration,and an abundance of regulatory and myeloid-derived suppressor cells,poses considerable barriers to effective immune reactivation.This review provides a comprehensive synthesis of the mechanistic barriers undermining ICI efficacy in GBM,including the blood-brain barrier,low tumor mutational burden,adaptive immune resistance,and iatrogenic immunosuppression.It also explores emerging predictive and prognostic biomarkers,such as programmed death-ligand 1(PD-L1)expression,immune gene signatures,tumor-infiltrating lymphocyte profiles,and circulating markers in cerebrospinal fluid and plasma,which hold promise for guiding patient selection and therapeutic monitoring.Importantly,recent breakthroughs in combinatorial immunotherapy strategies are highlighted,including the integration of ICIs with radiotherapy,anti-angiogenic agents,oncolytic viruses,personalized neoantigen vaccines,and tumor microenvironment reprogramming approaches.Innovative delivery platforms,such as nanoparticles,focused ultrasound,and convection-enhanced delivery,are also discussed for their potential to improve drug bioavailability and local immune activation in the CNS.This review hypothesizes that the therapeutic efficacy of ICIs in GBM can be considerably enhanced by disrupting immune exclusion and reversing immunosuppression through integrated,multimodal strategies guided by dynamic biomarker profiling and spatially resolved immunemapping.This hypothesisdriven approach aims to bridge translational gaps and inform next-generation clinical trial designs that may unlock the potential of immunotherapy for GBM.
