Colorectal cancer (CRC) is one of the most frequently diagnosed malignant tumors. However, clear evidence explaining the regulatory mechanisms of programmed death ligand 1 (PD-L1) in CRC has been limited. To illustrate the function of YTH N6-methyladenosine (m6A) RNA binding protein F2 (YTHDF2), we conducted a comprehensive evaluation of expression profiling datasets from online databases and clinical samples. We used a subcutaneous immunodeficient mouse model to investigate the impact of YTHDF2 on CRC. Western blots, flow cytometry, PD-1/PD-L1 binding assay, and cell killing assay were used to assess the relationship between YTHDF2 and PD-L1. We used RNA sequencing, along with methylated RNA immunoprecipitation (MeRIP) and RNA binding protein immunoprecipitation (RIP) sequencing to analyze mRNA expression, m6A methylation levels, and YTHDF2 target transcripts. The m6A methylation locations of mRNAs were verified using sequence-based RNA adenosine methylation site predictor (SRAMP), MeRIP-qRT-PCR, RIP-qRT-PCR, and a dual-luciferase reporter system. YTHDF2 was upregulated in CRC tissues, and patients with higher YTHDF2 expression had a worse prognosis. The in vivo model showed that YTHDF2 promoted CRC growth, whereas in vitro experiments showed that inhibiting YTHDF2 expression did not affect cell proliferation, migration, or invasion. Mechanistically, interference with YTHDF2 reduced PD-L1 expression and the binding ability between PD-1 and PD-L1. The use of RNA-seq, MeRIP-seq, RIP-seq, and bioinformatics tools confirmed that the speckle type BTB/POZ protein (SPOP) mRNA was a YTHDF2 target and validated its m6A methylation sites. After YTHDF2 knockdown, SPOP mRNA stability increased, causing an increase in SPOP expression and a decrease in PD-L1 expression. This study demonstrated that YTHDF2 might upregulate PD-L1 expression by destabilizing m6A-containing SPOP mRNA and promote CRC development. The biological effect of the YTHDF2-SPOP-PD-L1 axis presented a promising target for CRC treatment and provided an approach to enhance the efficacy of anti-PD-1/PD-L1 therapy.

Primary resistance to chimeric antigen receptor (CAR) T-cell therapies has limited their widespread application. Our prior genome-wide CRISPR/Cas9 screening revealed that the loss of CD58, a crucial intrinsic resistance factor in tumors, resulted in insufficient immune synapse formation and impaired CAR T-cell activation and cytotoxicity. However, the specific signaling pathway and transcriptional changes associated with CAR T-cell dysfunction have not been addressed. Here, we revealed that AP-1-mediated activation was attenuated in CAR T cells impaired by tumor CD58 loss, driving a decrease in mitochondrial biogenesis, metabolic kinetic impairment, mitochondrial membrane potential loss and ROS accumulation. Moreover, this AP-1 attenuation triggered death receptor-independent apoptosis through the intrinsic mitochondrial pathway. In seeking therapeutic strategies, we pharmacologically and genetically blocked three distinct inhibitory phosphatases positioned upstream of AP-1 signaling. Multifaceted validation has demonstrated that dual specificity phosphatase 6 (DUSP6) blockade is an effective approach to supplement AP-1 signaling while notably reducing CAR T-apoptosis and enhancing mitochondrial fitness, proliferation and long-term cytotoxicity. The transcriptomic profiles of DUSP6-ablated CAR T cells revealed markedly upregulated T-cell activation signatures and enriched metabolic pathways. Clinically, bulk and single-cell RNA-seq analyses revealed that DUSP6 was downregulated in patients who responded to T-cell-based immunotherapy, implying its relevance to patient outcomes. Our findings repositioned CD58 not merely as an immune synapse component but also a metabolic checkpoint in CAR T-cell biology, the loss of which triggers AP-1-dependent mitochondrial derangement and creates a permissive landscape for intrinsic apoptosis, which can be ameliorated by ablation of the inhibitory phosphatase DUSP6. Crucially, DUSP6 ablation represents a promising engineering target to potentiate CAR T-cell efficacy in broader applications.

Introduction:The management of advanced cholangiocarcinoma (CCA) remains a clinical challenge. Prognostic biomarkers are needed to guide treatment decisions. The C-reactive protein-albumin-lymphocyte (CALLY) index reflects nutritional, immune, and inflammatory status and has shown prognostic value in other cancers. However, its role in CCA patients receiving chemoimmunotherapy is unexplored. Methods:We conducted a retrospective propensity score-matched (PSM) cohort study involving advanced CCA patients who were treated with chemoimmunotherapy. Participants were stratified into high- or low-CALLY groups based on an optimal cut-off value of 1.42. PSM (1:1) was applied to balance baseline covariates. Overall survival (OS) and progression-free survival (PFS) were compared between groups using Kaplan-Meier analysis, and Cox regression analysis was employed to identify prognostic factors. The prognostic models underwent comprehensive internal validation, including bootstrap resampling (1,000 iterations) for calibration and discrimination assessment. Health-related quality of life (HRQoL) was assessed using a mixed model for repeated measures. Results:After 1:1 propensity matching, 55 patients were retained in each group, with balanced baseline characteristics. The high-CALLY group exhibited significantly longer median OS (13.00 months vs. 11.50 months; P = 0.019) and PFS (7.50 months vs. 6.00 months; P = 0.020). Cox analysis confirmed the CALLY index as a valuable prognostic factor for both OS (hazard ratio (HR), 0.68; 95% confidence interval (CI), 0.50 to 0.93; P = 0.014) and PFS (HR, 0.70; 95% CI, 0.58 to 0.85; P < 0.001). Internal validation demonstrated good model performance, with optimism-corrected C-indices of 0.704 for OS and 0.716 for PFS. Furthermore, patients with a high CALLY index showed significantly slower deterioration in HRQoL from week 18 onward (P < 0.05). Conclusion:The CALLY index is a robust prognostic biomarker for advanced CCA patients undergoing chemoimmunotherapy, associated with significantly improved survival and better-preserved quality of life. Its integration into clinical practice could enhance risk stratification and facilitate personalized treatment strategies.

Despite its remarkable efficacy in hematologic malignancies, chimeric antigen receptor T-cell (CAR-T) therapy is often limited by atypical clinical outcomes, necessitating predictive biomarkers. Here, prompted by a patient with low-tumor-burden non-Hodgkin lymphoma who develops severe and persistent cytokine release syndrome (CRS), we identify the germline TNFR2M196R single-nucleotide polymorphism, rs1061622, as a candidate genetic determinant of atypical responses. The TNFR2M196R variant is found to enhance the antitumor efficacy of CAR-T cells by reducing their apoptosis. Paradoxically, however, patients carrying the TNFR2M196R variant exhibit higher rates of early tumor progression. Further investigation reveals that while the TNFR2M196R mutation increases tumor susceptibility to CAR-T-mediated killing, it also accelerates tumor proliferation kinetics. Together, these results establish the TNFR2M196R SNP as a dual-function genetic determinant that modulates both CAR-T cell efficacy and intrinsic tumor behavior, highlighting the critical impact of germline genetics on cancer immunotherapy.

Chest X-ray (CXR) plays a pivotal role in clinical diagnosis, and a variety of task-specific and foundation models have been developed for CXR interpretation. However, these models often struggle to adapt to new diagnostic tasks and complex reasoning scenarios. Recently, LLM-based agents have emerged as a promising paradigm for CXR analysis, enhancing model’s capability via tool coordination, multi-step reasoning, and team collaboration, etc. However, existing agents often rely on a single diagnostic pipeline and lack mechanisms for assessing tools’ reliability, limiting their adaptability and credibility. To this end, we propose CXRAgent, a director-orchestrated, multi-stage agent for CXR interpretation, where a central director coordinates the following stages: (1) Tool Invocation: The agent strategically orchestrates a set of CXR-analysis tools, with outputs normalized and verified by the Evidence-driven Validator (EDV), grounding diagnostic outputs with visual evidence to support reliable downstream diagnosis; (2) Diagnostic Planning: Guided by task requirements and intermediate findings, the agent formulates a targeted diagnostic plan, assembles an expert team, defines member roles, and coordinates their interactions to enable adaptive collaborative reasoning; (3) Collaborative Decision-making: The agent integrates insights from the expert team with accumulated contextual memories, synthesizing them into an evidence-backed conclusion. Experiments on diverse tasks show that CXRAgent achieves strong performance with reliable visual grounding, attaining overall accuracies of 67.0% on CheXbench and 75.6% on Medical-CXR-VQA, and a RaTEScore of 0.569 on MIMIC-CXR for report generation. Code and data are available at this link.

Antigen heterogeneity substantially limits the efficacy of chimeric antigen receptor-modified T (CAR T) cell therapy against solid tumors. Our study highlights the potent antitumor activity of low-dose decitabine-primed CAR T (dCAR T) cells in solid tumor models, a benefit previously confirmed in hematologic malignancies. Notably, dCAR T cell infusion in immunocompetent mice led to substantial elimination of mixed tumor masses containing both antigen-positive and antigen-negative cells, without the need for prior lymphodepletion. Our analysis showed notable proinflammatory remodeling of the tumor immunosuppressive microenvironment. Crucially, antigen-activated dCAR T cells sustained high levels of interferon-γ production, which induced immunogenic cell death in tumor cells and activated conventional dendritic cells. This, in turn, stimulated endogenous CD8+ T cells, enhancing their antigen-spreading capacity and aiding in the clearance of abscopal antigen-negative tumors. These findings reveal the robust antigen-spreading capability of dCAR T cells, underscoring their clinical potential in addressing solid tumors with inherent antigen heterogeneity.

The YAP/TAZ-TEAD complex, a core downstream co-regulatory module of the Hippo pathway, drives tumorigenesis and progression upon aberrant activation. Targeting its protein-protein interaction (PPI) is a highly promising anticancer strategy. TEAD palmitoylation maintains TEAD homeostasis and facilitates YAP/TAZ binding, and this pocket has been successfully exploited to develop covalent and non-covalent TEAD inhibitors. To discover novel chemotype TEAD inhibitors, we established a strategy integrating virtual screening, drug repurposing and structure-based optimization, and recently reported several FDA-approved drugs as potential TEAD inhibitors. Herein, we identify NSAID diflunisal (DF) as a potent TEAD inhibitor, elucidating its undefined anticancer mechanism. Structure-based optimization of DF further yielded novel biphenyl-structured covalent TEAD inhibitors, and representative derivative DF-CIV-2 features a unique sagitta-shaped binding conformation, potently inhibiting TEAD palmitoylation, transcriptional activity and YAP-TEAD complex formation, downregulating Hippo target genes, suppressing YAP-dependent cancer cell proliferation and 3D spheroid growth, and synergizing with MEK/PI3Kα/mTORC1 inhibitors. Collectively, we established a feasible TEAD inhibitor discovery pipeline, and identified DF and its sagitta-shaped derivative DF-CIV-2 as promising TEAD inhibitors, providing novel molecular tools and insights for YAP/TAZ-TEAD-targeted cancer therapy.

Patients with inflammatory bowel disease (IBD) exhibit a dysregulated bile acid pool, characterized by increased primary and decreased secondary bile acids, largely due to gut microbiota dysfunction. However, the impact of colitis on hepatic bile acid synthesis remains poorly understood. In this study, analyses of public datasets, in-house patient samples, and an animal model revealed that colitis enhances flux through the classical bile acid synthesis pathway while suppressing the alternative pathway. Oral administration of chenodeoxycholic acid (CDCA) redirected bile acid synthesis toward the alternative pathway and alleviated colitis in mice. Single-cell RNA sequencing and adoptive transfer experiments demonstrated that CDCA administration reduced pro-inflammatory neutrophil accumulation in the colon by downregulating epithelial-derived CXCL2, a finding validated by in vitro assays and a transgenic mouse model. Mechanistic studies further demonstrated that lithocholic acid, a CDCA metabolite in the gut, activates colonic epithelial vitamin D receptor, thereby suppressing CXCL2 via NF-κB inhibition. Clinical sample analyses supported these findings, showing that a higher cholic acid to CDCA ratio positively correlates with neutrophil counts and CXCL2 levels in IBD patients. Together, these findings suggest a critical role of hepatic bile acid synthesis pathways in IBD pathogenesis and highlight CDCA as a potential therapeutic candidate.

Despite advances in cancer immunotherapies such as immune checkpoint blockade (ICB), durable patient responses remain constrained, which is largely due to the highly suppressive tumor immune microenvironment (TIME). In this study, by analyzing cohorts of patients with pan-cancer and experimental validation, we found that monocarboxylate transporter 1 (MCT1) expression is broadly upregulated in malignant and myeloid compartments within the TIME. MCT1 expression is also associated with worse survival, a suppressive TIME state, and poor treatment response to ICB therapy. Functionally, MCT1-mediated lactate uptake by tumor cells and tumor-associated macrophages (TAM) suppresses CD8+ T-cell activation and cytotoxicity in the ex vivo coculture models. Mechanistically, lactate exposure and uptake via MCT1 in tumor cells and TAMs induce IL10 production, which contributes to the inhibition of the antitumor response of CD8+ T cells. Moreover, in MC38 and LLC mouse cancer models, pharmacologic MCT1 inhibition reprograms the immunosuppressive myeloid populations, improves CD8+ T-cell infiltration and function, and triggers tumor regression. Therefore, these results indicate that MCT1 has the potential to be a biomarker for patients across cancer types and to be a promising therapeutic target for enhanced cancer immunotherapy.

Triple-negative breast cancer (TNBC) remains a lethal malignancy with limited targeted therapies and high metastatic rates. Cancer cells evade macrophage clearance by overexpressing anti-phagocytic cell surface proteins, termed “don’t eat me” signals. Blocking antibodies (e.g., anti-CD47) against these signals show therapeutic promise in multiple cancers, yet variable responses and limited durability of efficacy to such agents imply additional unknown “don’t eat me” signals exist. Here, we detected positive CD52 expression in tumors from TNBC patients and demonstrated that CD52 on TNBC cells facilitates immune evasion by engaging the inhibitory receptor sialic acid-binding Ig-like lectin G (Siglec-G) on tumor-associated macrophages. Genetic ablation of either CD52 or Siglec-G, as well as antibody-mediated blockade of their interaction restored macrophage phagocytic activity both in vitro and in vivo. This consequently suppressed tumor progression, improved survival, and promoted an immunologically active tumor microenvironment in TNBC mouse models. Additionally, cotreatment with anti-CD52 sensitized tumor cells to PD-1 blockade therapy in the spontaneous MMTV-PyMT TNBC model. Our findings identify CD52 as a prominently expressed anti-phagocytic checkpoint in TNBC and reveal the therapeutic potential of dual PD-1/CD52 blockade as a novel immunotherapeutic strategy.

Despite the success of CAR T therapy in non-Hodgkin lymphoma (NHL), recurrence remains challenging. Previously, we showed that ex vivo priming with decitabine (DAC) enhances CAR T persistence and efficacy. Here, we report on an open-label non-randomised phase I/II trial (NCT04697940) evaluating DAC-primed CD19/CD20 dual-targeted CAR T cells (dCAR T) in 23 patients with relapsed or refractory NHL. Primary endpoints are safety and dose-toxicity for phase I, and overall response rate and complete response rate (CRR) for phase II. Secondary endpoints include progression-free survival (PFS), overall survival, and duration of response. This trial has met pre-specified endpoints. Treatment is well tolerated and achieves durable responses, with an 87% CRR and a 2-year PFS of 77% (median follow-up, 24.3 months). Compared with historically unmodified CAR T cohorts, dCAR T cells exhibited robust in vivo expansion and sustained persistence. Single-cell sequencing indicates that DAC priming enriches for memory-like progenitors, which maintain cytotoxic and memory signatures, and upregulates genes associated with T cell fitness and engagement of endogenous immunity. These data establish DAC-priming as a clinically feasible epigenetic reprogramming strategy enhanceing CAR T durability and efficacy, offering a generalizable paradigm for engineered cell therapies in malignant tumors.

CD30-targeted chimeric antigen receptor (CAR) T-cell therapy faces clinical challenges in classical Hodgkin lymphoma (cHL). While current optimization strategies focus on CAR design, manufacturing protocols, and preconditioning regimens, tumor-intrinsic resistance mechanisms remain poorly understood. Our study revealed that CD30 in cHL cells is associated with N-glycans at Asn101 and Asn276, which are essential for protein stability but do not affect cell proliferation or apoptosis. Genetic ablation of these N-glycans or enzymatic deglycosylation significantly enhanced CD30-targeted CAR-T-cell accessibility to tumor cells, leading to improved T-cell activation and cytotoxic function. Notably, pretreatment with eliglustat, an FDA-approved glycosphingolipid synthesis inhibitor, selectively potentiated the antitumor activity of CD30-targeted CAR-T cells in wild-type CD30-expressing tumors but had minimal effects on CD30 glycosylation-deficient variants. Eliglustat combined with CD30-targeted CAR-T cells resulted in superior tumor control in xenograft models without additional toxicity. Mechanistically, eliglustat trimmed terminal sialic acids from CD30 N-glycans while preserving the core N-glycan structure. Furthermore, the addition of eliglustat also enhanced the tumor-killing activity of brentuximab vedotin (BV), a CD30-directed antibody-drug conjugate, both in vitro and in vivo. This glycoimmunotherapy paradigm represents a clinically actionable approach to overcome glycan-mediated immune evasion and enhance therapeutic efficacy in CD30-positive lymphomas.

In this single-arm, single-center, registrational phase 2 trial, tandem CD19/CD20 chimeric antigen receptor (CAR) T cell (TanCAR7) therapy showed promising efficacy and safety in patients with relapsed/refractory non-Hodgkin's lymphoma (r/r NHL). Here, we report 5-year follow-up results, including assessments of durable response, survival, and safety. We also investigated risk factors and biomarkers associated with treatment resistance or relapse and evaluated salvage therapies after CAR-T cell failure. Among 87 patients with r/r NHL treated with TanCAR7, the objective response rate was 78% (complete remission rate, 70%) with a median follow-up of 63.4 months. At data cut-off, 40% of patients remained in remission. Median overall survival (OS) was not reached, with an estimated 5-year OS rate of 60.1% and median progression-free survival (PFS) of 33 months. No new or unexpected TanCAR7-related serious adverse events or deaths were observed. High tumor burden and systemic inflammation were risk factors for resistance and relapse. In addition to CAR-T cell expansion in peripheral blood, high levels of endogenous CD8 + T cells and total lymphocytes after infusion correlated with treatment benefit. Salvage chemotherapy post TanCAR7 failure showed limited efficacy, whereas targeted therapy or secondary CAR-T cell therapy achieved clinical responses in a subset of patients. This first-ever 5-year follow-up analysis of a dual-targeted CAR T-cell therapy shows long-term remission potential and no new safety signals in patients with r/r NHL. Trial Registration: ClinicalTrials.gov: NCT03097770.

Objective: Clinical use of stimulator of interferon genes (STING) agonists has challenges due to poor responsiveness and variable efficacy. Therefore, identifying tumor types that are sensitive to these agents and clarifying the underlying mechanisms are essential. Methods: In vitro screening was performed to identify tumor types that are sensitive to STING agonists. The non-nucleotide agonist, SR-717, and the macrocyclic agonist, E7766, were compared for efficacy. Complementary in vivo and in vitro studies, including gene-knockout models, HMGN2-knockout Neuro-2A and CT-2A cells apoptosis assays, and murine tumor models, were then performed. These experiments focused on the mechanism by which SR-717 mediates antitumor effects and emphasized the role of STING signaling-induced high-mobility group nucleosome-binding protein 2 (HMGN2). In addition, the potential of HMGN2 as a prognostic biomarker was assessed. Results: Neuroblastomas and glioblastomas, two nervous system tumors, were shown to be sensitive to STING agonists. SR-717 exhibited greater antitumor efficacy compared to E7766. Mechanistic studies indicated that STING agonists promote apoptosis through activation of the intrinsic STING-signal transducer and activator of transcription 1 (STAT1)-HMGN2 axis within tumor cells. Ectopic expression of HMGN2 in melanoma cells, which naturally lack HMGN2, led to significant apoptosis. Furthermore, analysis of The Cancer Genome Atlas and Gene Expression Omnibus databases revealed positive correlation between elevated HMGN2 expression and patient survival, supporting the utility of HMGN2 as a prognostic biomarker. Conclusions: This study clarified the mechanism underlying the potent antitumor activity of SR-717 in nervous system tumors through activation of the STING-STAT1-HMGN2 signaling pathway and demonstrated that SR-717 has superior efficacy compared to E7766. In addition, HMGN2 was shown to exhibit translational potential as a prognostic biomarker for patient survival.

Purpose Preoperative assessment of pathologic complete response (pCR) to neoadjuvant therapy is an urgent need for anorectal preservation in patients with locally advanced rectal cancer (LARC). Artificial intelligence assistance remains challenging due to a lack of prospective validation and reliable interpretability. Methods and Materials Eligible patients with LARC were retrospectively collected. Radiomic features extracted from postneoadjuvant therapy magnetic resonance imaging were applied to train a Deep Residual Shrinkage Network (DRSN) to generate Radscore for pCR probability. DRSN was integrated with significant clinicopathological factors to construct a multimodality model, named as RAPIDS-II, in the training set. RAPIDS-II performance in pCR prediction was verified in a testing set and further confirmed in a multicenter, prospective validation trial (NCT number: 04278274). The improvements of radiologists’ visual assessment with RAPIDS-II assistance were evaluated in this prospective cohort. Area under curve (AUC) was used as primary endpoint for model performance. Results Retrospectively recruited 823 patients with LARC were divided into the training set (n = 575) and the testing set (n = 248). Compared with the DRSN model, RAPIDS-II showed a comparable AUC of 0.813 (95% CI, 0.736-0.874) in the testing set (P = 0.020). In the prospective validation cohort (n = 207), RAPIDS-II performed robustly with AUC of 0.795 (95%CI, 0.723-0.859) in identifying patients with pCR. Importantly, RAPIDS-II assistance improved in overall AUC and sensitivity of radiologists’ visual assessment, especially for junior radiologists. Interpretable SHapley Additive exPlanations analysis identified that Radscore attributed most to RAPIDS-II prediction. Conclusions The interpretable RAPIDS-II model demonstrates good performance in pCR evaluation and shows potential as a tool to assist clinicians, particularly those with less experience, in tailoring individualized therapy.

Background:Prognostic stratification for epidermal growth factor receptor (EGFR)-mutated non-small cell lung cancer (NSCLC) patients with brain metastases (BM) varies widely, and identifying drivers of early mortality remains challenging. This study systematically evaluated the impact of clinical and molecular factors on short-term survival (≤6 months) and treatment-related outcomes using a comprehensive real-world dataset, and characterized the leptomeningeal metastasis (LM) subgroup with its distinct clinical features and exceptionally poor prognosis. Methods:We retrospectively analyzed 261 consecutive EGFR‑mutated NSCLC patients with BM diagnosed at a single center between 2014 and 2024. Data on Eastern Cooperative Oncology Group (ECOG) performance status, timing of BM (synchronous initial vs. metachronous), presence of LM, generation of EGFR-tyrosine kinase inhibitor (TKI) administered, and other relevant covariates were extracted. Overall survival (OS) was defined from the date of BM diagnosis to death or last follow‑up; short‑term survival was defined as OS ≤6 months. Independent risk factors were identified by multivariable logistic regression with backward selection. Survival curves were generated using the Kaplan-Meier method and compared with the log‑rank test. Hazard ratios (HRs) and 95% confidence intervals (CIs) were estimated using Cox proportional hazards models. Results:Of the 261 patients, 79 (30.3%) had short‑term survival. Independent risk factors were ECOG ≥1 [odds ratio (OR) =4.11, 95% CI: 1.82-9.30, P<0.001], synchronous initial BM (OR =5.57, 95% CI: 2.56-12.13, P<0.001), and first‑generation EGFR‑TKI vs. third‑generation (OR =5.88, 95% CI: 2.19-15.81, P<0.001). Third‑generation TKIs (OR =0.17, 95% CI: 0.06-0.46) and ECOG 0 (OR =0.18, 95% CI: 0.07-0.50) were independently protective (both P<0.001). Median OS was longer with third‑generation vs. first‑generation TKIs (34.0 vs. 17.0 months; HR =0.424, 95% CI: 0.31-0.58, P<0.001). In the LM subgroup (n=37), third‑generation TKIs showed a trend toward improved survival (HR =0.464, 95% CI: 0.211-1.017, P=0.04). Conclusions:In this retrospective cohort of EGFR‑mutated NSCLC patients with BM, poor performance status, synchronous initial BM, and first‑generation EGFR‑TKIs independently predicted short‑term survival. Third‑generation EGFR‑TKIs demonstrated significantly longer OS, suggesting a potential survival benefit that warrants consideration in clinical decision‑making, including for patients with LM. However, given the retrospective design and single‑center setting, these findings should be validated in prospective studies. Our results reinforce the importance of early molecular testing and upfront central nervous system (CNS)‑penetrant therapy in this high‑risk population.

Chimeric antigen receptor T-cell (CAR T-cell) therapy represents a paradigm shift in treating refractory hematological malignancies. Despite its transformative potential, clinical application is frequently complicated by severe, potentially life-threatening toxicities, with cytokine release syndrome (CRS) being the most prominent. This comprehensive review synthesizes current therapeutic strategies for CRS, integrating recent advancements in multidisciplinary care models, refined risk stratification, and evolving prophylactic protocols. Furthermore, this paper contextualizes CRS management within the broader landscape of CAR T-cell-related adverse event (AE) mitigation, highlighting established best practices and exploring future directions aimed at enhancing the safety profile and optimizing the therapeutic index of CAR T-cell therapies.

Ischemic stroke (IS) remains a major clinical challenge due to the difficulty of early diagnosis and incomplete understanding of its pathological mechanisms. Post-translational modifications (PTMs) regulate key cellular processes in IS, but their roles as diagnostic biomarkers and therapeutic targets have not been fully elucidated. This study aimed to identify PTM-related genes (PTMRGs) associated with IS and evaluate their diagnostic and therapeutic potential using comprehensive bioinformatics and machine learning methods. Gene expression data from two GEO cohorts (GSE16561, training group n = 63; GSE58294, testing group n = 92) were analyzed. Differential expression analysis, weighted gene co-expression network analysis (WGCNA), and intersection with known PTMRGs were performed to screen candidate genes. Protein-protein interaction (PPI) networks and machine learning algorithms (Boruta, SVM-RFE, LASSO) were used to prioritize biomarkers. The expression of key genes was validated in clinical samples by RT-qPCR. An artificial neural network (ANN) was constructed to evaluate diagnostic performance. In addition, immune infiltration, gene set enrichment analysis (GSEA), gene-gene interaction (GGI), and molecular docking analyses were conducted to explore biological functions and therapeutic candidates. A total of 1465 upregulated genes and 1782 downregulated genes were identified. WGCNA revealed modules significantly associated with IS, yielding 75 key PTMRGs identified after intersection with DEGs and PTMRGs. Six genes (ATG7, KAT2A, RNF20, UBA1, UBE2I, and USP15) were identified as diagnostic markers with AUC > 0.7. RT-qPCR in 10 IS patients and 10 controls confirmed differential expression, consistent with bioinformatics results. The ANN model showed high diagnostic accuracy (AUC = 0.983 in training, 0.95 in testing). Functional enrichment linked these genes to ubiquitin-mediated proteolysis, DNA repair, and Myc signaling. Immune analysis showed associations with CD8 + T cells and neutrophils. Molecular docking suggested that ETYA may interact with UBE2I and KAT2A. In conclusion, the six diagnostic genes and the ANN model provide powerful diagnostic tools for IS, and ETYA may have a potential role in the treatment of IS.