Autoimmune diseases are comprised of several chronic inflammatory conditions characterized by activation of autoreactive T and B lymphocytes and produces autoantibodies. While most of the current treatment options mainly rely on non-specific immunosuppression, stem cell based therapies can be a promising candidate for promoting "immune reset" and restore long term tolerance against these complex diseases. Stem cell based therapy eliminate autoreactive clones and reconstitute immune tolerance. Besides, mesenchymal stromal cells (MSCs; also referred to as mesenchymal stem cells) have immunomodulatory effects via paracrine signalling, including anti-inflammatory cytokine and extracellular vesicles secretion. However, successful clinical translation is limited by complex high-dimensional datasets and patient heterogeneity. In this review, we have focused on the convergent journey of Artificial Intelligence (AI) and stem cell based therapies, which mainly evolved in recent times. We also discussed how machine learning and deep learning methods facilitate the analysis of single-cell multi-omics, spatial transcriptomics, and easily identify disease states. Furthermore, the role of AI in optimizing the design and potency of therapeutic modalities, including regulatory T cells (Tregs), Chimeric Antigen Receptor (CAR)-T cells, and mesenchymal stromal cells (MSCs) have also been explored. Altogether, this review highlights the emerging role of AI-guided approaches in bridging mechanistic immunology with future precision stem cell therapies, while several computational and preclinical studies are promising, prospective clinical validation remains necessary before widespread clinical implementation .

The potential carcinogenic effects of sodium-glucose cotransporter 2 (SGLT2) inhibitors in patients with type 2 diabetes mellitus (T2DM) remain controversial, particularly regarding site-specific gastrointestinal (GI) neoplasms. This systematic review and meta-analysis aimed to determine the relationship between SGLT2 inhibitors and the risk of GI neoplasms in patients with T2DM. We searched PubMed, EMBASE, Cochrane CENTRAL, Scopus, and Web of Science through March 17, 2025, for RCTs in T2DM comparing SGLT2 inhibitors with placebo or active comparators. Two reviewers independently screened studies, extracted data, and assessed the risk of bias. The primary outcome was GI neoplasms reported in publications, supplementary materials, or trial registries, usually as adverse events rather than centrally adjudicated cancer endpoints. Pooled odds ratios (ORs) with 95% confidence intervals (CIs) were calculated in Stata 17.0. In 48 RCTs (n = 48,765), SGLT2 inhibitor therapy was not associated with overall GI neoplasm risk (OR = 1.10, 95% CI: 0.84-1.44; p = 0.46; I[2] = 0%). Site-specific analyses showed no statistically significant association for esophageal (OR = 1.12, 95% CI 0.37-3.45), gastric (1.20, 0.65-2.23), hepatic (0.62, 0.31-1.22), pancreatic (0.91, 0.51-1.64), colonic (1.28, 0.78-2.08), colorectal (0.76, 0.27-2.17), and rectal neoplasms (0.98, 0.49-1.97), with all p-values > 0.05. Subgroup analyses by agents (e.g., canagliflozin, dapagliflozin, empagliflozin), baseline age, body mass index (BMI), HbA1c, treatment duration, and dose were also non-significant (all p > 0.05). Approximately half of the trials had follow-up of one year or less, limiting our ability to evaluate long-term risk. Available RCT evidence does not show a clear increase in GI neoplasm risk with SGLT2 inhibitors in T2DM. However, limited follow-up, low event counts, and non-cancer-specific outcome ascertainment, the findings should be interpreted as reassuring but not definitive evidence of long-term oncologic safety.Systematic review registration: PROSPERO No. CRD42024619019.

TATA-binding protein-associated factor 3 (TAF3), a member of the TAF family, plays a crucial role in safeguarding finely balanced transcriptional programs. Previous research identified TAF3 as a critical prognostic marker in hepatocellular carcinoma (HCC). This study aimed to elucidate TAF3's functional role and mechanistic underpinnings in liver cancer progression. Through the establishment of stable TAF3-knockdown and overexpression cell lines in human HCC cells (HepG2 and MHCC97H), comprehensive functional analyses revealed that TAF3 knockdown significantly inhibited cancer cell proliferation, migration, and invasion, while its overexpression promoted these malignant phenotypes. Clinically, elevated TAF3 expression correlated with aggressive clinicopathological features and poor prognosis in HCC patients. Mechanistic investigations demonstrated that TAF3 transcriptionally activates sterol regulatory element-binding protein 2 (SREBP2), a master regulator of cholesterol synthesis, leading to increased intracellular cholesterol accumulation. This cholesterol elevation functionally contributed to oncogenic processes, as exogenous cholesterol supplementation reversed the impaired malignancy in TAF3-deficient cells. In vivo validation using a subcutaneous xenograft mouse model confirmed that TAF3 knockdown suppressed tumor growth, an effect effectively counteracted by high-cholesterol dietary intervention. Collectively, these findings establish that TAF3 promotes liver cancer progression through transcriptional activation of SREBP2 and subsequent enhancement of cholesterol biosynthesis. The study identifies a novel TAF3/SREBP2/cholesterol axis as a promising therapeutic target in TAF3-overexpressing hepatocellular carcinomas, providing mechanistic insights into the interplay between transcriptional regulation and metabolic reprogramming in cancer progression.

Systemic lupus erythematosus (SLE) is a complex autoimmune disorder characterized by multi-organ involvement and a protracted clinical course. Current diagnostic strategies, which rely heavily on clinical symptoms and serology, are often insufficient for early detection. Therefore, highly accurate diagnostic biomarkers are urgently needed to facilitate early intervention and optimize personalized treatment strategies. D atasets GSE61635 and GSE135779 were integrated to identify differentially expressed genes. Weighted gene co-expression network analysis (WGCNA) was performed to isolate the module with the strongest clinical relevance. Mendelian randomization and single‑cell RNA‑seq were used to identify key disease‑relevant genes. A diagnostic model was then constructed, and gene set variation analysis (GSVA), along with gene set enrichment analysis (GSEA), was conducted to elucidate the underlying molecular pathways. IFIT1 and IFIT3 were identified as 2 core genes highly expressed in monocytes and T cells of SLE patients. Functional enrichment analysis revealed that these genes were enriched in immune-related pathways, metabolic pathways related to inflammation and genomic stability. The diagnostic model showed good accuracy, with an area under the curve (AUC) of 0.974 on the training set and 0.912 on the validation set. IFIT1 and IFIT3 represent promising biomarkers for diagnosing SLE and appear to mediate key immune and metabolic disturbances. Furthermore, the developed model serves as an accurate and reliable instrument for early diagnosis and personalized therapy. Large-scale clinical studies are warranted to further validate these findings and evaluate their clinical application.

Fusion transcripts are hybrid RNA molecules generated through genomic rearrangements or RNA-level fusion mechanisms. They represent important molecular features of many cancers and can function as oncogenic drivers, diagnostic biomarkers, prognostic indicators, and therapeutic targets. Since the discovery of the BCR::ABL1 fusion in chronic myeloid leukemia, numerous cancer-associated fusion transcripts have been identified across hematologic malignancies and solid tumors. These fusion events encompass diverse biological mechanisms, including constitutively active kinases, aberrant transcription factors, epigenetic regulators, and non-coding fusion RNAs. This review summarizes current knowledge of the mechanisms underlying fusion transcript formation, including genomic rearrangement-dependent and rearrangement-independent processes, as well as fusion circular RNAs. The functional roles of fusion transcripts in cancer biology and their clinical relevance as diagnostic, prognostic, and predictive biomarkers are discussed. In addition, recent advances in fusion transcript detection and characterization are reviewed, including next-generation sequencing, long-read sequencing, single-cell approaches, artificial intelligence-assisted computational methods, and CRISPR/Cas9-mediated strategies for functional modeling and functional validation of fusion transcripts. Despite the rapid expansion of fusion transcript catalogs, the biological and clinical significance of most identified fusion events remains incompletely understood. Future progress will depend on integrating advanced sequencing technologies, artificial intelligence-assisted computational prioritization, and systematic functional validation to distinguish clinically actionable fusion transcripts from biologically neutral events. Such multidisciplinary approaches will be essential for translating fusion transcript research into precision oncology and improving cancer diagnosis, patient stratification, and targeted therapy.

Cellular senescence (CS) plays a crucial role in various diseases, but its role in hepatocellular carcinoma (HCC) remains unclear. CS-related genes were clustered to identify subtypes. A risk score was constructed and validated in three independent cohorts. Associations with clinical characteristics, tumor immune microenvironment, mutation status, heterogeneity, and treatment efficacy were analyzed. Single-cell analysis was used to examine risk score distribution, and machine learning algorithms along with a nomogram were applied to assess prognostic value. Three CS subtypes were identified, with subtype 1 showing the worst prognosis. High risk score was associated with advanced clinical stage and grade, poor prognosis, and an immunosuppressive microenvironment driven by regulatory T cells. It also correlated with higher tumor mutations (notably TP53) and increased heterogeneity. High-risk patients showed poor response to sorafenib and transcatheter arterial chemoembolization (TACE) and may benefit less from immunotherapy. At single-cell level, the risk score was predominantly expressed in malignant hepatocytes and linked to cell stemness. The CS-related risk score is a potential prognostic indicator for poor outcomes in HCC, playing a significant role in tumor progression and offering potential value for clinical diagnosis and prediction of treatment response.

To develop a machine learning-based prediction model that integrating T cell subset data with clinical features to predict the 28-day mortality risk in sepsis patients. A retrospective cohort study was conducted using the MIMIC-IV database. Collected data included demographics, T cell subsets, laboratory results, SOFA score, GCS, and 28-day mortality. Feature selection was performed using LASSO regression combined with 10-fold cross-validation. We compared the performance of six machine learning models, namely RF, SVM, XGB, GLM, GBM, and LASSO logistic regression. Model performance was evaluated using the AUROC, calibration curves, and DCA, while the SHAP method was employed to interpret the optimal model. A total of 781 sepsis patients were included in our study, with a 28-day mortality rate of 18.5%. LASSO regression identified 12 key features: Age, CD + 4 Count, CD8 + T cell count, lactate, platelet, albumin, Na+, Bun, heart rate, anion gap, eosinophil count, monocyte count. Among the six compared machine learning models, GLM achieved the best comprehensive performance in the testing set, with an AUROC of 0.720, good calibration (Brier score: 0.134). SHAP analysis clarified the contribution degree and direction of each predictive factor to the model output. A GLM model integrating T cell subsets (notably CD8 + T cell count) and clinical features was successfully constructed. The model showed moderate discriminative ability and potential clinical application value in predicting the 28-day mortality risk of sepsis. SHAP-based interpretability clarifies individual risk factors, serving as an auxiliary prognostic tool for clinicians.

Following its discovery, radium was initially promoted as a therapeutic agent for a wide range of conditions, including skin disorders, allergies, visual problems, joint pain, and as a general health-enhancing remedy. However, subsequent research demonstrated its significant radiological toxicity and harmful effects on human health. Despite these risks, more than a century after its discovery, certain radium isotopes continue to have limited medical applications, particularly in the treatment of specific cancers. Following ingestion, radium is absorbed into the body and behaves similarly to calcium, resulting in its deposition in bone tissue. A proportion of absorbed radium is eliminated through feces and urine, while the remaining fraction may persist in the skeleton and contribute to long-term internal radiation exposure. The severity and onset of adverse health effects are closely related to the absorbed dose and duration of exposure. Chronic exposure to elevated radium levels has been associated with hematological abnormalities, cataracts, dental and jaw damage, bone cancer, and increased mortality. Advances in radiobiology, radiation protection, and personalized medicine continue to improve the safety and effectiveness of radiation-based treatments. Modern radiotherapy, including selected applications involving targeted radionuclide therapy, requires a detailed understanding of radiation dose distribution, biological responses, and patient-specific factors. The optimization of re-irradiation strategies remains an important clinical challenge that requires careful assessment of previous radiation exposure, tissue tolerance, and potential risks. Because radium exposure can occur through ingestion, inhalation, or dermal contact, continued monitoring, appropriate regulatory frameworks, and access to specialized analytical methods remain essential components of public health protection. International cooperation and strengthened surveillance programs are needed to ensure the safe management of radium-containing materials and minimize potential health risks.

To evaluate the long-term prognosis of patients with relapsed/refractory B-cell acute lymphoblastic leukemia (R/R B-ALL) who received sequential allogeneic hematopoietic stem cell transplantation (allo-HSCT) following CD19 Chimeric Antigen Receptor T-cell (CAR-T) therapy, and to identify the risk factors influencing overall survival (OS) and the severity of acute graft-versus-host disease (aGVHD). A retrospective analysis was conducted on 63 patients with R/R B-ALL who received sequential allo-HSCT after CD19 CAR-T therapy. Cox regression analysis was used to identify factors influencing OS. Patients were stratified according to the severity of aGVHD (low-grade vs. high-grade). Logistic regression was employed to identify predictors of severe aGVHD, and a nomogram was constructed based on these predictors. The model was internally validated using receiver operating characteristic (ROC) curves, calibration plots, and decision curve analysis (DCA). The two-year OS and progression-free survival (PFS) rates were 67.7% and 55.6%, respectively.The time interval from CAR-T therapy to transplantation, along with high-grade aGVHD, were identified as independent risk factors for OS, whereas low-grade aGVHD exerted a protective effect. A longer time interval from CAR-T therapy to transplantation and pre-transplant minimalresidual disease (MRD)-positive status were independent predictors of severe aGVHD. The nomogram developed for predicting the severity of aGVHD demonstrated good discrimination (areaunder the curve [AUC] = 0.827), satisfactory calibration, and clinical utility upon internal validation. We identified key prognostic factors and developed a validated nomogram that enables early, individualized prediction of severe aGVHD in this setting. This model can assist in early risk stratification and targeted intervention to improve outcomes.

Lung cancer remains a leading cause of cancer-related deaths worldwide. Despite advances in targeted therapy and immunotherapy, treatment outcomes are often constrained by tumor heterogeneity and the complex tumor microenvironment. Recent studies highlight the critical role of the nervous system in lung cancer pathogenesis through bidirectional interactions between the central/peripheral nervous system and cancer cells. This review systematically explores how psychological stress, sleep disorders, and smoking influence the initiation and progression of lung cancer via neuroendocrine-immune networks. We further detail the dual roles of key neurotransmitters-including catecholamines, GABA, serotonin, dopamine, glutamate, histamine, and acetylcholine-in regulating tumor proliferation, epithelial-mesenchymal transition, angiogenesis, and immune evasion, with an emphasis on their context-dependent and receptor subtype-specific effects. Emerging evidence indicates that neurotransmitters remodel the tumor microenvironment by modulating immune cell function and receptor-mediated signaling pathways. We also discuss the therapeutic potential of neuromodulatory agents, particularly β-blockers and neurotransmitter receptor-targeted drugs, and their possible synergistic effects with conventional therapies, while critically distinguishing between strategies supported by clinical evidence and those remaining at the preclinical stage. Challenges persist, including neurotransmitter concentration gradients, receptor subtype specificity, and limitations of preclinical models. The convergence of neuroscience and oncology offers promising avenues for innovative lung cancer therapies, underscoring the necessity for interdisciplinary collaboration to translate these discoveries into clinical practice.

Bexarotene has been established as an effective therapeutic agent for mycosis fungoides (MF), both in early, refractory disease and in advanced stages. It has a predictable, albeit significant, metabolic toxicity profile, raising concerns regarding long-term cardiovascular safety. We conducted a retrospective cohort study utilizing TriNetX to identify adults with MF. Patients were stratified by bexarotene exposure and 1:1 propensity score matched (PSM) for demographics, lifestyle factors, comorbidities, and baseline medications. After PSM, two well-balanced cohorts of 2,242 patients each were formed with a median follow-up of 4.0 years for both groups. Bexarotene treatment demonstrated no significant increase in the risk of the composite adapted-MACE endpoint (HR = 0.93, 95% CI: 0.78-1.11, p = 0.425). Similarly, no significant differences were observed for individual cardiovascular components, including acute myocardial infarction (HR = 0.95, p = 0.730), stroke (HR = 0.90, p = 0.514), or heart failure (HR = 0.93, p = 0.455). Conversely, bexarotene was associated with a significantly increased risk of metabolic and endocrine abnormalities, including hypertriglyceridemia (RR = 6.98, p < 0.001), hypothyroidism (RR = 4.93, p < 0.001), hyperlipidemia (RR = 2.18, p < 0.001), and hypercholesterolemia (RR = 1.99, p < 0.001). In the largest real-world cohort of MF patients treated with bexarotene to date, the substantial metabolic and endocrine abnormalities associated with treatment did not translate into an observed increase in cardiovascular events during a median follow-up of 4 years. These findings provide reassurance regarding the short- to intermediate-term cardiovascular profile of bexarotene when appropriate monitoring and management of lipid and thyroid abnormalities are maintained. However, longer follow-up is required to determine whether these metabolic effects influence long-term cardiovascular risk.

Extranodal involvement (ENI) is incorporated into routine risk assessment for diffuse large B-cell lymphoma (DLBCL), but treating ENI as a single binary adverse feature may conceal clinically relevant heterogeneity in extranodal burden, anatomical distribution, treatment feasibility, and immune-molecular context. We evaluated ENI as a burden- and site-aware phenotype rather than as a uniform descriptor. We retrospectively studied 710 adults with newly diagnosed DLBCL treated with first-line immunochemotherapy between June 2011 and December 2024. Outcomes were analyzed according to ENI status, number of extranodal sites, and involved organs. Targeted sequencing was available in 176 tumors, and RNA sequencing was performed in 107 quality-controlled tumor samples. Clinical models were interpreted as adjustment models because ENI burden overlaps with established risk factors; site-specific and molecular analyses were prespecified as exploratory and were interpreted with attention to available-case denominators, treatment heterogeneity, sparse subgroups, and multiplicity. ENI was associated with inferior overall survival (OS) and progression-free survival (PFS) in unadjusted analyses. Multisite ENI was enriched for adverse baseline features, including advanced Ann Arbor stage, elevated lactate dehydrogenase (LDH), higher International Prognostic Index (IPI), and impaired performance status. After multivariable adjustment, the independent prognostic contribution of ENI burden was attenuated, indicating substantial clinical overlap with systemic disease burden and host fitness. Several anatomical sites showed candidate adverse signals, but rare-site estimates were limited by small subgroup sizes, wide confidence intervals, and multiple testing. Targeted sequencing and RNA-seq suggested heterogeneous genomic and immune-transcriptional patterns across ENI categories; these results should be regarded as hypothesis-generating because of tissue availability, tumor-only sequencing in many cases, modest molecular sample sizes, and lack of orthogonal immune validation. ENI in DLBCL is best interpreted as a clinically heterogeneous disease descriptor rather than a single uniform prognostic category. ENI burden and anatomical distribution provide useful descriptive and prognostic context, but they should be interpreted alongside established clinical risk, treatment intensity, censoring patterns, tissue-sampling constraints, and molecular ascertainment limitations. Exploratory genomic and immune-transcriptional correlates identified in this study require prospective multicenter validation with harmonized staging, treatment-intensity annotation, matched-normal sequencing, and spatial or cellular immune profiling before ENI-informed biological or therapeutic stratification can be applied clinically.

Pancreatic ductal adenocarcinoma (PDAC) represents a highly aggressive form of pancreatic cancer. It is distinguished by a profound metabolic plasticity that supports its survival in a hypoxic and nutrient deprived microenvironment. Among the metabolic pathways altered in PDAC, fatty acid metabolism including synthesis, uptake, storage and β-oxidation serves as a key metabolic process that supports tumor growth, progression, and therapeutic resistance. This review discusses emerging evidence on fatty acid metabolic reprogramming and its relevance to the development and progression of pancreatic ductal adenocarcinoma. Key dysregulated enzymes as well as critical metabolic regulators are highlighted. Also, we discuss how genetic alterations drive lipogenic and oxidative pathways to sustain proliferation, redox balance, and adaptation to metabolic stress. Additionally, this review integrates current evidence to illustrate how the tumor microenvironment influences enhancing lipid availability and promoting metabolic symbiosis. Finally, this review outlines current pharmacological approaches that interfere with fatty acid metabolism, emphasizing lipid metabolism as a promising potential intervention strategy to inhibit tumor growth and sensitize cancer cells to existing treatments.