BackgroundGrowing evidence implicates gut microbiota (GM) in the pathogenesis of esophageal cancer (EC). However, the causal nature of this association-particularly the potential mediating role of circulating metabolites-remains insufficiently clarified, especially across EC subtypes.MethodsWe applied a two-sample Mendelian randomization (MR) framework to investigate the causal associations of GM and blood metabolites with EC and esophageal adenocarcinoma (EAC). The primary analysis was conducted using the inverse-variance weighted method, with complementary MR approaches and genetic risk score (GRS) validation to ensure robustness. Mediation analyses were further performed to assess whether metabolites mediate the effects of GM on EC and EAC.ResultsTwenty-five GM taxa showed significant associations with EC, including 11 with risk-promoting and 14 with protective effects. For EAC, 15 taxa were implicated, with 11 increasing and 4 decreasing disease risk. In addition, five metabolites were causally linked to EC (two risk-related and three protective), and six to EAC (five risk-related and one protective). Mediation analyses revealed that specific metabolites partially mediated the effects of GM on both EC and EAC.ConclusionsOur findings provide genetic evidence for a causal GM-metabolite-cancer axis in EC pathogenesis. These results highlight the role of circulating metabolites as potential intermediaries linking GM to EC and EAC, and suggest new avenues for biomarker discovery and microbiome-targeted interventions.

Pulmonary delivery of anti-inflammatory siRNA holds great potential for the management of severe pneumonia. However, conventional siRNA carriers, primarily cationic polymers, struggle to penetrate the mucus barrier, resulting in limited transfection efficiency. Herein, nanocomplexes (NCs) capable of penetrating both the mucus and cytomembrane barriers were developed to deliver TNF-α siRNA (siTNF-α) for effective pneumonia management. To construct the NCs, membrane-penetrating polypeptide (DPP) first condensed siTNF-α and formed a cationic inner core, which was further coated with a charge-reversal polymer (PD) followed by the adsorption of a RAGE-binding peptide (RBP) via electrostatic interactions. The resulting RDDsT NCs exhibited negatively charged surfaces and thus enabled efficient mucus layer penetration after intratracheal administration in lipopolysaccharide (LPS)-induced acute lung injury (ALI) mice. In the slightly acidic microenvironment of inflamed alveolar space, PD underwent charge reversal from negatively charged to positively charged, shedding off to facilitate the intracellular delivery of the DPP/siTNF-α core into alveolar macrophages. Meanwhile, the liberated RBP blocked RAGE-ligand interactions, further down-regulating pro-inflammatory factors. Consequently, the cooperative action of siTNF-α and RBP alleviated inflammation and propelled the recovery of pulmonary functions. This study renders an enlightened strategy to overcome the mucus/cytomembrane barrier against pulmonary siRNA delivery, and holds profound potential for gene/peptide co-therapy against pneumonia.

Epithelial cell adhesion molecule (EpCAM) serves a key biomarker for epithelial cancers detection in clinical diagnostics. There is a great need to develop highly sensitive methods for EpCAM-positive cell detection with a high throughput performance. Here we report an aptamer-mediated signal amplification strategy, that is, Endonuclease-Assisted Assay on Superwettable Droplet Microarrays (EASDM), for ultrasensitive EpCAM detection on single tumor cell surface. We have demonstrated that EASDM achieves sensitive detection of EpCAM on live tumor cell surface, with a limit of detection (LOD) as low as a single cell. It has been validated to distinguish MCF-7 cells (high EpCAM expression), MDA-MB-231 cells (low EpCAM expression), and white blood cells (negative control of EpCAM expression), or the tumor cells under chemically-induced epithelial-mesenchymal transition. We have further verified the robust performance of EASDM by using human whole blood samples spiked with cancer cells. This pilot platform integrating aptamer-based enzymatic amplification and superwettable droplet microarrays promises advancing cancer diagnostics/prognostics with high throughput, sensitivity, and accuracy for personalized medicine.

Objective We aimed to investigate the accuracy of a circulating exosome-based gene panel for predicting spread through air spaces (STAS). Methods A training cohort of 700 patients with stage I LUAD and a validation cohort of 178 patients were evaluated. Multivariable logistic regression was performed to identify exosome-derived genes and computed tomography (CT) features associated with STAS. The predictive performance of the gene panel, frozen sections (FS), and CT features was compared. Combined prediction models incorporating multiple diagnostic modalities were subsequently constructed and validated. Results STAS-positive tumors exhibited distinct exosome-derived gene signatures, including ALK rearrangement and elevated expression of MTA1, Twist, Slug, and β-Catenin. In the training cohort, the five-gene panel achieved an AUC of 0.802 (95 % CI: 0.770–0.834), with a sensitivity of 0.847 and a specificity of 0.666—superior to both FS (P < 0.001) and CT imaging features (P < 0.001). The AUC of the combined model integrating the gene panel with FS was 0.856 (95 % CI: 0.827–0.884), comparable to the model integrating the gene panel with radiology (AUC 0.846, 95 % CI 0.817–0.875; P = 0.471), and significantly higher than the combination of FS with radiology (AUC 0.803, 95 % CI: 0.767–0.840; P = 0.008 and P = 0.043, respectively). Similar results were also observed in the validation cohort. Conclusion ALK rearrangement and elevated expression of MTA1, Twist, Slug and β-Catenin in circulating exosomes were independent predictors of STAS. The five-gene panel serves as a promising adjunct to radiology and frozen sections for preoperative identification of STAS, potentially guiding optimal surgical decision-making.

Introduction:It remains controversial regarding the prognostic impact and therapeutic implications for immunotherapy of lymph node yield (LNY) during sublobar resection (SR) on stage I lung adenocarcinoma (LUAD). Methods:We analyzed a retrospective cohort of 400 patients with stage I LUAD who underwent SR, with peripheral blood samples prospectively collected for detecting inflammatory cytokines (IFCs). The effect of different LNY (≥4 vs. <4 nodes) on survival and IFC change was evaluated. Consensus clustering analyses were performed using data from The Cancer Genome Atlas (TCGA) and our validation cohort to explore associations between IFCs and immune/cell death profiles. A Bayesian meta-analysis was further conducted to assess the impact of LNY in LUAD undergoing SR. Results:The survival analysis of our cohort demonstrated that increased LNY during SR did not prolong RFS (≥4 vs. <4 nodes: HR = 1.15; 95%CI: 0.76-1.74). A lower LNY during SR was associated with significantly better RFS in stage I LUAD receiving adjuvant immunochemotherapy (≥4 vs. <4 nodes: HR = 0.41; 95%CI: 0.17-0.94). In terms of IFCs, extensive lymph node dissection led to significantly increased levels of IL-6, IL-4, IL-10 and TNF-α after SR (p < 0.05). Consensus clustering based on the IFCs identified two subgroups (Cluster 1 and 2) in TCGA cohort with distinct immune and cell death profiles, including differences in immunogenic cell death and damage-associated molecular patterns. Cluster 2 exhibited a higher Tumor Immune Dysfunction and Exclusion (TIDE) and tumor mutation burden scores. Similar findings were observed in our validation cohort, where Cluster 2 displayed higher number of neoantigens. The Bayesian meta-analysis also corroborated that increased LNY did not improve RFS (HR = 0.98; 95%CI: 0.20-2.94) in pathological stage I LUAD. Discussion:Increased LNY during SR might confer no additional benefits to RFS for p-stage I LUAD. Excessive removal of LNs might exert adverse impact on physical sensitivity to immunochemotherapy. Personalized lymph node management should be adopted for selected node-negative disease.

Tumor mutational burden (TMB) is a key biomarker for predicting immunotherapy benefit in lung adenocarcinoma. Deep learning–based TMB prediction from whole-slide images (WSIs) offers a non-invasive alternative to whole-exome sequencing (WES), yet conventional convolutional neural networks (CNNs) are limited by restricted receptive fields, resulting in insufficient global semantics and suboptimal generalization in high/medium/low TMB classification. To overcome these limitations, we propose a hybrid architecture that combines a Residual Attention Block (RAB) with a Multi-scale Attention Module (MSAM), and further integrates a lightweight Transformer branch to form the unified RMT framework for three-class TMB prediction. The RAB and MSAM components enhance local morphological representation through residual channel attention and multi-scale spatial attention, while embedded Transformer encoders capture global long-range dependencies via self-attention, yielding a comprehensive feature-learning scheme that unifies local extraction, multi-scale fusion, and global semantic modeling. This minimally intrusive Transformer design preserves training stability and reduces development cost while supporting fine-grained TMB stratification. Experiments on a large-scale lung adenocarcinoma WSI dataset show that RMT achieves 97.67

Objective:We aimed to comprehensively assess the accuracy of patient-derived organoids to recapitulate the genomic features of patient tumors, and to quantify the reliability of patient-derived organoid-based drug screening, by pooling the available evidence and the validation using our patient-derived lung cancer organoids (LCOs). Methods:A comprehensive online search was conducted to identify eligible studies. A total of 20 LCOs harvested from surgical samples were also established as our internal validation cohort, with whole-exome sequencing and transcriptomic sequencing performed, respectively. Pooled concordance in genetic profiles between LCOs and original tumors was calculated, along with the overall accuracy of LCOs for drug profiling. Results:The pooled concordance in mutational landscape between LCOs and original tumors reached 74% (95% confidence interval [CI]: 61%-84%) in overall populations without heterogeneity (I 2 = 0%, p = 0.918), and that in gene expression achieved 70% (95% CI: 51%-84%; I 2 = 8.5%, p = 0.437). Subgroup analyses of our internal LCOs by morphologic patterns suggested that LCOs retained genetic fidelity regardless of intratumoral histologic heterogeneity and pathologic features. The accuracy of LCOs for predicting drug sensitivity achieved 90.7% (95% CI: 77.4%-9.2%) in the pooled populations with mild heterogeneity (I 2 = 30.2%, p = 0.231). Conclusions:Our study indicates that LCOs can stably recapitulate the genomic features of patient tumors irrespective of histologic heterogeneity and be a promising precision medicine tool for anticancer drug screening or for predicting drug responses.

Objective: To evaluate a simplified surgical approach for Acute Type A Aortic Dissection (ATAAD). Methods: A retrospective analysis (Jan 2021–Oct 2025) compared 50 patients: an observation group undergoing Fontus stent graft with "MUSIC" strategy (Moderate hypothermia, Unilateral Selective antegrade cerebral perfusion, Ice Cap) versus a control group receiving Sun's procedure under deep hypothermic circulatory arrest (DHCA). Results: The Fontus/MUSIC group showed shorter circulatory arrest time, less postoperative drainage, faster recovery of coagulation and liver enzymes, better early renal and pulmonary function, and shorter ventilation. A graft bend >70.79° improved true lumen expansion. No differences were found in ECMO, renal replacement, stay, or 30-day mortality. Conclusion: The Fontus graft with “MUSIC” reduces technical complexity and perioperative risk, supporting its routine use in regional centers.

Objective Lung adenocarcinoma (LUAD) is the most prevalent subtype of lung cancer. Herein, we probed into the role of E3 ubiquitin protein ligase family member 1 (HERC1) in promoting ferroptosis and inhibiting LUAD cell proliferation by regulating RAF proto-oncogene serine/threonine-protein kinase (C-RAF). Methods In cultured human normal lung epithelial cells and non-small cell lung adenocarcinoma cell lines, HERC1 expression was determined by RT-qPCR and Western blot tests. PC-9 and Calu-3 cells were transfected with oe-HERC1, oe-C-RAF or their negative controls. Reactive oxygen species (ROS), malondialdehyde (MDA), glutathione (GSH), and Fe2+ levels were assessed by biochemical assays. Cell viability, death, and proliferation were evaluated by CCK-8, LDH and colony formation assays, followed by assessments of HERC1-C-RAF interaction, C-RAF ubiquitin level, and C-RAF protein stability. Results HERC1 was poorly expressed in LUAD cells. HERC1 promoted LUAD cell ferroptosis and repressed their proliferation and migration, corresponding to reduced levels of system xc-, GPX4, and GSH, as well as elevated levels of ROS, MDA, Fe2+, and ACSL4. LUAD cells overexpressing HERC1 displayed decreased C-RAF protein level, HERC1-C-RAF interaction, elevated C-RAF ubiquitin level, and accelerated C-RAF protein degradation, indicating that HERC1 facilitated C-RAF ubiquitin degradation and attenuated C-RAF protein stability via interaction with C-RAF. C-RAF overexpression partially abrogated the regulatory impact of HERC1 on LUAD cell ferroptosis and proliferation. Conclusion HERC1 expedites C-RAF ubiquitin degradation by interacting with C-RAF, which consequently promotes ferroptosis, thereby inhibiting LUAD cell proliferation.

Background: Patients with esophageal cancer (EC) who have undergone esophagectomy are at risk of developing hypoxemia and encountering postoperative complications. It is essential to ascertain whether the high-flow nasal cannula (HFNC) therapy offers superior clinical efficacy compared to conventional oxygen therapy (COT). Methods: Clinical data from 80 patients who experienced hypoxemia subsequent to radical esophagectomy were retrospectively collected at our institution spanning January 2020 to December 2022. The whole cohort was divided into two groups: the HFNC group and the COT group. Following oxygen administration, we evaluated the variations in arterial blood gas parameters and infection indices within each group, in addition to scrutinizing the occurrence of postoperative pulmonary complications. Results: The HFNC group was associated with a better oxygenation index (F group=41.779, p < 0.001) and partial pressure of carbon dioxide (F group=16.760, p < 0.001) compared with the COT group. Moreover, there were statistically significant differences in the reduction of C-reactive protein (F group = 17.603, p < 0.001) and neutrophil count (F group = 4.395, p=0.039) in the HFNC group compared with the COT group after 3 days of oxygen therapy. Notably, patients treated with HFNC exhibited a markedly reduced risk of developing postoperative complications, especially pneumonia (p=0.039). Conclusion: HFNC outperformed COT in enhancing oxygenation and reducing carbon dioxide levels and infection indices among patients with hypoxemia after radical resection of EC and also lowered the risk of postoperative pneumonia.

Background Globally, lung cancer is a malignant tumor with the highest morbidity and mortality rate, of which lung adenocarcinoma (LUAD) is the most common pathological type, accounting for about 40% of all cases. This study aims to investigate the impact and molecular mechanisms of mitotic spindle positioning (MISP) on the occurrence and progression of LUAD. Methods The expression level of MISP were evaluated by qPCR and Western blotting. Hippo signaling-related protein levels were measured by Western blotting. Co-immunoprecipitation assay was applied to assess the interaction between MISP and MST1. The roles of MISP in LUAD cells were determined through CCK-8 assay, transwell migration/invasion assay, and flow cytometric analysis. A xenograft tumor model (n = 6) was established to calculate the effect of MISP in vivo. Results MISP expression was significantly higher in LUAD tissues compared to normal tissues. MISP overexpression promoted LUAD cell proliferation and migration in in vitro models, as well as tumor growth and metastasis in LUAD animal models. Western blotting demonstrated that MISP overexpression increased the expression of YAP, TAZ and CYR61, and decreased the expression of MST1 and phosphorylated YAP, thereby blocking the Hippo signaling pathway. While MISP knockdown had the opposite effect. Mechanistic experiments revealed that MISP directly interacts with MST1, thereby inhibiting Hippo signaling. Conclusions In conclusion, our study reveals a novel discovery that MISP promotes the growth and invasion of LUAD cells by interacting with MST1 and suppressing the Hippo pathway.

ABSTRACT Background The existing diagnostic criteria for cancer cachexia do not meet clinical needs. We aimed to establish novel comprehensive evaluation scales for cachexia specific to patients with solid tumours. Methods This study included 12 651 patients (males: 6793 [53.7%]; females: 5858 [46.3%]; medium age: 58 [interquartile range:50/66] years; medium follow‐up duration: 24.16 [13.32/44.84] months; 4271 [33.8%] patients died; mean survival: 55.53 [95% confidence interval, 54.87/56.10] months; 3344 [26.4%], 4184 [33.1%] and 5123 [40.5%] patients with Stage I–II, III and IV tumour, respectively; derivation set: 10022, validation set: 2629 patients) with 14 types of solid tumours, including lung, gastric, liver, breast, oesophageal, cervical, bladder, pancreatic, prostate, ovarian, colorectal cancer, nasopharyngeal and endometrial carcinoma and cholangiocarcinoma, from an open and ongoing multicentre cohort study in China. Risk factors for cachexia, including tumour characteristics and nutritional parameters, were examined to develop diagnostic scales using Cox proportional hazards models and Kaplan–Meier analysis. Results Ten nutrition items (body mass index, weight loss, intake reduction, physical activity function, fatigue, handgrip strength, anorexia, albumin level, albumin/globulin ratio and neutrophil/lymphocyte ratio) with different weighted scores were identified to construct a nutrition‐weighted scoring scale (NWSS) for nutrition risk. Tumour type and tumour burden status (tumour‐node‐metastasis stage and radical or non‐radical tumour) were determined to construct a disease‐weighted scoring scale (DWSS) for disease risk. A lumped scale (5 × 5 matrix) established using a five‐grade classification of nutrition and disease risk was used to determine a five‐grade classification of comprehensive cachexia risk: A, no cachexia risk (reference; lowest disease and nutrition risks); B, cachexia risk (hazard ratio [HR] = 4.517 [4.033/5.058]); C, pre‐cachexia (HR = 9.755 [8.73/10.901], medium survival = 21.21 months); D, cachexia (HR = 16.901 [14.995/19.049], medium survival = 11.61 months); and E, refractory cachexia (HR = 31.879 [28.244/35.981], medium survival = 4.83 months, highest disease and nutrition risks) (p < 0.001). Patients in Categories A–D benefited from nutrition therapy and anti‐tumour treatments to varying degrees. Patients in Category E were clinically refractory to nutrition therapy without prolonged survival compared with patients without nutrition therapy (medium survival, pre‐hospitalization nutrition therapy vs. hospitalization nutrition therapy vs. without nutrition therapy, 2.89 [1.91/3.88] vs. 4.04 [3.21/4.88] vs. 5.89 [4.73/7.04] months, p = 0.015) and anti‐tumour treatments without prolonged survival compared with patients receiving palliative care (medium survival, radical anti‐tumour treatments vs. adjuvant anti‐tumour treatments vs. palliative anti‐tumour treatments vs. and palliative care, 6.48 [4.42/8.53] vs. 6.48 [3.23/9.73] vs. 4.83 [4.22/5.44] vs. 2.70 [1.09/4.30] months, p = 0.263). Conclusion We systematically developed a novel definition and grading diagnostic criteria for tumour‐type‐specific comprehensive cancer cachexia risk.

It remains undetermined regarding the impact of neoadjuvant therapy on immunogenic cell death (ICD) and subsequent tumor microenvironment (TME) remodeling in esophageal squamous cell carcinoma (ESCC). And it is of paramount significance to identify beneficiaries from neoadjuvant therapy in treatment-naïve ESCC. In this study, 88 ESCC samples undergoing neoadjuvant therapy plus surgery (NA+S) or surgery alone (SA) were subjected to bulk-RNA sequencing. A five-gene RINscore incorporating ICD-related signature genes with TME-based hub genes was established to predict clinical outcomes and pharmacological responses, in which SLAMF7 and IL1R1 were selected out as co-expressed genes. The regulatory mechanism of the repressive co-transcription factor BATF of SLAMF7 and IL1R1 was further demonstrated. Our data demonstrated that NA+S led to high abundance in kinds of T helper cells, nature killer T cells and M1-like macrophages with increased CD8+T cells infiltration compared with SA. ICD phenotypes were further characterized in treatment-naïve ESCC to determine their differences in TME and potential benefits from NA. Our findings not only offered novel insights into the distinct TME and ICD profiles of ESCC undergoing different therapeutic modes, but also provided the RINscore, which may aid oncologists in determining individualized (neo)adjuvant immunotherapy regimen.

It has not been determined which descriptor spread through air spaces (STAS) should be incorporated into the context of the ninth Tumor, Node and Metastasis (TNM) staging system: the T or the uncertain resection [R(un)] category. A multicenter retrospective cohort of 807 patients with pathological stage I lung adenocarcinoma was included in this study to assess the feasibility of incorporating STAS into the T descriptor or the R(un) category by analyzing recurrence-free survival (RFS) and overall survival (OS). Decision curve analysis (DCA) was performed to evaluate the standardized net benefit of the proposed T (nT) and that of the proposed residual tumor classification (nR) versus the current staging systems. Log-rank tests indicated that patients with pT1/STAS-positive lung adenocarcinoma had similar RFS and OS to patients with pT2a disease irrespective of R status. Regarding STAS as an indicator for upgrading R0 to R(un), comparable survival was observed between pT1-2a/STAS-positive patients undergoing R0 segmentectomy and pT1-2a patients undergoing R(un) segmentectomy. We further assessed the effects of the combination of STAS with either T or R category on survival in a validation cohort. Subgroup analyses stratified by surgical procedures further identified the consistency of the nT category in discriminating RFS and OS. However, the separation of nR0 and nR(un) disease in pT2a tumors treated by lobectomy or segmentectomy was not sufficiently distinguished. DCA further corroborated a greater predictive capability of nT versus the current T category. In conclusion, STAS might be preferentially considered as an indicator for upgrading pT1 disease into pT2a in the future TNM staging system.

Esophageal squamous cell carcinoma (ESCC) is a highly aggressive malignancy with a dismal prognosis. Hitherto, little has been known regarding the clinical implications of tertiary lymphoid structures (TLS) and its biological mechanisms of antitumor effect on treatment-naïve ESCC. We herein identified the presence of TLS as an independent factor for favorable survival. By characterizing the immune infiltration and genomic profiles based on transcriptomic datasets, we found TLS abundant in enriched B cells with IRF4 as a signature gene. Increased expression of IRF4 and its positive correlation with STING in activating tumor-infiltrating B cells were also investigated using a single-cell RNA sequencing dataset. CD40 as a co-regulator of IRF4 and TLS formation, in vitro experiments were conducted to further demonstrate the competitive binding relationships between CD40 and STING with TRAF2 in promoting IRF4 expression and B cell activation via the non-canonical NF-kB signaling pathway, in which CD40 reduced STING ubiquitination while promoting its phosphorylation. Our data provided deeper insights into the potential role of activated B cells and TLS in ESCC, with implications for the development of biomarkers and therapeutic targets.

Background The purpose of this study was to assess the prognostic factors for non-lepidic invasive adenocarcinoma presenting as subsolid nodules. The feasibility of detecting non-lepidic predominant patterns on frozen section (FS) was also evaluated. Methods A multicenter retrospective cohort of 614 patients with clinical T1N0M0 non-lepidic invasive adenocarcinoma presenting as subsolid nodule was included. Two subgroups were divided based on the consolidation-to-tumor ratio (CTR) on lung window: ground glass opacity (GGO)-dominant subgroup (CTR<0.5), solid-dominant subgroup (CTR≥0.5). Kaplan-Meier approach and multivariable Cox models were used to identify risk factors for recurrence-free survival (RFS) and overall survival (OS). FS and final pathology (FP) of 100 specimens were also reviewed by five pathologists for tumor grading synchronously. Results Multivariate analysis indicated that segmentectomy was a risk factor for shortened RFS and OS in the solid-dominant subgroup rather than in the GGO-dominant one. Subset analysis demonstrated survival disadvantages of segmentectomy for high-grade adenocarcinoma but not for intermediate-grade one in the solid-dominant subgroup. However, segmentectomy exhibited non-inferiority to lobectomy in the GGO-dominant subgroup irrespective of tumor grade. The overall accuracy of identifying non-lepidic patterns was 84 % with a good interobserver agreement. Multivariable logistic analysis identified presence of complex glandular pattern and acinar pattern as independent predictors of the discrepancy between FS and FP. Conclusions Segmentectomy should be cautiously performed for patients with radiologically solid-dominant non-lepidic invasive adenocarcinoma, especially for those with high-grade patterns. FS had high diagnostic accuracy and satisfactory interobserver agreement for tumor grading, which might aid surgeons in determining the appropriate surgical procedure.

This study is aimed to develop predictive models for classifying thymic epithelial tumor (TET) histological subtypes (A/AB/B1, B2/B3, C) and WHO stages (I-IV) using radiomics features derived from contrast-enhanced CT scans. These models were validated on multicenter external datasets to improve preoperative diagnosis and guide treatment decisions. A total of 257 patients diagnosed with TET between January 2013 and April 2024 were retrospectively analyzed, with 181 cases from the First Affiliated Hospital of Soochow University served as the training cohort and 76 cases from the Second Affiliated Hospital used as an external test set. All patients underwent preoperative enhanced CT scans. After manual segmentation of the volume of interest (VOI), 1,038 radiomic features were extracted. Feature selection was performed using PCA and LASSO methods. Three models (clinical semantic, radiomics, and a fusion model combining both) were built using random forest algorithms. The fusion model achieved the highest performance in the external test set, with an accuracy of 0.908 and F1 score of 0.896 for histological subtype classification, and an accuracy of 0.803 and F1 score of 0.833 for WHO staging. The radiomics model shows slightly lower performance, while the clinical semantic model performs the weakest. Our findings suggest that machine learning models integrating radiomics and clinical features can effectively predict TET subtypes and stages, offering a non-invasive tool for accurate preoperative assessment with strong generalization ability.

BACKGROUND:This study aims to investigate the prognostic value of baseline and absolute changes of inflammatory markers in thymic epithelial tumors (TETs). METHODS:This real-world study enrolled 147 patients. Inflammatory markers, including baseline and absolute changes of platelet-to-lymphocyte (PLR), platelet-to-monocyte (PMR), lymphocyte-to-monocyte (LMR), and neutrophil-to-lymphocyte ratio (NLR), were retrospectively collected. Survival analysis was conducted using Kaplan-Meier analysis. Lasso and multivariate Cox regression were conducted to assess the relationship between disease-free survival (DFS), overall survival (OS) and inflammatory markers. Predictive nomograms were constructed by variables identified in Cox regression. RESULTS:Higher baseline NLR and PLR, lower baseline LMR and PMR, a greater change in PLR and smaller changes in LMR, PMR, and PNR related to inferior DFS and OS. Multivariate Cox regression identified higher baseline PLR and advanced Masaoka Stage were risk, while higher baseline PMR was protective OS factors. Age, higher baseline PLR, and advanced Masaoka Stage were risk DFS factors. A greater absolute PLR change and metastasis were risk, while a lower absolute PMR change was protective OS factors. Age and a greater absolute PLR change were risk, while a greater absolute LMR change was favorable DFS factors. CONCLUSIONS:The present study is the first to reveal the prognostic role of the baseline and absolute changes of inflammatory markers in TETs. The prognostic models based on these inflammatory markers offer promising tools for predicting the prognosis of TET patients.

Background Lung cancer remains one of the most prevalent cancer types worldwide, with a high mortality rate. Upregulation of programmed cell death protein 1 (PD-1) and its ligand (PD-L1) may represent a key mechanism for evading immune surveillance. Immune checkpoint blockade (ICB) antibodies against PD-1 or PD-L1 are therefore widely used to treat patients with lung cancer. However, the mechanisms by which lung cancer and neutrophils in the microenvironment sustain PD-L1 expression and impart stronger inhibition of CD8(+) T cell function remain unclear. Methods We investigated the role and underlying mechanism by which PD-L1(+) lung cancer and PD-L1(+) neutrophils impede the function of CD8(+) T cells through magnetic bead cell sorting, quantitative real-time polymerase chain reaction (RT-PCR), western blotting, enzyme-linked immunosorbent assays, confocal immunofluorescence, gene silencing, flow cytometry, etc. In vivo efficacy and safety studies were conducted using (Non-obeseDiabetes/severe combined immune deficiency) SCID/NOD mice. Additionally, we collected clinical and prognostic data from 208 patients who underwent curative lung cancer resection between 2017 and 2018. Results We demonstrated that C-X-C motif chemokine ligand 5 (CXCL5) is markedly overexpressed in lung cancer cells and is positively correlated with a poor prognosis in patients with lung cancer. Mechanistically, CXCL5 activates the phosphorylation of the Paxillin/AKT signaling cascade, leading to upregulation of PD-L1 expression and the formation of a positive feedback loop. Moreover, CXCL5 attracts neutrophils, compromising CD8(+) T cell-dependent antitumor immunity. These PD-L1(+) neutrophils aggravate CD8(+) T cell exhaustion following lung cancer domestication. Combined treatment with anti-CXCL5 and anti-PD-L1 antibodies significantly inhibits tumor growth in vivo. Conclusions Our findings collectively demonstrate that CXCL5 promotes immune escape through PD-L1 upregulation in lung cancer and neutrophils chemotaxis through autocrine and paracrine mechanisms. CXCL5 may serve as a potential therapeutic target in synergy with ICBs in lung cancer immunotherapy.