Objective.Accelerated single photon emission computed tomography (SPECT) imaging, achieved by reducing either the number of projection angles or the acquisition time per angle, enhances clinical workflow efficiency but introduces elevated noise. This study aims to develop and validate a universal DL-based reconstruction framework that effectively generalizes across diverse, clinically-realistic SPECT acceleration protocols by overcoming the data scarcity challenge.Approach.SPECT bone scans from 103 patients were acquired under a standard scan (60 views, 12 s/view, 60v12s) followed by a fast scan using one of five acceleration protocols (60v6s, 60v3s, 30v12s, 30v6s, 16v12s). AU-Net-based reconstruction framework was implemented using three strategies: (1) single model, trained on individual acceleration protocol), (2) base model, trained on aggregated datasets from five acceleration protocols, and (3) transfer model, fine-tuned from the base model for protocol-specific optimization. Pixel-level accuracy and structural similarity were assessed using peak signal-to-noise ratio (PSNR), structural similarity index (SSIM) metrics, and maximum pixel value of lesions. Clinical evaluation of image quality, radionuclide detail, artifacts, and diagnostic confidence was conducted using a 5-point system.Main results.Quantitative evaluation showed the transfer model achieved better PSNR and SSIM across all protocols (highest 48.02 PSNR and 0.9918 SSIM in 30v6 s protocol). Qualitative analysis confirmed enhanced structural fidelity. Clinical evaluations rated the transfer model highest across metrics, with scores of 4.667 ± 0.508 (image quality, 30v12 s), 4.800 ± 0.250 (radionuclide detail, 60v6 s), 1.150 ± 0.173 (artifact reduction, 60v6 s), and 4.800 ± 0.250 (diagnostic confidence, 60v6 s), surpassing full-scan results in most cases.Significance. The proposed transfer learning (TL) framework effectively addressed data scarcity and improved reconstruction performance across diverse SPECT acceleration scenarios. The adoption of a TL strategy mitigates data scarcity by utilizing shared features and fine-tuning for specific protocols. The framework demonstrates potential for integration into fast SPECT workflows, facilitating reliable use across diverse imaging scenarios.

Background The prognostic relevance of type 2 diabetes in esophageal cancer patients undergoing neoadjuvant therapy followed by surgery remains unclear. This study assessed whether type 2 diabetes influences perioperative outcomes and long-term survival. Methods A total of 1099 patients who received neoadjuvant therapy followed by surgery at three high-volume institution in China from 2013 to 2024 were included. The diagnosis of type 2 diabetes was based on American Diabetes Association criteria. A 1:4 propensity score matching was performed using the R software. Baseline characteristics, perioperative outcomes, overall survival and progression-free survival were compared. Results After 1:4 matching, 114 patients in the diabetic group and 456 patients in the non-diabetic group were compared. No significant differences were found between non-diabetic and diabetic group regarding length of stay (9 days vs 9 days, P = 0.856), re-operation (2.2% vs 0.9%, P = 0.594), re-admission (1.5% vs 1.8%, P > 0.990), total complications (45.0% vs 51.8%, P = 0.231), or specific complications. At a median follow-up of 1086 days, the 2-year overall survival between non-diabetic and diabetic group was 73.9% vs 77.2%, and the 3-year overall survival was 62.4% vs 67.8% [hazard ratio: 0.840, 95% CI: 0.570–1.220, P = 0.356]. After a median follow-up of 927 days, the 2-year progression-free survival of the two groups was 73.2% vs 65.1%, and the 3-year progression-free survival was 65.2% vs 62.4% [hazard ratio: 1.130, 95% CI: 0.780–1.630, P = 0.529]. Conclusions Type 2 diabetes does not adversely affect perioperative outcomes or long-term survival in esophageal cancer patients undergoing neoadjuvant therapy and surgery.

CD4(+) T cells are indispensable for CD8(+) T cells-mediated anti-tumor immunity, while little is known about how CD4(+) T cells coordinate with other cells to promote CD8(+) T cells activity. In this study, by coupling single-cell RNA sequencing (scRNA-seq) with multiplex immunohistochemistry staining (mIHC), CXCL13-expressing CD4(+) T cells were explored to recruit CXCR5(+) B cells to form CXCL13(+)CD4(+) T cells:: CXCR5(+) B cells:: CD8(+) T cells triad, promoting the anti-tumor immunity and heralding a favorable prognosis. A CXCL13-expressing subset of CD4(+) T cells was identified to be associated with better prognosis and tumor-reactive hallmarks. The further cell types interaction analysis revealed a specific interaction between CXCL13(+)CD4(+) T cells and CXCR5(+) B cells in comparison with other cell types. The mIHC verified that an elevated level of CXCR5(+) B cells infiltration in the proximity of CXCL13(+)CD4(+) T cells. The subsequent Entropy analysis, which characterizes the colocalization among more than two cell types, was employed to reveal the spatial distribution pattern. The results revealed that CXCL13(+)CD4(+) T cells and CXCR5(+) B cells co-localized with cytotoxic CD8(+) T cells, forming a functional lymphocyte triad, whose existence heralded a better prognosis. Collectively, by integrating scRNA-seq with mIHC, we demonstrated that the recruitment of CXCR5(+) B cells by neoantigen reacting to CXCL13(+)CD4(+) T cells through CXCL13-CXCR5 signaling contributed to the lymphocyte triad, leading to an enhanced anti-tumor immunity and heralding a favorable prognosis in non-small-cell lung cancer (NSCLC).

CASE PRESENTATION:A 68-year-old man presented to our thoracic surgery department for evaluation of a left upper-lobe pulmonary nodule detected on chest CT scan. The oncologic history extended over 7 years. In December 2014, the patient was diagnosed with colorectal adenocarcinoma (ileocecal junction) by colonoscopy, and he underwent a laparoscopic right hemicolectomy and lymph node dissection. Postoperative pathological examination revealed a grade II-III adenocarcinoma of the ileocecal region (ulcerative type), infiltrating into the muscularis. Surgical margins were free of tumor. No metastasis was identified in the examined lymph nodes, including paracolic (19), paracolic (7), intermediate mesenteric (23), and central vascular (2) lymph nodes, corresponding to a pathologic stage of pT2N0M0. The patient recovered well and was placed on regular follow-up.

Background:Immune checkpoint inhibitors (ICIs) have transformed the treatment of esophageal cancer, yet only a subset of patients derive durable clinical benefit. The neutrophil-to-lymphocyte ratio (NLR) is a readily accessible inflammatory biomarker derived from routine blood tests; however, no interpretable machine learning model tailored to esophageal squamous cell carcinoma (ESCC) currently integrates NLR with multidimensional clinical features for response prediction. We developed an NLR-centered model to predict ICI response in patients with ESCC. Methods:This was a multicenter retrospective cohort study of 419 ESCC patients treated with ICIs at four medical centers. Eighteen pretreatment clinical and hematological features were used as input features; the outcome was ICI response. Six machine learning algorithms spanning the principal tabular-learning families were compared. Random Forest hyperparameters were selected by RandomizedSearchCV; the remaining models used pre-specified regularized hyperparameters to control overfitting. Internal validation used 5-fold stratified cross-validation (CV). Interpretability used SHapley Additive exPlanations (SHAP), and the optimal NLR cutoff was determined by the Youden index of the receiver operating characteristic (ROC) curve. Results:Under 5-fold stratified CV, Gradient Boosting Machine (GBM) achieved the highest area under the curve (AUC) (0.783); eXtreme Gradient Boosting (XGBoost) and Random Forest tied second (both 0.771), followed by Logistic L1 (0.747), Multilayer Perceptron (0.700), and K-Nearest Neighbors (0.694). XGBoost was selected as the primary interpretable model based on the overall balance of discrimination, calibration, clinical utility, and SHAP interpretability rather than on AUC alone. At the Youden-optimal threshold (0.612), XGBoost reached sensitivity 0.777, specificity 0.680, accuracy 74.2%, F1-score 0.795, positive predictive value 0.813, negative predictive value 0.630, and Brier 0.178. SHAP ranked NLR first by both mean |SHAP| (0.633) and gain (0.185), followed by Eastern Cooperative Oncology Group performance status and neutrophil count. NLR was lower in responders than non-responders (P<0.001; optimal cutoff 3.40). Conclusions:An interpretable NLR-centered XGBoost model built on routine clinical and hematological variables provided moderate-accuracy prediction of ICI response in ESCC, with an optimal NLR cutoff of 3.40 for initial clinical screening, pending external validation.

Abstract Topic Esophageal Cancer: Surgical Treatment of Esophageal Cancer – early outcomes and complications Background The prognostic impact of type 2 diabetes in esophageal cancer patients receiving neoadjuvant therapy followed by surgery remains unclear. We evaluated its association with perioperative outcomes and long-term survival. Methods We retrospectively analyzed 1,099 patients who underwent neoadjuvant therapy followed by surgery between 2013 and 2024 across three high-volume centers. Type 2 diabetes was defined according to American Diabetes Association criteria. A 1:4 propensity score matching analysis was performed to minimize baseline imbalance. Perioperative outcomes, overall survival (OS), and progression-free survival (PFS) were compared between diabetic and non-diabetic patients. Results After matching, 114 diabetic patients were compared with 456 non-diabetic patients. No significant differences were observed in length of hospital stay, re-operation, re-admission, total complications, or specific postoperative morbidities. After a median follow-up of 1,086 days, 2- and 3-year OS rates were 73.9% vs 77.2% and 62.4% vs 67.8%, respectively (HR 0.84, 95% CI 0.57–1.22, P=0.356). After a median follow-up of 927 days, 2- and 3-year PFS rates were 73.2% vs 65.1% and 65.2% vs 62.4%, respectively (HR 1.13, 95% CI 0.78–1.63, P=0.529). Type 2 diabetes was not associated with inferior survival outcomes. Conclusion Type 2 diabetes was not associated with increased perioperative morbidity or impaired long-term survival following neoadjuvant therapy and surgery.

BACKGROUND: Cancer cells, such as non-small cell lung cancer (NSCLC) cells, exhibit remarkable phenotypic plasticity and undergo epigenetic reprogramming, characteristics that enable them to evade targeted therapies. However, how NSCLC cells use epigenetic regulatory mechanisms (including endogenous DNA base damage) to drive transcriptome reprogramming and develop resistance to EGFR-tyrosine kinase inhibitor (EGFR-TKI) therapy remains unclear. METHODS: We employed an integrated multi-omics approach in erlotinib-sensitive and resistant NSCLC cells, including genome-wide mapping of abasic sites (AP sites) via Single-Strand Break Mapping at Nucleotide Genome Level-AP (SSiNGLe-AP), characterization of binding landscapes of 8-Oxoguanine DNA Glycosylase 1 (OGG1), Apurinic/Apyrimidinic Endonuclease 1 (APE1), and genome-wide profiling of G-quadruplex (G4) structures by Cleavage Under Targets and Tagmentation sequencing (CUT&Tag-seq). Combined with RNA-seq and ATAC-seq analysis, we performed functional validation using overexpression, RNAi, qPCR, western blot, immunofluorescence, sphere formation, and flow cytometry. High-resolution microscopy revealed that oxidized base lesions and repair complexes orchestrated the spatiotemporal dynamics of G4 structures. Co-localization of G4 structures with CpG islands (CGIs) was assessed by Whole-Genome Bisulfite Sequencing (WGBS). Finally, the therapeutic relevance of targeting this pathway was evaluated using a mouse xenograft model treated with APE1 inhibitors. RESULTS: We identified the DNA base lesion repair proteins OGG1 and APE1 as key mediators of transcriptional reprogramming that promote cancer cell plasticity and resistance to EGFR-TKIs. Mechanistically, genome-wide profiling revealed their dynamic redistribution to specific genomic loci in resistant cells. Beyond canonical repair, APE1 stabilizes G4 structures at the promoters of epithelial-mesenchymal transition (EMT) and stemness-associated genes, facilitating their transcriptional activation. This APE1-driven, G4-dependent transcriptional program occurs preferentially at hypomethylated CGIs, revealing a mechanism by which DNA secondary structures shape the epigenetic landscape to promote cellular plasticity. CONCLUSION: Our study elucidates a novel pathway in which OGG1/APE1-mediated processing of oxidative damage orchestrates a G4-dependent transcriptional program to drive EMT and stemness in EGFR-TKI-resistant NSCLC. The therapeutic potential of targeting this axis is demonstrated by the efficacy of BER inhibitors in suppressing tumor growth in vivo, establishing APE1-G4 targeting as a promising anti-resistance strategy.

Insufficient T-cell infiltration is a major barrier to the efficacy of immune checkpoint inhibitors (ICIs) in lung adenocarcinoma (LUAD). We aimed to investigate how the tumor-intrinsic m5C methyltransferase NSUN2 shapes the immune landscape of LUAD. Nsun2 conditional knockout mice and syngeneic mice models were employed. Single-cell RNA sequencing (scRNA-seq) and m5C sequencing were performed to elucidate the downstream pathway regulated by NSUN2. LUAD specimens from patients receiving neoadjuvant immunotherapy were used to assess the correlation between NSUN2 expression and immune infiltration. Elevated tumoral NSUN2 expression was correlated with a marked lack of CD8+ T-cell infiltration and immunotherapy resistance. Mechanistically, NSUN2 enhances the translation of histone deacetylase HDAC8 in an m⁵C-YBX1-dependent manner, which in turn directly represses the transcription of the chemoattractant CCL5, thereby impairing CD8+ T-cell recruitment into the tumor. NSUN2 depletion reversed this immunosuppressive axis and converted immunologically “cold” tumors into “hot”. Tumor-targeted liposomes loaded with the NSUN2 inhibitor synergized with anti-PD-1 therapy to induce significant tumor regression. Our findings identify NSUN2 as a critical orchestrator of T-cell exclusion in LUAD. It serves as both a candidate predictive biomarker for ICI failure and a promising druggable target. Targeting NSUN2 offers a potential strategy to overcome immunotherapy resistance and improve clinical outcomes in LUAD patients.

The design of antigen-binding domains (binders) is emerging as a decisive frontier in chimeric antigen receptor (CAR)-T cell engineering. Rather than serving as passive recognition elements, binders actively shape antigen selectivity, signaling thresholds, exhaustion propensity, persistence, and toxicity. We propose that optimal CAR performance requires a multidimensional design strategy. Success cannot be achieved by maximizing a single variable like affinity. Designers must instead integrate epitope position, binding kinetics, avidity, molecular geometry, and biophysical stability. In this review, we synthesize recent advances showing how these parameters collectively govern immunological synapse formation, antigen-density discrimination, and functional durability. Importantly, we highlight that the consequences of binder design are increasingly evident not only in mechanistic and preclinical studies, but also in real-world clinical translation and commercial trajectories. We further examine how the binder landscape has expanded beyond conventional scFvs to include VHHs, monobodies, DARPins, D-domains, peptides, natural ligands, TCR-mimic binders, and de novo AI-designed proteins. Together, these advances support a shift from empirical binder selection to rational binder orchestration as a foundational principle for next-generation CAR-T immunotherapy.

Non-small cell lung cancer (NSCLC) patients carrying HER2 exon 20 insertion (HER2 20ins) are respond poorly to conventional therapies and pan HER inhibitors. Trastuzumab Deruxtecan (T-DXd) has been approved in second line treatment in advanced NSCLC, but its efficacy in operable patients remains unclear. Here, we report a case of a patient with operable, HER2 20ins NSCLC who received neoadjuvant T-DXd, along with a second HER2 20ins patient who received conventional chemotherapy as a clinical reference. We utilized whole-genome sequencing (WGS) and immune microenvironment analysis to investigate the treatment's impact. WGS analysis revealed marked reduction in tumor clones with genome instability and extrachromosomal DNA (ecDNA) following T-DXd treatment. Meanwhile, immune profiling demonstrated substantially increased CD8 + T cell infiltration in tumor cores with elevated PD-1 expression. In contrast, such changes in CD8 + T cell infiltration and PD-1 expression were not evident in another HER2 20ins patient who received conventional chemotherapy. These findings suggest that neoadjuvant T-DXd may represent a promising therapeutic option for locally advanced HER2-mutant NSCLC, warranting further investigation in larger cohorts.

Type 2 diabetes mellitus (T2DM) is a well-recognized risk factor for various chronic and acute conditions. The primary aim of this study was to investigate the influence of T2DM on medical complications and long-term survival following minimally invasive esophagectomy (MIE). A total of 906 consecutive patients with esophageal squamous cell carcinoma (ESCC) who underwent MIE were included in this retrospective study. Patients were categorized into two groups based on preoperative comorbidity with T2DM. Propensity score matching was utilized to align baseline information between the two groups. After PSM, 324 patients in the without-T2DM group (No-T2DM) and 118 patients in the Coexisting-T2DM group (Coe-T2DM) were ultimately included in the study. Compared with the No-T2DM group, the Coe-T2DM group exhibited longer total operative duration (225 vs 215; P = 0.039), more intraoperative blood loss (100 vs 60; P = 0.001), higher incidence of postoperative complications (56.78 vs 39.20

Background:While curative-intent resection for gallbladder cancer (GBC) is being increasingly performed in elderly patients, perioperative morbidity and long-term oncological outcomes in this population remain unclear. Methods:Consecutive patients with GBC who underwent curative-intent resection (2016-2020) were identified from a multicentre database and stratified as elderly (>70 years) or younger (≤70 years). Propensity score matching (PSM) and inverse probability of treatment weighting (IPTW) were used to minimise selection bias. The outcomes compared included overall survival (OS), cancer-specific survival (CSS) and short-term outcomes. Logistic regression was used to identify risk factors for major morbidity and Cox regression was used for CSS, with the Fine-Gray competing risk model further applied to account for non-cancer-related death in the analysis of CSS. Results:Among the 575 patients enrolled, 432 were younger and 143 were elderly. After 1:1 PSM and IPTW, elderly patients had significantly higher 90-day major morbidity rates than younger patients (p=0.029 and p=0.010, respectively), but also demonstrated better CSS in both cohorts (p=0.038 and p<0.001, respectively). OS was significantly longer in elderly patients only after PSM (p<0.001), with no significant difference after IPTW (p=0.684). Multi-adjusted analysis confirmed that advanced age was an independent risk factor for major morbidities (original cohort: OR 1.71 (95% CI 1.12 to 2.59); PSM: OR 1.88 (95% CI 1.08 to 3.30); IPTW: OR 1.83 (95% CI 1.17 to 2.82)) but was associated with longer CSS (original cohort: HR 0.65 (95% CI 0.44 to 0.97); PSM: HR 0.29 (95% CI 0.18 to 0.45); IPTW: HR 0.62 (95% CI 0.42 to 0.90); Fine-Gray model: HR 0.34 (95% CI 0.30 to 0.39)). Conclusion:Despite a higher risk of major postoperative morbidity, advanced age itself might not be considered a contraindication for curative-intent resection. Selected elderly patients with GBC may achieve superior CSS compared to their younger counterparts.

BACKGROUND:Several high-quality clinical trials have shown promising results compared with neoadjuvant chemotherapy (NCT) or chemoradiotherapy, further supporting the efficacy of neoadjuvant immunochemotherapy (NICT) for locally advanced esophageal cancer (AEC). This study evaluated the efficacy of NICT in different segments of esophageal squamous cell carcinoma (ESCC) and aimed to identify potential prognostic factors. METHODS:A retrospective cohort study analyzed 248 patients. Before undergoing radical surgery, the patients received two to four cycles of NICT. The relationship between different segments of ESCC and both clinical and pathologic characteristics were evaluated. RESULTS:The patients with tumors in the upper thoracic esophagus had a higher proportion of lower tumor regression grade (TRG) scores. A significantly higher proportion of patients with TRG 0-1 was observed in the upper and middle thoracic tumor groups than in the lower thoracic group (P = 0.001). In esophageal cancer, TRG 2-3 was significantly associated with lower thoracic tumor location (odds ratio [OR], 3.298; P = 0.003), clinical T3 stage (OR, 2.834; P < 0.001), and clinical N2 stage (OR, 2.557; P = 0.045). CONCLUSIONS:The patients with upper-third ESCC appeared to derive greater clinical benefits from NICT than those with tumors located in the lower third of the esophagus in this post hoc analysis. This could potentially serve as a predictor of NICT efficacy.

Pancreatic adenocarcinoma (PAAD) is a common malignancy with a very low survival rate. More and more studies have shown that SPTAN1 may be involved in the development and progression of a variety of tumors, including rectal cancer, Pancreatic adenocarcinoma, etc., and may affect their prognosis. Bioinformatics technology was used to analyze the relationship between SPTAN1 expression in PAAD and immune cell infiltration, immune regulatory factors and chemokines, and cell experiments were used to verify the relationship between SPTAN1 knock down and migration, invasion, apoptosis and cycle changes of PAAD cell lines. In addition, immunohistochemical staining of SPTAN1 was performed by tissue microarray (TMA) to study the relationship between high expression of SPTAN1 and clinicopathological features and overall survival rate. The expression of SPTAN1 is significantly correlated with immune cell infiltration, immunomodulators, chemokines and their receptors. In addition, it was found that the knock-down of SPTAN1 inhibited the migration and invasion ability of PAAD cell lines, promoted the apoptosis of cell lines, and also affected the changes of cell cycle. Immunohistochemical staining using tissue microarray (TMA) showed that the high expression of SPTAN1 was associated with M stage (P = 0.004) and CA199 (P = 0.012), and the overall survival rate of the high expression group was significantly lower than that of the low expression group (P = 0.043). Our results suggest that up-regulation of SPTAN1 is related to cell migration, invasion, apoptosis and cycle changes, and is associated with tumor immune invasion and poor prognosis of PAAD.

There is no consensus regarding whether primary tumour resection (PTR) should be performed in non-small cell lung cancer (NSCLC) patients with unexpected pleural dissemination (PD) discovered at thoracotomy. Consecutive NSCLC patients with surgically confirmed PD were retrospectively enrolled from two high-volume centres between January 2016 and December 2023. Patients were divided into the primary tumour resection (PTR) and exploratory thoracotomy (ET) group. PTR included wedge resection, segmentectomy and lobectomy. Patients in the ET group received biopsy only. Propensity score matching (PSM) was used to reduce selection bias from confounding factors. Disease-specific survival (DSS) and progression-free survival (PFS) were analysed using the Kaplan‒Meier method, and comparisons were made using the log-rank test. Multivariate Cox regression analyses were performed to identify the independent prognostic factors. A total of 223 patients were identified: 167 (74.9

OBJECTIVE:Robotic systems have enhanced thoracic surgery through 3D visualization, manoeuvrability, and tremor filtration. This study evaluated the effectiveness and safety of the da Vinci single-port surgical system for thoracic surgery in a Chinese population (ClinicalTrials.gov; NCT06246617). DESIGN:Patients with highly malignant pulmonary nodules suitable for minimally invasive lung resection underwent anatomical lung resection using a single-port robotic system between December 2023 and May 2024. Perioperative and short-term follow-up outcomes were also assessed. Clinical data from another clinical trial (RVlob Trial: NCT03134534) were also obtained for comparison. RESULTS:The procedure was successful in all 15 patients without intraoperative conversion, although 3 required an additional auxiliary port owing to a high body mass index and complex anatomy. Additionally, the median intraoperative blood loss and operative time were 15 (interquartile range [IQR], 10-37.5) mL and 100 (IQR, 86-119) minutes, respectively. The median length of hospital stay and duration of thoracic catheter retention were 8 (IQR, 7-9) and 2 (IQR, 2-2) days, respectively. No rehospitalizations or readmissions occurred. In comparison to the traditional robotic-assisted and video-assisted surgery groups, the single-port robotic surgery group showed a shorter operative time and less blood loss. CONCLUSIONS:Single-port robotic thoracic surgery using the da Vinci single-port surgical system is safe and feasible for anatomical lung resection. Moreover, larger controlled studies are required to validate its effectiveness and compare its advantages with those of other surgical platforms. CLINICAL TRIAL REGISTRATION NUMBER:ClinicalTrials.gov; NCT06246617.