BMC Cancer最新文献

Background: Lung cancer, particularly adenocarcinoma and squamous cell carcinoma, remains a leading cause of cancer-related deaths globally. The diagnosis of multiple primary lung cancers (MPLCs) has become more frequent due to advanced chest CT technology and improved health surveillance. However, differentiating MPLCs from intrapulmonary metastases (IPMs) and multiple benign pulmonary lesions (MBPLs) remains challenging.

Objectives: Distinguishing multiple primary lung cancers from metastases and benign lesions on CT remains challenging yet critical for treatment planning. Current approaches rely on subjective interpretation and invasive procedures. This study aims to develop and validate an automated deep learning classification system to provide rapid, objective diagnoses for optimizing patient management.

Materials and methods: We studied 260 patients (MPLC = 83, IPM = 81, MBPL = 96; 881 axial CT slices). Six pretrained architectures (DenseNet-121, EfficientNet-B1, MambaOut-Kobe, ResNet-50, SwinV2-CR-Tiny-224, ViT-Tiny-Patch16-224) were compared in a five-seed ablation (seeds 42, 789, 1011, 2025, 2048). Pairwise one-vs-rest DeLong tests were aggregated across seeds to compare AUCs. Clinical utility was assessed using decision curve analysis (DCA). The final model (MambaOut-Kobe) underwent stratified five-fold cross-validation.

Results: Considering efficiency, MambaOut-Kobe combined competitive accuracy with the lowest memory (~ 100 ± 14 MB) and low latency (~ 0.0093 ± 0.0017 s/image). Aggregated DeLong testing found no significant AUC differences among these models after multiplicity control. On five-fold cross-validation, MambaOut-Kobe achieved a macro-AUC of 0.946 ± 0.004 (95% CI 0.942-0.950), and an accuracy 0.829 ± 0.029 (95% CI 0.800-0.858). DCA demonstrated a positive net benefit across clinically relevant threshold probabilities compared with treat-all and treat-none strategies. Grad-CAM visualizations highlighted diagnostically relevant regions in CT images, providing interpretable decision-making support.

Conclusions: The MambaOut Kobe model demonstrates outstanding potential for clinical application in classifying MPLC, IPM, and MBPL. Its combination of high accuracy and computational efficiency makes it a promising tool for lung cancer diagnosis and treatment planning. This automated approach could reduce diagnostic uncertainty, minimize unnecessary invasive procedures, and facilitate timely, personalized treatment decisions for patients with multiple lung lesions. Future studies should focus on validating the model on larger, multicenter datasets and enhancing its discriminatory capacity between MPLC and IPM.

Background: Patients with a tumor regression grade of 0 or 1 (CAP), are classified as well-responders after neoadjuvant chemoradiotherapy (NCRT). Although well-responders are expected to have superior prognosis, occurrences of recurrence and metastasis still exist.

Methods: Clinical data of patients from January 2017 to 2022 were analyzed.

Inclusion criteria: adenocarcinoma, cT3-4N0 or cTanyN+, receiving NCRT and surgery, CAP 0 ~ 1. The primary endpoint was three-year disease-free survival (3y-DFS). According to occurrence of DFS events, patients were divided into DFS (470, 91.8%) and non-DFS group (42, 8.2%). Cox regression analysis was applied to identify risk factors affecting prognosis of well-responders.

Results: A total of 512 well-responders were included, with mean age of 59.1 ± 10.6 years. Compared to DFS group, patients in non-DFS group had advanced mrT4 stage (33.3% vs. 18.1%, P = 0.017), higher positive rates of mesorectal fascia (52.4% vs. 35.1%, P = 0.026) and extramural vascular invasion (59.5% vs. 36.6%, P = 0.003), advanced ypT4 stage (56.2% vs. 23.8%, P < 0.001), ypN+ (23.8% vs. 9.4%, P = 0.014), and tumor deposits (14.3% vs. 3.6%, P = 0.005). The follow-up was end up to May 2024, with a duration of 36 (18 to 53) months. The tumor recurrence and metastasis rates were 0.8% (4) and 7.0% (36). The estimated 3y-DFS and overall survival in well-responders were 90.1% and 97.4%. Cox analysis identified ypT3 ~ 4 stage as independent risk factor resulting inferior DFS.

Conclusion: Well-responders are expected to have superior prognosis. Special attention should be given to patients with advanced stages or those exhibiting positive mesorectal fascia or extramural vascular invasion, or adverse pathological features.

Genomic instability is a hallmark of nearly all human cancers, yet the genetic regulators that govern this instability, their functional roles, and their therapeutic potential remain incompletely understood in breast cancer. In this study, we conducted a comprehensive analysis of the expression profiles of genomic instability-maintaining genes (GIMGs) in breast cancer and revealed that elevated expression of GIMGs was associated with adverse clinical outcomes and an altered tumor immune microenvironment. Through further analysis, we identified a core set of GIMGs and pinpointed RPGRIP1L as a critical driver of genomic instability in breast cancer. Notably, RPGRIP1L expression was significantly upregulated in breast cancer tissues and strongly correlated with poor patient prognosis. Functional studies demonstrated that RPGRIP1L knockdown reduced genomic instability in tumor cells, as evidenced by decreased Phosphorylated Histone H2AX (γH2AX) expression, and suppressed tumor growth in mouse models. Additionally, high RPGRIP1L expression was linked to elevated expression of immune checkpoint Programmed Death Ligand 1(PD-L1) in human breast cancer samples. Mechanistically, we found that RPGRIP1L promoted the activation of the Hedgehog signaling pathway, which in turn drived tumor proliferation and upregulated PD-L1 expression. Collectively, these findings highlight RPGRIP1L as a key genomic instability-maintaining gene in human breast cancer, offering critical insights into the molecular mechanisms underlying disease progression. Furthermore, targeting RPGRIP1L may represent a promising therapeutic strategy for breast cancer.

Objective: Investigate gut microbiota's role in chemotherapy resistance development among pancreatic cancer patients and evaluate a nomogram prediction model.

Methods: 510 pancreatic cancer patients who received chemotherapy in our hospital from January 2023 to December 2024 were randomly divided into training set (n = 357) and validation set (n = 153). Risk factors for chemotherapy resistance in pancreatic cancer were screened through multiple logistic regression analysis, and a Nomogram model was constructed. The predictive efficacy of the model was evaluated by drawing the receiver operating characteristic curve and calibration curve, and was verified in the validation set. The clinical application value of the model was evaluated using decision curve analysis.

Results: The incidence of resistance in the training set was 64.99% (232/357), and that in the validation set was 65.36% (100/153). Multiple logistic regression analysis showed that Escherichia coli, Enterococcus faecalis, and Fusobacterium nucleatum were independent risk factors affecting the occurrence of chemotherapy resistance in pancreatic cancer patients (P < 0.05), while Akkermansia and Bifidobacterium were independent protective factors affecting the occurrence of chemotherapy resistance in pancreatic cancer patients (all P < 0.05). The C-indexes of the constructed Nomogram model in the training set and the validation set were 0.767 and 0.762 respectively. The areas under the curve (AUC) were 0.767 (95% CI: 0.708-0.827) and 0.762 (95% CI: 0.669-0.854) respectively. The sensitivity and specificity were 0.455, 0.865 and 0.511, 0.873 respectively.

Conclusion: The Nomogram prediction model constructed based on gut microbiota change indicators has a high predictive efficacy for the occurrence of chemotherapy resistance in pancreatic cancer patients, providing a basis for clinical early screening and intervention.

Background: Exosomes, extracellular vesicles pivotal in cancer intercellular communication, encapsulate biomolecules with potential as diagnostic and prognostic biomarkers. Efficient isolation is essential for accurate molecular profiling. This study compares three exosome isolation methods-size exclusion chromatography (SEC), lectin-binding, and TIM4-binding-for proteomic and miRNA analysis of plasma exosomes in cancer.

Methods: Plasma exosomes from patients with non-small cell lung cancer (NSCLC, N = 22), castration-resistant prostate cancer (CRPC, N = 7), and healthy controls (N = 15) were analyzed. Liquid chromatography-tandem mass spectrometry profiled exosomal proteins, and small RNA sequencing identified miRNAs.

Results: SEC, lectin-binding, and TIM4-binding methods identified 122, 153, and 87 proteins, and 335, 89, and 181 miRNAs, respectively. SEC detected the most unique miRNAs (183), while lectin-binding excelled in unique protein detection (56). In CRPC, 69 proteins and 21 miRNAs differed significantly (p < 0.05) from controls, with SEC identifying 11 proteins and 6 miRNAs, lectin-binding detecting 40 proteins and 1 miRNA, and TIM4-binding revealing 18 proteins and 14 miRNAs. In NSCLC, 33 proteins and 15 miRNAs showed differential expression (p < 0.05), with SEC detecting 14 proteins and 3 miRNAs, lectin-binding identifying 2 proteins, and TIM4-binding uncovering 17 proteins and 12 miRNAs.

Conclusions: The choice of exosome isolation method profoundly influences molecular profiling, with SEC optimizing miRNA detection, lectin-binding enhancing protein capture, and TIM4-binding enriching cancer-specific miRNAs. These findings underscore the need for tailored isolation strategies to unlock exosomes' potential as precise, multi-omic biomarkers for cancer diagnosis and monitoring.

Background: Pancreatic ductal adenocarcinoma (PDAC) primarily affects older adults and has a poor prognosis. Although tools like geriatric assessments and electronic patient-reported outcomes (ePRO) can guide treatment, they are underutilized in clinical practice. This secondary analysis of the INSPIRE pilot intervention, a pilot intervention that assessed the utility of incorporating data on patient preferences and frailty into multidisciplinary tumor board (MDTB) discussions, evaluated the clinical impact of incorporating patient preferences and frailty data into MDTB discussions.

Methods: The study included patients aged ≥ 60 years with PDAC enrolled in the INSPIRE intervention at the University of Alabama at Birmingham. Patients discussed at MDTBs with adequate medical records who did not forgo treatment initially were included. A control group comprised patients who completed preference and frailty surveys but whose ePRO data were not presented at MDTBs. Outcomes analyzed included treatment consistency with preferred National Comprehensive Cancer Network (NCCN) regimens based on fitness, unplanned treatment modifications, and healthcare utilization within six months of treatment initiation. Data were extracted from medical records and statistical analysis employed log-rank tests from cumulative incidence functions.

Results: Among 121 patients (24 intervention, 97 controls), the median age was 70 years. Compared to controls, the intervention group had fewer comorbidities (8% vs. 25% with no comorbidities, V = 0.22), a higher proportion of non-White patients (42% vs. 25%, V = 0.15), more resectable disease (48% vs. 35%, V = 0.14), and higher frailty rates (42% vs. 31%, V = 0.11). Intervention patients showed slightly higher consistency with NCCN preferred regimens based on fitness (63% vs. 60%, V = 0.02), fewer unplanned treatment modifications (54% vs. 68%, V = 0.12), and lower hospital admissions (33% vs. 50%, V = 0.13).

Conclusion: The INSPIRE intervention demonstrated promising signals when aligning treatment regimens with patient capabilities and preferences, potentially reducing unplanned treatment modifications and hospital admissions. Larger studies are needed to confirm these exploratory results and assess broader applicability.

Background: Numerous studies in the field of cancer indicate that long non-coding RNAs (lncRNAs) play a significant role in the development and malignancy of various cancers. In this study, the expression changes of TMEM105 in gastric cancer (GC) and its association with malignancy were examined.

Methods: In the in silico section, TMEM105 expression in GC and its correlation with patient prognosis were analyzed using publicly available datasets. Using WGCNA analysis on TCGA data, modules and potential pathways associated with TMEM105 were identified. Expression changes of TMEM105 in GC tissues compared with adjacent healthy tissues were analyzed by RT-qPCR. The expression level of TMEM105 in GC cell lines, including AGS and MKN-45, was modulated using siRNA. The relationship between TMEM105 expression and malignant characteristics, including cell proliferation, migration, and colony formation, was examined in these cell lines.

Results: TMEM105 was significantly upregulated in public datasets, including TCGA, GSE19826, and GSE54129. TMEM105 expression was associated with advanced disease stage and poor prognosis. WGCNA analysis revealed that TMEM105 expression clustered with genes related to cell proliferation, such as E2F target genes, within the same module. Silencing TMEM105 reduced the viability of GC cell lines and decreased mRNA expression of proliferation-related genes, including E2F1, Cyclin D1, and Ki-67. Silencing TMEM105 markedly reduced migration rates in GC cells. Moreover, TMEM105 expression affected the colony-forming ability of these cell lines.

Conclusion: High TMEM105 expression in GC is associated with poorer patient outcomes. It appears to influence cellular proliferation and contribute to the development and malignancy of gastric carcinoma. Consequently, TMEM105 could be a valuable therapeutic and prognostic target.

Background: Pancreatic neuroendocrine tumors (pNETs) are rare malignant tumors. While surgical resection is the primary therapeutic approach, postoperative survival outcomes remain suboptimal. This study aims to investigate the prognostic value of inflammation and nutrition-related indicators, such as Prognostic nutritional index (PNI), Neutrophil-lymphocyte ratio (NLR), Platelet-lymphocyte ratio (PLR), Systemic immune-inflammation index (SII), and Lymphocyte-monocyte ratio (LMR), in patients with pNETs undergoing surgery. Furthermore, we sought to develop and validate a prognostic nomogram incorporating independent predictors of postoperative overall survival (OS).

Methods: A retrospective analysis was conducted on the clinical data from 102 patients with pathologically confirmed pNETs. Univariate and multivariate Cox regression analyses were performed to identify prognostic factors. A nomogram was developed based on independent risk factors, and its predictive performance was internally validated.

Results: The analysis revealed that low preoperative levels of PNI, low NLR, low LMR, low SII, and low PLR were significantly associated with reduced postoperative survival. Multivariate Cox regression analysis identified PNI (HR = 0.153; 95% CI, 0.033-0.720), NLR (HR = 0.153; 95% CI, 0.037-0.633), and AJCC stage (HR = 20.126; 95% CI, 5.036-80.422) as independent prognostic factors. The developed nomogram demonstrated excellent discriminative ability, with time-dependent AUC values of 0.878 (3-year), 0.860 (5-year), and 0.830 (10-year), along with a C-index of 0.824 (95% CI 0.748, 0.901) upon internal validation. Risk stratification using the nomogram effectively distinguished low- and high-risk groups, and its validity was confirmed using Kaplan-Meier survival curves.

Conclusion: Despite limitations inherent to its retrospective, single-center design and small sample size, this study establishes a novel nomogram integrating inflammatory-nutritional biomarkers with clinicopathological staging. This tool provides clinicians with a personalized prognostic assessment framework to guide postoperative management strategies for pNETs patients.