Ferroptosis exerts a remarkable influence on the progression of non-small cell lung cancer (NSCLC). Although the contribution of lipocalin 2 (LCN2) to NSCLC pathogenesis has been demonstrated, further elucidation of the molecular determinants driving LCN2 dysregulation is essential for developing NSCLC interventions. mRNA levels were performed by quantitative polymerase chain reaction (PCR), and protein expression was assessed by immunoblotting and immunohistochemical assays. Reactive oxygen species (ROS), glutathione (GSH), malondialdehyde (MDA), and Fe2 + levels were measured to analyze cell ferroptosis. The relationship between zinc-finger E-box-binding homeobox 1 (ZEB1) and the LCN2 promoter was verified by chromatin immunoprecipitation (ChIP) and luciferase assays. Our data showed that LCN2 and ZEB1 levels were upregulated in NSCLC. LCN2 depletion reduced the growth, motility, and invasiveness of NSCLC cells, while promoting apoptosis and ferroptosis. LCN2 depletion also inhibited the tube formation of human umbilical vein endothelial cells (HUVECs) in vitro. Mechanistically, ZEB1 enhanced the transcription and expression of LCN2 in NSCLC cells. ZEB1 downregulation diminished HUVEC tube formation, suppressed the growth, motility, and invasiveness of NSCLC cells, and enhanced apoptosis and ferroptosis. Notably, these effects were counteracted by re-expression of LCN2. Additionally, ZEB1 downregulation inhibited the growth of xenograft tumors in vivo. Our study demonstrates the pro-tumorigenic role of the ZEB1/LCN2 cascade in NSCLC by promoting malignant progression and impeding ferroptosis. Such molecular insights may help devise novel candidates for NSCLC treatment.
Objective: Tumor microenvironment composition significantly influences tumor progression. This study aimed to explore the distribution of M2 tumor-associated macrophages (TAMs), reticular fibers (RFs), and collagen fibers (CFs) within the tumor microenvironment of pulmonary sarcomatoid carcinoma (PSC) and assess their clinicopathological significance.
Methods: Formalin-fixed paraffin tissue sections of 127 PSC patients from two medical centers were collected and analyzed by immunohistochemistry and the Gomori method. HALO software was used to analyze the distributions of M2TAMs, RFs, and CFs, and statistically analyzed for clinicopathological significance.
Results: Kaplan-Meier analysis showed that overall survival (OS) was longer in patients with low density of M2TAMs (P = 0.038) and high density of CFs (P = 0.046) and RFs (P = 0.010). Patients classified within the low-risk group, based on the combined factors MR and MC, experienced significantly longer OS than those in the high-risk group. Multivariate analysis identified the densities of M2TAMs, RFs, and CFs, along with MC and MR, as independent prognostic factors for patient OS. Nomogram models 1 and 2, with C-indices of 0.74 and 0.73, respectively, were highly effective in predicting OS. Decision curve analysis demonstrated that the Nomogram model outperformed pTNM staging in predicting medium- and long-term survival.
Conclusion: High densities of M2TAMs and low densities of RFs and CFs are associated with poor prognosis in PSC patients and are independent prognostic factors. The Nomogram model proved was more effective than pTNM staging in predicting medium- and long-term survival, offering a valuable tool for the individualized clinical treatment of PSC patients.
Introduction: The application of general-purpose large language models (LLMs) in cytopathology remains largely unexplored. This study aims to evaluate the accuracy and consistency of a custom version of ChatGPT-4 (GPT), ChatGPT o3, and Gemini 2.5 Pro as diagnostic support tools for cervical cytology.
Materials and methods: A total of 200 Papanicolaou-stained cervical cytology images were acquired at 40x magnification, each measuring 384 × 384 pixels. These images consisted of 100 cases classified as negative for intraepithelial lesion or malignancy (NILM) and 100 cases across various abnormal categories: 20 low-grade squamous intraepithelial lesion (LSIL), 20 high-grade squamous intraepithelial lesion (HSIL), 20 squamous cell carcinoma (SCC), 20 adenocarcinoma in situ (AIS), and 20 adenocarcinoma (ADC). Diagnostic accuracy and consistency were evaluated by submitting each image to a GPT, ChatGPT o3, and Gemini 2.5 Pro 5-10 times.
Results: When distinguishing normal from abnormal cytology, LLMs showed mean sensitivity between 85.4 % and 100 %, and specificity between 67.2 % and 92.7 %. ChatGPT o3 was more accurate in identifying NILM (mean 89.2 % vs. 67.2 %) but less accurate in detecting LSIL (34 % vs. 85 %), HSIL (6 % vs. 63 %), and ADC (28 % vs. 91 %). Chain-of-thought prompting and submitting multiple images of the same diagnosis to ChatGPT o3 and Gemini 2.5 Pro did not significantly improve accuracy. Both models also performed poorly in identifying cervicovaginal infections.
Conclusions: ChatGPT o3 and Gemini 2.5 Pro demonstrated complementary strengths in cervical cytology. Due to their low accuracy and inconsistency in abnormal cytology, general-purpose LLMs are not recommended as diagnostic support tools in cervical cytology.
Purpose: To evaluate the clinical, pathological, and prognostic significance of HER2-low and HER2-ultralow expression in breast cancer.
Methods: In this retrospective study, 171 HER2-negative breast cancer cases were reclassified by immunohistochemistry as HER2-low, HER2-ultralow, or HER2 IHC 0. Clinicopathological variables, Ki-67 index, hormone receptor (HR) status, and treatment data were analyzed. Welch's ANOVA, Kaplan-Meier survival with a 60-month landmark, and Cox regression were applied.
Results: The cohort included 49.1 % HER2-low, 27.5 % HER2-ultralow, and 23.4 % HER2 IHC 0 cases. No significant differences were found in clinical stage, histologic grade, or treatment distribution. Ki-67 index was lower in HER2 IHC 0 tumors (p = 0.047). Overall survival (OS) differed significantly: HER2-ultralow patients had the best outcomes (median OS: 47.5 months), followed by HER2 IHC 0 (37.5) and HER2-low (33.0) (log-rank p < 0.05). In multivariable Cox analysis, HER2-low status was independently associated with worse OS versus HER2-ultralow (HR = 1.86; p = 0.006), while HER2 IHC 0 showed no significant difference (HR = 1.42; p = 0.212). Higher Ki-67 and negative ER status were also linked to worse prognosis. Subgroup analysis revealed the poorest outcomes in HER2-low/HR-negative tumors.
Conclusion: HER2-ultralow tumors were associated with better survival than HER2-low, suggesting clinical relevance in further subclassifying HER2-negative breast cancers. Although the retrospective design and minor proportional hazard violations warrant caution, these findings may guide treatment for more individualized approaches. Future multicenter studies, supported by molecular profiling and AI-assisted HER2 scoring, are essential to validate these findings and improve reproducibility in low-level HER2 assessment.

