Oxygen sensing signaling is mainly mediated by EGLN prolyl hydroxylases, which plays important roles in physiology and diseases. Among them, EGLN2 plays an oncogenic role in breast cancer. However, the development of isoform-selective EGLN2 inhibitors remains a major challenge. Here, we report the identification of A10P3C, a small-molecule inhibitor that exhibits high potency and exceptional selectivity toward EGLN2. Guided by conformational sampling, ensemble-docking-based functional-group-residue contact analysis and simulation-based binding selectivity design, A10P3C was designed and strategically improved from an initially screened lead compound to selectively inhibit EGLN2. Unlike previously reported pan-hydroxylase inhibitors, A10P3C achieves potent isoform selectivity. Functionally, A10P3C effectively suppresses breast cancer cell proliferation, migration, and invasion, while inducing G1 cell cycle arrest and cellular senescence through on-target EGLN2 inhibition. These on-target effects are corroborated by transcriptomic profiling, which shows significant modulation of cell cycle and senescence pathways. Notably, A10P3C exhibits robust antitumor efficacy in patient-derived organoids and xenograft models, accompanied by favorable metabolic stability and minimal systemic toxicity. Collectively, these findings identify A10P3C as a selective chemical probe for EGLN2 biology and a promising therapeutic lead for precision treatment of EGLN2-driven breast cancers.

Complications occurring early after robot-assisted radical prostatectomy (RARP) can hinder postoperative recovery. Existing risk instruments, however, generally incorporate few predictors and assume linear associations. We aimed to derive and externally validate an interpretable machine-learning model that combines clinical, laboratory, and intraoperative information and to translate the model into an online risk calculator. Records were reviewed for 1,254 patients with prostate cancer (PCa) who received RARP at the First, Third, or Fifth Medical Center of the Chinese PLA General Hospital. Among them, 1,001 patients from the First and Fifth centers were randomly allocated 7:3 to training (n = 701) and internal testing (n = 300), while 253 patients from the Third center provided an independent external cohort. Sixty-four candidate variables underwent univariate screening and least absolute shrinkage and selection operator regression. The tested algorithms were gradient boosting decision trees (GBDT), random forest, logistic regression, support vector machine, AdaBoost, Gaussian naive Bayes, and multilayer perceptron. We examined discrimination with the area under the receiver operating characteristic curve (AUC), agreement between predicted and observed risk with calibration plots, and clinical usefulness with decision curve analysis (DCA). SHapley Additive exPlanations (SHAP) were used to interpret predictions. The leading model was also benchmarked against the Comorbidity Score for Robotic Surgery (CRS). Feature selection yielded eight predictors for model training. For the multilayer perceptron (MLP), the AUC was 0.880 on internal testing and 0.843 in the independent external cohort. Across both evaluations, this algorithm offered the most balanced combination of performance measures, good calibration, and favorable clinical net benefit. SHAP ranked crystalloid infusion as the leading predictor; operative time and thrombin time (TT) ranked next. In an exploratory comparison in the external cohort, MLP showed greater potential than CRS in discrimination, calibration, and net benefit. We visualized the final model and deployed it as a web-based calculator. The interpretable MLP model accurately estimated the likelihood of early postoperative complications among patients with PCa undergoing RARP. Its accompanying online calculator may assist clinicians in recognizing high-risk patients as soon as surgery is completed and may inform individualized risk classification, postoperative monitoring, and clinical care.

Chemotherapy-induced toxicities represent a major clinical challenge in gastric cancer treatment, with limited effective management strategies. Emerging evidence suggests that gut microbiota dysbiosis is a key driver of these adverse effects, while traditional Chinese medicine (TCM) shows potential in modulating gut microbial composition. This review aims to systematically summarize the molecular mechanisms and clinical evidence of TCM in alleviating chemotherapy-induced toxicities in gastric cancer through gut microbiota regulation and to identify current research gaps.

A systematic literature search was conducted using SCI-indexed studies, primarily published in recent years, with inclusion of foundational studies where necessary. The review focused on chemotherapy-induced gut microbiota dysbiosis and its role in toxicity, as well as the effects of TCM-including polysaccharides, flavonoids, and herbal formulations-on microbial composition, metabolic products such as short-chain fatty acids (SCFA), LPS, and inflammatory signaling pathways.

Gut microbiota dysbiosis contributes to systemic inflammation primarily through activation of the LPS/TLR4/NF-κB/NLRP3 inflammasome pathway and reduced SCFA-mediated histone deacetylase inhibition. TCM regulates gut microbiota through a "structure-metabolite-immune" framework, restoring microbial diversity, increasing SCFA production, reducing LPS levels, and suppressing inflammatory signaling. Clinically, oral and injectable TCM formulations such as SFI and HCS have shown moderate- to high-quality evidence in improving immune function, alleviating myelosuppression, and reducing gastrointestinal toxicities, whereas evidence for external therapies remains limited.

TCM represents a promising microbiota-targeted strategy for mitigating chemotherapy-induced toxicities in gastric cancer. Future studies should focus on STING signaling and the gut-brain axis, conduct large-scale randomized controlled trials (RCTs), and develop microbiota-based personalized TCM interventions to advance evidence-based integrative oncology.

Pancreatic ductal adenocarcinoma (PDAC) is a fast-evolving malignancy, with resection being the only chance of cure. We aim to determine which imaging technique provides the best diagnostic performance in detecting and assessing the spread of PDAC.

Our study was registered on PROSPERO (CRD42024567008). On July 9, 2024, a systematic search was conducted in three databases, focusing on direct comparative studies that reported on patients suspected of PDAC who underwent at least two imaging methods. When available, the reference test was represented by pathological examinations, and the diagnostic performance of each diagnostic method was assessed. A bivariate model was used to calculate the pooled sensitivity and specificity, along with 95% confidence intervals (CI) for the meta-analysis.

Forty direct comparative studies were eligible for meta-analysis. CT, MRI, and endoscopic ultrasound (EUS) had no significant differences in diagnostic performance for detecting PDAC. MRI showed higher sensitivity [0.89 [0.67-0.97]) and specificity (0.91 [0.76-0.97]) compared to CT (sensitivity 0.60 (CI: 0.44-0.74) and specificity 0.84 [CI: 0.70-0.93]) in distinguishing liver metastases. CT outperformed MRI for vascular invasion assessment. For lymph node detection, EUS proved to have higher sensitivity than CT or FDG-PET/CT. CT had a higher specificity (0.88 [CI: 0.76-0.94]) compared to MRI (0.63 [CI: 0.48-0.76]) when assessing resectability.

MRI has a higher pooled sensitivity and specificity for detecting liver metastases of PDAC compared to CT, while the specificity of CT proves to be higher than that of MRI for assessing resectability.

QuestionGiven the pancreatic cancer's tight therapeutic window and the need for precise diagnosis, we aim to determine how to best assess it with imaging techniques. FindingsMRI has a significantly higher pooled sensitivity for liver metastases detection than CT. CT has a significantly higher pooled sensitivity for vascular invasion and specificity for resectability assessment than MRI. Clinical relevanceA liver MRI may provide additional value before potentially curative surgery to rule out small metastases. CT should be performed to search for additional metastatic sites and exclude non-resectable cases. Both imaging techniques provide valuable information for assessing pancreatic cancer.

This study aimed to characterize long-term intracranial control, recurrence patterns, radiation necrosis (RN) risk, and treatment intensity among patients with ≥ 24 months of follow-up after stereotactic cranial radiotherapy (SRT).

We conducted a retrospective review of patients treated with SRT at a single institution from 2018 to 2023. Eligible patients received at least one course of SRT and had ≥ 104 weeks of follow-up. Clinical, treatment, and radiologic data were collected as well as in-field and out-of-field failure, RN incidence and management, survival, and treatment patterns over time.

Fifty-one patients met the inclusion criteria, contributing 130 SRT courses treating 241 lesions. Median follow-up was 33.8 months. Most patients (51%) had non-small cell lung cancer, and 80% received systemic therapy after initial SRT. In-field failure occurred in 16.2% of lesions and 47.1% of patients, with higher failure rates in resection cavities than in intact metastases. Out-of-field failure occurred in 47.1% of patients, most commonly managed with additional SRT. The 2‑year in-field and out-of-field recurrence-free rates were 47.7% and 56.6%, respectively. Radiation necrosis developed in 21 lesions (8.7%), with substantially higher incidence in resection cavities (29.4%) and re-irradiated targets. Median time to RN was 11.2 months, and 38% required surgical intervention. Five-year overall survival was 67.8%.

Among long-term survivors treated with SRT, both intracranial recurrence and RN remain clinically significant. As patient survival improves, individualized surveillance, careful consideration of cumulative radiation exposure, and optimized re-irradiation strategies are essential to minimize toxicity while maintaining durable intracranial control.

Pyroptosis is an inflammatory form of programmed cell death characterized by rapid plasma membrane rupture and the release of intracellular mediators, which distinguishes it from apoptosis and necrosis. Pyroptosis plays important roles in infection, autoimmunity, and cancer, highlighting its significance in immune homeostasis. Gasdermin E (GSDME), a key executor of pyroptosis, has been increasingly implicated in cancer. This study examined the expression and prognostic significance of GSDME in stomach adenocarcinoma (STAD).

GSDME was significantly upregulated in STAD tissue and associated with advanced stage and poor prognosis according to data from The Cancer Genome Atlas (TCGA) and the Xiantao Academic Online tool. Gene set enrichment analysis (GSEA) revealed that high GSDME expression was associated with genes enriched in immunosuppressive-related pathways in the TCGA-STAD cohort stratified by GSDME expression. Immune analysis revealed that GSDME expression was positively correlated with immune cell infiltration and the expression of immune checkpoint molecules, lymphocyte markers, immunomodulators, and major histocompatibility complex (MHC) molecules, as determined using R software and the Tumor-Immune System Interactions and Drug Bank (TISIDB). Western blotting and immunohistochemistry confirmed the elevated expression of GSDME in gastric cancer (GC) tissues. Functional analysis revealed that GSDME knockdown inhibited the proliferation and migration of AGS and MKN-45 GC cells while promoting cell death.

These findings suggest that GSDME may serve as a prognostic biomarker in GC and as a potential immune regulator.

Plants are important sources of bioactive compounds with demonstrated anticancer and cancer-preventive properties. However, traditional methods for discovering and deciphering the mechanisms of these compounds are often slow and labor-intensive. The integration of machine learning (ML) provides a computational approach for bioactivity prediction, candidate prioritization, and mechanistic hypothesis generation. Algorithms including support vector machines (SVM) and graph neural networks (GNN) are increasingly employed to support multiple stages from compound screening to experimentally guided mechanistic investigation. Despite these advances, challenges such as data heterogeneity and structural complexity of natural products remain. This review systematically outlines the natural anticancer compounds derived from plants, summarizes the application of machine learning in bioactivity prediction, candidate prioritization, and mechanistic investigation, compares functional-group prevalence and co-occurrence patterns in plant-derived anticancer compounds and FDA-approved anticancer drugs, and identifies current challenges along with future directions for AI-assisted discovery in this field.

Recent advancements in our understanding of the molecular underpinnings of acute myeloid leukemia (AML) disease biology have translated into breakthroughs in clinical management. In this review, we summarize the latest data incorporating precision approaches to treatment of FLT3-mutated AML with a focus on augmenting the curative role allogeneic hematopoietic cell transplantation (alloHCT) plays in the current treatment paradigm.

Incorporation of the next-generation FLT3 inhibitors quizartinib and gilteritinib into upfront therapy, whether in combination with intensive induction chemotherapy or less intensive hypomethylating agent/venetoclax combinations, have led to deeper and more durable remissions and resulted in greater transplant eligibility. AlloHCT candidacy has also expanded due to increased use of alternative donors, made possible with post-transplant cyclophosphamide-containing graft-versus-host disease prophylaxis regimens, and more tolerable reduced intensity conditioning. Additionally, risk stratification in the peri-transplant setting using ultra-sensitive FLT3-ITD measurable residual disease (MRD) assays has refined post-transplant maintenance approaches allowing escalation or de-escalation of therapy guided by each patient's risk of relapse. These advances have resulted in improved clinical outcomes for this historically high-risk disease. AlloHCT remains an integral component of the treatment of FLT3-mutated AML. Novel FLT3 inhibitors used in frontline combination regimens have led to more numerous and deeper remissions, and as post-transplant maintenance they have shown success in preventing the early relapse characteristic of this disease. MRD has emerged as an integral component of FLT3-ITD AML risk stratification and surveillance and is likely to further revolutionize our current treatment approaches.

Ovarian cancer remains one of the most lethal gynecological malignancies worldwide due to late-stage diagnosis, poor prognosis, and frequent chemoresistance. The limited sensitivity and specificity of current diagnostic approaches, particularly for early-stage disease, highlight the urgent need for improved biomarkers and diagnostic strategies. Nowadays, ovarian cancer diagnostics are based mainly on the established biomarkers CA125 and HE4, which in some countries most commonly combined in the Risk of Ovarian Malignancy Algorithm (ROMA). Despite their widespread clinical use, these biomarkers are of limited specificity and sensitivity, particularly in early-stage disease, underscoring the need for more efficient screening and diagnostic approaches. This review summarizes current evidence on the molecular mechanisms linking oxidative stress, inflammation, and platelet activation with ovarian cancer progression and discusses the diagnostic and prognostic potential of related biomarkers. Oxidative stress and inflammation contribute to dysregulated signaling pathways and tumor progression, providing a mechanistic basis for the identification of novel biomarkers. Increasing evidence indicates that platelet activation contributes to tumor growth, angiogenesis, and metastasis. This justifies the strategy of exploring platelet-derived mediators, including thromboxane B₂ and other eicosanoids, as potential biomarkers of ovarian cancer. Overall, current evidence indicates that oxidative stress-, inflammation-, and platelet-associated molecular pathways represent promising sources of novel biomarkers and potential therapeutic targets in ovarian cancer.