Predicting clinical and radiological outcomes of epidermal growth factor receptor tyrosine kinase inhibitors (EGFR-TKIs) in patients with non-small cell lung cancer (NSCLC) and brain metastases (BMs) is crucial for effective patient management. Machine learning (ML)-based models have increasingly been utilized to predict EGFR-TKI response in patients with lung cancer brain metastasis (LCBM). In this study, we aimed to evaluate the predictive performance of ML-based models for EGFR-TKI response prediction.

A comprehensive literature search was conducted using PubMed, Embase, Scopus, and Web of Science from database inception to April 25, 2025. Studies that developed ML-based models to predict EGFR-TKI response were included.

Eight studies involving 1322 LCBM patients were included. The included studies used logistic regression (LR), LR with least absolute shrinkage and selection operator (LASSO), decision tree (DT), and a Cox-based deep learning model (DL-Cox). The meta-analysis revealed a pooled area under the curve (AUC) of 0.84 (95% CI 0.78-0.91) and accuracy (ACC) of 0.75 (95% CI 0.62-0.88) with a sensitivity (SEN) of 0.82 (95% CI 0.77-0.87) and a specificity (SPE) of 0.73 (95% CI 0.66-0.80) for prediction of EGFR-TKI response. The meta-analysis of diagnostic odds ratios (DOR) exhibited a pooled DOR of 12.41 (95% CI 7.32-21.04).

ML-based models show promising ability to predict EGFR-TKI response in LCBM, supporting their potential to guide treatment selection. However, their use in clinical practice remains limited by small retrospective datasets and lack of external validation.

Richter transformation (RT), the progression of chronic lymphocytic leukemia (CLL) to aggressive lymphomas, poses a significant therapeutic challenge with historically poor outcomes. Chemoimmunotherapy (CIT) regimens have demonstrated limited efficacy with short durations of response. This review aims to evaluate the evolving treatment landscape for RT, with a focus on recent advances in targeted therapies, immunotherapies, and cellular therapies that are redefining the current and future standards of care.

The treatment paradigm for RT is rapidly shifting away from cytotoxic chemotherapy. The combination of the B-cell lymphoma 2 inhibitor venetoclax with CIT has emerged as a new first-line benchmark with promising response rates and overall survival. Covalent Bruton tyrosine kinase (BTK) inhibitors had modest activity as monotherapy but showed improved responses when given with an immune checkpoint inhibitor. Pirtobrutinib has demonstrated responses even in heavily pretreated patients. Furthermore, advancement in immunotherapy has expanded treatment options for this patient population with bispecific T-cell engagers achieving high response rates and chimeric antigen receptor T-cell therapy providing deep, durable responses and favorable median overall survival in the relapsed/refractory (R/R) setting. The therapeutic landscape for RT has broadened with the introduction of targeted agents and immunotherapy. Venetoclax-based regimens represent a new standard for chemotherapy-eligible patients, allowing for a more effective bridge to potentially curative consolidation with transplantation. For R/R disease, novel BTK inhibitors, bispecific antibodies, and cellular therapies are demonstrating substantial efficacy. Ongoing trials investigating combinations of these agents are poised to further transform RT management.

In our study, we identified a novel, ruxolitinib-sensitive, CCDC6::JAK2 fusion gene as a driver of atypical JAK2-unmutated MPN with a polycythemic phenotype. The CCDC6::JAK2 chimeric protein retains the CCDC6 coiled-coil domain and the JAK2 kinase domain. Dimerization of chimeric proteins through coiled-coil domains promotes JAK2 autophosphorylation leading to constitutive activation of the JAK/STAT signaling pathway.

This study aimed to investigate the role of minichromosome maintenance complex component 6 (MCM6), a DNA replication licensing factor, in retinoblastoma progression and its impact on melphalan chemosensitivity.

MCM6 expression patterns were analyzed using single-cell RNA sequencing (scRNA-seq) of retinoblastoma and validated in patient tumors, including specimens obtained after failed melphalan therapy. Stable MCM6 knockdown cell lines were established for proliferation and cell-cycle assays, DNA damage analyses, and chemosensitivity testing. In vivo xenograft models were employed to evaluate the therapeutic efficacy of MCM6 knockdown combined with melphalan.

The scRNA-seq revealed that MCM6 was highly expressed in retinoblastoma cells and embedded in a proliferation-associated gene network. Elevated expression was also confirmed in human retinoblastoma, particularly in tumors from patients with failed melphalan therapy. MCM6 knockdown suppressed cell proliferation and cell-cycle progression while enhancing melphalan-induced DNA damage, thereby sensitizing retinoblastoma cells to melphalan. In vivo, MCM6 depletion synergized with melphalan to significantly inhibit intraocular tumor growth.

MCM6 acts as a critical regulator of retinoblastoma growth and modulates response to melphalan. Targeting MCM6 may offer a therapeutic approach to improve outcomes of chemotherapy in retinoblastoma.

Colorectal cancer (CRC) is among the most prevalent malignancies, with an increasing incidence in China, posing increasing challenges to current screening strategies. This study aimed to evaluate the feasibility of a screening strategy based on the stool-based methylated Syndecan-2 gene (mSDC2) test for detecting advanced colorectal neoplasia in the Chinese population.

In this multicenter randomized study, participants aged 45-80 years and identified through community screening were randomized to take either the mSDC2 test or a fecal immunochemical test (FIT), and those with positive results were referred for colonoscopy between 2020 and 2022. The primary outcome was the detection rate of advanced colorectal neoplasia, including advanced adenoma, advanced serrated polyps, and CRC. The secondary outcome was adherence to colonoscopy. Univariable and multivariable logistic regression analyses were conducted to evaluate colonoscopy adherence among participants with positive stool-based test results.

A total of 44,475 eligible participants were enrolled from 189 communities (93 in the mSDC2 test group and 96 in the FIT group). In the mSDC2 test group (n = 21,854), the detection rate of advanced neoplasia was 0.53% (n = 115), including 93 (0.43%) advanced adenomas, 8 (0.04%) advanced serrated polyps, and 17 (0.08%) CRC cases. In the FIT group (n = 22,621), the detection rate of advanced neoplasia was 0.47% (n = 106), including 71 (0.31%) advanced adenomas, 9 (0.04%) advanced serrated polyps, and 29 (0.13%) CRC cases. In addition, adherence to colonoscopy was greater in the mSDC2 test group (35.6% vs. 26.4%; odds ratio: 1.59, 95% confidence interval: 1.36-1.87; P <0.001).

This study revealed the effectiveness of the mSDC2 test and FIT in detecting advanced colorectal neoplasia in a large Chinese population in a real-world setting, and indicated a possible improvement in adherence to colonoscopy for the mSDC2 test. Further studies with improved comparability and longer follow-up are warranted to clarify its clinical utility.

ClinicalTrials.gov, No. NCT04221854.

Liver hepatocellular carcinoma (LIHC), a lethal malignancy with rising global incidence, urgently requires robust biomarkers for early diagnosis. Although dysregulated RNA-binding proteins contribute to oncogenesis, the role of UTP3 in LIHC prognosis, epigenetic modulation, and immune regulation has not been fully investigated. We conducted an integrative multi-omics analysis of TCGA and GTEx datasets to characterize UTP3 expression across cancers using DESeq. 2. Protein-level alterations were validated with CPTAC and HPA data. Clinical associations were evaluated via Cox regression and MethSurv for methylation analysis. The tumor immune microenvironment was assessed using ssGSEA and TIMER algorithms. CeRNA networks were built with RNA22, miRmap, and starBase tools, and functional studies were performed in Huh7 and HCCLM3 cell lines. Elevated UTP3 expression independently predicted unfavorable overall survival (HR = 1.574, p = 0.007) and was associated with advanced TNM stages (p < 0.05). Mechanistically, hypomethylation at cg26775961 (p = 0.011) promoted UTP3 upregulation, whereas the SNHG15/AL024498.1-miR-483-3p axis regulated its posttranscriptional levels (p < 0.001). Tumors with high UTP3 showed immunosuppressive features, including reduced cytotoxic cells infiltration (p = 0.008) and increased PD-L1 expression (p < 0.001). UTP3-interacting proteins (GRSF1 and NFKB1) enriched in NF-κB signaling (NES = 2.1, FDR = 0.03). A nomogram centered on UTP3 yielded a 1-year AUC of 0.693. This multi-omics study establishes UTP3 as a key regulator connecting epigenetic alterations, immune suppression, and tumor progression in LIHC.

PurposeAlthough human lysine oxidase-like 3 (LOXL3) is associated with various cancers, its role in pleural mesothelioma (PM) remains uncharacterized. This study investigated the expression level and prognostic association of LOXL3 in PM.MethodsTissue specimens were collected from patients with PM. The expression levels of LOXL3 were assessed using immunohistochemistry, Western blot analysis, and quantitative reverse transcription PCR. The clinical correlation analysis was conducted using R software (version 3.6.3), incorporating data from both The Cancer Genome Atlas and Chuxiong cohorts. Univariate and multivariate Cox proportional hazards regression models alongside Kaplan-Meier survival curve analysis were performed to evaluate prognostic significance. Additionally, gene expression correlation studies between LOXL3 and other members of the LOX family were performed using the Gene Expression Profiling Interactive Analysis platform. Finally, Gene Set Enrichment Analysis was conducted to identify the signaling pathways associated with LOXL3.ResultsLOXL3 exhibited significant upregulation in both sarcomatoid and biphasic PM subtypes compared to the control samples. Clinico-pathological analysis revealed the correlations between LOXL3 expression levels and cancer type, and Wilms tumor protein 1 (WT-1) status. Cox regression analysis identified cancer type as an independent prognostic factor. Kaplan-Meier analysis demonstrated obviously poorer survival rates in cohorts with high LOXL3 expression. Notably, coordinated expression patterns were observed between LOXL3 and LOXL4. The protein expression level of LOXL3 exhibits a positive correlation with CD68, CD206, and programmed death-ligand 1 (PD-L1), with this correlation being particularly pronounced in sarcomatoid mesothelioma. Functional enrichment analysis indicated that high LOXL3 expression was primarily associated with pathways related to oxidative phosphorylation, late and early estrogen response, and adipogenesis.ConclusionLOXL3 is highly expressed in PM and associated with poor prognosis, and is involved in tumor immune evasion. The expression level of LOXL3 is correlated with cancer types and the expression level of WT-1. Cancer type is an independent prognostic factor for PM. LOXL3 expression is positively associated with LOXL4, and high LOXL3 expression is enriched in oxidative phosphorylation, estrogen response, and adipogenesis pathways, while the low-expression group is enriched in apoptosis, interleukin-2/signal transducer and activator of transcription 5, mammalian target of rapamycin complex 1, and transforming growth factor-β pathways. CD68, CD206, PD-L1, and LOXL3 may collaboratively contribute to the regulation of the PM microenvironment and are closely linked to the invasion and metastasis of PM. Therefore, LOXL3 can be used as both a prognostic marker and a potential therapeutic target for PM.

Non-small-cell lung cancer (NSCLC) is the leading cause of cancer-related mortality worldwide. Although radiotherapy (RT) is used to treat over half of NSCLC patients, about 30% have inherent or acquired radioresistance leading to treatment failure. There's a clinically unmet need to investigate mechanisms of radioresistance that can be targeted in combination with RT. Among these, HMGB1 has been shown to play a key role in tumor progression. Our research investigates TLR4, a receptor for HMGB1, highly expressed in NSCLC tissues, as a mediator of radioresistance.

The TLR4 inhibitor, TAK242, was tested in NSCLC cell lines (murine: LLCI, KLN205; human: H1975, SW900). Cells were irradiated at 2 and 10 Gy. In vivo, KLN205 cells were implanted in DBA/2 mice and tumors were irradiated at 10Gy. Gene and protein expression of TLR4 and MyD88 were assessed in vitro and in vivo. HMGB1 secretion was quantified after RT. Clonogenic assays were performed to evaluate the effect of TAK242 on radiosensitivity in vitro. The combination of TAK242 and RT was investigated in vivo in mice bearing KLN205 tumors.

TAK242 significantly decreased NSCLC cell proliferation and migration. Radiation at 2 and 10 Gy increased TLR4 gene expression in vitro and in vivo in a dose-dependent manner. In vitro, TLR4 and HMGB1 protein expression was upregulated following radiation. TAK242 in combination with radiation enhanced radiosensitivity in vitro. TAK242 decreased the percentage of cells in the G1 phase, coupled with an increase in late S and G2/M, suggesting radiosensitization via cell cycle modulation. In vivo, the combination of RT and TAK242 significantly reduced growth of KLN205 tumors.

These findings show that TLR4 inhibition enhances RT sensitivity in NSCLC.

The JAK2-V617F mutation is the most common genetic alteration in myeloproliferative neoplasms (MPN), which can progress to secondary acute myeloid leukemia (sAML), a chemotherapy-resistant disease with limited treatment options and a poor prognosis. Although the JAK1/2 inhibitor Ruxolitinib is clinically approved, its efficacy is limited by toxicity to normal cells and the development of drug resistance. Here, the deSUMOylase DESI2 is identified as a novel component of the JAK2-V617F complex by mass spectrometry-based proteomics. Mechanistically, DESI2 selectively binds to and stabilizes JAK2-V617F by mediating its deSUMOylation and deubiquitination at lysine 962 (K962). Importantly, DESI2 protein is specifically and highly expressed in JAK2-mutant-driven cell lines and MPN primary clinical samples, suggesting its potential role in JAK2-V617F regulation and disease progression. Genetic depletion of DESI2 suppresses both JAK2 mutant cell growth and MPN disease onset in vitro and in vivo. Moreover, through a compound screen, followed by chemical proteomics and compound optimization, WWQ-03-012 is discovered, which selectively degrades mutant JAK2, induces primary leukemia cells death, and inhibits MPN progression through targeting DESI2 enzymatic activity in vitro and in vivo. These studies provide a novel therapeutic strategy against mutated JAK2 signaling in MPN and sAML.