Predicting recurrence of pancreatic cancer after surgery could inform clinical decision making, including adjuvant therapies and follow-up. This study aimed to develop and validate a deep learning model using digitized whole-slide images (WSI) of histopathology.

Publicly available WSI of pancreatic ductal adenocarcinoma resections from three cohorts were used for training. The model consisted of a pan-cancer foundation model to generate embeddings, mean-pooling across tissue patches, and then a fully connected neural network. Model predictions were compared with human-labeled histopathologic features and genomic alterations. The model was externally validated in a meta-analysis of a single-center cohort from Princess Margaret Cancer Centre, a multicenter cohort from France, and the PRODIGE 24 trial of adjuvant chemotherapy.

The deep learning model was trained on 12,594 tissue patches from 257 patients. High-risk classifications were associated with squamous morphology, reactive stroma, tumor cellularity, and necrosis, whereas low-risk classifications were associated with tubulopapillary and conventional morphologies, as well as deserted stroma. High-risk cancers were enriched for basal-like gene expression profiles and distinct oncogenic pathways. In a meta-analysis of the external cohorts, the hazard ratio (HR) for death comparing high-versus low-risk cancers was 1.49 (95% CI, 1.25 to 1.79, P < .001), whereas the HR for recurrence or death was 1.41 (95% CI, 1.19 to 1.68, P < .001). The classifications remained prognostic among moderately differentiated cancers.

An open-source deep learning model using WSI from pancreatic cancer resections generated risk classifications that correlated with histopathologic and genomic features. Classifications were externally validated in a meta-analysis of three cohorts. This model could be applied to WSI to provide individualized prognostic information for patients.

Medical microrobots have strong potential for targeted therapeutic delivery; however, current systems achieve only physical targeting, and once at the target site, they are unable to distinguish healthy cells from cancerous ones because of the lack of biological selectivity. Here, we present a biohybrid microrobot system that combines magnetic targeting with biological selectivity. The microrobots are derived from human embryonic kidney cells genetically engineered to produce tumor necrosis factor-related apoptosis-inducing ligand (TRAIL), a molecule that induces cancer cell death in multiple tumor types without damaging healthy cells. Engineered cells are then conjugated to biocompatible magnetic Janus particles-silica beads half-coated with FePt nanofilms-to enable external magnetic control. With magnetic fields, the microrobots accumulate around the tumor spheroids and continuously release TRAIL for several days, leading to selective cancer cell death while avoiding damage to healthy cells. This study combines microrobotics with genetically engineered cell therapies to achieve a targeted, prolonged, and cancer-selective therapeutic delivery.

Cancer therapeutics frequently fail in clinical trials because of poor therapeutic index (efficacy-to-toxicity ratio). We systematically identified targets likely to have a good therapeutic index, revealing insulin receptor substrate 4 (IRS4) as a dependency in IRS4-expressing cancers. Pan-cancer analysis of pediatric-enriched cancers revealed IRS4 expression consistent with dependency in 68% of choroid plexus, 37% of malignant rhabdoid, 31% of NUT midline, and 5% of osteosarcomas, while in adult cancers, it was expressed in 8% of uterine leiomyosarcomas and 1 to 2% of lung squamous, stomach, and breast carcinomas. IRS4 expression in adult tumors was associated with enhancer hijacking rearrangements, including recurrent GATA3-IRS4 and ANKRD30A-IRS4 in breast cancer, while rhabdoid and NUT midline cancers expressed IRS4 epigenetically. IRS4 fueled cancer dependency through PI3K-Akt activation, and domain analysis revealed the PH and PTB domains, which have a predicted drug pocket, to be dispensable, suggesting degradation-based modalities. These data reveal IRS4 as a target in IRS4-expressing cancers and suggest inhibitory approaches.

Accurate detection of microhepatocellular carcinoma (HCC) remains a major clinical challenge owing to the limited specificity and sensitivity of current imaging modalities. Herein, we present a dual-injection magnetic resonance imaging (MRI) peptidic probe based on in vivo membrane engineering, achieving in situ signal amplification and molecularly precise imaging. The first injection of programmable nanoparticles coassembled from two peptide monomers, incorporating a GPC3-targeting ligand, a β sheet-forming motif, a dibenzocyclooctyne (DBCO) handle, and porphyrin IX (PpIX) for fluorescence tracking. Following systemic administration, the nanoparticles high-specifically bind to GPC3-overexpressing tumor membranes and transform into surface-anchored nanofibrils, exposing confined DBCO groups. A second injection of azide-modified Gd-DOTA enables rapid copper-free click conjugation on the nanofibrillar scaffold, yielding a nearly fourfold increase in longitudinal relaxivity. MRI demonstrated strong T₁-weighted signal enhancement and high tumor-to-liver contrast in Hepa1-6 tumor-bearing mice. In vivo membrane engineering strategy establishes a generalizable platform for receptor-guided molecular imaging and early cancer detection.

KRAS mutations are a hallmark of pancreatic ductal adenocarcinoma (PDAC), driving tumor initiation and progression in the vast majority of cases, with KRAS^G12D being the most prevalent variant. Recent advances have led to the development of mutation-specific KRAS inhibitors (KRASi), yet their clinical impact is hindered by the rapid onset of drug resistance. In this study, we identify Fos-related antigen-2 (Fra-2), a stress-responsive transcription factor of the AP-1 family, as a key mediator of adaptive resistance to the KRAS^G12D selective inhibitor MRTX-1133. Using a combination of established PDAC cell lines, xenograft models, and patient-derived organoids, we demonstrate that Fra-2 expression is consistently upregulated following MRTX-1133 treatment. Functional assays reveal that Fra-2 overexpression promotes resistance by reprogramming the transcriptional landscape, directly enhancing mTOR expression and signaling. Consistently, FRA2 and MTOR levels strongly correlate in PDAC patient samples. Collectively, these findings uncover a mechanistic interplay between Fra-2 and the mTOR pathway in MRTX-1133-resistant PDAC, highlighting that targeting Fra-2 may represent a valuable approach to enhance the efficacy of KRASi.

Autoimmune encephalitides (AE) comprise a group of autoimmune diseases of the nervous system that primarily affect the brain's gray matter. The pathogenesis of AE is largely driven by the production of autoantibodies targeting intracellular and extracellular structures within the nervous system. The most extensively studied forms are those associated with antibodies targeting the NMDAR, AMPAR-, GABAβR-, LGI1, and Caspr2 extracellular antigens. Evidence suggests that anti-NMDA receptor encephalitis occurs more frequently than viral encephalitis in individuals under 30 years of age. Anti-NMDA receptor encephalitis is typically considered in young patients presenting with a primary psychotic episode, catatonia, or epileptic syndrome. However, literature indicates that anti-NMDA receptor encephalitis may present with only a single group of clinical manifestations, rather than the full spectrum of symptoms outlined in the diagnostic criteria, which include abnormal behavior, mental disorders or cognitive dysfunction, impaired consciousness, speech disorders, epileptic syndrome, motor disorders, and autonomic dysfunction. Such monosymptomatic variants can be challenging to diagnose due to mild, nonspecific symptoms and the limited diagnostic value of paraclinical data. We present a case of a patient with rapidly progressive cognitive and behavioral disorders, with no other clinical symptoms, who was diagnosed with anti-NMDA receptor encephalitis and was suspected of having a high titer of anti-NMDA receptor antibodies in both blood and cerebrospinal fluid. A good clinical response to immunosuppressive therapy, along with a recurrent course and repeated detection of autoantibodies in the cerebrospinal fluid, confirmed the primary assumption that the patient had a rare monosymptomatic variant of anti-NMDA receptor encephalitis.

Diffuse large B-cell lymphoma (DLBCL) is the most common type of non-Hodgkin lymphoma in the world. It exhibits high heterogeneity and invasiveness and is prone to developing treatment resistance. Therefore, there is an urgent need for good prognostic evaluation indicators and therapeutic targets. In recent years, immunotherapy has become a research hotspot for DLBCL. Tumor-associated neutrophil (TAN) is widely expressed in various tumors and is an important component of the immune microenvironment. However, there have been few studies on the role of TAN in DLBCL. This study has demonstrated that CD66b, which is a marker of TAN, is a good prognostic marker of DLBCL and its expression is related to the prognosis of DLBCL patients. The expression level of CD66b is also closely correlated with the objective response rate of the R-CHOP treatment regimen in DLBCL patients with non-GCB subtype. The expression level of CD66b has a high reference value for the determination of the treatment plan. The combined detection of CD66b and PD-L1/PD-L2 is of significance to predict the prognosis of DLBCL patients.

To investigate the effects of Mucin 1 (MUC1) in human triple-negative breast cancer MDA-MB-231cells, the MDA-MB-231 cell line with MUC1 knockout (231-MUC1-KO) was constructed by CRISPR/Cas9 gene editing. Cell proliferation was evaluated using EDU and colony formation assays, and cell migration and invasion were detected by transwell assay. Autophagy flow was assessed by western blot and Ad-mCherry-GFP-LC3B dual-fluorescence system and validated by lysosome inhibitor barfimycin A1 and autophagy inducer rapamycin. Key proteins of autophagosomes and lysosomal fusion (STXl7, SNAP29) and lysosomal tagged protein (LAMP1) were detected by western blot, and lysosomal pH was evaluated by Lysotracker Red fluorescence. MUC1 expression was low in human normal breast epithelial cells MCF-10A, but was highly expressed in human MDA-MB-231 cells and tissues. Successful MUC1 knockout was confirmed by gene sequencing, RT-qPCR, and western blot. Loss of MUC1 gene expression in 231-MUC1-KO significantly reduced proliferation, migration, and invasion. Compared with the control group, MUC1 knockout led to a significant increase of autophagy-related proteins LC3-II and p62, which is consistent with the effect of lysosome inhibitor bleomycin A1. After adding the autophagy inducer rapamycin, compared with the control group, the accumulation of LC3-II and p62 proteins also further increased. The expression level of LAMP1 was downregulated and the lysosome pH increased, but the expression levels of STXl7 and SNAP29 were not affected. These findings suggest that MUC1 promotes malignant behaviors in MDA-MB-231 cells by regulating autophagic flow, likely through lysosomal dysfunction-mediated autophagy blockade.

Approximately 12% of patients with cancer annually visit the emergency department (ED) for disease- or treatment-related issues. These patients often face delays in care, including prolonged wait times and extended length of stay (LOS), contributing to ED crowding, delayed treatment, and increased mortality. Numerous studies have investigated interventions to reduce LOS and prevent ED visits for patients with cancer. However, a systematic overview of these interventions is currently lacking. In this review we aimed to present interventions that optimize input, throughput and output in ED care by reducing ED LOS or ED visits for patients with cancer.

We searched five electronic library databases: Medline ALL via Ovid; Embase.com; Web of Science Core Collection; the Cochrane Central Register of Controlled Trials via Wiley; and Google Scholar. Inclusion criteria for this review were as follows: 1) research on (a subset of) patients with cancer; 2) conducted in or in collaboration with the ED; 3) the introduction of an intervention aimed at optimizing ED input, throughput, and output; and 4) performance of the intervention was measured using outcomes, such as ED LOS, number of ED visits or hospitalizations, use of acute-care services, or time to antibiotics.

The literature search yielded 11,357 articles. After removing duplicates, 7,315 unique articles remained for screening. Of these, 109 were selected for detailed abstract review. Following this second screening, 35 articles underwent full-text analysis, and 16 articles met all inclusion criteria. These studies identified four categories of interventions: scoring systems (n=5); dedicated cancer urgent care facilities (n=5); protocolized care (n=3); and staffing optimization (n=3). Among scoring systems, use of the Edmonton Symptom Assessment Scale reduced ED visits (relative rate (RR) = 0.92) and hospitalizations (RR = 0.86), while the Clinical Index of Stable Febrile Neutropenia score showed higher specificity (98.3%) than the Multinational Association for Supportive Care in Cancer score (54.2%) for identifying low-risk febrile neutropenia.

We identified four categories of intervention that could potentially reduce ED visits and ED LOS, of which scoring systems showed the most potential. Rather than developing new tools, future efforts should prioritize the implementation, validation, and refinement of these existing strategies to optimize treatment of cancer patients in the emergency department.

Hepatocellular carcinoma (HCC) is a highly heterogeneous malignancy with poor prognosis due to drug resistance and recurrence. Tumor-associated macrophages (TAMs) are pivotal in the HCC tumor microenvironment, but their prognostic and therapeutic relevance remains incompletely defined. This study aimed to identify macrophage-related genomic signatures, delineate HCC molecular subtypes, and construct a prognostic model to predict survival and therapy response.

We integrated scRNA-seq (GSE151530) and bulk RNA-seq (HCCDB18, TCGA-HCC) data. Macrophage-related genes were identified via differential expression analysis of scRNA-seq data. Consensus clustering (Euclidean distance, hierarchical clustering) was used for subtype delineation. A prognostic model was constructed using PCA (Principal Component Analysis) on 25 OS-related differentially expressed genes (DEGs; univariate Cox regression), with z-scored normalization and 3 principal components. Immune infiltration (ssGSEA) and drug sensitivity [immunophenoscore (IPS) scores, pRRophetic] were analyzed.

Four HCC subtypes were identified; Cluster C showed the most favorable survival. The PCA-derived score strongly correlated with OS (p<0.001) and immunotherapy responsiveness (higher scores=enhanced sensitivity). High scores were associated with increased effector T cell infiltration and reduced T cell exhaustion. Drug sensitivity analyses revealed divergent responses to immunotherapy and conventional agents across subgroups.

Macrophage-related genomic signatures are critical for HCC prognosis and therapy response. The PCA-based model holds promise as a biomarker for personalized therapy, warranting larger cohort validation and mechanistic exploration.