Find the papers that actually matter

Radiology report generation (RRG) aims to automatically produce clinically accurate textual reports from medical images. Existing methods predominantly rely on autoregressive (AR) language models, whose causal dependency structure restricts generation to a unidirectional left-to-right process. This paradigm can induce sequence bias, where models tend to follow stereotypical token orders and high-frequency report templates rather than fully grounding generation in image-specific evidence. In this paper, we propose AnchorDiff, the first masked-diffusion framework for RRG that integrates knowledge-graph-derived clinical anchors into diffusion language modeling. By leveraging bidirectional context and iterative refinement, AnchorDiff mitigates the limitations of fixed-order autoregressive decoding. Specifically, we introduce a topology-aware training strategy that uses RadGraph-derived entity hierarchies to assign clinically important tokens differentiated masking protection and loss weights. We further design an inference-time rewriting strategy that detects unstable committed tokens through perturbation-based testing and selectively revises them during denoising. Extensive experiments on the MIMIC-CXR and MIMIC-RG4 benchmarks demonstrate that AnchorDiff achieves state-of-the-art (SOTA) performance, showing the effectiveness of clinically anchored masked diffusion for radiology report generation.

Multimodal magnetic resonance imaging (MRI) is crucial for brain tumor segmentation, with many methods leveraging its four key modalities to capture complementary information for effective sub-region analysis. However, the absence of several modalities is very common in practice, leading to severe performance degradation in existing full-modality segmentation methods. Limited by the structured data model, recent works often adopt a multi-stage training strategy for full-modality and missing-modality scenarios, which increases training costs and inadequately addresses the interference of miss. In this work, we propose a graph-based one-stage framework for robust brain tumor segmentation with missing modalities. Specifically, we introduce modality-specific virtual nodes that serve as supplementary information sources to compensate for missing modalities. To enhance model robustness against arbitrary modality combinations, we leverage the inherent flexibility of graph networks to devise a dynamic connection strategy. This mechanism dynamically adjusts the adjacency matrix based on modality availability, preserving beneficial information flow while mitigating interference effects caused by missing modalities. Furthermore, we enhance the graph network through heterogeneous weight matrices, enhancing its adaptability to multimodal scenarios. Extensive experiments on the BRATS-2018 and BRATS-2020 datasets demonstrate that our method outperforms the state-of-the-art methods on almost all subsets of incomplete modalities.

Accurate polyp segmentation in colonoscopy is essential for early colorectal cancer detection, yet real-world clinical environments pose persistent challenges such as motion blur, specular reflections, and illumination instability. Most existing methods are optimized on clean benchmark images and suffer noticeable performance degradation when deployed in authentic surgical scenarios. We propose DepthPolyp, a lightweight and robust segmentation framework based on pseudo-depth-guided multi-task learning and efficient feature modulation. The architecture combines hierarchical Ghost factorization for compact feature generation, Interleaved Shuffle Fusion for low-cost cross-scale interaction, and Dynamic Group Gating for adaptive group-wise feature weighting. Extensive experiments demonstrate that DepthPolyp achieves strong cross-dataset generalization when trained on degraded data and evaluated on both clean and noisy target domains, consistently outperforming lightweight baselines and remaining competitive with substantially larger models. In real surgical video evaluation on PolypGen, DepthPolyp achieves better segmentation performance than models up to $20\times$ larger while preserving real-time inference speed. With only 3.57M parameters and 0.86 GMACs, the proposed method runs at over 180 FPS on mobile devices, making it well suited for real-time deployment in resource-constrained clinical environments. Code and pretrained weights are available at: https://github.com/ReaganWu/DepthPolyp/

Brain tumors exhibit high heterogeneity in morphology and multimodal contrast, making manual slice-by-slice de lineation time-consuming and experience-dependent, thus necessitating efficient and stable automated segmentation methods. To address the limitations of CNNs in modeling long-range dependencies, and the heavy computational and memory overhead and inter-block contextual in coherence of Transformers in 3D MRI, this paper proposes Multi-Head Mamba (MHMamba). This method combines a U-shaped architecture with a multi-head state-space model (Mamba), splitting the channel dimension into parallel SSM heads and aggregating them with residuals. This enhances long-range representation and improves the stability of multimodal training while maintaining linear complexity. To further align statistics and enhance lesion response, we designed a channel-space calibration module for multi-head outputs and introduced an adaptive fusion mechanism at skip connections to dynamically connect global semantics with local details, thereby improving boundary consistency and the detection of small-volume lesions. We conducted experiments and ablations on BraTS2021 and BraTS2023. The results showed that MHMamba achieved stable and significant improvements in overall accuracy, boundary smoothness, and sensitivity to tumor core and small-volume enhancement areas, while preserving the linear-complexity advantage of Mamba-based modeling, thus verifying the effectiveness and versatility of the method.

Multimodal 3D MRI brain tumor segmentation is a pivotal step in radiotherapy target delineation, surgical planning and post-treatment assessment. Existing methods often assume artifact-free MRI images. However, inevitable patient motion during scanning introduces artifacts and blur that degrade boundary and texture features, leading to poor segmentation performance. To bridge this gap, we introduce Degradation-Aware Blur-Segmentation Net (DABSeg), a synchronous deblurring 3D multimodal MRI segmentation network that unifies blur removal and accurate segmentation. Specifically, we propose a feature-domain motion-deblurring stem to compensate for blur and rebalance intensity. Concurrently, the backbone network embeds a blur-aware cross-modal cross-attention module and multi-scale residual aggregation to yield effective modality complementarity. Notably, we optimize a joint loss that combines weighted Dice with a clear-reference reconstruction term, where imbalanced weights are applied to small targets to boost learning intensity and predictive stability for small lesions and border regions. Systematic comparisons and ablation experiments on the BraTS2020 dataset under both clear and degenerative conditions consistently demonstrate that DABSeg surpasses state-of-the-art methods in tumor Dice score and boundary precision. These results validate the effectiveness of degenerative-aware cross-task collaborative learning in improving the robustness and clinical utility of multi-modal 3D brain tumor segmentation under realistic degenerative conditions. The source code is available at https://github.com/YuchunWang24/DABSeg_ICPR

Accurate intraoperative and postoperative diagnosis of spread through air spaces (STAS) is essential for guiding surgical decisions and postoperative management in lung cancer. However, histopathological assessment is labor-intensive and is prone to missed or incorrect diagnoses. We propose a Diffusion Attention Expert Model (DAEM) to detect STAS in frozen sections (FSs) and paraffin sections (PSs). Its diffusion attention expert module leverages full attention aggregation to learn multi-scale features from histopathological images, while a dual-branch architecture strengthens multi-scale feature representation. On an internal dataset, DAEM achieves AUCs of 0.8946 for FSs and 0.9112 for PSs. Validation on external multi-center datasets from eight institutions demonstrates strong generalizability and interpretability. Using tumor microenvironment (TME) features in PSs, we further enable semi-automatic measurement of STAS location and its distance from the primary tumor. Several quantitative TME metrics are identified as potential biomarkers for STAS, including micropapillary-type STAS. Overall, DAEM offers a clinically actionable framework for STAS assessment by enabling accurate and interpretable detection on FSs and PSs, supporting postoperative risk stratification through quantitative TME-based analysis.

Langerhans cell histiocytosis (LCH) is a rare, heterogeneous inflammatory myeloid neoplasm. This study focuses to summarize real-world experiences of using mitogen-activated protein kinase (MAPK) inhibitors in managing pediatric LCH. Thirty-six children with LCH received vemurafenib (n = 11), dabrafenib (n = 19), or trametinib (n = 6), either as monotherapy (n = 24) or combined with chemotherapy (n = 12). The median Disease Activity Score declined from 5 at baseline to 1 after 3 months of MAPK inhibitor therapy (p < .0001). Elevated interleukin-2 receptor (IL-2R) and tumor necrosis factor-α (TNF-α) levels were observed in 90.1% and 81.8% of patients, respectively. Median IL-2R and TNF-α levels decreased from 1212 U/mL and 15 ρg/mL at baseline to 494 U/mL and 4.18 ρg/mL at 3 months (p < .0001, p = .002), with no significant differences between MAPK inhibitor alone and combined with chemotherapy groups (p = .7791, p = .7503). The objective response rate at 3 months was 94.4% (34/36). At last follow-up, 20 patients had no active disease (NAD); 13 had NAD with residual diabetes insipidus and/or sclerosing cholangitis; 1 had NAD with neurodegeneration; and 2 died of progressive disease. Among 13 patients who discontinued treatment, relapse occurred more frequently in the monotherapy group (75%) than in the combination group (20%), but the difference was not statistically significant (p = .103). [Correction added on 07 February 2026, after first online publication: The p value has been revised from .047 to .103.]. Grade ≥3 adverse events occurred in 11.1% and resolved with dose adjustment. MAPK inhibitors are effective and well tolerated in pediatric LCH. Combination therapy may reduce relapse risk. Further prospective studies are warranted.

Lung cancer is a leading cause of cancer-related mortality worldwide, accounting for approximately 13% of all diagnoses. Non-small cell lung cancer (NSCLC) comprises 85% of lung cancer cases, with recent genetic profiling revealing the critical role of driver mutations in its pathogenesis. Advances in molecular stratification have redefined NSCLC classification, enabling targeted therapeutic approaches based on specific genetic abnormalities. Targeted therapies have shown significant clinical benefits, particularly in NSCLC patients with mutations in the epidermal growth factor receptor (EGFR), Kirsten rat sarcoma oncogene, and B-Raf proto-oncogene. These mutations are among the most prevalent actionable drivers, and treatments utilizing tyrosine kinase inhibitors (TKIs) have demonstrated substantial improvements in progression-free survival compared to conventional chemotherapy. Emerging data suggest that TKI therapy may represent the optimal approach for advanced NSCLC with EGFR and KRAS mutations. This review highlights FDA-approved therapies for NSCLC and examines recent clinical and translational research on EGFR and KRAS mutations. The findings underscore the pivotal role of personalized medicine and the potential of targeted therapies to transform treatment paradigms and improve patient outcomes in NSCLC.

Gastric cancer is a heterogeneous and complicated epithelial cancers. Chronic H. pylori and EBV infection, as well as intestinal microbiota exposure make gastric cancer encountered a complex tumor immune microenvironment. Mitophagy and m6A are deeply involved in immune microenvironment in the development of tumors. We used integrating machine learning of bulk and single cell RNA sequencing to explore the immune mechanisms of m6A related mitophagy genes (MRMGs) in gastric cancer. RT-qPCR and immunochemistry were used to verify gene expression. Prognostic model that involves a total of 20 DE-MRMGs exhibited a performance property in prognosis, immunotherapy prediction and tumor mutation burden in patients with gastric cancer. And significant difference between high-risk group and low-risk group focus on T cells which clarified in both bulk RNA and single cell RNA data. In terms of mechanism, vimentin may participates in T cell differentiation of malignant gastric cancer. Meanwhile, vimentin expression in patients display a significant increasing in low differentiated gastric cancer than high differentiated gastric cancer. Vimentin may be a diagnostic marker to draw the distinction between low and high differentiated gastric cancer in the mechanism of probably affecting T cell differentiation.

Wearable microneedles (MNs) biosensors offer significant potential for minimally invasive biomarker detection in interstitial fluid (ISF). However, suboptimal fluid collection efficiency, dermal irritation due to rigid structural components, and inadequate long-term operational stability collectively represent critical barriers to their clinical translation for early disease diagnosis. Here, we propose a fully integrated dual-extraction-driven flexible microneedles (DFMN) sensor for active detection of melanoma-associated miRNA. The DFMN sensor comprises hydrogel MNs, hollow Au-Ag NPs functionalized flexible electrodes, and a miniaturized wireless module. The synergistic combination of passive diffusion driven by hydrogel MN swelling and active transport via reverse iontophoresis enhanced ISF extraction volume by 1.6-fold within 5 min. This dual-extraction mechanism enabled real-time wireless detection of miRNA-221 with an ultra-low limit of detection (LOD) of 43 aM. Furthermore, to enhance reliability in complex biological environments, an artificial neural network algorithm was implemented, achieving a correlation coefficient (R2) of 0.985. The DFMN sensor holds broad application prospects in interdisciplinary fields such as human-machine interaction, smart healthcare, and artificial intelligence, providing an innovative solution for real-time sensitive detection of wearable flexible sensors.