Find the papers that actually matter

Radiation pneumonitis (RP) is a serious complication in lung cancer patients with pre-existing interstitial lung disease (ILD) undergoing radiotherapy. Accurate risk stratification is crucial for individualized management. But predictive models integrating multimodal data are lacking. This study aimed to develop a novel machine learning-based nomogram integrating clinical, dosimetric, and inflammatory predictors for RP risk assessment in this high-risk population. This retrospective study of 424 ILD patients collected clinical, dosimetric, and inflammatory data. Machine learning algorithms created composite dosimetric (D score) and inflammatory (Inflamm score) scores. A multivariable logistic regression nomogram was built incorporating these scores with clinical risk factors. Model performance was assessed using area under the curve (AUC), calibration curve, and decision curve analysis (DCA). RP occurred in 200 (47%) patients. Independent risk factors included higher performance status, Charlson comorbidity index (CCI), usual interstitial pneumonia (UIP) pattern, immunotherapy, concurrent chemoradiotherapy, more radiation sessions, and lower lung volume. The D score and Inflamm score were both independent predictors. The integrated nomogram (AUC = 0.929) showed excellent discrimination, significantly outperforming the clinical model (AUC = 0.86), D score (AUC = 0.758) (both p < 0.001), and Inflamm score (AUC = 0.910, p = 0.168). Calibration curve and DCA confirmed its strong calibration ability and clinical utility to identify high-risk patients early. The integrated nomogram combining clinical, dosimetric, and inflammatory predictors enables accurate, individualized RP risk assessment in lung cancer patients with ILD. It can guide adjustments to individualized radiotherapy plans or preventive interventions, supporting better patient selection, treatment decisions, and proactive follow-up.

The serine/threonine kinase IKKε is overexpressed or activated in various cancers, making it a promising therapeutic target. Through a large-scale virtual screening of over 12 million compounds, we identified N8 as a novel IKKε inhibitor, selected for its favourable docking score and drug-likeness profile. The inhibitory activity of N8 on IKKε was validated in vitro across several cancer cell lines, including HCT116 (colorectal), HepG2 (liver), T24 (bladder), MDA-MB-231 (breast), A549 (lung), and HeLa (cervical). N8 demonstrated significant reductions in cell viability, colony formation, and migration, particularly in HCT116 colorectal cancer cells, where it exhibited superior efficacy compared to established IKKε inhibitors. Mechanistically, N8's anticancer activity appears to be mediated through modulation of autophagy rather than apoptosis.

To explore the predictive value of machine learning-based multimodal MRI radiomics combined with clinical features in the efficacy of high-intensity focused ultrasound (HIFU) ablation of uterine fibroids. This study included 390 patients with uterine fibroids who underwent HIFU ablation. Patients were stratified into high and low ablation groups based on an 80% non-perfused volume ratio (NPVR) and randomly divided into training (70%) and test (30%) sets. Radiomics features were extracted from T1-weighted imaging (T1WI), T2-weighted imaging (T2WI), and diffusion-weighted imaging (DWI). The most predictive features were selected via Recursive Feature Elimination (RFE) and the Least Absolute Shrinkage and Selection Operator (LASSO), and combined with clinical characteristics. Logistic Regression (LR), Random Forest (RF) and eXtreme Gradient Boosting (XGBoost), were constructed to predict ablation efficacy, with performance assessed using the area under the receiver operating characteristic curve (AUC). The results indicated that age, uterine fibroid location, and T2WI signal intensity were independent predictive factors (p < 0.05). The multimodal-clinical fusion XGBoost model exhibited the optimal performance. In the test set, this model achieved an AUC of 0.894, with an accuracy of 82.1%, sensitivity of 88.9%, and specificity of 74.1%.The calibration curve and decision curve analysis (DCA) confirmed that the predicted probabilities of the model were highly consistent with the actual risks, and stable clinical net benefits were achieved. The XGBoost model based on multimodal MRI and clinical features may serve as a reference for predicting HIFU ablation efficacy and optimizing treatment strategies.

To develop and validate a risk prediction model for malignant transformation in patients with gallbladder polyps (GBPs) using an interpretable machine learning framework and evaluate its predictive accuracy. A retrospective cohort of 1,027 surgical patients was enrolled from Yantai Yuhuangding Hospital (training set: n = 933) and Shanghai Eastern Hepatobiliary Surgery Hospital (validation set: n = 94). Feature selection for the training set was performed using the least absolute shrinkage and selection operator (LASSO) regression method. A predictive model was constructed with the XGBoost machine learning algorithm and evaluated using Shapley Additive exPlanation (SHAP). LASSO regression identified five significant risk factors for malignant transformation in GBPs: presence of concomitant cholecystitis, polyp count, polyp base width, age, and maximum polyp diameter. The area under the receiver operating characteristic curve (AUC) was 0.862 (95% confidence interval [CI]: 0.8342-0.8893) in the training set and 0.777 (95% CI: 0.6804-0.8737) in the validation set. SHAP analysis illustrated the contribution of each factor. This study developed and validated a risk prediction model for malignant transformation in patients with GBPs. The model demonstrated favorable discrimination, calibration, accuracy, and clinical applicability. Integration with SHAP technology may assist clinicians in optimizing treatment and management strategies.

Anti-PD-(L)1 agents, inhibitors of programmed cell death protein 1 (PD-1) or its ligand (PD-L1), are established therapies that improve cancer management as well as the disease and societal burden of specific metastatic and early-stage cancers. The aim of the study was to determine the impact of adopting anti-PD-(L)1 agents for the treatment of all eligible patients with early-stage cancers versus reserving anti-PD-(L)1 agents for patients with metastatic disease alone in Hungary. This study evaluated two scenarios, one where anti-PD-(L)1 agents were used to treat all eligible early-stage disease cases (ESD scenario) of melanoma (stage IIB-C and III), renal cell carcinoma (RCC), and triple-negative breast cancer (TNBC) versus a reference scenario where anti-PD-(L)1 agents were only used to treat metastatic disease cases in Hungary (2024-2033). A Markov-modeling approach estimated the health outcomes and productivity losses from each scenario from a societal perspective. Outcomes included recurrence-/event-/disease-free life-years, total life-years, quality-adjusted life-years (QALYs), productive years (patients and caregivers), recurrences/events, active treatments for metastatic disease, and deaths. The cumulative health and productivity impact of ESD treatment with anti-PD-(L)1 agents in Hungary was the difference in health and productivity outcomes between the ESD and reference scenarios for the time horizon of the model. ESD treatment with anti-PD-(L)1 agents was estimated to increase recurrence-/event-/disease-free life-years (+13.8%), total life-years (+3.7%), and QALYs (+4.7%), as well as productive work years for patients (+39.6%) and caregivers (+27.6%). Concurrently, there would be fewer recurrences/events (-31.0%), active treatments for metastatic disease (-34.0%), post-recurrence deaths (-30.3%), and total deaths (-23.1%). Investing in anti-PD-(L)1 agents for early-stage disease may not only increase the life expectancy and QALYs for patients in Hungary but also increase productive work years for both patients and caregivers in Hungary. In addition, it may also help to reduce metastatic disease treatments and cancer-related deaths.

Janus kinase (JAK) dysregulation plays a central role in the pathogenesis of inflammatory, autoimmune, and malignant disorders, making the JAK family an essential therapeutic target across multiple disease domains. Over the past two decades, the field has progressed from the identification of early JAK2 inhibitors to the approval of several first-generation agents, including ruxolitinib, tofacitinib, baricitinib, and fedratinib, which validated the clinical feasibility of JAK blockade. However, limitations related to safety, isoform selectivity, long-term tolerability, and off-target kinase interactions continue to restrict their broader application and highlight the need for next-generation molecules. In this review, we provide a comprehensive and strategic assessment of the molecular features underpinning JAK2 and JAK3 selectivity, including signaling features directly relevant to inhibitor design, mutational landscapes, and structural determinants such as the uniquely targetable Cys909 residue in JAK3. Although the JAK family comprises four kinases, this review intentionally focuses on JAK2 and JAK3, where structural divergence, disease relevance, and emerging selectivity strategies provide the strongest opportunities for next-generation precision inhibitor design. We integrate recent advances in synthetic chemistry, including hinge-binding optimization, heterocyclic diversification, multicomponent reactions, and scaffold-hopping strategies, with computational methodologies such as molecular docking, molecular dynamics simulations, QM/MM calculations, and machine-learning-based predictive modelling. Together, these multidisciplinary approaches have accelerated hit discovery, refined selectivity, and improved the pharmacokinetic and safety profiles of emerging JAK inhibitors. By consolidating progress across medicinal chemistry, structural biology, and computational design, this review outlines key opportunities and remaining challenges in developing next-generation JAK inhibitors with enhanced precision and therapeutic value for oncology, immunology, and chronic inflammatory diseases.

The fungal priority pathogen and basidiomycete, Cryptococcus neoformans (Cn), causes lung and brain infection in predominantly immuno-compromised individuals and there is an urgent need for new treatment options. The pyrazolopyrimidine-based cyclin dependent kinase (CDK)7 inhibitor, BS-181, has anticancer properties, but its antifungal activity has not been investigated. We show that cryptococcal CDK7 more closely resembles the human enzyme than that of ascomycetes, and that BS-181 inhibits its activity. BS-181 inhibited growth of both Cn and Cryptococcus gattii (Cg), but not ascomycete fungi and delayed progression through the G2/M phase of the cell cycle. Transcriptomic analysis revealed that BS-181 induces splicing defects leading to elevated intron retention within the transcriptome and also suppresses translational processes. BS-181 displayed additive or synergistic activity with licensed antifungals against laboratory and clinical Cn and Cg strains, most notably with amphotericin B where synergy (2-4-fold reduction in the amphotericin B MIC) was achieved using low-sub micromolar concentrations of BS-181. Compared with either drug alone, BS-181-AmB combination therapy provided greater protection against Cn infection in a wax moth model (p ≤ 0.032) and extended survival of Cn-infected mice. These findings demonstrate that CDK7 inhibitors, already of interest as anticancer agents, could be repurposed to prevent or treat opportunistic fungal infections in cancer patients when combined with licensed antifungals limited by either toxicity or resistance.

Artificial intelligence based methods for anatomy segmentation can help automate characterization of large imaging datasets. The growing number of similar functionality models raises the challenge of evaluating them on datasets that do not contain ground truth annotations. We introduce a practical framework to assist in this task. We harmonize the segmentation results into a standard, interoperable representation, which enables consistent, terminology-based labeling of the structures. We extend 3D Slicer to streamline loading and comparison of these harmonized segmentations and demonstrate how standard representation simplifies review of the results using interactive summary plots and browser-based visualization using the OHIF Viewer. To demonstrate the utility of the approach, we apply it to evaluating segmentation of 31 anatomical structures (lungs, vertebrae, ribs, and heart) by 6 open-source models-TotalSegmentator 1.5 and 2.6, Auto3DSeg, MOOSE, MultiTalent, and CADS-for a sample of computed tomography scans from the publicly available National Lung Screening Trial dataset. We demonstrate the utility of the framework in enabling automating loading, structure-wise inspection, and comparison across models. Preliminary results ascertain the practical utility of the approach in allowing quick detection and review of problematic results. The comparison shows excellent agreement segmenting some (e.g., lung) but not all structures (e.g., some models produce invalid vertebrae or rib segmentations). The open-source resources developed include segmentation harmonization scripts, interactive summary plots, and visualization tools. These resources assist in segmentation model evaluation in the absence of ground truth, ultimately enabling informed model selection.

Follicular thyroid neoplasms represent a common subtype of thyroid tumor that has become the most prevalent endocrine neoplasms in recent decades. Accurate diagnosis and grading of these tumors are critical for clinical management of follicular thyroid neoplasms, yet remain challenging due to the limitations of conventional imaging modalities without histopathological molecular information. There is growing interest in analytical techniques that can provide metabolic and molecular insights without the need for exogenous reagents. In such a context, Fourier transform infrared (FTIR) microspectroscopy has emerged as a promising label-free approach for detecting intrinsic disease biomarkers. In this work, we proposed a label-free framework for screening and grading follicular thyroid neoplasm tissues using FTIR microspectroscopy and machine learning. A total of 32 clinical samples in the form of tissue sections were collected from a real-world cohort and measured for the spectral data, from patients diagnosed with follicular thyroid adenoma, follicular tumor with uncertain malignant potential, and follicular thyroid carcinoma. For tumor screening, disease-specific features were extracted from the FTIR mapping data and further imaged through integration, principal component analysis, and clustering, enabling visual and quantitative assessment of tissue abnormalities. For neoplasms grading, a deep neural network trained with an adversarial learning strategy achieved a grading accuracy of 94.0% on an independent test set. These findings collectively demonstrate the potential of FTIR microspectroscopy as a powerful, reagent-free tool for the diagnosis, pathological evaluation, and clinical management of follicular thyroid neoplasms.

Raman spectroscopy has been proved to have the potential to accurately diagnose a variety of diseases, and novel Raman probes or instruments for clinical applications have been constantly developed. However, biological tissues are usually structurally complex. The Raman signals collected in vivo may be a mixture of various chemical components, even different tissues, which poses challenges for disease analysis and diagnosis. This work proposed a Raman spectral unmixing approach to separate the signals of different tissues from their mixed spectra. Specifically, continuous wavelet transform was performed to extract the multi-scale time-frequency domain features of Raman spectra. TransUnet model was introduced to analyze the multi-scale features from high-frequency to low-frequency through the convolution and transformer modules, and predict the Raman signals of target components. Breast cancer tissues were selected as the research subject, the Raman signals of stroma and adipocyte were successfully separated from their mixed tissues, and multiple biochemical changes in breast cancer tissues were revealed through further analysis of the unmixing signals. This work will contribute to biological in vivo detection of Raman probes or instruments, enabling them to separate signals from different tissues, structures, and even biochemical molecular components for more detailed and accurate analysis of diseases.