Although perioperative stroke is a recognized risk in thoracic endovascular aortic arch repair, the long-term impact of altered cervical flow resulting from different device configurations is underappreciated. We compared angiographic data from two aortic stents, with anatomical (in-line) and nonanatomical (retrograde inner branch) configurations, and demonstrated a persistent filling delay in the retrograde inner branch device. Using computational fluid dynamic modeling, we demonstrated that parameters such as turbulence, vorticity, and Q-criterion are elevated in an idealized retrograde inner branch geometry. By way of a fluid mechanic scaling analysis, we show that the retrograde inner branch configuration results in a larger Dean number, thereby explaining our observed angiographic and computational modeling results. Our work should prompt further investigation regarding the long-term impact of branch configuration and altered cervical blood flow. Thoracic endovascular aortic arch repair is growing in prevalence. Perioperative stroke risk remains a major risk factor. A less studied phenomenon is the impact of differing branch configurations on long-term hemodynamics. This work compared angiographic data from two devices with differing branch flow configurations: antegrade vs retrograde. A persistent cervical branch filling was noted in the retrograde branch device. We applied fluid mechanic scaling analysis and validated with computational modeling to show that the retrograde branch configuration is promoting of disturbed flow. This work should promote further clinical investigation regarding long-term impact of branch configuration.
All oligodendrogliomas have a characteristic 1p/19q co-deletion that alters the expression of hundreds of genes on both affected chromosomal arms. The search for genes on 1p and 19q that drive oligodendroglioma development has only made little progress over the last years. Therefore, a computational network-based approach for the analysis of single-cell oligodendroglioma transcriptomes is developed to predict potential driver gene candidates within the region of the 1p/19q co-deletion purely based on tumor cells. Nine genes with strong impact on signaling pathways (ATP6V0B, F3, FUCA1, FTL, HNRNPR, ID3, JUN, MIIP, and PGM1) and 6 partially overlapping genes with strong impact on immune pathways (F3, FTL, FOSB, IFI6, ISG15, and SPINT2) were consistently predicted in at least 2 of the 3 analyzed oligodendrogliomas. Almost all of these genes are known to play important roles in growth, proliferation, and stem cells of closely related gliomas, but also roles in migration or reprogramming of the microenvironment had been reported in experimental glioma studies. Comparisons to a previous network-based bulk oligodendroglioma analysis and additional evaluations of the expression behavior of candidate genes in related normal brain cells further strengthen the study. Additional validations based on 2 independent oligodendrogliomas support the candidate genes. Robustness of the predictions is shown for imputed and nonimputed data. Strengths of the network-based approach are demonstrated by comparisons to related approaches. All findings clearly suggest that the developed network-based approach for the analysis of single-cell tumor transcriptomes is able to predict novel potential driver gene candidates for oligodendrogliomas. These are very valuable information for future experimental studies. The computational network-based approach can also be transferred to the analysis of single-cell transcriptomes of other types of cancer.
Digital twins of mammalian cell cultures hold great potential for predictive bioprocess modeling, yet their development is challenged by the nonlinear dynamics and metabolic complexity of these systems. We present a hybrid computational framework that integrates mechanistic and data-driven modeling to construct predictive digital twins for Chinese hamster ovary (CHO) cell cultures producing monoclonal antibodies. The framework couples ordinary differential equation (ODE) models with constraint-based metabolic modeling and machine learning components trained on Bayesian-estimated metabolic rates. Applied to 23 CHO fed-batch cultures, viable cell density, product titer, and key metabolite concentrations are accurately predicted under varying feeding and media conditions within a unified simulation engine, where empirical variability is incorporated through multivariate statistical constraints derived from experimental data. Cross-validation analyses demonstrated strong generalization across process variations, highlighting the framework's capacity to capture both biochemical constraints and adaptive cellular behavior. This hybrid modeling approach provides a mechanistically interpretable yet data-adaptive foundation for constructing bioprocess digital twins. By bridging statistical, mechanistic, and machine learning methodologies, it advances the computational representation of CHO cell culture systems and offers a generalizable strategy for predictive modeling in complex biological production processes.
Quantum technologies-quantum computing, quantum sensing, and quantum-enabled materials-are increasingly proposed as tools to accelerate drug discovery. Yet "quantum advantage" is frequently asserted without standardized benchmarks, clinically meaningful endpoints, or controlled comparisons against modern classical workflows. This review separates (i) quantum computing for molecular simulation and optimization, (ii) quantum sensing for structural/biophysical characterization and diagnostics, and (iii) quantum nanotechnologies for imaging and sensing, and then extends the framework to include device-led and physical therapies that increasingly co-evolve with drug development: photobiomodulation (red/NIR), focused ultrasound for blood-brain barrier opening and delivery enhancement, noninvasive neuromodulation devices (tDCS/TMS), and optogenetic therapies. We summarize demonstrated capabilities and constraints of NISQ-era computing, outline algorithmic classes for quantum chemistry and hybrid variational methods, evaluate quantum error-mitigation strategies and their limits, and contrast claimed performance with classical baselines in computational chemistry and machine learning. We conclude that near-term translational value is most substantial for quantum sensing and for device/physical platforms with established clinical evidence. In contrast, quantum computing remains principally hypothesis-generating until fault tolerance and reproducible advantage are established. Device-based modalities-including transcranial photobiomodulation for neuropsychiatric indications, focused ultrasound enabling CNS drug delivery, and home-supervised neuromodulation-are already reshaping therapeutic landscapes and clinical trial design. For drug discovery, the central requirement is not quantum novelty but validated decision impact, demonstrated under controlled benchmarks aligned with reproducibility expectations comparable to those evolving for AI/ML-driven methods in regulated contexts.
Inconsistent links between arterial stiffness and cognition may reflect limited cognitive tests and unaccounted diurnal pulse wave velocity variation. To bridge this knowledge gap, we investigated 24-hour ambulatory estimated pulse wave velocity (ePWV) and its association with dementia-related neuroimaging and cognitive function in hypertension. We assessed 893 patients with hypertension aged ≥50 years (mean age, 67.2 years; 52.3% women), including brain magnetic resonance imaging (n=545), global cognitive testing (n=623), and ambulatory ePWV measurements. White matter hyperintensity and hippocampus were quantified via Computational Anatomy Toolbox 12 and Statistical Parametric Maps 12. Cognition was assessed via the Mini-Mental State Examination and Montreal Cognitive Assessment. Among 623 tested participants, the prevalence of mild cognitive impairment was 10% (Mini-Mental State Examination, n=62) and 18.5% (Montreal Cognitive Assessment, n=115). Cognitive scores decline with higher white matter hyperintensity burden and lower hippocampal volume (P≤0.024). Higher 24-hour ePWV quartiles showed graded associations with higher white matter hyperintensity volume and lower hippocampal volume (both P<0.001) and lower cognitive scores (P≤0.037). Multivariable models showed each 1-SD (+1.2 m/s) increment in 24-hour ePWV were associated with 2.00±1.74 mL greater white matter hyperintensity volume (P=0.004), and 0.54±0.14 mL smaller hippocampal volume (P<0.001), independent of age, systolic blood pressure, and other confounders. These associations persisted after further adjustment for carotid-femoral PWV, which itself showed no independent association (P≥0.18). Results were consistent for daytime and nighttime ePWV and across key subgroups. Ambulatory ePWV is an independent risk factor for dementia-related brain pathology. Targeting arterial stiffness represents a promising strategy for dementia prevention.
Pathogenic variants in the COL4A4 gene lead to Alport syndrome, a hereditary kidney disorder characterized by deficiencies in the glomerular basement membrane (GBM), progressive renal failure, and associated visual and auditory dysfunctions. This study is aimed at detecting genetic variants and conducting an integrative network analysis to elucidate their involvement in Alport syndrome and chronic kidney disease (CKD). The enrolled patients were biological sisters. Clinical and pathological data were collected using laboratory tests. Genetic testing was performed to identify the variants via whole-exome sequencing (WES). The pathogenicity of the detected variant was confirmed using different computational approaches. The detected variants were classified according to the ACMG criteria. The interaction networks of CKD, Alport syndrome, and COL4A4 were examined using STRING v11.5, whereas pathways were analyzed by KEGG and STRING pathways analysis. Patients were diagnosed with CKD of Stage 5. The ultrasound results were not clear due to kidney fibrosis. WES revealed an 18-base pair deletion in the COL4A4 gene (chr2: g.227958874_227958891del). The detected variant was classified as pathogenic by ACMG criteria. Online Mendelian Inheritance in Man (OMIM) reveals that the variant in COL4A4 is due to autosomal recessive Alport syndrome 2. Sensorineural hearing loss was observed in two family members of the patients. These results show the existence of Alport syndrome in their family. The disease-protein interaction by STRING reveals that the COL4A4 protein is strongly associated with both CKD and Alport syndrome diseases. KEGG analysis indicates a consistent association of PI3K-Akt and AGE-RAGE pathways among both diseases. Dysfunction of COL4A4 leads to disruptions in GBM structure and signaling, which hasten CKD in Alport syndrome. Genetic testing is essential for identifying the exact cause of the disease, which aids in early diagnosis and control to spread within families.
The spleen is the most commonly injured solid organ in patients sustaining blunt abdominal trauma. The widespread use of computed tomography (CT) has significantly improved its diagnosis and grading. Delayed splenic rupture (DSR) refers to hemorrhage occurring more than 48 hours after blunt abdominal trauma in a patient who was initially hemodynamically stable. Although uncommon, it remains a recognized complication of splenic injury and can occur days to weeks after the initial trauma and carries significant mortality. We present a case of DSR presented one week after the injury, which was managed nonoperatively. The aim is to shed light on this uncommon but potentially fatal complication of spleen injury.
Panic attacks (PAs) are acute anxiety episodes that are pervasive, with one in 10 individuals having experienced a PA in the past year. PAs impair daily functioning and are associated with an increase in emergency room visits and suicide attempts. Despite their impact, the unpredictable nature of PAs makes them challenging to manage. PAs are transdiagnostic, occurring in individuals across and without a mental health diagnosis. However, prior work has largely focused on PA indications within individuals with panic disorder. This study identifies PA risk factors from over 6 months of passive sensing data recorded by Oura Rings in 182 young adults with and without adverse childhood experiences and psychiatric diagnoses, beyond just panic disorder. Our findings reveal that changes in Oura Ring-derived measures are associated with next-day PAs, with distinct associations observed across different mental health diagnoses. For individuals with panic disorder, the likelihood of PA increases with time spent inactive. For those with depression, the likelihood of PA increases with decreased variation in nightly respiratory rate, decreased rapid eye movement sleep, and increased time spent in high-intensity activity. For those without a mental health diagnosis, the likelihood of PA increases with decreased heart rate variability. Data aggregation window sizes that capture the associations with PA risk vary by diagnosis and the type of feature, suggesting that cumulative physiological patterns from windows up to 7 days before a PA contribute to onset. These findings point to the possibility that continuous monitoring of panic attack risk could one day support preventive mental health intervention.
Many inverse problems and signal processing problems involve low-rank regularizers based on the nuclear norm. Commonly, proximal gradient methods (PGM) are adopted to solve this type of non-smooth problems as they can offer fast and guaranteed convergence. However, PGM methods cannot be simply applied in settings where low-rank models are imposed locally on overlapping patches; therefore, heuristic approaches have been proposed that lack convergence guarantees. In this work we propose to replace the nuclear norm with a smooth approximation in which a Huber-type function is applied to each singular value. By providing a theoretical framework based on singular value function theory, we show that important properties can be established for the proposed regularizer, such as: convexity, differentiability, and Lipschitz continuity of the gradient. Moreover, we provide a closed-form expression for the regularizer gradient, enabling the use of standard iterative gradient-based optimization algorithms (e.g., nonlinear conjugate gradient) that can easily address the case of overlapping patches and have well-known convergence guarantees. In addition, we provide a novel step-size selection strategy based on a quadratic majorizer of the line-search function that leverages the Huber characteristics of the proposed regularizer. Finally, we assess the proposed optimization framework by providing empirical results in dynamic magnetic resonance imaging (MRI) reconstruction in the context of local low-rank models with overlapping patches.
Prior studies investigating the relationship between lingual plate (LP) fracture risk and mandibular third molar (M3M) anatomy have produced inconsistent findings. Factors such as age, sex, impaction depth, and buccolingual angulation have shown variable associations with fracture risk. Importantly, M3M rotational angulation has rarely been evaluated. While most research has focused on the middle and apical thirds of the root, one clinical study suggested the cemento-enamel junction may represent a structurally vulnerable site. This study aimed to identify anatomical and demographic risk factors associated with LP fracture across the coronal, middle, and apical thirds of the M3M root. Cone beam computed tomography (CBCT) images of 300 mandibular third molars were retrospectively reviewed. Data collected included patient sex, age, tooth side, and spatial alignment. M3M root length was measured and divided into coronal, middle, and apical thirds. The thinnest LP region at each level was recorded, and fracture risk was classified into three categories: low, moderate, and high. Ordinal logistic regression was used to determine significant risk factors. Fracture risk was significantly lower in the coronal third compared to the middle and apical thirds. Younger patients exhibited higher fracture risk at all levels. Greater root length increased risk at the apical third. Horizontal impaction level II, mesial and horizontal angulations, and buccal inclination were associated with elevated risk, while lingual inclination and rotational angulation reduced risk. Age, root morphology, impaction depth, and three-dimensional M3M angulations are key predictors of LP fracture risk at distinct root levels.
Heart failure with reduced ejection fraction (HFrEF) affects millions worldwide and is characterized by chronic cardiac dysfunction, impaired perfusion, altered skeletal muscle energetics, and, thus, exercise intolerance. Efficient therapeutic strategies reducing the burden of the impaired muscle metabolism in HFrEF are currently lacking. Hence, in the present study, we sought to determine whether myosin dynamics and its important role in ATP consumption can constitute a potent biochemical target to optimize skeletal muscle energy usage in HFrEF. We used skeletal muscle tissue from 11 human patients with HFrEF and 10 controls with comparable age, sex, and body mass index. We isolated individual myofibres and incubated them ex vivo with varying concentrations of a myosin inhibitor, mavacamten. We then performed 2'-(or-3')-O-(N-Methylanthraniloyl) adenosine 5'-triphosphate chase experiments, together with LC/MS-based proteomics profiling. We observed a distinct regulation of acetyl-lysine sites and higher myosin energy consumption in resting muscle fibers from patients with HFrEF than in controls. When exposed to mavacamten, we found a dose-dependent reduction in myosin ATP consumption in myofibres of patients with HFrEF, reversing the pathological over-consumption. Skeletal muscle myosin becomes inefficient in HFrEF. Pharmacological inhibition of myosin ATPase activity offers an inventive strategy to lower muscle energy demand and potentially address metabolic disturbances in HFrEF.
Clear cell renal cell carcinoma (ccRCC) can transition from indolent, low-grade lesions to high-grade, lethal disease through a layered cascade of genomic, epigenomic, metabolic, and immune remodeling. The initiating event in ∼90% of ccRCC is loss of chromosome 3p, enabling biallelic inactivation of VHL and frequent co-loss of chromatin regulators PBRM1, BAP1, and SETD2. The order and combination of genetic alterations shape distinct evolutionary trajectories in ccRCC. PBRM1 loss, observed in approximately 55% of cases, is linked to angiogenic, initially low-grade tumors that may later progress to higher-grade disease. In contrast, BAP1 loss (∼15%) drives early high-grade, inflammatory, immune-enriched phenotypes associated with aggressive behavior and worse prognosis. Progression is further shaped by structural and copy-number events including, chromothripsis coupling 3p loss with 5q gain, and recurrent 9p and 14q losses and 8q gain further promote cell-cycle dysregulation, genomic instability, and metastatic competence. Functionally, VHL loss stabilizes HIF-2α, driving VEGF signaling and Carbonic Anhydrase IX (CA9) expression and coupling pseudohypoxia to metabolic reprogramming and redox protection (glutathione/SLC7A11). Proteogenomic and metabolomic studies further highlight nutrient addiction with GLUT1/ASCT2 upregulation and a stress-resistant metabolic shield linked to grade and therapy resistance. Single-cell and spatial atlases place these programs in anatomic setting. They show that invasive fronts with high epithelial-mesenchymal transition (EMT) activity co-localize with myeloid and regulatory T-cell niches dominated by IL-1β, NF-κB, IL-10, STAT3, and TGF-β, along with exhausted CD8+ T cells, thereby promoting immune escape and invasion. Integrating these layers yields mechanism-based biomarkers and therapeutic nodes for risk-adapted precision treatment.
Prenatal detection of corpus callosum (CC) abnormalities is essential for assessing fetal neurodevelopment, yet conventional ultrasound diagnosis faces challenges from operator variability and suboptimal fetal positioning. We developed a novel deep learning framework CC-FocusNet that integrates automated region localization with an anatomy-aware dual-stream architecture for multi-view analysis. The model was trained on 496 cases and validated on an independent external cohort of 93 cases. We assessed both diagnostic performance and clinical interpretability through attention visualization. Our framework achieved 97.36% accuracy on the external test set. Grad-CAM++ heatmaps revealed that model attention consistently focused on clinically relevant anatomical landmarks, demonstrating strong interpretability. When integrated into clinical workflows, the AI system enhanced diagnostic accuracy and efficiency, particularly reducing misdiagnosis rates in challenging cases. This interpretable AI system provides accurate and efficient prenatal detection of CC abnormalities, offering substantial potential to support clinical decision-making and enable timely intervention for at-risk pregnancies.
Transthyretin amyloidosis (ATTR) is a degenerative disease affecting the heart and other organs. Transthyretin (TTR) aggregation is a driver of ATTR pathology, but the mechanism is poorly understood. We used proteomics and tissue clearing technology on wild-type (WT) human cardiac (WT/WT) and V122I human cardiac (V122I/WT) tissue to better understand TTR cardiomyopathy. Flash-frozen cardiac tissue slices from human subjects with end-stage WT-TTR cardiomyopathy, end-stage V122I TTR cardiomyopathy, and an age-matched control were used. Fibrils and tissue proteomes were extracted and assessed by bottom-up proteomics. Tissue clearing was performed using a lauryl sulfate-based lipid removal strategy. Slices were stained using indirect immunofluorescence against targets identified by proteomics. TTR deposits were imaged by antibody and AmyTracker 480 staining. Structures of ATTR fibrils were characterized using cryogenic electron microscopy. Proteomic analysis revealed high abundance of TTR, proteins associated with amyloid fibrils, as well as angiogenic, hemostatic, and complement cascade-associated proteins. Three-dimensional imaging revealed loss of normal microvascular architecture, regions of hypervascularization and hypovascularization, and microvascular obstruction by capillary thrombosis. ATTR fibrils adopted the spearhead fold and were decorated with collagen VI, an extracellular matrix component. Based on our imaging and proteomic data, we hypothesize that ATTR cardiomyopathy is a microangiopathy driven by capillary bed thromboinflammation and dysregulated angiogenic revascularization. In this model, increased capillary permeability exposes components of the vascular basement membrane to misfolded TTR. These components promote aggregation and stabilize amyloid fibrils. Congestion of the vascular basement membrane prevents appropriate revascularization, reducing cardiac exertional capacity over time, leading to heart failure.
Ill-defined centrilobular nodules, the most common radiological sign of welding fume (WF) exposure, are also frequently seen in smoking-related respiratory bronchiolitis due to their common pathophysiological features. Thus, smoking is often a confounding factor in the diagnosis of welder's pneumoconiosis. This study aimed to investigate the quantitative computed tomography (QCT) patterns associated with WF exposure, while accounting for the effects of cigarette smoking. Additionally, the relationship between WF exposures and pulmonary function tests (PFTs) was also evaluated as the secondary outcome. This study involved 136 male employees, comprising 66 welders and 70 controls. The WF exposure index was retrospectively estimated by using a semiquantitative exposure assessment, and cigarette pack-years were noted. Two radiologists, blinded to exposure information, performed QCT analysis, an objective and reproducible method. PFT results were also compared. A significant decrease in the histogram's skewness and kurtosis was observed with WF exposure, in contrast to the effects of smoking. In multivariate analysis, the percentage of voxels in the range of -949 to -850 HU decreased with WF exposure (p = 0.037), whereas the percentage of voxels greater than -750 HU increased significantly. Each incremental WF exposure index was associated with a 0.39% decrease in predicted FEV1% (forced expiratory volume in the first second) when adjusted for smoking. We identified functional losses and distinct QCT patterns associated with exposure to WF and cigarette smoke, both of which may cause chronic lung inflammation. These findings may offer valuable insights for further research.
The drivers between host plant, associated rhizosphere microbiome functions, and related plant health implications are complex and a field of continuous development. Furthermore, understanding of the interplay between soil, plant, and microbiome across different plant species and contrasting geographical areas is scarce. The United Kingdom (UK) Crop Microbiome Cryobank project, the world's first open crop/soil microbiome resource can fill this research gap. It utilizes contrasting UK soil types, with comprehensive environmental and agronomic metadata and has generated associated rhizosphere and bulk soil microbiome information for six crops (wheat, barley, oats, fava beans, oilseed rape, and sugar-beet) including a bacterial culture collection and 16S rRNA gene datasets. Here, using functional and taxonomic data from 24 000 bacterial cultures and 315 16S rRNA gene metabarcoded soil libraries, we show that geographical location and soil environment primarily influence the phylogeny of rhizosphere bacterial communities, whereas crop genotype is key in determining the function of associated rhizosphere microbiota. Sugar-beet and oilseed rape predominantly select for drought tolerant microbes, barley for Zn-solubilizing microbes and fava bean has a reduced selection of N-mineralizing microbes. These findings emphasize the need to consider the host plant's developmental requirements and edaphic factors for successful deployment of microbiome facilitated agriculture.
This literature review critically examines the design, validation, and application of non-invasive in-ear electroencephalography (ear-EEG) systems as emerging wearable platforms for long-term neurophysiological monitoring and intervention. Following PRISMA guidelines, studies published between 2010 and 2025 were systematically selected from four major databases and organized into four thematic domains: in-ear wearable system design and validation, multimodal sensing and stimulation, embedded intelligence, and brain-state monitoring and rehabilitation. The review focuses exclusively on wearable, ear-centered EEG technologies, explicitly excluding cochlear implants and other invasive or behind-the-ear systems. We analyze key engineering challenges unique to ear-EEG, including electrode placement constraints, mechanical-electrical coupling, motion robustness, power efficiency, and long-term wearability. The review highlights a growing transition toward compact, wireless ear-EEG systems with on-device signal processing and embedded machine learning, enabling real-time brain-state estimation under ambulatory conditions. Multimodal integration, combining ear-EEG with complementary sensors such as EOG, inertial units, and cardiovascular signals is shown to improve artifact awareness, contextual interpretation, and closed-loop capability. Beyond summarizing existing technologies, this review identifies critical gaps limiting clinical translation, including the lack of standardized validation protocols, limited embedded autonomy, and underexplored closed-loop neurofeedback and neuromodulation architectures. By synthesizing advances across hardware design, signal processing, and intelligent system integration, this work provides a systems-level roadmap for the future development of wearable, intelligent, and clinically robust ear-EEG platforms for mental health, neurorehabilitation, and continuous brain monitoring.
Surgical revascularization, alongside percutaneous intervention, is a viable therapeutic option with satisfactory long-term results. Accumulating evidence indicates that biological sex modulates the predisposition to coronary artery disease and perioperative risk. The study aimed to identify potential long-term prognostic factors after surgical revascularization in patients undergoing off-pump coronary artery bypass (OPCAB) using bilateral mammary arteries. In total, 276 consecutive patients were operated on due to complex stable coronary disease with off-pump surgical revascularization using bilateral mammary arteries. The long-term survival rates, including the Kaplan-Meier survival curve, were compared based on all-cause mortality risk between the female and male populations. Among the 276 analyzed participants, 64 (23%) were deceased, with a median follow-up time of 3,307 days (range: 1,703-5,414 days). The 5-, 10-, and 15-year survival rates in the male and female populations were 90.5% vs. 96.5%, 78.1% vs. 91.4%, and 62.9% vs. 81.3%, respectively. Female sex may be related to superior long-term survival in off-pump revascularization with bilateral mammary artery grafts. While early perioperative risks may be higher in women due to anatomical and referral delay factors, the long-term survival in women may surpass that of men once the immediate postoperative period is successfully navigated. Large-scale studies are required to confirm the suggested association.
Pancreatic adenocarcinoma typically presents late, most often with nonspecific symptoms such as weight loss, jaundice, and abdominal pain. Presentation with prolonged fever and weakness mimicking an infectious etiology is exceedingly rare. We report a case of a 54-year-old man with no known comorbidities who presented with persistent fever, cough, and weakness of 20 days' duration. Initial investigations suggested hepatic abscesses on ultrasonography. However, further evaluation with a contrast-enhanced computed tomography scan revealed a pancreatic uncinate process mass with multiple hepatic and skeletal metastases, suggestive of advanced pancreatic malignancy. Positron-emission tomography computed tomography confirmed the presence of a pancreatic mass with widespread metastatic lesions. Increased serum cancer antigen (CA) 19-9 and CA-125 levels also supported the diagnosis of metastatic pancreatic adenocarcinoma. This case highlights the diagnostic challenge of pancreatic carcinoma presenting atypically as fever with incidentally detected liver abscess-like lesions.
Radiodensity of subcutaneous adipose tissue (SAT), measurable on routine computed tomography (CT), may reflect metabolic status and cachexia, both of which influence cancer outcomes. However, its prognostic role in metastatic non-small cell lung cancer (NSCLC) treated with immune checkpoint inhibitors (ICI) remains unclear. This study aimed to evaluate the prognostic value of SAT radiodensity in this patient population. The retrospective analysis included 92 patients with stage IV NSCLC receiving ICI. Subcutaneous adipose tissue radiodensity (Hounsfield units) was measured from pre-treatment CT at the L3 level and categorized into quartiles. Kaplan- Meier analysis, log-rank test, and Cox proportional hazards models were used. Nonlinear associations were assessed using restricted cubic splines. Cox models were? adjusted for demographic, clinical, and treatment factors. A p-value < 0.05 was considered statistically significant. Median overall survival for Q1, Q2, Q3, and Q4 was 13.4, 26.3, 18.4, and 14.2 months, respectively (log-rank p = 0.0226). Compared with Q1, Q2 showed a significantly reduced mortality risk across all models (fully adjusted hazard ratios = 0.32, 95% CI: 0.15-0.64, p = 0.002). Q3 and Q4 were not significantly different from Q1. Restricted cubic spline analysis revealed a mild U-shaped relationship (p for nonlinearity = 0.0094), with intermediate SAT density linked to best outcomes. Programmed death ligand 1 expression significantly modified the SAT-survival association (p for interaction < 0.0001). Moderate SAT radiodensity was associated with improved survival in metastatic NSCLC patients on ICI, potentially reflecting an optimal metabolic-immune balance. Subcutaneous adipose tissue density, easily obtained from routine imaging, warrants further prospective validation as a scalable prognostic biomarker.