Tag: pubmed

Comput Biol Med. 2025 Apr 17;191:110166. doi: 10.1016/j.compbiomed.2025.110166. Online ahead of print.

ABSTRACT

Early detection of brain tumors in MRI images is vital for improving treatment results. However, deep learning models face challenges like limited dataset diversity, class imbalance, and insufficient interpretability. Most studies rely on small, single-source datasets and do not combine different feature extraction techniques for better classification. To address these challenges, we propose a robust and explainable stacking ensemble model for multiclass brain tumor classification. To address these challenges, we propose a stacking ensemble model that combines EfficientNetB0, MobileNetV2, GoogleNet, and Multi-level CapsuleNet, using CatBoost as the meta-learner for improved feature aggregation and classification accuracy. This ensemble approach captures complex tumor characteristics while enhancing robustness and interpretability. The proposed model integrates EfficientNetB0, MobileNetV2, GoogleNet, and a Multi-level CapsuleNet within a stacking framework, utilizing CatBoost as the meta-learner to improve feature aggregation and classification accuracy. We created two large MRI datasets by merging data from four sources: BraTS, Msoud, Br35H, and SARTAJ. To tackle class imbalance, we applied Borderline-SMOTE and data augmentation. We also utilized feature extraction methods, along with PCA and Gray Wolf Optimization (GWO). Our model was validated through confidence interval analysis and statistical tests, demonstrating superior performance. Error analysis revealed misclassification trends, and we assessed computational efficiency regarding inference speed and resource usage. The proposed ensemble achieved 97.81% F1 score and 98.75% PR AUC on M1, and 98.32% F1 score with 99.34% PR AUC on M2. Moreover, the model consistently surpassed state-of-the-art CNNs, Vision Transformers, and other ensemble methods in classifying brain tumors across individual four datasets. Finally, we developed a web-based diagnostic tool that enables clinicians to interact with the proposed model and visualize decision-critical regions in MRI scans using Explainable Artificial Intelligence (XAI). This study connects high-performing AI models with real clinical applications, providing a reliable, scalable, and efficient diagnostic solution for brain tumor classification.

PMID:40249992 | DOI:10.1016/j.compbiomed.2025.110166

Comput Biol Med. 2025 Apr 17;191:110137. doi: 10.1016/j.compbiomed.2025.110137. Online ahead of print.

ABSTRACT

In this work, we developed deep neural networks for the fast and comprehensive estimation of the most salient features of aortic blood flow. These features include velocity magnitude and direction, 3D pressure, and wall shear stress. Starting from 40 subject-specific aortic geometries obtained from 4D Flow MRI, we applied statistical shape modeling to generate 1,000 synthetic aorta geometries. Complete computational fluid dynamics (CFD) simulations of these geometries were performed to obtain ground-truth values. We then trained deep neural networks for each characteristic flow feature using 900 randomly selected aorta geometries. Testing on remaining 100 geometries resulted in average errors of 3.11% for velocity and 4.48% for pressure. For wall shear stress predictions, we applied two approaches: (i) directly derived from the neural network-predicted velocity, and, (ii) predicted from a separate neural network. Both approaches yielded similar accuracy, with average error of 4.8 and 4.7% compared to complete 3D CFD results, respectively. We recommend the second approach for potential clinical use due to its significantly simplified workflow. In conclusion, this proof-of-concept analysis demonstrates the numerical robustness, rapid calculation speed (less than seconds), and good accuracy of the CFD-based machine learning approach in predicting velocity, pressure, and wall shear stress distributions in subject-specific aortic flows.

PMID:40249990 | DOI:10.1016/j.compbiomed.2025.110137

J Aging Health. 2025 Apr 18:8982643251336801. doi: 10.1177/08982643251336801. Online ahead of print.

ABSTRACT

ObjectivesThis study aims to examine the relationship between social connectedness and health-related quality of life (HRQoL) among older adults, focusing on the impact of technology accessibility and geographic context (urban vs. rural).MethodsData from the 2021 National Health and Aging Trends Study (NHATS) with 2303 participants aged 65 and older were used. Confirmatory factor analysis validated measures of social connectedness and HRQoL, followed by regression analysis to explore their relationship, including the moderating roles of technology accessibility and geographic context.ResultsThe findings indicate a significant positive relationship between social connectedness and HRQoL. However, technology accessibility moderates this relationship only in rural areas, where lower technology access enhances the positive effects of social connectedness on HRQoL.DiscussionThe results suggest that interventions to improve HRQoL among older adults should consider different geographical locations. Notably, promoting in-person interactions is crucial for enhancing the HRQoL of rural older adults.

PMID:40249964 | DOI:10.1177/08982643251336801

Phlebology. 2025 Apr 18:2683555251336447. doi: 10.1177/02683555251336447. Online ahead of print.

ABSTRACT

BackgroundPrevious research has indicated a correlation between chronic venous insufficiency (CVI) and cardiovascular disease. However, whether CVI is an independent risk factor for coronary artery disease (CAD) remains underexplored. This study aimed to investigate this relationship by utilizing coronary artery calcium score (CACS) assessment during CVI screening and comparing it with CACS in patients undergoing cardiac ablation treatment.Methodsand Subjects: A retrospective cohort study was conducted, approved by the ethical committee (IRB 2024012). From February to July 2023 Simultaneous non-contrast lower limb vein CT and CACS measurements were performed on CVI patients aged 50 and above and less than 90, excluding cases with history of heart failure, post-thrombotic syndrome, percutaneous coronary intervention (PCI), cardiac ablation, cardiac surgery, peripheral arterial disease, and renal failure. Parameters included coronary risk factors and CACS. Control group was composed of sex- and age-matched patients receiving cardiac ablation treatment from April 2020 to December 2023. A comparison between the two groups was made, and univariate and multivariate analyses were conducted. Statistical significance was set at p < .05.ResultsComparison between CVI group (n = 234) and cardiac ablation group (n = 234) were as follows: mean age (71 ± 9:71 ± 9, not significant (NS)), females (154:145, NS), body mass index (BMI) (23.6 ± 3.9:22.6 ± 3.5, p = .004), hypertension (103:121, NS), dyslipidemia (100:66, p = .001), diabetes (30:24, NS), Creatinine (0.76 ± 0.25:0.87 ± 0.64, p = .02), respectively. The total CACS was 214 ± 578 in the CVI group and 64.8 ± 233 in the cardiac ablation group (p < .001). The median CACS values were 14.8 (IQR: 0-178) and 0 (IQR: 0-16), respectively. CVI group included 35% with CACS >100 and the cardiac ablation group did 12%, respectively (p < .001). Univariate analysis identified age [beta 9.6 (95% CI 5.2 to 13.9), p < .001], hypertension [beta 142.4 (95% CI 62.5 to 222), diabetes [beta 179.1 (95% CI 53.5 to 305), p = .005], dyslipidemia [beta 164.5 (95% CI 81.2 to 248), p < .001],creatinine [beta 85.4 (95% CI 2.6 to 168), p = .04], and CVI [beta 149 (95% CI 69.4 to 229), p < .001], p = .001] as risk factors. Multivariate analysis revealed age [beta 7.1 (95% CI 2.7 to 11.5), p = .002], hypertension [beta 86.5 (95% CI 1.7 to 171), p = .046], dyslipidemia [beta87.5 (95% CI 1.4 to 174), p = .047], and CVI [beta 143.6 (95% CI 63.7 to 223), p < .001] as strong correlates of CACS.ConclusionsMultivariate analysis indicated that CVI is an independent risk factor for coronary artery disease, even after adjusting for age, hypertension, and dyslipidemia as confounding factors.

PMID:40249957 | DOI:10.1177/02683555251336447

JMIR Form Res. 2025 Apr 18;9:e69113. doi: 10.2196/69113.

ABSTRACT

BACKGROUND: Acute respiratory infections (ARI) are a significant challenge in military settings due to close communal living, which facilitates the rapid transmission of pathogens. A variety of respiratory pathogens contribute to ARI, each varying in prevalence, severity, and impact on organizational productivity. Understanding and mitigating the impact of ARI is critical for optimizing the health of military personnel and maintaining organizational productivity.

OBJECTIVE: This retrospective study of surveillance data aims to identify pathogens causing ARI among servicemen and determine which pathogens contribute most to medical absenteeism, defined as the combined duration of the issued medical certificate and light duty.

METHODS: From September 2023 to August 2024, anonymous nasopharyngeal swabs (BioFire FilmArray Respiratory Panel) were collected from Singapore Armed Forces servicemen presenting with ARI symptoms after a doctor’s consultation at a local military camp’s medical centre. The presence of fever and duration of medical certificate and light duty were self-reported by Singapore Armed Forces servicemen.

RESULTS: A total of 1095 nasopharyngeal swabs were collected, of which 608 (55.5%) tested positive. The most common respiratory pathogen was human rhinovirus/enterovirus (HRV/HEV) in 303 (27.7%) individuals. The highest proportions of fever were observed in servicemen with influenza (62.8%, 27/43), SARS-CoV-2 (34.3%, 12/35), and parainfluenza (31.6%, 12/38). The odds of patients with influenza that have fever was 5.8 times higher than those of patients infected with HRV/HEV (95% CI 2.95-11.40, P<.001). The median duration of medical certificate, light duty, and medical absenteeism were 0 (IQR 0), 2 (IQR 2) and 2 (IQR 0) days, respectively. The odds of patients with influenza having a medical certificate with duration ≥1 day was 5.34 times higher than those in patients with HRV/HEV (95% CI 2.63-10.88, P<.001). No significant differences in the duration of medical absenteeism were found between HRV/HEV and other pathogens.

CONCLUSIONS: Compared to HRV/HEV, influenza infections were significantly associated with longer medical certificate duration. Nonetheless, there were no significant differences in the overall duration of medical absenteeism across pathogens, as servicemen infected with other pathogens were given light duty instead. These findings emphasize the need for pathogen-agnostic ARI measures. While influenza vaccinations are already mandatory for servicemen in local military camps, encouraging additional public health measures (eg, mask-wearing among symptomatic servicemen, COVID-19 vaccinations, therapeutics) can further reduce ARI incidence, minimize the duration of medical absenteeism, and mitigate the impact on organizational productivity.

PMID:40249956 | DOI:10.2196/69113

J Am Acad Orthop Surg. 2025 Apr 15. doi: 10.5435/JAAOS-D-25-00011. Online ahead of print.

ABSTRACT

INTRODUCTION: Tibiotalocalcaneal (TTC) arthrodesis is a critical surgical intervention for advanced hindfoot and ankle pathologies, offering pain relief, stabilization, and functional alignment restoration. Intramedullary nail fixation, particularly with dynamic compression (DC) nails, has emerged as a promising solution for addressing high nonunion rates associated with standard static compression (SC) nails. This study compares union and complication rates between DC and SC nails in TTC arthrodesis using the Nationwide Readmissions Database.

METHODS: This retrospective cohort study used the Nationwide Readmissions Database to identify cases of TTC fusion with DC and SC nails based on ICD-10-PCS codes. Propensity score matching (1:1) controlled for confounders, including age, sex, and comorbidities. Primary outcomes included complications such as thromboembolism, wound dehiscence, cellulitis, implant-related complications, nonunion, malunion, and infections. Secondary outcomes included 30-day and 31-90-day readmission rates. Statistical significance was set at P < 0.05.

RESULTS: The study analyzed 311 cases (149 with DC, 162 with SC). Demographic and comorbidity distributions were balanced after matching. Nonunion rates were significantly lower in the DC group (6.0%) compared with the SC group (17.3%; P = 0.002). Overall complication rates were comparable (DC: 30.2% vs. SC: 35.2%, P = 0.350).

DISCUSSION: DC devices demonstrated markedly reduced nonunion rates compared with SC nails, likely because of the continuous compression provided by the nitinol-based design. This novel finding validates the biomechanical advantages of devices using DC in TTC fusion and aligns with previous research advocating for such devices.

CONCLUSION: DC nails offer an advancement in TTC arthrodesis by markedly reducing nonunion rates. Future studies should focus on cost-effectiveness, long-term outcomes, and patient-specific optimization to further refine treatment protocols.

PMID:40249946 | DOI:10.5435/JAAOS-D-25-00011

J Med Internet Res. 2025 Apr 18;27:e71057. doi: 10.2196/71057.

ABSTRACT

BACKGROUND: Knee osteoarthritis (OA) presents a significant burden in China due to its high prevalence, aging population, and rising obesity rates. Despite clinical guidelines recommending evidence-based care, limited practitioner training and inadequate telehealth integration hinder effective OA management.

OBJECTIVE: The aim of this study was to evaluate the effectiveness of an e-learning program in improving the confidence and knowledge of Chinese physical therapists in managing knee OA and to explore their perceptions of the program.

METHODS: This was a randomized controlled trial with 2 parallel arms involving 81 rehabilitation practitioners from 18 Chinese provinces. The intervention group completed a 4-week web-based training program (Physiotherapy Exercise and Physical Activity for Knee Osteoarthritis [PEAK]-Chinese), while the control group received no training. The primary outcome was self-reported confidence in OA management (11-point scale). Secondary outcomes included knowledge (Chinese Knee Osteoarthritis Knowledge Scale [KOAKS]) and likelihood of clinical application of core OA treatments. Process measures and semistructured interviews captured participants’ training perceptions. Quantitative data were analyzed using regression models, 2-sided t tests, and descriptive statistics, while thematic analysis was performed on the interview data of 10 participants.

RESULTS: A total of 80 participants completed the outcome measures at 4 weeks. The intervention group demonstrated significant improvements in confidence compared to the control group, including managing OA with exercise-based programs (adjusted mean difference=3.27, 95% CI 2.72-3.81), prescribing exercise (adjusted mean difference=3.13, 95% CI 2.55-3.72), and delivering telehealth (adjusted mean difference=4.41, 95% CI 3.77-5.05). KOAKS scores also improved significantly (mean change=9.46); however, certain belief bias related to OA concepts and the use of scans remained unchanged (25/41, 61% and 27/41, 66%, respectively). Approximately 73% (30/41) of the intervention participants rated the course as extremely useful. Interviews emphasized the need for cultural adaptation and practical telehealth training with real-life scenarios to enhance program applicability.

CONCLUSIONS: The PEAK program improved Chinese practitioners’ confidence and knowledge in managing knee OA, underscoring e-learning’s potential to support evidence-based OA care in China. To optimize future implementations, further research strategies could include enhancing cultural relevance, addressing misconceptions, and incorporating practical, real-world training.

TRIAL REGISTRATION: Chinese Clinical Trial Register ChiCTR2400091007; https://www.chictr.org.cn/showproj.html?proj=239680.

PMID:40249943 | DOI:10.2196/71057

JMIR Form Res. 2025 Apr 18;9:e66791. doi: 10.2196/66791.

ABSTRACT

BACKGROUND: Smartphone apps are a growing field supporting the prevention of chronic diseases. The user experience (UX) is an important predictor of app use and should be considered in mobile health research. Long-term skin protection behavior is important for those with work-related skin diseases. However, altering health behavior is complex and requires a high level of self-management. We developed a maintenance program consisting of the Mein Hautschutz im Alltag (MiA; “My skin protection in everyday life”) app combined with an individual face-to-face goal-setting interview to support patients in the implementation of skin protection behavior after inpatient rehabilitation.

OBJECTIVE: The objectives of this paper are to (1) describe the intervention in a standardized manner; (2) evaluate the UX, subjective quality, and perceived impact of the MiA app; and (3) evaluate the adherence to the MiA app.

METHODS: We followed a user-centered and multistage iterative process in 2 steps that combined qualitative and quantitative data. The maintenance program was tested over 12 weeks after discharge from rehabilitation. The UX, subjective quality, and perceived impact were evaluated formatively based on the user version of the Mobile Application Rating Scale after 12 weeks (T2). Adherence was measured using the frequency of interactions with the app.

RESULTS: In total, 42 patients took part (with a dropout rate of n=18, 43% at T2). The average age was 49.5 (SD 13.1) years, and 57% (24/42) were male. We found high ratings for the UX, with an average score of 80.18 (SD 8.94) out of a theoretical maximum of 100, but there were a few exceptions in the usability and interaction with the app. The app was most frequently rated with 4 out of 5 stars (15/24, 65%), which indicates a high subjective quality. Furthermore, the app seemed to influence important determinants to implement skin protection behavior. Adherence to skin protection tracking was higher over the study period than adherence to skin documentation and goal assessment. The number of adherent participants to skin protection tracking was higher in the skin care and skin cleansing categories (28/42, 67% each) compared to the skin protection category (13/42, 31%) on day 1 and decreased until day 84 in all dimensions (12/42, 29% each for skin care and skin cleansing; 9/42, 21% for skin protection).

CONCLUSIONS: The results in terms of adherence met the expectations and were consistent with those of other studies evaluating the use of apps for chronic diseases. Interaction with the app could be increased using artificial intelligence to determine eczema severity via photos. It should be investigated which subgroups have difficulties with usability to individualize the support to a greater degree during onboarding. There is a need for further research regarding the effectiveness of the MiA app on skin protection behavior, quality of life, and eczema severity.

PMID:40249942 | DOI:10.2196/66791

Blood Adv. 2025 Apr 18:bloodadvances.2024015734. doi: 10.1182/bloodadvances.2024015734. Online ahead of print.

ABSTRACT

The prognostic impact of monocytic differentiation in AML patients receiving Venetoclax (Ven) and azacitidine (Aza) remains unclear. In a prospective cohort of 86 newly diagnosed AML patients treated with Ven-Aza, we used multiparametric flow cytometry (MFC) to define mono-blasts as AML blasts co-expressing ≥2 monocytic markers (CD4, CD36, CD64) per ELN guidelines. Patients with higher mono-blasts/CD45+ proportions had lower complete response rates (OR=0.24, p=0.005) and significantly shorter overall survival (OS, 4.0 versus 14.9 months, p=0.003). A ≥10% mono-blasts/CD45+ threshold, identified via maximally selected rank statistics, stratified patients into mono-blasthigh (≥10%) and mono-blastlow (<10%) groups. MFC reclassified 20% of FAB non-M4/5 and 15% of FAB M4/5 cases into mono-blasthigh and mono-blastlow groups, respectively. Multivariable analysis confirmed mono-blasthigh status as an independent adverse prognostic factor for OS (HR=1.95, p=0.023), with a particularly strong impact in ELN 2024 favorable-risk patients (HR=2.81, p=0.024). Our findings highlight monocytic differentiation, assessed via MFC, as a key predictor of Ven-Aza resistance and poor survival, independent of genetic classification. Given its availability in routine diagnostics, MFC-based monocytic assessment could improve AML risk stratification and treatment decisions in patients eligible for less intensive therapies.

PMID:40249917 | DOI:10.1182/bloodadvances.2024015734

Rhinology. 2025 Apr 18. doi: 10.4193/Rhin24.420. Online ahead of print.

ABSTRACT

BACKGROUND: The aim of this study was to evaluate the efficacy of mepolizumab as add-on therapy to intranasal corticosteroids for the treatment of severe, uncontrolled Chronic Rhinosinusitis with Nasal Polyps (CRSwNP) in a real-life setting in the Triveneto region of northeast Italy.

METHODS: Patients with severe CRSwNP receiving mepolizumab were followed up at 1, 3, 6, 9 and 12 months from the first administration. At baseline and at each follow-up, patients underwent nasal endoscopy, completed the Sinonasal Outcome Test 22, Visual Analogue Scales for smell, nasal obstruction, rhinorrhoea and facial pain, the Nasal Congestion Score and the Asthma Control Test. Peak nasal inspiratory flow, Sniffin’ Sticks Identification Test and blood eosinophil count were also evaluated.

RESULTS: Ninety patients from twelve different rhinological units were enrolled in the study. Both patient- and physician- derived outcome measures significantly improved within the first month after biological treatment initiation, maintaining the benefit at subsequent follow-ups. Nine percent of patients discontinued the treatment due to lack of benefit within the first year. No major adverse events were reported.

CONCLUSIONS: Mepolizumab is effective in improving nasal obstruction and the sense of smell in patients with severe uncontrolled CRSwNP, based on both patient- and physician derived outcome measures.

PMID:40249916 | DOI:10.4193/Rhin24.420