Category: Statistics

Alzheimers Dement. 2025 Dec;21 Suppl 2:e097150. doi: 10.1002/alz70856_097150.

ABSTRACT

BACKGROUND: We assessed differences in retinal structure and microvasculature among individuals with Alzheimer’s disease (AD), APOE ε4 carriers, and APOE ε4 noncarriers.

METHOD: Participants underwent cognitive evaluation with Mini-Mental State Examination. Optical coherence tomography (OCT) angiography (Zeiss Cirrus HD-5000 AngioPlex) vessel density (VD) and perfusion density (PD) in the Early Treatment Diabetic Retinopathy Study 3-mm and 6-mm circles and rings were analyzed. OCT angiography peripapillary measurements (capillary perfusion density (CPD), capillary flux index (CFI)) and OCT retinal thickness parameters (retinal nerve fiber layer (RNFL), ganglion cell-inner plexiform layer (GC-IPL), central subfield thickness (CST)) were also compared among groups. Participants with diabetes, glaucoma, or vitreoretinal pathology were excluded. Age and sex-adjusted generalized estimating equations were used for statistical analysis.

RESULT: 275 eyes from 150 AD participants, 373 eyes from 189 cognitively normal APOE ε4 carriers, and 422 eyes from 214 cognitively normal APOE ε4 noncarriers were enrolled. 6mm circle PD and 6mm inner ring VD were significantly lower in APOE ε4 carriers vs noncarriers (p = 0.037 and p = 0.047, respectively). AD did not have significantly different retinal microvasculature measurements vs APOE ε4 carriers except for peripapillary CPD, which was significantly higher in AD vs APOE ε4 carriers or noncarriers (p = 0.003 & p < 0.001, respectively). Comparing AD to noncarriers, CST (p = 0.014), 6mm circle PD (p = 0.025), 6mm circle VD (p = 0.036), and 6mm inner ring VD (p = 0.042) were significantly lower in AD vs APOE ε4 noncarriers. GC-IPL thickness, RNFL thickness, 3mm circle and ring VD and PD, and peripapillary CFI did not differ significantly among or between groups (all p > 0.05).

CONCLUSION: APOE ε4 carriers generally did not have significant differences in retinal microvasculature compared to AD, but showed differences compared to ε4 noncarriers, which may indicate that ε4 carriers have begun to show biomarker signs of AD without clinical symptoms. Peripapillary CPD was significantly higher in AD vs APOE ε4 carriers, which may be due to imaging time point during disease course. Longitudinal studies may better elucidate the role of these potential biomarkers along the AD spectrum.

PMID:41442156 | DOI:10.1002/alz70856_097150

Appl Neuropsychol Child. 2025 Dec 24:1-12. doi: 10.1080/21622965.2025.2606108. Online ahead of print.

ABSTRACT

Attention-deficit/hyperactivity disorder (ADHD) is a common neurodevelopmental disorder, with clinical diagnosis primarily relying on behavioral manifestations and lacking objective biological markers. Electroencephalography (EEG), due to its high temporal resolution and low cost, has been considered to hold great potential for assisting ADHD diagnosis. However, existing methods mostly focus on modeling two-dimensional features such as time-space, frequency-space, or time-frequency representations of EEG, while very few approaches simultaneously capture the joint three-dimensional time-frequency and spatial features. In this paper, we propose a novel deep learning framework-TFS-FENet (Time-Frequency Spatial Feature Extraction Net)-which integrates convolutional neural networks with an attention mechanism to effectively model the time-frequency-spatial characteristics of EEG. Using a resting-state EEG dataset collected at Southeast University, consisting of three categories (ADHD-Inattentive, ADHD-Combined, and typically developing children), experimental results show that TFS-FENet achieves an accuracy of 96.89% in the ADHD three-class classification task, significantly outperforming comparison models such as ShallowConvNet, EEGNet, ResNet18, and DeepConvNet. In the binary classification task (ADHD vs. typically developing children), accuracy further improved to 99.36%. Moreover, interpretability analysis based on channel weights revealed that the model relied more heavily on the prefrontal, temporal, and occipital regions, which is highly consistent with existing neurobiological findings of ADHD.

PMID:41442154 | DOI:10.1080/21622965.2025.2606108

Alzheimers Dement. 2025 Dec;21 Suppl 1:e105561. doi: 10.1002/alz70855_105561.

ABSTRACT

BACKGROUND: Positron emission tomography (PET) radiotracers targeting amyloid-beta (Aβ) plaques, glucose metabolism, and glial reactivity can be used to diagnose and monitor Alzheimer’s disease (AD) progression over time. However, whether animal models recapitulate the temporal progression of AD biomarker abnormalities is still unclear. This study aimed to longitudinally evaluate brain Aβ accumulation, glucose metabolism, and glial responses in a rat model of amyloidosis using PET-imaging.

METHOD: Longitudinal PET imaging with [¹⁸F]FDG, [¹¹C]PK11195, and [¹¹C]PIB was performed at 6, 9, 12, and 15 months of age in wild-type (WT) and TgF344-AD (Tg) rats. Images were manually co-registered to a rat magnetic resonance imaging template. Standardized uptake values (SUV) were calculated for [¹⁸F]FDG and [¹¹C]PK11195, while SUV ratios (SUVR) for [¹¹C]PIB were calculated using the pons as the reference region. Results were normalized using Z-scores, with statistical significance defined as Z > 2 (p < 0.05).

RESULT: No significant differences were observed between wild-type (WT) and transgenic (Tg) animals at 6 months. However, by 9 months, Tg animals exhibited increased amyloid deposition (max Z-value=3.9, entorhinal cortex), enhanced glucose metabolism (max Z-value=5,1, frontal cortex), and heightened glial reactivity (max Z-value=2,7, hippocampus). At 12 months, the levels of Aβ load (max Z-value=5.1, entorhinal cortex) and glucose hypermetabolism (max Z-value= 3.7, frontal cortex) continued to rise, while glial reactivity remained unchanged. By 15 months, Aβ load increased even further (max-Z value=5.9, entorhinal cortex), glucose metabolism stay elevated (max Z-value= 4.9, frontal cortex), and glial reactivity showed another increase (max Z-value=2.7, hippocampus).

CONCLUSION: These findings suggest that during the early stages of Aβ deposition, there is a simultaneous increase in glucose metabolism and glial reactivity. This is followed by a transient decrease in glial reactivity, with a resurgence of glial activation in later stages, while glucose metabolism remains elevated. This supports the hypothesis that amyloid deposition triggers both early and late neuroinflammatory responses. The persistent hypermetabolism observed reinforces that murine models are more resilient to amyloid pathology. These results underscore the value of an imaging platform to study the temporal progression of amyloid pathology in genetically modified rodent models with strong translational relevance.

PMID:41442147 | DOI:10.1002/alz70855_105561

Alzheimers Dement. 2025 Dec;21 Suppl 2:e099404. doi: 10.1002/alz70856_099404.

ABSTRACT

BACKGROUND: Anti-amyloid therapy trials have reported changes in some Alzheimer’s disease (AD) blood biomarkers (BBMs) suggesting that these BBMs may be useful minimally invasive measures of amyloid clearance and treatment outcomes. However, a better understanding of how AD BBMs changes at the individual level reflect treatment response in relation to other relevant clinical endpoints is necessary before they can be adopted for treatment monitoring. Here, several AD BBMs were assessed longitudinally in patients undergoing lecanemab therapy.

METHOD: Patients undergoing lecanemab therapy at Mayo Clinic, Rochester, MN were recruited to participate in this study. Patients were asked to provide an EDTA-plasma sample before treatment initiation during the first infusion visit and thereafter at 3-, 6-, 12-, and 18 months post-treatment. p-tau217, p-tau181, GFAP, NfL, Aβ42, Aβ40, Aβ42/40, and p-tau217/Aβ42 measurements were obtained using Fujirebio Lumipulse assays. Biomarker median differences were assessed between all time points. Patients were grouped into three categories: a decreased biomarker (median decrease of at least 10%), an increased biomarker (median increase of at least 10%), and a stable biomarker (median change < +/-10%.) Biomarker data up to 6 months after treatment is presented.

RESULT: Of forty-one patients enrolled since November 2023, twenty had time points up to 6 months. None of the biomarkers showed a statistically significant median change from baseline at 3 months (p >0.05). At 6 months, only Aβ42/40 showed a statistically significant median change (decrease) between 0 and 6 months (n = 20, p = 0.0058). Table 1 shows patients grouped by median biomarker changes from baseline to 6 months. GFAP, p-tau217, and Aβ42/40 exhibited the greatest number of patients with decreased biomarker concentrations (80%, 60%, and 60% of patients with median decreases of -25%, -26%, and -18% respectively), followed by p-tau181 (50% of patients; median decrease -22%), p-tau217/Aβ42 (45% of patients; median decrease -40%) and Aβ42 (45% of patients; median decrease -18%). None of the patients showed a decreased NfL or Aβ40 concentration.

CONCLUSION: Preliminarily analysis revealed that GFAP, p-tau217, and Aβ42/40 had the most frequent post-lecanemab biomarker decreases. Additional sample collection is ongoing to assess the association of AD BBM changes with Amyloid PET over time.

PMID:41442140 | DOI:10.1002/alz70856_099404

Hosp Pract (1995). 2025 Dec 24:2599079. doi: 10.1080/21548331.2025.2599079. Online ahead of print.

ABSTRACT

BACKGROUND: Posterior spinal fixation (PSF) of the lumbosacral region is a commonly performed procedure for managing various spinal pathologies. Deep vein thrombosis (DVT) is a potential complication that can lead to serious outcomes such as thromboembolism. This study aimed to determine the prevalence of DVT and identify associated risk factors in patients undergoing lumbosacral PSF at Firoozgar Hospital, Tehran.

METHODS: This prospective cohort study included patients who underwent lumbosacral PSF for degenerative diseases or trauma. All participants underwent lower limb color Doppler ultrasonography before surgery to rule out preexisting DVT. Postoperatively, they were monitored for clinical signs of DVT for two weeks and underwent a follow-up Doppler ultrasound. Demographic and clinical data were collected and analyzed using univariate and multivariate statistical methods to identify risk factors associated with DVT.

RESULTS: DVT occurred in 5 of 109 patients (4.6%), of which 3 (2.8%) were symptomatic and 2 (1.8%) asymptomatic on routine postoperative ultrasound. DVT occurrence was significantly associated with factors including motor impairment, neurological deficits, duration of preoperative hospitalization, intraoperative blood loss, and the need for transfusion. Additional factors such as level of consciousness, severity of pain, time to postoperative mobilization, duration of surgery, age, underlying medical conditions, surgical history and cause, number of spinal fusion levels, and BMI also showed significant associations with DVT. No significant correlation was found with gender or preoperative anticoagulant use.

CONCLUSION: Identifying risk factors for DVT in patients undergoing lumbosacral PSF can help inform targeted preventive strategies and improve patient outcomes. These findings underscore the importance of early mobilization, careful perioperative management, and individualized risk assessment in spinal surgery patients.

PMID:41442124 | DOI:10.1080/21548331.2025.2599079

Phys Eng Sci Med. 2025 Dec 24. doi: 10.1007/s13246-025-01687-y. Online ahead of print.

ABSTRACT

Phantom experiments are widely used for standardisation in positron emission tomography (PET), but current practices do not necessarily reflect clinical reality and require meticulous phantom preparation for repeatability. 3D printing can reduce these limitations by optimizing preparatory methods and improving phantom features. This work proposes employing 3D-printed porous grids as an alternative mechanism to emulate targets with contrast. Acrylonitrile butadiene styrene (ABS) cubic grids (4 cm/side) with varying design characteristics and targets were printed. Grids were immersed in a [¹⁸F]FDG solution with soap within a conventional phantom. Five consecutive acquisitions were repeated on five different days (Day 0, 1, 4-6) using a Discovery MI PET/CT. Target representation index (TRI) (analogous to recovery coefficient) and dilution coefficient (DC) were the metrics used for the analysis. Friedman test was utilized for statistical inference across days. PET images resulted in clear demarcation of various contrast regions produced by the dilution grid. Quantitative metrics showed consistent results across trials, confirming robustness. Dilution coefficient achieved (mean ± std. dev.) were 0.55 ± 0.05, 0.41 ± 0.06, and 0.33 ± 0.03 versus 0.5, 0.4 and 0.3 (theoretical), falling within 10% threshold. Observed TRI_{max, mean} were in range of 0.4-1.2. Correlation across days was strong for TRI_{max, mean} (p-values ≥ 0.67) but the DC_max (p-values ~ 0.03) values denoted minor bias in generated contrast due to noise. 3D-printed grids offer a reliable, reproducible alternative for PET/CT assessment. 27 hot targets with varying contrasts and size were produced with a single tracer administration and the metrics stayed stable across different acquisitions.

PMID:41442113 | DOI:10.1007/s13246-025-01687-y

Cardiovasc Interv Ther. 2025 Dec 24. doi: 10.1007/s12928-025-01235-1. Online ahead of print.

ABSTRACT

Drug-coated balloons (DCB) are increasingly utilized for treating in-stent restenosis (ISR), yet the comparative efficacy of limus-coated balloons (LCBs) versus paclitaxel-coated balloons (PCBs) remains uncertain. The aim of this study is to compare the clinical and angiographic outcomes of LCB versus PCBs in treating ISR. We searched PubMed, Embase, and ClinicalTrials.gov through July 2025 for RCTs comparing LCBs versus PCBs in patients with ISR. The primary outcome was late lumen loss (LLL). Secondary outcomes included percentage diameter stenosis (%DS), minimal lumen diameter (MLD), and binary restenosis at 6-12 months and target lesion revascularization (TLR), target lesion failure (TLF), target vessel myocardial infarction, cardiac death, and stent thrombosis at 12 months. Mean differences (MDs) were calculated for continuous outcomes and relative risks (RRs) for binary outcomes. Six RCTs with 968 patients (512 LCB, 456 PCB) showed statistical non-inferiority for LLL with an MD of 0.06 mm (-0.07 to 0.18, P for non-inferiority < 0.001, I² = 65%) based on the prespecified 0.20 mm margin. No significant differences were found in other angiographic outcomes: MD of 3.13 (-1.07 to 7.33, p = 0.14) for %DS, – 0.07 (-0.17 to 0.03, p = 0.15) for MLD, and RR of 1.32 (0.86 to 2.03, p = 0.21) for binary restenosis. Clinical outcomes were comparable with a non-significant trend toward higher TLR (RR: 1.23 [0.87 to 1.75], P = 0.24) and TLF (1.19 [0.88 to 1.63], P = 0.26) in LCB arm. LCBs are statistically non-inferior to PCBs for ISR treatment regarding late lumen loss, with considerable heterogeneity. Given the inconclusiveness of angiographic outcomes and marginally better clinical outcomes in PCBs, the conduct of larger trials seems necessary.

PMID:41442108 | DOI:10.1007/s12928-025-01235-1

Bull Math Biol. 2025 Dec 24;88(1):11. doi: 10.1007/s11538-025-01578-z.

ABSTRACT

Calibrating mathematical models of biological processes is essential for achieving predictive accuracy and gaining mechanistic insight. However, this task remains challenging due to limited and noisy data, significant biological variability, and the computational complexity of the models themselves. In this method’s article, we explore a range of approaches for parameter inference in partial differential equation (PDE) models of biological systems. We introduce a unified mathematical framework, the Correlation Integral Likelihood (CIL) method, for parameter estimation in systems exhibiting heterogeneous or chaotic dynamics, encompassing both pattern formation models and individual-based models. Departing from classical Bayesian inverse problem methodologies, we motivate the development of the CIL method, demonstrate its versatility, and highlight illustrative applications within mathematical biology. Furthermore, we compare stochastic sampling strategies, such as Markov Chain Monte Carlo (MCMC), with deterministic gradient flow approaches, highlighting how these methods can be integrated within the proposed framework to enhance inference performance. Our work provides a practical and theoretically grounded toolbox for researchers seeking to calibrate complex biological models using incomplete, noisy, or heterogeneous data, thereby advancing both the predictive capability and mechanistic understanding of such systems.

PMID:41442086 | DOI:10.1007/s11538-025-01578-z

Cancer Causes Control. 2025 Dec 24;37(1):3. doi: 10.1007/s10552-025-02088-y.

ABSTRACT

PURPOSE: Accurate preoperative differentiation between benign and malignant adnexal masses is essential for guiding optimal surgical management. This study aimed to assess the diagnostic performance, calibration, and clinical utility of serum biomarkers (CA-125, CEA), the O-RADS MRI risk score, and Risk of Malignancy Indices (RMI-I-V) in both premenopausal and postmenopausal women.

METHODS: This retrospective study included data from consecutive patients who underwent surgical management for ovarian masses at a rural tertiary care center in Southern India over 2 years. Preoperative ultrasonography, serum CA-125, CEA levels, and O-RADS MRI risk scores were recorded. RMI-I-V were calculated for each case. Statistical analyses included Receiver Operating Characteristic (ROC) curves, calibration plots, and decision curve analysis to assess discrimination and clinical utility across decision thresholds (5-50%).

RESULTS: A total of 129 women were evaluated-98 (75.9%) had benign, 5 (3.9%) borderline, and 26 (20.2%) malignant ovarian masses. At recommended cut-offs, all RMI models and serum biomarkers significantly differentiated between benign, borderline, and malignant cases. RMI-IV and RMI-V demonstrated the best sensitivity (92.31%), specificity (90.82% and 92.86%), and negative predictive values (97.80% and 97.85%), whereas CEA showed the poorest sensitivity (23.08%). Calibration was most accurate for RMI-V, with RMI-II and RMI-IV also performing well. Decision curve analysis confirmed the highest net clinical benefit for RMI-II and RMI-IV across thresholds of 5-50%.

CONCLUSION: RMI-based models, especially RMI-IV, demonstrated excellent diagnostic accuracy and clinical utility, supporting their use as a reliable, cost-effective tool for adnexal mass evaluation.

PMID:41442084 | DOI:10.1007/s10552-025-02088-y

Drugs Real World Outcomes. 2025 Dec 24. doi: 10.1007/s40801-025-00537-3. Online ahead of print.

ABSTRACT

INTRODUCTION: Opioid use is common in rheumatoid arthritis (RA) for pain management; however, evidence of opioid-associated adverse events is increasing. While biological (b) or targeted synthetic (ts) disease-modifying antirheumatic drugs (DMARDs) improve functional outcomes such as pain, little is known about their impact on opioid utilization patterns. This study investigated opioid utilization before and after b/tsDMARD initiation and assessed effect modification by sex.

METHODS: Using 5% Medicare claims data from 2012 to 2020, this cohort study included older adults (≥ 65 years) with RA who initiated b/tsDMARDs (first prescription = index date), and had continuous Medicare Parts A, B, and D, but not Part C enrollment, during 12 months before and after initiation. The outcomes of interest were any opioid use and long-term opioid therapy (LTOT). McNemar’s test was performed to compare outcomes before and after b/tsDMARD initiation. Sex-based differences in changes in opioid use after b/tsDMARD initiation were also evaluated.

RESULTS: The study cohort included 3585 individuals with RA initiating b/tsDMARDs; most were female (75.87%) with a mean (SD) age of 73.15 (5.99) years. Following b/tsDMARD initiation, any opioid use decreased significantly from 2094 (58.41%) to 2017 (56.26%) (p = 0.015). However, LTOT use increased significantly from 733 (20.45%) to 900 (25.10%) (p < 0.001), following b/tsDMARD initiation. No evidence of sex differences in the association between b/tsDMARD initiation and opioid utilization was identified.

CONCLUSIONS: Initiating b/tsDMARDs was associated with a modest reduction in any opioid use. However, LTOT use in RA remained persistently high. The impact of different b/tsDMARD initiation on opioid utilization patterns needs further investigation.

PMID:41442071 | DOI:10.1007/s40801-025-00537-3