Category: Statistics

Med Image Comput Comput Assist Interv. 2023 Oct;14222:346-356. doi: 10.1007/978-3-031-43898-1_34. Epub 2023 Oct 1.

ABSTRACT

Statistical shape modeling (SSM) enables population-based quantitative analysis of anatomical shapes, informing clinical diagnosis. Deep learning approaches predict correspondence-based SSM directly from unsegmented 3D images but require calibrated uncertainty quantification, motivating Bayesian formulations. Variational information bottleneck DeepSSM (VIB-DeepSSM) is an effective, principled framework for predicting probabilistic shapes of anatomy from images with aleatoric uncertainty quantification. However, VIB is only half-Bayesian and lacks epistemic uncertainty inference. We derive a fully Bayesian VIB formulation and demonstrate the efficacy of two scalable implementation approaches: concrete dropout and batch ensemble. Additionally, we introduce a novel combination of the two that further enhances uncertainty calibration via multimodal marginalization. Experiments on synthetic shapes and left atrium data demonstrate that the fully Bayesian VIB network predicts SSM from images with improved uncertainty reasoning without sacrificing accuracy.

PMID:39503046 | PMC:PMC11536909 | DOI:10.1007/978-3-031-43898-1_34

Front Aging Neurosci. 2024 Oct 21;16:1468602. doi: 10.3389/fnagi.2024.1468602. eCollection 2024.

ABSTRACT

INTRODUCTION: Alzheimer’s disease (AD) is marked by the accumulation of fibrillary aggregates composed of pathological tau protein. Although neuroinflammation is frequently observed in conjunction with tau pathology, current preclinical evidence does not sufficiently establish a direct causal role in tau tangle formation. This study aimed to evaluate whether chronic Toll-like receptor 4 (TLR4) stimulation, induced by a high dose of lipopolysaccharide (LPS, 5 mg/kg), exacerbates neurofibrillary tangle (NFT) pathology in a transgenic mouse model of tauopathy that expresses human truncated 151-391/3R tau, an early feature of sporadic AD.

METHODS: We utilized a transgenic mouse model of tauopathy subjected to chronic TLR4 stimulation via weekly intraperitoneal injections of LPS over nine consecutive weeks. Neurofibrillary tangle formation, microglial activation, and tau hyperphosphorylation in the brainstem and hippocampus were assessed through immunohistochemistry, immunofluorescence, and detailed morphometric analysis of microglia.

RESULTS: Chronic LPS treatment led to a significant increase in the number of Iba-1⁺ microglia in the LPS-treated group compared to the sham group (p < 0.0001). Notably, there was a 1.5- to 1.7-fold increase in microglia per tangle-bearing neuron in the LPS-treated group. These microglia exhibited a reactive yet exhausted phenotype, characterized by a significant reduction in cell area (p < 0.0001) without significant changes in other morphometric parameters, such as perimeter, circumference, solidity, aspect ratio, or arborization degree. Despite extensive microglial activation, there was no observed reduction in tau hyperphosphorylation or a decrease in tangle formation in the brainstem, where pathology predominantly develops in this model.

DISCUSSION: These findings suggest that chronic TLR4 stimulation in tau-transgenic mice results in significant microglial activation but does not influence tau tangle formation. This underscores the complexity of the relationship between neuroinflammation and tau pathology, indicating that additional mechanisms may be required for neuroinflammation to directly contribute to tau tangle formation.

PMID:39503044 | PMC:PMC11536299 | DOI:10.3389/fnagi.2024.1468602

Transl Gastroenterol Hepatol. 2024 Oct 14;9:63. doi: 10.21037/tgh-24-40. eCollection 2024.

ABSTRACT

BACKGROUND: Liver transplantation (LT) is the most effective and radical treatment for hepatocellular carcinoma (HCC). Most LT criteria are based on the morphological characteristics of tumors, which are not enough to predict the risk of tumor recurrence. It is found that some serological biomarkers can predict tumor recurrence and may be a good indicator for selecting suitable HCC patients for LT. This article aims to evaluate the predictive effect of preoperative serological indicators on long-term overall survival (OS) and tumor recurrence-free survival (TFS) of patients with HCC after LT, and to explore its significance for expanding the Milan criteria.

METHODS: Clinical data of 253 patients after LT in HCC were collected retrospectively. The receiver operating characteristic curve was used to calculate the best cut-off value. χ² test was used to analyze the correlation between preoperative serological indicators and tumor pathological features. Univariate and multivariate analyses were used to analyze the risk factors affecting the OS and TFS rates and the predictive values of different LT criteria were compared. Nomogram model was used to predict the OS and TFS rates of patients exceeding Milan criteria.

RESULTS: Independent risk factors for poor OS and TFS rates were alpha-fetoprotein (AFP) >200 ng/mL, gamma-glutamyl transpeptidase (GGT) >80 IU/L, total tumor diameter (TTD) >8 cm and microsatellite lesions. Nomogram model showed patients beyond Milan criteria had better survival when AFP ≤200 ng/mL and GGT ≤80 IU/L or AFP >200 ng/mL, GGT ≤80 IU/L and TTD ≤8 cm. According to Milan criteria, AFP, GGT and TTD, Milan-AFP-GGT-TTD (M-AGT) criteria was established. There was no significant difference in OS and TFS rates among patients in M-AGT, Milan, Hangzhou, Malaya and the University of California at San Francisco (UCSF) criteria.

CONCLUSIONS: Preoperative serological indicators AFP and GGT can effectively predict long-term OS and TFS in HCC patients after LT. Establishing M-AGT criteria based on serological indicators is helpful to supplement the Milan criteria.

PMID:39503026 | PMC:PMC11535812 | DOI:10.21037/tgh-24-40

Transl Gastroenterol Hepatol. 2024 Aug 21;9:72. doi: 10.21037/tgh-24-24. eCollection 2024.

ABSTRACT

BACKGROUND AND OBJECTIVE: Liver transplantation is the gold standard treatment for patients with hepatocellular carcinoma (HCC). Current allocation systems face a complex issue due to the imbalance between available organs and recipients. The prioritization of HCC patients remains controversial, leading to potential disparities in access to transplantation. Factors such as tumor size, alpha-fetoprotein (AFP) levels, Model of End-Stage Liver Disease (MELD) score, and response to locoregional therapy (LRT) contribute to determining waitlist dropout risk in HCC patients. Several statistical and machine learning (ML) models have been proposed to predict waitlist dropout, incorporating variables related to tumor and patient factors, underlying liver disease, and waitlist time. This narrative review aims to summarize the evidence regarding different prediction models of HCC waitlist dropout.

METHODS: All published articles up to December 25, 2023, were considered. Articles not based on prediction models using conventional statistical methods or ML models were excluded.

KEY CONTENT AND FINDINGS: Factors such as tumor size, AFP levels, MELD score, and LRT response have been shown to impact disease progression in these patients, influencing waitlist dropout. Most articles in the literature are based on statistical models. Both ML and statistical models may offer promising results, but their application is currently limited. Several attempts have been made to find the best model to stratify the risk of waitlist dropout in HCC patients. However, to date, none of the explored models have been implemented. The allocation of HCC recipients is still based on supplementary scoring systems or geographical criteria.

CONCLUSIONS: Improving methodology and databases in future research is essential to obtain accurate and reliable models for clinicians. This is the only way to achieve real applicability.

PMID:39503025 | PMC:PMC11535785 | DOI:10.21037/tgh-24-24

Transl Gastroenterol Hepatol. 2024 Aug 9;9:65. doi: 10.21037/tgh-24-14. eCollection 2024.

ABSTRACT

BACKGROUND: Acute pancreatitis (AP) is a complex inflammatory condition with rising incidence globally. Despite various known causes, early diagnosis remains challenging due to limitations in existing biomarkers. Metabolomics offers a promising avenue for identifying novel biomarkers and elucidating underlying pathophysiological mechanisms. Previous AP metabolomics studies primarily focused on analyzing serum, urine, and pancreatic tissues from patients or animal models. However, systematic metabolomics studies that analyze multiple tissues simultaneously are still lacking. The primary aim of our study is to obtain valuable clues to explore the pathophysiological mechanisms of AP and discover novel biomarkers to enable early detection.

METHODS: Using a mouse model of AP induced by cerulein, we conducted gas chromatography-mass spectrometry (GC-MS) metabolomic analysis on serum, pancreas, liver, spleen, colon, and kidney samples. Twelve male C57BL/6J mice were randomly divided into AP and control (CON) groups. Serum and tissue samples were collected, processed, and analyzed using established protocols. Multivariate statistical analysis was employed to identify differential metabolites and impacted metabolic pathways.

RESULTS: Distinct metabolic profiles were observed between AP and CON groups across multiple tissues. Elevated levels of ketone bodies, amino acids, citric acid, and lipids were noted, with significant differences in metabolite levels identified. Notably, 3-hydroxybutyric acid (3-HBA), branched-chain amino acids (BCAAs), phenylalanine, and L-lysine showed consistent alterations, suggesting their potential as early diagnostic biomarkers for AP. Pathway analysis revealed perturbations in several metabolic pathways, providing insights into the pathophysiological mechanisms underlying AP.

CONCLUSIONS: Our study highlights the utility of metabolomics in identifying potential biomarkers for early diagnosis of AP and elucidating associated metabolic pathways. 3-HBA, BCAAs, phenylalanine and L-lysine emerge as promising biomarkers for further clinical validation. These findings contribute to a better understanding of AP pathophysiology and underscore the potential of metabolomics in precision medicine approaches for AP management.

PMID:39503020 | PMC:PMC11535808 | DOI:10.21037/tgh-24-14

MedEdPORTAL. 2024 Nov 5;20:11466. doi: 10.15766/mep_2374-8265.11466. eCollection 2024.

ABSTRACT

INTRODUCTION: Military families face unique stressors beyond civilian life, such as deployments, frequent relocations, and the potential for combat, all of which can significantly impact well-being. A trauma-informed care (TIC) approach to military medicine is paramount; however, a critical gap exists, with no published curricula to guide practitioners in employing TIC in the care of military-connected individuals.

METHODS: We delivered a 50-minute interactive and virtual session to second-year medical students at the Uniformed Services University (USU) that reviewed the neurobiology of adversity and the relevance of TIC in caring for military-connected populations. Participants completed a 14-question pre- and posttest on perceived knowledge, attitudes, practice, and confidence, as well as posttest questions evaluating session quality. The USU Institutional Review Board approved this evaluation.

RESULTS: One hundred sixty medical students participated in the session, with 78 matched pre- and posttest responses. We observed a statistically significant pre-post improvement (p ≤ .05) in all category scores, with the largest changes in knowledge (1.33) and confidence (1.33). On a 5-point Likert scale, with 5 being best, mean scores for overall quality of the session and relevance of the material to participants’ learning and future practice were 3.95 and 4.20, respectively.

DISCUSSION: By equipping health care providers with knowledge and confidence to apply TIC in military medicine, we can improve the well-being of service members and their families across both military and civilian health care settings. Broader implementation of this program has potential to improve patient outcomes and overall health care delivery for this population.

PMID:39503007 | PMC:PMC11534622 | DOI:10.15766/mep_2374-8265.11466

Cureus. 2024 Oct 6;16(10):e70924. doi: 10.7759/cureus.70924. eCollection 2024 Oct.

ABSTRACT

Objective This study aims to evaluate the MRI morphology of corpus callosum malformations and the associated anomalies frequently observed in a tertiary care center in northern India. Methods We conducted a retrospective cross-sectional study using MRI reports, images, and clinical records from January 2020 to July 2024. A total of 19 patients with corpus callosum agenesis or hypoplasia were identified, with 17 patients included after excluding those with incomplete records. We analyzed MRI findings for agenesis type (complete or partial), commissural abnormalities, midline anomalies, cortical abnormalities, and posterior fossa abnormalities. Statistical analysis was performed using chi-square tests to compare associations between complete and partial agenesis. Results Of the 17 patients, 52.9% had complete agenesis and 47.1% had partial agenesis. Complete agenesis was associated with higher rates of commissural involvement (44.4% vs. 12.5%), midline anomalies (22.2% vs. 0%), and Probst bundle formation (88.9% vs. 37.5%). Ventricular distortion was more common in complete agenesis (88.9% vs. 40%), and cortical malformations were also more prevalent (44.4% vs. 12.5%). Other anomalies included holoprosencephaly, Dandy-Walker malformation, and hypoxic-ischemic encephalopathy changes. Conclusion Complete agenesis of the corpus callosum is significantly associated with other commissural abnormalities, midline cysts, Probst bundle formation, ventricular distortions, and cortical malformations compared to partial agenesis. This study highlights the varied imaging presentations of corpus callosum malformations and their associated anomalies.

PMID:39502996 | PMC:PMC11536608 | DOI:10.7759/cureus.70924

Cureus. 2024 Oct 6;16(10):e70914. doi: 10.7759/cureus.70914. eCollection 2024 Oct.

ABSTRACT

One of the most essential elements of providing high-quality palliative care is good symptom control, which is guided through an assessment of the evolution of the patient’s level of distress over time. The Edmonton Symptom Assessment System (ESAS) was created to measure diverse symptoms in palliative care patients. It is a user-friendly, validated, and reliable multi-item instrument. Globally, ESAS is extensively used to direct daily clinical care and foster communication across multidisciplinary teams, ensuring coordinated patient management, facilitating referrals to specialized programs, and evaluating the quality of care. In this systematic review, we aim to evaluate the influence of using ESAS on symptom assessment and control among palliative care patients and settings. We have employed both manual and electronic search strategies among databases to determine relevant studies. This systematic review included original studies published between 2013 and 2023 that implemented the ESAS as part of palliative care for adult patients with terminal illness or advanced disease. Studies recruiting pediatric patients, case reports with limited sample sizes and no descriptive statistics, and nonhuman or laboratory studies were excluded. The ROBINS-I (Risk Of Bias In Non-randomized Studies-of Interventions) tool was used to assess the quality of the included studies, which assessed the methodological quality and potential risk of bias in non-randomized clinical studies. We included eight studies that recruited a total of 3184 patients. Fatigue followed by pain were the symptoms with the highest score on ESAS among all the studies, while the lowest score was recorded for nausea and dyspnea in most of the studies. ESAS scores showed improvement in the follow-up visits. Two of the included studies reported satisfaction with the utilization of the ESAS tool. As stated in our results, it can improve the overall quality of life for patients receiving palliative care by assisting healthcare professionals, family members, and caregivers in methodically assessing and treating symptoms over time. By fostering better communication among care teams and involving caregivers in the process, the use of ESAS promotes a more patient-centered approach to care, ensuring that both the patient’s needs and the perspectives of their loved ones are considered.

PMID:39502990 | PMC:PMC11535720 | DOI:10.7759/cureus.70914

Cureus. 2024 Oct 6;16(10):e70929. doi: 10.7759/cureus.70929. eCollection 2024 Oct.

ABSTRACT

INTRODUCTION: A persistent systemic hyperglycemia as observed in patients with type 2 diabetes mellitus (DM) has been associated with periodontal inflammation. Therefore, the primary objective of the present study was to assess the levels of thromboxane B2 (TXB2) and prostaglandin E2 (PGE2) in the gingival crevicular fluid (GCF) of type 2 DM and non-diabetic patients with chronic periodontitis. The secondary objectives were to correlate these levels with various parameters such as age, gender, glycated hemoglobin (HbA1c) levels, probing depth (PD), and fasting blood sugar levels and to determine the significant predictors for PGE2 and TXB2 levels. Materials and methods: This case-control, cross-sectional study was conducted on 60 patients who were divided into three groups: group 1 (n=20) comprising type 2 DM patients with chronic periodontitis, group 2 (n=20) composed of non-diabetic patients with chronic periodontitis, and group 3 (n=20) as controls comprising of periodontally and systemically healthy individuals. HbA1c, fasting blood glucose levels, PD, and PGE2 and TXB2 levels were checked in GCF for all the patients. The data was subjected to statistical analysis.

RESULTS: The two-way analysis of variance test results revealed statistically significant differences across groups for all parameters. The levels of both biomarkers showed a positive correlation with HbA1c, PD, fasting blood sugar levels, and duration of type 2 DM. Furthermore, PD and fasting blood sugar levels showed the strongest influence on both PGE2 and TXB2 levels. For PGE2, fasting blood sugar levels (p=0.006) and PD (p<0.001) were significant predictors. For TXB2, significant predictors included HbA1c (p=0.003), fasting blood sugar levels (p=0.015), and PD (p<0.001).

CONCLUSION: PGE2 and TXB2 levels were significantly increased in type 2 DM with chronic periodontitis, compared with non-diabetic patients with chronic periodontitis and periodontally healthy patients.

PMID:39502985 | PMC:PMC11537775 | DOI:10.7759/cureus.70929

Cureus. 2024 Oct 6;16(10):e70921. doi: 10.7759/cureus.70921. eCollection 2024 Oct.

ABSTRACT

Introduction Optimisation of patients with type 2 diabetes mellitus (T2DM) prior to metabolic surgery aims to achieve tight glycaemic control by the time of surgery. Little is known about the influence of altering preoperative glycated haemoglobin (HbA1c) on postoperative weight loss and glycaemic control. The aim of this study was to determine whether a change in HbA1c during the preoperative period correlated with long-term weight maintenance and HbA1c in patients undergoing metabolic surgery. The quantity of glucose-lowering medication used prior to and following surgery was also examined. Methods A retrospective analysis was conducted on patients with T2DM who underwent metabolic surgery between 2013 and 2017. Preoperative HbA1c change was measured as a change in glycaemic control during the one-year pre-surgery. The primary outcomes were % excess weight loss (EWL) and HbA1c at five-year post-surgery. Secondary outcomes were % EWL and HbA1c at one-year post-surgery and the use of glucose-lowering medications post-surgery. The Pearson correlation coefficient (r) was used to determine the relationship between the pre-surgery HbA1c change and postoperative % EWL and HbA1c. A chi-squared test was used to calculate the statistical impact of changes in medication use post-surgery. Results Sixty-nine patients with complete data were included in the study. The mean change in HbA1cin the one-year pre-surgery, the one-year post-surgery and five-year post-surgery was -0.9% (1.5), -0.7% (1.2) and 0% (0 1.8), respectively. A change in HbA1cin the one-year pre-surgery did not correlate with % EWL at one-year and five-year post-surgery or with HbA1cat one-year and five-year post-surgery. At one-year and five-year post-surgery, there was a significant decrease in the proportion of patients requiring glucose-lowering medications compared to patient use prior to surgery (p < 0.001). Conclusion This study demonstrated a significant reduction in the proportion of glucose-lowering medication required long-term following metabolic surgery. Altering preoperative glycaemic control was not associated with long-term weight maintenance or glycaemic control.

PMID:39502975 | PMC:PMC11537774 | DOI:10.7759/cureus.70921