A Staged Approach using Machine Learning and Uncertainty Quantification to Predict the Risk of Hip Fracture.

Hip fractures present a significant healthcare challenge, especially within aging populations, where they are often caused by falls. These fractures lead to substantial morbidity and mortality, emphasizing the need for timely surgical intervention. Despite advancements in medical care, hip fractures impose a significant burden on individuals and healthcare systems. This paper focuses on the prediction of hip fracture risk in older and middle-aged adults, where falls and compromised bone quality are predominant factors. We propose a novel staged model that combines advanced imaging and clinical data to improve predictive performance. By using convolutional neural networks (CNNs) to extract features from hip DXA images, along with clinical variables, shape measurements, and texture features, our method provides a comprehensive framework for assessing fracture risk. The study cohort included 547 patients, with 94 experiencing hip fracture. A staged machine learning-based model was developed using two ensemble models: Ensemble 1 (clinical variables only) and Ensemble 2 (clinical variables and DXA imaging features). This staged approach used uncertainty quantification from Ensemble 1 to decide if DXA features are necessary for further prediction. Ensemble 2 exhibited the highest performance, achieving an Area Under the Curve (AUC) of 0.9541, an accuracy of 0.9195, a sensitivity of 0.8078, and a specificity of 0.9427. The staged model also performed well, with an AUC of 0.8486, an accuracy of 0.8611, a sensitivity of 0.5578, and a specificity of 0.9249, outperforming Ensemble 1, which had an AUC of 0.5549, an accuracy of 0.7239, a sensitivity of 0.1956, and a specificity of 0.8343. Furthermore, the staged model suggested that 54.49% of patients did not require DXA scanning. It effectively balanced accuracy and specificity, offering a robust solution when DXA data acquisition is not always feasible. Statistical tests confirmed significant differences between the models, highlighting the advantages of the advanced modeling strategies. Our staged approach offers a cost-effective holistic view of patients' health. It could identify individuals at risk with a high accuracy but reduce the unnecessary DXA scanning. Our approach has great promise to guide interventions to prevent hip fractures with reduced cost and radiation.

A new method of modeling the multi-stage decision-making process of CRT using machine learning with uncertainty quantification.

Aims: Current machine learning-based (ML) models usually attempt to utilize all available patient data to predict patient outcomes while ignoring the associated cost and time for data acquisition. The purpose of this study is to create a multi-stage machine learning model to predict cardiac resynchronization therapy (CRT) response for heart failure (HF) patients. This model exploits uncertainty quantification to recommend additional collection of single-photon emission computed tomography myocardial perfusion imaging (SPECT MPI) variables if baseline clinical variables and features from electrocardiogram (ECG) are not sufficient. Methods: 218 patients who underwent rest-gated SPECT MPI were enrolled in this study. CRT response was defined as an increase in left ventricular ejection fraction (LVEF) > 5% at a 6±1 month follow-up. A multi-stage ML model was created by combining two ensemble models: Ensemble 1 was trained with clinical variables and ECG; Ensemble 2 included Ensemble 1 plus SPECT MPI features. Uncertainty quantification from Ensemble 1 allowed for multi-stage decision-making to determine if the acquisition of SPECT data for a patient is necessary. The performance of the multi-stage model was compared with that of Ensemble models 1 and 2. Results: The response rate for CRT was 55.5% (n = 121) with overall male gender 61.0% (n = 133), an average age of 62.0±11.8, and LVEF of 27.7±11.0. The multi-stage model performed similarly to Ensemble 2 (which utilized the additional SPECT data) with AUC of 0.75 vs. 0.77, accuracy of 0.71 vs. 0.69, sensitivity of 0.70 vs. 0.72, and specificity 0.72 vs. 0.65, respectively. However, the multi-stage model only required SPECT MPI data for 52.7% of the patients across all folds. Conclusions: By using rule-based logic stemming from uncertainty quantification, the multi-stage model was able to reduce the need for additional SPECT MPI data acquisition without sacrificing performance.

A blueprint for success in real-world evidence: "glocal" approach to building capabilities and generating impactful evidence.

The past decade has seen the increasing influence and relevance of real-world data (RWD) and real-world evidence (RWE) in healthcare decision making. The value added by RWD/RWE has prompted the pharmaceutical industry to develop high performing systems and practices to harness the power of evidence generated at the global level. However, this worldwide transformation provides outstanding opportunities to support capability building within local affiliates and to impact key country-level stakeholders through resulting evidence. Therefore, we present an Evidence Blueprint Initiative, which links the global and local ("glocal") skills, and furthermore addresses the opportunities and gaps in evidence generation capabilities at the local level. Cross-functional experts were recruited at the local, regional, and global level to define best practices. A framework was developed to characterize the foundational expertise needed and to assess markets' existing capabilities. Subsequently, targeted roadmaps were developed and implemented to build capabilities in specific areas within each affiliate. The impact from the Blueprint is encouraging, resulting in improved local evidence plans, established evidence teams, enhanced RWD use and strategic implementation of patient centric science in local affiliates. The success of the Blueprint resides in empowering affiliates to realise their local evidence generation ambitions and to match them to their local context. It strengthens and expands the ties between various parts of the organisation and the external environment while building fit-for-future evidence capabilities from local affiliates.

A woman in her thirties with facial swelling and necrotic lesions. / En kvinne i 30-årene med hevelse og nekrotiske lesjoner i ansiktet.

A woman in her thirties developed multiple necrotic lesions on her face in a short space of time. A dermatologist recognised clinical features associated with a rare diagnosis, which was eventually confirmed by diagnostic workup. Known atopic eczema increased the risk of a severe course of what was initially a mild disease, and a multidisciplinary approach proved necessary.

A machine learning method integrating ECG and gated SPECT for cardiac resynchronization therapy decision support.

PURPOSE: Cardiac resynchronization therapy (CRT) has been established as an important therapy for heart failure. Mechanical dyssynchrony has the potential to predict responders to CRT. The aim of this study was to report the development and the validation of machine learning models which integrate ECG, gated SPECT MPI (GMPS), and clinical variables to predict patients' response to CRT. METHODS: This analysis included 153 patients who met criteria for CRT from a prospective cohort study. The variables were used to model predictive methods for CRT. Patients were classified as "responders" for an increase of LVEF ≥ 5% at follow-up. In a second analysis, patients were classified as "super-responders" for an increase of LVEF ≥ 15%. For ML, variable selection was applied, and Prediction Analysis of Microarrays (PAM) approach was used to model response while Naïve Bayes (NB) was used to model super-response. These ML models were compared to models obtained with guideline variables. RESULTS: PAM had AUC of 0.80 against 0.72 of partial least squares-discriminant analysis with guideline variables (p = 0.52). The sensitivity (0.86) and specificity (0.75) were better than for guideline alone, sensitivity (0.75) and specificity (0.24). Neural network with guideline variables was better than NB (AUC = 0.93 vs. 0.87) however without statistical significance (p = 0.48). Its sensitivity and specificity (1.0 and 0.75, respectively) were better than guideline alone (0.78 and 0.25, respectively). CONCLUSIONS: Compared to guideline criteria, ML methods trended toward improved CRT response and super-response prediction. GMPS was central in the acquisition of most parameters. Further studies are needed to validate the models.

Necrotizing Bacterial Myositis as the Initial Presentation of Severe Aplastic Anaemia.

INTRODUCTION: Necrotizing soft tissue infections are rapidly progressing infections associated with severe inflammation and cytokine release. Early recognition and surgical intervention are key factors to secure survival. The current case presents a patient with multifocal necrotizing soft tissue infection as the initial presentation of severe aplastic anaemia. Case Presentation. A man in his fifties was admitted with septic shock with multiorgan failure and severe pancytopenia, after two days of malaise with high fever and right flank pain. The diagnosis streptococcal necrotizing myositis was significantly delayed due to atypical clinical findings. After initial surgical exploration, the decision was made to defer from surgical debridement due to extensive involvement of several muscle groups, grave pancytopenia, and suspected dismal prognosis. Surprisingly, the patient stabilized after antibiotics and intensive care treatment. Based on severe pancytopenia and hypocellular bone marrow, with no evidence of other bone marrow disorders, the patient was diagnosed with aplastic anaemia. Treatment for aplastic anaemia with antithymocyte globulin, cyclosporine, and eltrombopaq was started, and 2 months later, a partial haematological recovery was observed. The patient could be discharged from hospital without antibiotic treatment. CONCLUSIONS: This case illustrates the crucial role of a multidisciplinary approach on admission and further during the clinical course. Clinical improvement despite severe neutropenia and stabilization during immunosuppressive therapy suggest that immunological factors modulate clinical course in necrotizing soft tissue infections.

Pathological airway remodelling in inflammation.

INTRODUCTION: Airway remodelling refers to a wide pattern of pathophysiological mechanisms involving smooth muscle cell hyperplasia, increase of activated fibroblasts and myofibroblasts with deposition of extracellular matrix. In asthma, it includes alterations of the epithelial cell layer with goblet cell hyperplasia, thickening of basement membranes, peri-bronchial and peri-bronchoalveolar fibrosis. Moreover, airway remodelling occurs not only in asthma but also in several pulmonary disorders such as chronic obstructive pulmonary disease, idiopathic pulmonary fibrosis and systemic sclerosis. Asthma treatment with inhaled corticosteroids does not fully prevent airway remodelling and thus have restricted influence on the natural course of the disease. OBJECTIVES: This review highlights the role of different fibroblast phenotypes and potential origins of these cells in airway remodelling. RESULTS: During inflammatory conditions, such as asthma, fibroblasts can differentiate into an active, more contractile phenotype termed myofibroblast, with expression of stress fibres and alpha-smooth muscle actin. The origin of myofibroblasts has lately been debated, and three sources have been identified: recruitment and differentiation of resident tissue fibroblasts; fibrocytes - circulating progenitor cells; and epithelial-mesenchymal transition. CONCLUSION: It is clear that airway mesenchymal cells, including fibroblasts/myofibroblasts, are more dynamic in terms of differentiation and origin than has previously been recognised. Considering that these cells are key players in the remodelling process, it is of utmost importance to characterise specific markers for the various fibroblast phenotypes and to explore factors that drive the differentiation to develop future diagnostic and therapeutic tools for asthma patients.

BicaudalD actively regulates microtubule motor activity in lipid droplet transport.

BACKGROUND: A great deal of sub-cellular organelle positioning, and essentially all minus-ended organelle transport, depends on cytoplasmic dynein, but how dynein's function is regulated is not well understood. BicD is established to play a critical role in mediating dynein function-loss of BicD results in improperly localized nuclei, mRNA particles, and a dispersed Golgi apparatus-however exactly what BicD's role is remains unknown. Nonetheless, it is widely believed that BicD may act to tether dynein to cargos. Here we use a combination of biophysical and biochemical studies to investigate BicD's role in lipid droplet transport during Drosophila embryogenesis. METHODOLOGY/PRINCIPAL FINDINGS: Functional loss of BicD impairs the embryo's ability to control the net direction of droplet transport; the developmentally controlled reversal in transport is eliminated. We find that minimal BicD expression (near-BicD(null)) decreases the average run length of both plus and minus end directed microtubule (MT) based transport. A point mutation affecting the BicD N-terminus has very similar effects on transport during cellularization (phase II), but in phase III (gastrulation) motion actually appears better than in the wild-type. CONCLUSIONS/SIGNIFICANCE: In contrast to a simple static tethering model of BicD function, or a role only in initial dynein recruitment to the cargo, our data uncovers a new dynamic role for BicD in actively regulating transport. Lipid droplets move bi-directionally, and our investigations demonstrate that BicD plays a critical-and temporally changing-role in balancing the relative contributions of plus-end and minus-end motors to control the net direction of transport. Our results suggest that while BicD might contribute to recruitment of dynein to the cargo it is not absolutely required for such dynein localization, and it clearly contributes to regulation, helping activation/inactivation of the motors.

Specific haptoglobin expression in bronchoalveolar lavage during differentiation of circulating fibroblast progenitor cells in mild asthma.

Haptoglobin is an acute-phase glycoprotein considered to be involved in tissue repair and is produced by fibroblasts and inflammatory cells. By using a gel-based proteomic approach, we show for the first time a possible role for haptoglobin in the differentiation of fibroblast progenitor cells, termed fibrocytes, in patients with mild asthma. Bronchoalveolar lavage fluid (BALF) was performed to sample circulating fibrocytes from patients with mild asthma and nonasthmatic control subjects. Fibrocytes from the airway lumen were characterized by triple staining of the markers CD34/CD45R0/alpha-smooth muscle actin, and subjected to confocal microscopy. The protein expression pattern was analyzed using two-dimensional electrophoresis (2-DE) and matrix-assisted laser desorption ionization time-of-flight (MALDI-TOF). Elevated levels of haptoglobin expression in BALF was reported in a sub-group of patients with mild asthma (p < 0.05) when compared to the other subjects. In addition, this increase in haptoglobin was accompanied by differentiation of fibrocytes into fibroblast-like cells. When further analyzing the expression pattern of haptoglobin isoforms, a heterozygous expression was detected in the patients where fibrocyte differentiation could be observed. These data raise the possibility that an acute and specific inflammatory state facilitates the differentiation of fibroblast progenitor cells into activated fibroblasts. Furthermore, this study proposes a novel role for haptoglobin in airway remodeling in patients with asthma.

Tissue fibrocytes in patients with mild asthma: a possible link to thickness of reticular basement membrane?

BACKGROUND: Myofibroblasts, proposed as being derived from circulating fibrocytes, are considered to be important cells in thickening of the basement membrane in patients with asthma. We have studied the correlation of tissue fibrocyte levels to basement membrane thickness and the presence of fibrocytes in bronchoalveolar lavage fluid (BALF) in steroid-naive patients with mild asthma and controls. METHODS: Patients with mild asthma (n = 9) were recruited and divided into two categories based on whether or not fibroblast-like cells could be established from BALF. Non-asthmatic healthy subjects (n = 5) were used as controls. Colocalization of the fibrocyte markers CD34, CD45RO, procollagen I, and alpha-smooth muscle actin (alpha-SMA) were identified in bronchial biopsies from patients and controls by confocal microscopy. Kruskall-Wallis method was used to calculate statistical significance and Spearman coefficient of rank correlation was used to assess the degree of association. RESULTS: In patients with BALF fibroblasts, a 14-fold increase of tissue cells expressing CD34/CD45RO/alpha-SMA and a 16-fold increase of tissue cells expressing CD34/procollagen I was observed when compared to controls (p < 0.05). In contrast, patients without BALF fibroblasts displayed a 2-fold increase when compared to controls (p < 0.05). Fibrocytes were localized close to the basement membrane which was significantly thicker in patients with BALF fibroblasts when compared to the other two groups of subjects. Furthermore, basement membrane thickness could be correlated to the number of fibrocytes in tissue (r = 0.711). Fibroblasts-like cells were cultured from BALF where 17.6% of these cells expressed CD34, CD45RO and alpha-SMA. CONCLUSION: These findings indicate a correlation between recruited fibrocytes in tissue and thickness of basement membrane. Fibroblast progenitor cells may therefore be important in airway remodeling in steroid-naive patients with mild asthma.

Functional and phenotypical comparison of myofibroblasts derived from biopsies and bronchoalveolar lavage in mild asthma and scleroderma.

BACKGROUND: Activated fibroblasts, which have previously been obtained from bronchoalveolar lavage fluid (BALF), are proposed to be important cells in the fibrotic processes of asthma and scleroderma (SSc). We have studied the motility for BALF derived fibroblasts in patients with SSc that may explain the presence of these cells in the airway lumen. Furthermore, we have compared phenotypic alterations in activated fibroblasts from BALF and bronchial biopsies from patients with mild asthma and SSc that may account for the distinct fibrotic responses. METHODS: Fibroblasts were cultured from BALF and bronchial biopsies from patients with mild asthma and SSc. The motility was studied using a cell migration assay. Western Blotting was used to study the expression of alpha-smooth muscle actin (alpha-SMA), ED-A fibronectin, and serine arginine splicing factor 20 (SRp20). The protein expression pattern was analyzed to reveal potential biomarkers using two-dimensional electrophoresis (2-DE) and sequencing dual matrix-assisted laser desorption ionization time-of-flight mass spectrometry (MALDI-TOF-TOF). The Mann-Whitney method was used to calculate statistical significance. RESULTS: Increased migration and levels of ED-A fibronectin were observed in BALF fibroblasts from both groups of patients, supported by increased expression of RhoA, Rac1, and the splicing factor SRp20. However, these observations were exclusively accompanied by increased expression of alpha-SMA in patients with mild asthma. Compared to BALF fibroblasts in mild asthma, fibroblasts in SSc displayed a differential protein expression pattern of cytoskeletal- and scavenger proteins. These identified proteins facilitate cell migration, oxidative stress, and the excessive deposition of extracellular matrix observed in patients with SSc. CONCLUSION: This study demonstrates a possible origin for fibroblasts in the airway lumen in patients with SSc and important differences between fibroblast phenotypes in mild asthma and SSc. The findings may explain the distinct fibrotic processes and highlight the motile BALF fibroblast as a potential target cell in these disorders.

Presence of activated mobile fibroblasts in bronchoalveolar lavage from patients with mild asthma.

Activated fibroblasts are suggested to be involved in the deposition of extracellular matrix in the formation of peribronchial fibrosis in asthma. We report the novel finding of activated elongated fibroblasts accompanied by elevated numbers of eosinophils in bronchoalveolar lavage fluid from 5 out of 12 patients with mild asthma (= 42%), whereas no fibroblasts were observed in the control subjects without asthma (n = 17). The elongated fibroblasts migrated twice as far when compared with fibroblasts from corresponding bronchial biopsies from the same patients, accompanied by an induced expression of RhoA and Rac1, indicating that the increased expression of these proteins are linked to increased migratory capabilities. Moreover, the elongated fibroblasts had an elevated production of the proteoglycans biglycan, versican, perlecan, and decorin, which correlated to an active cytoplasm in these cells. Differential expression patterns between the two fibroblast groups in motility-regulating proteins, such as cofilin, nuclear chloride ion channel protein, and heat-shock protein 20, were identified by two-dimensional electrophoresis and mass spectrometry. These findings indicate the presence of activated and mobile fibroblasts accompanied by an induced inflammatory response outside the airway epithelium in patients with mild asthma, results that may play a role in formation of airway fibrosis.

Proteome annotations and identifications of the human pulmonary fibroblast.

We hereby report on a three year project initiative undertaken by our research team encompassing large-scale protein expression profiling and annotations of human primary lung fibroblast cells. An overview is given of proteomic studies of the fibroblast target cell involved in several diseases such as asthma, idiopatic pulmonary disease, and COPD. It has been the objective within our research team to map and identify the protein expressions occurring in both activated-, as well as resting cell states. The JGGL database www.2DDB.org has been built around these data, allowing advanced hypothesis building using the interactive query bioinformatic tools developed. Gene ontology has been applied to these annotations, classifying and correlating protein expressions to function. The localization as well as the biological processes involved for the annotations are being presented including an annotation-, and sequence-identification strategy, resulting in close to 2000 protein identities. Both gel based, high resolution 2D-gels, and liquid-phase separation (three-dimensional HPLC), as well as the combination of gel- and LC-based approaches (1D-gels and nano-capillary LC, reversed-phase) were utilized. Protein sequencing and structure identities were acquired by a combination of MALDI-, and electrospray-mass spectrometry techniques. Phenotypical and morphological characterizations were also made for this human disease target cell in both stimulated- and resting-cell states. The use of functional assays that demonstrate the key regulating role of growth factors and cytokine stimuli such as PDGF, TGF-beta, and EGF and the effect of ECM molecules such as Biglycan, are also presented and discussed.

Nanocapillary liquid chromatography interfaced to tandem matrix-assisted laser desorption/ionization and electrospray ionization-mass spectrometry: mapping the nuclear proteome of human fibroblasts.

Miniaturized liquid chromatography nanoseparation in combination with minigel fractionation of human primary cell nuclei is presented. We obtained high-sensitivity and high-throughput identification of expressed proteins by subcellular fractionation and nanocapillary liquid chromatography interfaced to both electrospray ionization (ESI)- and matrix-assisted laser desorption/ionisation (MALDI) tandem mass spectrometry. The reversed-phase nanocapillary eluents were applied directly onto the MALDI target plate as discrete crystal spots using in-line matrix infusion. When working with primary cells, only a limited amount of sample is available. To maximize the number of identified proteins from a restricted amount of sample, miniaturized sample preparation protocols and nanoflow separation is a necessity, especially when working with low-abundant proteins. From the same isolated nuclear sample, complementary separation of intact proteins by two-dimensional (2-D) gel electrophoresis was made. In total 594 gene products from the nuclear preparations were identified out of which 261 were unique. Several proteins involved in transcriptional events were identified such as TATA-binding protein, EBNA-co-activator, and interleukin enhancer binding proteins, indicating that sufficient proteomic depth is obtained to study transcriptional controlling events. Our results suggest that by sample prefractionation and downscaled nanoflow separation along with a combined mass spectrometry strategy, it is possible to identify a large number of nuclear proteins from human primary cells. These findings are of particular importance due to the disease link of these targets cells.

Proteoglycan and proteome profiling of central human pulmonary fibrotic tissue utilizing miniaturized sample preparation: a feasibility study.

The objective of this study was to isolate fibrotic cells from human lung biopsies taken from different central pulmonary locations. A comparison was made of cell morphology, proteoglycan- and protein-expression in mesenchymal cell cultures obtained from human bronchial biopsies from patients with asthmatic-like disorders. We isolated viable cells from 10 out of the 12 biopsies. The fibroblast-like cells were positive for the biomarker a-smooth muscle actin, indicating that the cells were in an activated state. Two different types of fibroblast-like cells were observed from human pulmonary connective tissue; one of contractile type with lamellipodia that facilitate migration and a second cell type with an increased cell size, which most likely is of a synthetic phenotype. This is the first evidence of alterations in the proteoglycan expression pattern of versican, perlecan, biglycan and decorin which can be linked to the pathophysiological state of asthmatics proven by a combination of solid-phase extraction by reversed phase and by peptide mass fingerprinting using matrix-assisted laser desorption/ionization-time of flight mass spectrometry. Protein expression analysis using two-dimensional electrophoresis was interfaced to miniaturized sample preparation techniques using microcapillary extraction. Four protein groups were identified; cytoskeletal, adhesion, scavenger and metabolic proteins. These patient's proteomes showed a high degree of heterogeneity between patients but larger homogeneity within biopsies derived from different locations of the same patient.

A comparison of technologies for database-driven websites for medical education.

Static web pages may be easy to setup using text processors or user-friendly web editing software. However, some basic knowledge of the implementation (HTML) is usually needed for final editing and maintenance. As a result many static web pages are left without appropriate updating. With a database driven, dynamic approach web contents may change through user interaction, the pages are usually easier to maintain, and design elements are separated from contents. Database driven solutions (or content management systems) for larger organizations may be purchased at many different levels of complexity. For smaller systems, such as separate courses or projects, they may, however, be developed locally. The present study compares four alternative technologies for database driven web pages. Each technology offers advantages and disadvantages and many issues need to be considered when selecting platform. The present study demonstrates that database driven web pages may be effectively deployed using free Open Source software. Details may be found at www.intermed.dk/datadriven.