Predicting Visit Cost of Obstructive Sleep Apnea Using Electronic Healthcare Records With Transformer.

BACKGROUND: Obstructive sleep apnea (OSA) is growing increasingly prevalent in many countries as obesity rises. Sufficient, effective treatment of OSA entails high social and financial costs for healthcare. OBJECTIVE: For treatment purposes, predicting OSA patients' visit expenses for the coming year is crucial. Reliable estimates enable healthcare decision-makers to perform careful fiscal management and budget well for effective distribution of resources to hospitals. The challenges created by scarcity of high-quality patient data are exacerbated by the fact that just a third of those data from OSA patients can be used to train analytics models: only OSA patients with more than 365 days of follow-up are relevant for predicting a year's expenditures. METHODS AND PROCEDURES: The authors propose a translational engineering method applying two Transformer models, one for augmenting the input via data from shorter visit histories and the other predicting the costs by considering both the material thus enriched and cases with more than a year's follow-up. This method effectively adapts state-of-the-art Transformer models to create practical cost prediction solutions that can be implemented in OSA management, potentially enhancing patient care and resource allocation. RESULTS: The two-model solution permits putting the limited body of OSA patient data to productive use. Relative to a single-Transformer solution using only a third of the high-quality patient data, the solution with two models improved the prediction performance's [Formula: see text] from 88.8% to 97.5%. Even using baseline models with the model-augmented data improved the [Formula: see text] considerably, from 61.6% to 81.9%. CONCLUSION: The proposed method makes prediction with the most of the available high-quality data by carefully exploiting details, which are not directly relevant for answering the question of the next year's likely expenditure. Clinical and Translational Impact Statement: Public Health- Lack of high-quality source data hinders data-driven analytics-based research in healthcare. The paper presents a method that couples data augmentation and prediction in cases of scant healthcare data.

Evolution of sleep-disordered breathing and blood pressure during menopausal transition.

The purpose of this study was to investigate how the blood pressure increase observed during menopausal transition is affected by sleep-disordered breathing and the menopause itself. Further, we aimed to find new sleep-disordered breathing related markers that would predict the development of hypertension. Sixty-four community-dwelling premenopausal women aged 45-47 years were studied. Polysomnography, serum follicle stimulating hormone, forced expiratory volume in 1 s, and a physical examination were performed at baseline and again after 10 years of follow-up. Indices for sleep apnea/hypopnea and inspiratory flow-limitation were determined. Regression models were used to study the relationships between variables. Changes in the apnea-hypopnea index or serum follicle stimulating hormone were not significant for blood pressure change. An increase in morning blood pressure during the follow-up period was associated with a body mass-index increase. An increase in evening blood pressure was associated with an increase in inspiratory flow-limitation during non-rapid eye movement sleep. Incident hypertension during the follow-up was associated with hypopnea (median hypopnea index 7.6/h, p = 0.048) during rapid eye movement sleep at baseline. Users of menopausal hormone therapy had a lower rapid eye movement sleep apnea-hypopnea index (1.6/h vs. 6.9/h, p = 0.026) at baseline whereas at follow-up users and non-users did not differ in any way. The progression of menopause or the use of menopausal hormone therapy had a minimal effect on blood pressure in our population. The effects of inspiratory flow-limitation on blood pressure profile should be studied further.

Automated detection of pulmonary embolism from CT-angiograms using deep learning.

BACKGROUND: The aim of this study was to develop and evaluate a deep neural network model in the automated detection of pulmonary embolism (PE) from computed tomography pulmonary angiograms (CTPAs) using only weakly labelled training data. METHODS: We developed a deep neural network model consisting of two parts: a convolutional neural network architecture called InceptionResNet V2 and a long-short term memory network to process whole CTPA stacks as sequences of slices. Two versions of the model were created using either chest X-rays (Model A) or natural images (Model B) as pre-training data. We retrospectively collected 600 CTPAs to use in training and validation and 200 CTPAs to use in testing. CTPAs were annotated only with binary labels on both stack- and slice-based levels. Performance of the models was evaluated with ROC and precision-recall curves, specificity, sensitivity, accuracy, as well as positive and negative predictive values. RESULTS: Both models performed well on both stack- and slice-based levels. On the stack-based level, Model A reached specificity and sensitivity of 93.5% and 86.6%, respectively, outperforming Model B slightly (specificity 90.7% and sensitivity 83.5%). However, the difference between their ROC AUC scores was not statistically significant (0.94 vs 0.91, p = 0.07). CONCLUSIONS: We show that a deep learning model trained with a relatively small, weakly annotated dataset can achieve excellent performance results in detecting PE from CTPAs.

A relational database to identify differentially expressed genes in the endometrium and endometriosis lesions.

Endometriosis is a common inflammatory estrogen-dependent gynecological disorder, associated with pelvic pain and reduced fertility in women. Several aspects of this disorder and its cellular and molecular etiology remain unresolved. We have analyzed the global gene expression patterns in the endometrium, peritoneum and in endometriosis lesions of endometriosis patients and in the endometrium and peritoneum of healthy women. In this report, we present the EndometDB, an interactive web-based user interface for browsing the gene expression database of collected samples without the need for computational skills. The EndometDB incorporates the expression data from 115 patients and 53 controls, with over 24000 genes and clinical features, such as their age, disease stages, hormonal medication, menstrual cycle phase, and the different endometriosis lesion types. Using the web-tool, the end-user can easily generate various plot outputs and projections, including boxplots, and heatmaps and the generated outputs can be downloaded in pdf-format.Availability and implementationThe web-based user interface is implemented using HTML5, JavaScript, CSS, Plotly and R. It is freely available from https://endometdb.utu.fi/ .

ePCR: an R-package for survival and time-to-event prediction in advanced prostate cancer, applied to real-world patient cohorts.

Motivation: Prognostic models are widely used in clinical decision-making, such as risk stratification and tailoring treatment strategies, with the aim to improve patient outcomes while reducing overall healthcare costs. While prognostic models have been adopted into clinical use, benchmarking their performance has been difficult due to lack of open clinical datasets. The recent DREAM 9.5 Prostate Cancer Challenge carried out an extensive benchmarking of prognostic models for metastatic Castration-Resistant Prostate Cancer (mCRPC), based on multiple cohorts of open clinical trial data. Results: We make available an open-source implementation of the top-performing model, ePCR, along with an extended toolbox for its further re-use and development, and demonstrate how to best apply the implemented model to real-world data cohorts of advanced prostate cancer patients. Availability and implementation: The open-source R-package ePCR and its reference documentation are available at the Central R Archive Network (CRAN): https://CRAN.R-project.org/package=ePCR. R-vignette provides step-by-step examples for the ePCR usage. Supplementary information: Supplementary data are available at Bioinformatics online.

Evaluation of tobacco (Nicotiana tabacum L. cv. Petit Havana SR1) hairy roots for the production of geraniol, the first committed step in terpenoid indole alkaloid pathway.

The terpenoid indole alkaloids are one of the major classes of plant-derived natural products and are well known for their many applications in the pharmaceutical, fragrance and cosmetics industries. Hairy root cultures are useful for the production of plant secondary metabolites because of their genetic and biochemical stability and their rapid growth in hormone-free media. Tobacco (Nicotiana tabacum L. cv. Petit Havana SR1) hairy roots, which do not produce geraniol naturally, were engineered to express a plastid-targeted geraniol synthase gene originally isolated from Valeriana officinalis L. (VoGES). A SPME-GC-MS screening tool was developed for the rapid evaluation of production clones. The GC-MS analysis revealed that the free geraniol content in 20 hairy root clones expressing VoGES was an average of 13.7 µg/g dry weight (DW) and a maximum of 31.3 µg/g DW. More detailed metabolic analysis revealed that geraniol derivatives were present in six major glycoside forms, namely the hexose and/or pentose conjugates of geraniol and hydroxygeraniol, resulting in total geraniol levels of up to 204.3 µg/g DW following deglycosylation. A benchtop-scale process was developed in a 20-L wave-mixed bioreactor eventually yielding hundreds of grams of biomass and milligram quantities of geraniol per cultivation bag.

Transcutaneous CO(2) plateau as set-point for respiratory drive during upper airway flow-limitation.

Upper airway flow-limitation is often but not always associated with prolonged gradually increasing respiratory effort. We investigated the changes in transcutaneous carbon dioxide tension (tcCO(2)) during episodes of upper airway flow limitation during sleep with or without respiratory effort response. Seventy-seven episodes of progressive flow-limitation were analyzed in 36 patients with sleep-disordered breathing. TcCO(2) and arterial oxyhaemoglobin saturation (SaO2) were measured during steady breathing and during episodes of flow-limitation with and without effort response. After lights-off tcCO(2) increased and leveled-off at plateau, when breathing stabilized. During flow-limitation tcCO(2) increased at rate of 4.0kPa/h. Flow-limitation with increasing respiratory effort associated with tcCO(2) increase above the plateau (terminating at 105.2%, p<0.001), whereas flow-limitation without effort response associated with tcCO(2) increase starting below the plateau (95.8%, p<0.001). We conclude that the nocturnal tcCO(2) plateau indicates the level above which the increasing respiratory effort is triggered as response to upper airway flow-limitation. We propose that flow-limitation below the tcCO(2) plateau is an event related to stabilization of sleep and breathing.

Design and application of a generic clinical decision support system for multiscale data.

Medical research and clinical practice are currently being redefined by the constantly increasing amounts of multiscale patient data. New methods are needed to translate them into knowledge that is applicable in healthcare. Multiscale modeling has emerged as a way to describe systems that are the source of experimental data. Usually, a multiscale model is built by combining distinct models of several scales, integrating, e.g., genetic, molecular, structural, and neuropsychological models into a composite representation. We present a novel generic clinical decision support system, which models a patient's disease state statistically from heterogeneous multiscale data. Its goal is to aid in diagnostic work by analyzing all available patient data and highlighting the relevant information to the clinician. The system is evaluated by applying it to several medical datasets and demonstrated by implementing a novel clinical decision support tool for early prediction of Alzheimer's disease.

The temporal relationship between growth hormone and slow wave sleep is weaker after menopause.

OBJECTIVE: To study the temporal association between growth hormone (GH) and slow wave sleep (SWS) in middle-aged women. METHODS: Seventeen premenopausal and 18 postmenopausal women were studied using all-night polygraphic sleep recordings and blood sampling at 20-min intervals. The postmenopausal women were re-studied after six months on hormone therapy (HT) according to a randomized, double-blind, placebo-controlled protocol. RESULTS: The total sleep time (premenopausal 361.9±81.5 min, postmenopausal 358±67.7 min) and the percentages of the sleep stages did not differ between pre- and postmenopausal women. In postmenopausal women the first GH peak after sleep onset occurred later and with a more variable time interval compared to premenopausal women. The percentage of SWS was highest 40-20 min prior to the first GH peak after sleep onset in both groups with a higher SWS proportion in premenopausal women (p=0.048), although the total SWS percent for night did not differ. HT did not affect the distribution of SWS in postmenopausal women. CONCLUSIONS: The temporal relationship between GH and SWS in premenopausal women is less robust after menopause and is not improved with HT.

A disease state fingerprint for evaluation of Alzheimer's disease.

Diagnostic processes of Alzheimer's disease (AD) are evolving. Knowledge about disease-specific biomarkers is constantly increasing and larger volumes of data are being measured from patients. To gain additional benefits from the collected data, a novel statistical modeling and data visualization system is proposed for supporting clinical diagnosis of AD. The proposed system computes an evidence-based estimate of a patient's AD state by comparing his or her heterogeneous neuropsychological, clinical, and biomarker data to previously diagnosed cases. The AD state in this context denotes a patient's degree of similarity to previously diagnosed disease population. A summary of patient data and results of the computation are displayed in a succinct Disease State Fingerprint (DSF) visualization. The visualization clearly discloses how patient data contributes to the AD state, facilitating rapid interpretation of the information. To model the AD state from complex and heterogeneous patient data, a statistical Disease State Index (DSI) method underlying the DSF has been developed. Using baseline data from the Alzheimer's Disease Neuroimaging Initiative (ADNI), the ability of the DSI to model disease progression from elderly healthy controls to AD and its ability to predict conversion from mild cognitive impairment (MCI) to AD were assessed. It was found that the DSI provides well-behaving AD state estimates, corresponding well with the actual diagnoses. For predicting conversion from MCI to AD, the DSI attains performance similar to state-of-the-art reference classifiers. The results suggest that the DSF establishes an effective decision support and data visualization framework for improving AD diagnostics, allowing clinicians to rapidly analyze large quantities of diverse patient data.

Understanding sleep-disordered breathing through mathematical modelling.

Recent studies have uncovered high prevalence of undiagnosed sleep-disordered breathing, and its linkage to metabolic or cardiovascular disorders which represent increasing health hazard. However, the mechanistic links behind these disorders as well as their contribution to the experimental observations and treatment responses remain poorly understood. Therefore, the screening of clinical measurements still relies upon relatively simple diagnostic features, such as signal averages or event frequencies, which may represent suboptimal or surrogate markers of the underlying abnormality. Consequently, most patients are being treated with general therapies regardless of the cause of their key dysfunction. Combining experimental measurements with mathematical modelling has the potential to provide mechanistic insights into the individual factors underlying the disease progression, which may finally enable tailored treatment alternatives for each patient. This review depicts a number of modelling approaches to elucidate sleep-related dysfunctions of the human respiratory system, and how these models are being used to translate the measurements first into new ideas and then into testable hypotheses. Such model-based investigations can provide systematic strategies towards better understanding, predicting or even preventing these dysfunctions. Along with the brief description of the modelling approaches, we discuss their relative merits and potential implications especially for clinical research.

24-hour serum levels of growth hormone, prolactin, and cortisol in pre- and postmenopausal women: the effect of combined estrogen and progestin treatment.

OBJECTIVE: Our objective was to study the 24-h profiles of GH, prolactin (PRL), and cortisol concentrations in older postmenopausal and middle-aged premenopausal women, before and after estrogen-progestin treatment (EPT). DESIGN: The study was a randomized, placebo-controlled, double-blind trial. GH, PRL, and cortisol were sampled every 20 min for 24 h in 18 postmenopausal (aged 58-70 yr) and 17 premenopausal (aged 45-51 yr) women before and after 6 months of EPT. RESULTS: The mean 24-h GH (1.0 vs. 1.8 mU/liter, P = 0.033) and PRL (6.8 vs. 10.0 ng/ml, P = 0.009) concentrations were lower in postmenopausal than in premenopausal women. After EPT, the postmenopausal GH and PRL did not differ from premenopausal baseline levels. Postmenopausal mean 24-h GH (P < 0.001) and PRL (P = 0.002), daytime GH (P < 0.001) and nighttime PRL (P = 0.004) were higher during EPT compared with placebo. Cortisol levels did not differ. Premenopausal mean nighttime PRL (P = 0.026) and cortisol (P = 0.018) were higher during EPT compared with placebo. Postmenopausal PRL and premenopausal GH and PRL concentrations were higher at night than during the day. EPT did not alter this pattern. CONCLUSIONS: Menopause was associated with decreased 24-h levels of GH and PRL, which were reversible with EPT. In contrast, cortisol levels were not affected by menopause or EPT. In middle-aged premenopausal women, the studied effects of EPT were limited to nighttime increases of PRL and cortisol.

Overnight variability in transcutaneous carbon dioxide predicts vascular impairment in women.

While there are a number of studies demonstrating association between arterial oxyhaemoglobin saturation events during sleep and markers of vascular impairment, the contribution of peripheral carbon dioxide to the development of atherosclerosis is poorly understood. We used ultrasound imaging to measure carotid artery intima-media thickness (IMT), as well as flow-mediated dilatation (FMD) and nitroglycerin-mediated dilatation (NMD) of brachial artery, in 103 generally healthy 46-year-old (+/-2 years) women. Characteristic event patterns were extracted from their overnight recordings of arterial oxyhaemoglobin saturation , end-tidal partial pressure of carbon dioxide and transcutaneous partial pressure of carbon dioxide . Importance of the event patterns was evaluated through predictive modelling of classes of the ultrasound measurements while controlling for potential confounders. Prediction accuracy was assessed with cross-validation and reported as the area under the receiver operating characteristic curve (AUC). Overnight patterns predicted each of the ultrasound measurements with high accuracy (IMT, AUC = 0.70; FMD, AUC = 0.75; and NMD, AUC = 0.81; all with P < 0.001). Adding the or patterns into the models did not significantly increase their predictive powers (AUC = 0.72, AUC = 0.77 and AUC = 0.83, respectively). The most important patterns reflected overnight variability in . These results suggest a novel link between overnight carbon dioxide events and early signs of vascular impairment in middle-aged women. Non-invasive measurements combined with non-linear modelling techniques could be used to reveal potential markers of vascular impairment present in relatively healthy subjects.

Overnight features of transcutaneous carbon dioxide measurement as predictors of metabolic status.

OBJECTIVE: To systematically investigate whether overnight features in transcutaneous carbon dioxide (P(TcCO(2)) measurements can predict metabolic variables in subject with suspected sleep-disordered breathing. METHODS: The features extracted from the P(TcCO(2)) signal included the number of abrupt descents per hour and attributes that characterize the recovery after such an event. For each outcome variable, the subgroup of the 108 study subjects with the particular variable present was divided into two representative classes, and the optimal features that can predict the classes were learned. Overfitting was avoided by evaluating the classification algorithms using 10-fold cross-validation. RESULTS: (P(TcCO(2)) signal has a key role in determining the classes of high-density lipoprotein cholesterol and thyroid-stimulating hormone concentrations, and it improves the classification accuracy of glycosylated hemoglobin A1c and fasting plasma glucose values. CONCLUSIONS: The features learned from the (P(TcCO(2)) signal reflected the state of the selected metabolic variables in a subtle, but systematic, way. These findings provide a step towards understanding how metabolic disturbances are connected to carbon dioxide exchange during sleep.

Can carotid body perfusion act as a respiratory controller?

The carotid bodies contain chemoreceptor cells that respond to hypoxia and hypercapnia/acidosis of the arterial blood. Since the carotid bodies receive exceptionally high blood perfusion through branches of the external carotid artery, their impulse activity to the respiratory center is thought to be determined mainly by the arterial partial pressures of oxygen (O(2)) and carbon dioxide (CO(2)). However, this paradigm explains the observed increase in ventilation neither during mentally agitated states nor physical exercise. The objective of the work was to test whether physiologically feasible reductions in carotid body perfusion could explain such respiratory overdrive using a flow-sensitive mathematical model of the carotid body chemoreception. The model is based on the law of mass balance and on the description of the chemical reactions in the arterial blood and inside the receptor cells. The neural response to the arterial O(2) and CO(2) levels is assumed to be mediated via the controller's intracellular O(2) partial pressure and pH. The model predicts that the O(2) response is affected even by moderate changes in blood flow, whereas the CO(2) response is not altered until blood flow is severely limited. Reducing blood flow increases neural stimulus but decreases sensitivity to changes in the partial pressures of arterial O(2) and CO(2). An example is given in which relatively small changes in blood flow significantly modify the carotid body sensitivity to CO(2). These results suggest that limiting perfusion of the carotid bodies through vasoconstriction can offer a powerful mechanism to drive breathing beyond metabolic needs. This observation may provide important insight into the control of ventilation, e.g., during transition from wakefulness to sleep, before physical exercise or during panic attack.

Non-invasive respiratory monitoring during wakefulness and sleep in pre- and postmenopausal women.

Menopause and aging cause hormonal changes with respiratory consequences. The aim of the present study was to investigate the physiological changes in respiration during wakefulness and sleep across menopause in non-patient population using non-invasive measurements of blood and tissue gases. The arterial oxyhemoglobin saturation (SaO2), heart rate, end-tidal partial carbon dioxide tension (EtCO2) and transcutaneous partial carbon dioxide tension (TcCO2) were measured during wakefulness and sleep in thirteen pairs of BMI-matched pre- and postmenopausal women. Postmenopausal women had lower SaO2 during sleep than during wakefulness, whereas premenopausal women maintained their wakefulness SaO2 levels also during sleep. EtCO2 levels did not change either between wakefulness and sleep or between premenopausal and postmenopausal groups. TcCO2 levels increased from wakefulness to sleep in both groups and the increase was greater in the postmenopausal group. The impact of sleep on the non-invasive measurements of blood and tissue gases is stronger in postmenopausal women.

Optimization of multi-photon event discrimination levels using Poisson statistics.

In applications where random multi-photon events must be distinguishable from the background, detection of the signals must be based on either analog current measurement or photon counting and multi-level discrimination of single and multi-photon events. In this paper a novel method for optimizing photomultiplier (PMT) pulse discrimination levels in single- and multi-photon counting is demonstrated. This calibration method is based on detection of photon events in coincidence to short laser pulses. The procedure takes advantage of Poisson statistics of single- and mult-iphoton signals and it is applicable to automatic calibration of photon counting devices on production line. Results obtained with a channel photomultiplier (CPM) are shown. By use of three parallel discriminators and setting the discriminator levels according to the described method resulted in a linear response over wide range of random single- and multi-photon signals.