Dermal features derived from optoacoustic tomograms via machine learning correlate microangiopathy phenotypes with diabetes stage

Skin microangiopathy has been associated with diabetes. Here we show that skin-microangiopathy phenotypes in humans can be correlated with diabetes stage via morphophysiological cutaneous features extracted from raster-scan optoacoustic mesoscopy (RSOM) images of skin on the leg. We obtained 199 RSOM images from 115 participants (40 healthy and 75 with diabetes), and used machine learning to segment skin layers and microvasculature to identify clinically explainable features pertaining to different depths and scales of detail that provided the highest predictive power. Features in the dermal layer at the scale of detail of 0.1-1 mm (such as the number of junction-to-junction branches) were highly sensitive to diabetes stage. A 'microangiopathy score' compiling the 32 most-relevant features predicted the presence of diabetes with an area under the receiver operating characteristic curve of 0.84. The analysis of morphophysiological cutaneous features via RSOM may allow for the discovery of diabetes biomarkers in the skin and for the monitoring of diabetes status.

Opening a window to skin biomarkers for diabetes stage with optoacoustic mesoscopy

Being the largest and most accessible organ of the human body, the skin could offer a window to diabetes-related complications on the microvasculature. However, skin microvasculature is typically assessed by histological analysis, which is not suited for applications to large populations or longitudinal studies. We introduce ultra-wideband raster-scan optoacoustic mesoscopy (RSOM) for precise, non-invasive assessment of diabetes-related changes in the dermal microvasculature and skin micro-anatomy, resolved with unprecedented sensitivity and detail without the need for contrast agents. Providing unique imaging contrast, we explored a possible role for RSOM as an investigational tool in diabetes healthcare and offer the first comprehensive study investigating the relationship between different diabetes complications and microvascular features in vivo. We applied RSOM to scan the pretibial area of 95 participants with diabetes mellitus and 48 age-matched volunteers without diabetes, grouped according to disease complications, and extracted six label-free optoacoustic biomarkers of human skin, including dermal microvasculature density and epidermal parameters, based on a novel image-processing pipeline. We then correlated these biomarkers to disease severity and found statistically significant effects on microvasculature parameters as a function of diabetes complications. We discuss how label-free RSOM biomarkers can lead to a quantitative assessment of the systemic effects of diabetes and its complications, complementing the qualitative assessment allowed by current clinical metrics, possibly leading to a precise scoring system that captures the gradual evolution of the disease.

Tritium Beta Spectrum Measurement and Neutrino Mass Limit from Cyclotron Radiation Emission Spectroscopy

The absolute scale of the neutrino mass plays a critical role in physics at every scale, from the subatomic to the cosmological. Measurements of the tritium end-point spectrum have provided the most precise direct limit on the neutrino mass scale. In this Letter, we present advances by Project 8 to the cyclotron radiation emission spectroscopy (CRES) technique culminating in the first frequency-based neutrino mass limit. With only a cm^{3}-scale physical detection volume, a limit of m_{β}<155  eV/c^{2} (152  eV/c^{2}) is extracted from the background-free measurement of the continuous tritium beta spectrum in a Bayesian (frequentist) analysis. Using ^{83m}Kr calibration data, a resolution of 1.66±0.19  eV (FWHM) is measured, the detector response model is validated, and the efficiency is characterized over the multi-keV tritium analysis window. These measurements establish the potential of CRES for a high-sensitivity next-generation direct neutrino mass experiment featuring low background and high resolution.

Audiological phenotyping evaluation in KBG syndrome: Description of a multicenter review

OBJECTIVES

Our objective was to reinforce clinical knowledge of hearing impairment in KBG syndrome. KBG syndrome is a rare genetic disorder due to monoallelic pathogenic variations of ANKRD11.The typical phenotype includes facial dysmorphism, costal and spinal malformation and developmental delay. Hearing loss in KBG patients has been reported for many years, but no study has evaluated audiological phenotyping from a clinical and an anatomical point of view.

METHODS

This French multicenter study included 32 KBG patients with retrospective collection of data on audiological features, ear imaging and genetic investigations.

RESULTS

We identified a typical audiological profil in KBG syndrome: conductive (71%), bilateral (81%), mild to moderate (84%) and stable (69%) hearing loss, with some audiological heterogeneity. Among patients with an abnormality on CT imaging (55%), ossicular chain impairment (67%), fixation of the stapes footplate (33%) and inner-ear malformations (33%) were the most common abnormalities.

CONCLUSION

We recommend a complete audiological and radiological evaluation and an ENT-follow up in all patients presenting with KBG Syndrome. Imaging evaluation is necessary to determine the nature of lesions in the middle and inner ear.

Bone morphogenetic protein 10 as predictor for adverse outcomes in patients with atrial fibrillation

Background

Patients with atrial fibrillation (AF) face an increased risk of death and major adverse cardiovascular events (MACE). Bone morphogenetic protein 10 (BMP10) is a novel atrial-specific biomarker, but data about its prognostic value in AF patients are lacking.

Purpose

We aimed to assess the predictive value of BMP10 for death and MACE in AF patients in comparison to N-terminal prohormone of B-type natriuretic peptide (NT-proBNP).

Methods

Baseline concentrations of BMP10 and NT-proBNP were measured in stable patients with AF enrolled in Swiss-AF, a prospective multicenter observational cohort study. Primary outcomes were all-cause death and MACE (composite of heart failure hospitalization, cardiovascular death, stroke, systemic embolism, myocardial infarction). Measures of discriminative power were used to compare multivariable Cox proportional hazard models using the different biomarkers.

Results

A total of 2219 AF patients were included with a median follow-up of 4.3 years (IQR 3.9, 5.1). Mean age was 73±9 years and 27% were women. Incidence rate per 100 patient-years of all-cause death and MACE increased across BMP10 quartiles (Figure 1). In the multivariable adjusted Cox proportional hazard model, the hazard ratio (HR) and 95% confidence interval (CI) of BMP10 was 1.60 (1.37; 1.87) to predict all-cause death, and 1.54 (1.35; 1.76) to predict MACE. For all-cause death, the C-index (95% CI) was 0.783 (0.763; 0.809) for BMP10, 0.784 (0.765; 0.810) for NT-proBNP, and 0.789 (0.771; 0.815) for both biomarkers combined. For MACE, the C-index (95% CI) was 0.732 (0.715; 0.754) for BMP10, 0.747 (0.731; 0.768) for NT-proBNP, and 0.750 (0.734; 0.771) for both biomarkers combined. When grouping patients according to clinical used NT-proBNP categories (&lt;300, 300–900, &gt;900 ng/l), higher incidence rates and adjusted HRs were observed for the primary outcomes in patients with high BMP10 in the categories of low NT-proBNP (all-cause death aHR 2.28 [1.15; 4.52], MACE aHR 1.88 [1.07; 3.28]) and high NT-proBNP (all-cause death aHR 1.61 [1.14; 2.26], MACE aHR 1.38 [1.07; 1.80]) (Figure 2).

Conclusion

The novel atrial-specific biomarker BMP10 strongly predicts all-cause death and MACE in patients with AF. BMP10 provides additional prognostic information in low- and high-risk patients according to NT-proBNP stratification.

Funding Acknowledgement

Type of funding sources: Foundation. Main funding source(s): Swiss National Science Foundation, Swiss Heart Foundation

Associations of SCORE2, circulating cardiovascular biomarkers and carotid intima-media-thickness in a population-based cohort

Background

The updated SCORE2 model to estimate the 10-year risk of fatal and non-fatal cardiovascular disease (CVD) was recently introduced, which is based solely on traditional cardiovascular risk factors.

Purpose

We aimed to investigate the associations of SCORE2-predicted risk with four circulating cardiovascular biomarkers (high-sensitivity troponin I [hs-cTnI], N-terminal pro B-type natriuretic peptide [NT-proBNP], cystatin C-derived estimated glomerular filtration rate [eGFR] and high-sensitivity C-reactive protein [hs-CRP]) as well as mean carotid intima-media-thickness (cIMT) in a prospective, population-based German cohort.

Methods

In the first set of 10,000 participants, who were aged 45–74 years and recruited between 2016 and 2019, the SCORE2 model was applied in a cross-sectional manner. Individuals with prevalent CVD were excluded for this purpose. Eligible individuals were then categorized into five risk groups (&lt;2.5%, 2–&lt;5%, 5–&lt;10%, 10–&lt;15%, and ≥15%) according to SCORE2. To test for the associations of circulating biomarkers and cIMT with SCORE2, we created box plots and computed Pearson's correlation coefficients (R). Considering cIMT as a biological surrogate for incident CVD, we explored the incremental utility of circulating biomarkers to predict cIMT beyond SCORE2 by multivariable logistic regression analysis with stepwise selection of variables, quantified by Beta-coefficients per one standard deviation (SD) increase with respective 95% confidence intervals (CI). Discrimination (C-index) and category-free net reclassification improvement (NRI) for predicting mean cIMT &gt;1mm were calculated for this extended model in comparison to SCORE2 alone as the reference.

Results

In 8,518 individuals free of CVD, median estimated 10-year risk of CVD based on SCORE2 was 6.1 (interquartile range [IQR] 3.2, 9.9) %. All four investigated biomarkers (hs-cTnI, R=0.41; NT-proBNP, R=0.21; hs-CRP, R=0.22; eGFR, R=−0.44; all P&lt;0.001), and mean cIMT (R=0.42, P&lt;0.001) correlated strongly with the continuous SCORE2 risk and the respective SCORE2 risk categories (Figure 1). In multivariable regression analysis, all four circulating biomarkers remained significant independent predictors of mean cIMT when added to SCORE 2 (Figure 2). When compared to SCORE2 as the reference model (C-index 0.763, 95% CI 0.742–0.784) to predict mean cIMT &gt;1mm, the joint addition of all four stepwise-selected investigated biomarkers led to a small but significant improvement of discrimination (C-index 0.770, 95% CI 0.749–0.791, P&lt;0.001) and reclassification yield (NRI 0.154, 95% CI 0.059–0.250, P=0.002).

Conclusions

All four investigated circulating biomarkers depicting different pathophysiological pathways and mean cIMT correlate strongly with the cardiovascular risk estimated by SCORE2. Circulating biomarkers may further improve CVD-risk prediction when added to the traditional cardiovascular risk factors currently considered by SCORE2.

Funding Acknowledgement

Type of funding sources: None.

Diagnostic validation of a high-sensitivity troponin I assay and its use in a guideline-consistent rapid diagnostic protocol

Background

Current ESC guidelines on management of non-ST-elevation myocardial infarction (NSTEMI) recommend rapid diagnostic protocols using validated high-sensitivity cardiac troponin (hs-cTn) assays (1). While established protocols are available for several hs-cTn assays, data on the diagnostic performance of the Siemens Atellica IM High-Sensitivity Cardiac Troponin I assay is limited.

Methods

In a cohort study including 1,800 patients presenting with suspected acute MI. Final diagnosis of MI was adjudicated by two cardiologists separately in accordance with the fourth universal definition of MI (2). We developed and validated a 0/1h diagnostic algorithm using the Siemens Atellica assay. The algorithm was established in the first 928 patients and validated in the following 872 patients.

Results

ROC analyses for the diagnosis of NSTEMI revealed high discriminatory ability of the Siemens hs-cTnI assay with an area under the curve of 0.88 (95% confidence interval (CI): 0.86–0.90) at 0h, 0.93 (CI: 0.91–0.94) after 1h and 0.95 (CI: 0.93–0.96) after 3h (Figure 1).

The derived algorithm consisted of a baseline rule-out of non-ST elevation MI using a cutoff &lt;3 ng/L in patients with a symptom onset ≥3h or an admission troponin I &lt;6 ng/L with a delta change from 0h to 1h &lt;3 ng/L. For rule-in, an admission troponin I ≥120 ng/L or an increase within the first hour ≥12 ng/L was required. Application of the algorithm to the validation cohort showed a negative predictive value of 99.8% (CI 98.7%-100.0%), a sensitivity of 99.1% (CI 95.1%–100.0%) and 48.3% of patients ruled out, whereas 15.1% were ruled in with a positive predictive value of 68.0% (CI 59.1%-75.9%) and a specificity of 94.4% (92.5%–96.0%). The diagnostic performance was comparable to guideline-recommended application of an established hs-cTnI assay in a rapid 0/1h strategy (1).

Conclusion

The Siemens hs-cTnI assay is well-suited for application in rapid diagnostic stratification of patients with suspected MI and showed similar performance when compared with an established hs-cTnI assay using an algorithm recommended by recent ESC guidelines.

Funding Acknowledgement

Type of funding sources: Foundation. Main funding source(s): German Heart foundation Figure 1

Cross-sectional and longitudinal characterisation of cognitive function and outcomes in patients presenting to hospital with cardiovascular risk factors

Background/Introduction

Cardiovascular (CV) diseases including atrial fibrillation and arteriosclerosis are associated with impaired cognitive function. Cognitive dysfunction can impact the process of shared clinical decision making, reduce adherence to polypharmacy, and decrease quality of life. The prevalence of cognitive dysfunction in contemporary patients with CV diseases and its implication on future CV events is not well known.

Purpose

We 1) quantified cognitive function in patients presenting to hospital with CV diseases, 2) identified clinical variables and blood biomarkers associated with cognitive dysfunction, and 3) quantified the hazard of abnormal cognitive function for predicting MACCE (major adverse CV and cerebrovascular events).

Methods and results

Of 1625 consecutive patients presenting acutely to a large teaching hospital with CV diseases, 614 patients (median age [Q1, Q3] 68 [58, 76] years; 66% male) who completed the Montreal Cognitive Assessment (MoCA) were analysed. The median [Q1, Q3] MoCA score was 25 points [21, 27]. 360 patients (59%) had an abnormal score (&lt;26). At baseline, patients with abnormal scores were more likely to be female (odds ratio, OR [95% confidence intervals], 1.874 [1.287, 2.728]), have BMI&lt;30 (OR 0.584 [0.410, 0.831]), heart failure (OR 1.492 [1.043, 2.135]), diabetes (OR 2.212 [1.529, 3.199]), chronic kidney disease (CKD-EPI&lt;60 ml/min, OR 1.553 [1.021, 2.361]), and have more CV co-morbidities (OR per additional co-morbidity 1.415 [1.246, 1.605]). Amongst 12 CV biomarkers tested, elevated Bone Morphogenetic Protein 10 (OR 1.325 [1.022, 1.719]) and Growth Differentiation Factor 15 (OR 1.419 [1.054, 1.912]) increased odds of abnormal scores.

Cox proportional hazards model adjusted for competing risk of non-CV death assessed the relationship between abnormal cognitive function and MACCE (stroke, TIA, myocardial infarction, hospitalisation for heart failure, CV death). Follow-up time ranged from 2.7 to 6.1 years. Patients were censored at 2.5 years for this analysis. 130 out of 614 patients experienced a MACCE (21%) and 71 had a non-CV death (12%). Patients with abnormal MoCA scores were at higher risk for MACCE (subhazard ratio, sHR [95% CI] 1.827 [1.253, 2.664]). The hazard remained significant after adjustment for age, sex, obesity, atrial fibrillation, stroke, heart failure, hypertension, coronary artery disease, diabetes, peripheral artery disease and renal dysfunction (sHR 1.367 [1.056, 2.326]; Figure). All-cause mortality was 1.785 times higher for those with abnormal MoCA scores [1.061, 3.002].

Conclusion

In this study, 3 out of 5 patients with CV diseases had abnormal MoCA scores at baseline. Abnormal cognitive scores significantly predicted patients who went on to experience a MACCE within 2.5 years of follow-up. These observations call for further research and action to provide additional diagnostics, support and early intervention to address cognitive dysfunction in CV patients.

Funding Acknowledgement

Type of funding sources: Public grant(s) – EU funding. Main funding source(s): EU H2020 CATCH ME Cumulative incidence function

Calcite crystal orientation patterns in the bilayers of laminated shells of benthic rotaliid foraminifera

Shells of calcifying foraminifera play a major role in marine biogeochemical cycles; fossil shells form important archives for paleoenvironment reconstruction. Despite their importance in many Earth science disciplines, there is still little consensus on foraminiferal shell mineralization. Geochemical, biochemical, and physiological studies showed that foraminiferal shell formation might take place through various and diverse mineralization mechanisms. In this study, we contribute to benthic foraminiferal shell calcification through deciphering crystallite organization within the shells. We base our conclusions on results gained from electron backscattered diffraction (EBSD) measurements and describe microstructure/texture characteristics within the laminated shell walls of the benthic, symbiontic foraminifera: Ammonia tepida, Amphistegina lobifera, Amphistegina lessonii. We highlight crystallite assembly patterns obtained on differently oriented cuts and discuss crystallite sizes, morphologies, interlinkages, orientations, and co-orientation strengths. We show that: (i) crystals within benthic foraminiferal shells are mesocrystals, (ii) have dendritic-fractal morphologies and (iii) interdigitate strongly. Based on crystal size, we (iv) differentiate between the two layers that comprise the shells and demonstrate that (v) crystals in the septa have different assemblies relative to those in the shell walls. We highlight that (vi) at junctions of different shell elements the axis of crystal orientation jumps abruptly such that their assembly in EBSD maps has a bimodal distribution. We prove (vii) extensive twin-formation within foraminiferal calcite; we demonstrate (viii) the presence of two twin modes: 60°/[001] and 77°/~[6 -6 1] and visualize their distributions within the shells. In a broader perspective, we draw conclusions on processes that lead to the observed microstructure/texture patterns.

[Efficacy of Disease Management Programs Asthma and COPD? Results of a Cross-Sectional Study]

BACKGROUND

 The efficacy of the German disease management programs (DMP) asthma and chronic obstructive pulmonary disease (COPD) cannot be shown with the legally bound documentations. Studies with control groups are rare. Aim of this work was to investigate in a cross-sectional study whether the disease control differs in participants (DMP+) and non-participants (DMP - ) of the DMPs asthma and COPD.

METHODS

 The study was a prospective multicenter cross-sectional study. Primary endpoints were the Asthma Control Test™ (ACT) in the asthma part of the study and the COPD Assessment Test™ (CAT) for the COPD part.

RESULTS

 A total of 1038 asthma patients and 846 COPD patients were included, of whom about 70 % participated in the corresponding DMP. The ACT total score was higher in asthma DMP+ patients than in DMP- patients (mean difference 0.86; 95 % CI: 0.29 - 1.43;p = 0.003), but not clinically relevant. For COPD there was no clinically relevant difference in COPD disease impact (0.52; 95 % CI: - 0.71 - 1.75; p = 0.405). Although DMP patients had to be enrolled in the respective DMP for at least one year, only 60 % of these patients had participated in a structured education. We did not observe a difference in disease control in DMP patients who respectively participated and did not participate in a structured education.

DISCUSSION

 There was no clinically relevant difference in disease control between DMP+ and DMP- patients. The efficacy of DMPs has been demonstrated internationally in randomized controlled trials. Randomized controlled trials should be conducted in Germany for demonstrating efficacy of DMPs asthma and COPD.

REGISTRATION

 drks.de, DRKS00007664, Registration date: Jan 15, 2015.

Terebratulide brachiopod shell biomineralization by mantle epithelial cells

To understand mineral transport pathways for shell secretion and to assess differences in cellular activity during mineralization, we imaged with TEM and FE-SEM ultrastructural characteristics of outer mantle epithelium (OME) cells. Imaging was carried out on Magellania venosa shells embedded/etched, chemically fixed/decalcified and high-pressure frozen/freeze-substituted samples from the commissure, central shell portions and from puncta. Imaging results are complemented with morphometric evaluations of volume fractions of membrane-bound organelles. At the commissure the OME consists of several layers of cells. These cells form oblique extensions that, in cross-section, are round below the primary layer and flat underneath fibres. At the commissure the OME is multi-cell layered, in central shell regions it is single-cell layered. When actively secreting shell carbonate extrapallial space is lacking, because OME cells are in direct contact with the calcite of the forming fibres. Upon termination of secretion, OME cells attach via apical hemidesmosomes to extracellular matrix membranes that line the proximal surface of fibres. At the commissure volume fractions for vesicles, mitochondria and lysosomes are higher relative to single-cell layered regions, whereas for endoplasmic-reticulum and Golgi apparatus there is no difference. FE-SEM, TEM imaging reveals the lack of extrapallial space between OME cells and developing fibres. In addition, there is no indication for an amorphous precursor within fibres when these are in active secretion mode. Accordingly, our results do not support transport of minerals by vesicles from cells to sites of mineralization, rather by transfer of carbonate ions via transport mechanisms associated with OME cell membranes.

Multimessenger observations of a flaring blazar coincident with high-energy neutrino IceCube-170922A

Previous detections of individual astrophysical sources of neutrinos are limited to the Sun and the supernova 1987A, whereas the origins of the diffuse flux of high-energy cosmic neutrinos remain unidentified. On 22 September 2017, we detected a high-energy neutrino, IceCube-170922A, with an energy of ~290 tera–electron volts. Its arrival direction was consistent with the location of a known γ-ray blazar, TXS 0506+056, observed to be in a flaring state. An extensive multiwavelength campaign followed, ranging from radio frequencies to γ-rays. These observations characterize the variability and energetics of the blazar and include the detection of TXS 0506+056 in very-high-energy γ-rays. This observation of a neutrino in spatial coincidence with a γ-ray–emitting blazar during an active phase suggests that blazars may be a source of high-energy neutrinos.

Search for γ-Ray Line Signals from Dark Matter Annihilations in the Inner Galactic Halo from 10 Years of Observations with H.E.S.S

Spectral lines are among the most powerful signatures for dark matter (DM) annihilation searches in very-high-energy γ rays. The central region of the Milky Way halo is one of the most promising targets given its large amount of DM and proximity to Earth. We report on a search for a monoenergetic spectral line from self-annihilations of DM particles in the energy range from 300 GeV to 70 TeV using a two-dimensional maximum likelihood method taking advantage of both the spectral and spatial features of the signal versus background. The analysis makes use of Galactic center observations accumulated over ten years (2004-2014) with the H.E.S.S. array of ground-based Cherenkov telescopes. No significant γ-ray excess above the background is found. We derive upper limits on the annihilation cross section ⟨σv⟩ for monoenergetic DM lines at the level of 4×10^{-28}  cm^{3} s^{-1} at 1 TeV, assuming an Einasto DM profile for the Milky Way halo. For a DM mass of 1 TeV, they improve over the previous ones by a factor of 6. The present constraints are the strongest obtained so far for DM particles in the mass range 300 GeV-70 TeV. Ground-based γ-ray observations have reached sufficient sensitivity to explore relevant velocity-averaged cross sections for DM annihilation into two γ-ray photons at the level expected from the thermal relic density for TeV DM particles.

MED13L-related intellectual disability: involvement of missense variants and delineation of the phenotype

Molecular anomalies in MED13L, leading to haploinsufficiency, have been reported in patients with moderate to severe intellectual disability (ID) and distinct facial features, with or without congenital heart defects. Phenotype of the patients was referred to "MED13L haploinsufficiency syndrome." Missense variants in MED13L were already previously described to cause the MED13L-related syndrome, but only in a limited number of patients. Here we report 36 patients with MED13L molecular anomaly, recruited through an international collaboration between centers of expertise for developmental anomalies. All patients presented with intellectual disability and severe language impairment. Hypotonia, ataxia, and recognizable facial gestalt were frequent findings, but not congenital heart defects. We identified seven de novo missense variations, in addition to protein-truncating variants and intragenic deletions. Missense variants clustered in two mutation hot-spots, i.e., exons 15-17 and 25-31. We found that patients carrying missense mutations had more frequently epilepsy and showed a more severe phenotype. This study ascertains missense variations in MED13L as a cause for MED13L-related intellectual disability and improves the clinical delineation of the condition.

Deconfounding Microarray Analysis

Objectives: Microarray analysis requires standardized specimens and evaluation procedures to achieve acceptable results. A major limitation of this method is caused by heterogeneity in the cellular composition of tissue specimens, which frequently confounds data analysis. We introduce a linear model to deconfound gene expression data from tissue heterogeneity for genes exclusively expressed by a single cell type.

Methods: Gene expression data are deconfounded from tissue heterogeneity effects by analyzing them using an appropriate linear regression model. In our illustrating data set tissue heterogeneity is being measured using flow cytometry. Gene expression data are determined in parallel by real time quantitative polymerase chain reaction (qPCR) and microarray analyses. Verification of deconfounding is enabled using protein quantification for the respective marker genes.

Results: For our illustrating dataset, quantification of cell type proportions for peripheral blood mononuclear cells (PBMC) from tuberculosis patients and controls revealed differences in B cell and monocyte proportions between both study groups, and thus heterogeneity for the tissue under investigation. Gene expression analyses reflected these differences in celltype distribution. Fitting an appropriate linear model allowed us to deconfound measured transcriptome levels from tissue heterogeneity effects. In the case of monocytes, additional differential expression on the single cell level could be proposed. Protein quantification verified these deconfounded results.

Conclusions: Deconfounding of transcriptome analyses for cellular heterogeneity greatly improves interpretability, and hence the validity of transcriptome profiling results.

Two Models for Outcome Prediction

Objectives: Accurately predicting disease progress from a set of predictive variables is an important aspect of clinical work. For binary outcomes, the classical approach is to develop prognostic logistic regression (LR) models. Alternatively, machine learning algorithms were proposed with artificial neural networks (ANN) having become popular over the last decades. Although some studies have compared predictive accuracies of LR and ANN models, some concerns regarding their methodological quality have been voiced. Our comparison has the advantage of being based on two large independent data sets allowing for elaborate model development and independent validation.

Methods: From the German Stroke Database, a learning data set including 1754 prospectively recruited patients with acute ischemic stroke was used. Utilizing LR and ANN, two prognostic models were developed predicting restitution of functional independence and survival after 100 days. The resulting models were applied to classify 1470 patients with acute ischemic stroke; this test data set was collected independently from the learning data. Error fractions in the test data were determined, and differences in error fractions between the algorithms were calculated with 95% confidence intervals.

Results: For most prognostic models, error fractions in the test data were below 40%. There was no difference between the algorithms except for the model predicting completely versus incompletely restituted or deceased patients (difference in error fractions = 4.01% [2.10-5.96%], p = 0.0001).

Conclusions: The conscientiously applied LR remains the gold standard for prognostic modelling; however, ANN can be an alternative automated “quick and easy” multivariate analysis.

Independence Estimating Equations for Controlled Clinical Trials with Small Sample Sizes

Objectives: The application of independence estimating equations (IEE) for controlled clinical trials (CCTs) has recently been discussed, and recommendations for its use have been derived for testing hypotheses. The robust estimator of variance has been shown to be liberal for small sample sizes. Therefore a series of modifications has been proposed. In this paper we systematically compare confidence intervals (CIs) proposed in the literature for situations that are common in CCTs.

Methods: Using Monte-Carlo simulation studies, we compared the coverage probabilities of CIs and non-convergence probabilities for the parameters of the mean structure for small samples using modifications of the variance estimator proposed by Mancl and de Rouen [7], Morel et al. [8] and Pan [3].

Results: None of the proposed modifications behave well in each investigated situation. For parallel group designs with repeated measurements and binary response the method proposed by Pan maintains the nominal level. We observed non-convergence of the IEE algorithm in up to 10% of the replicates depending on response probabilities in the treatment groups. For comparing slopes with continuous responses, the approach of Morel et al. can be recommended.

Conclusions: Results of non-convergence probabilities show that IEE should not be used in parallel group designs with binary endpoints and response probabilities close to 0 or 1. Modifications of the robust variance estimator should be used for sample sizes up to 100 clusters for CI estimation.

Tutorial on Microarray Gene Expression Experiments

Objectives: With the collection of articles presented in this special issue, we aim at educating interested statisticians and biometricians on the one hand as well as biologists and medical researchers on the other with respect to basic necessities in planning, conducting and analyzing microarray gene expression experiments. The reader should get comprehensive directions to understand both the overall structure of this approach as well as the decisive details, which enable – or thwart – a meaningful data analysis.

Methods: For a one-day workshop with tutorial character we brought together experts in design, conduct and analysis of microarray gene expression experiments who prepared a series of comprehensive lessons. These contributions were then reworked into a series of introductory articles and bundled in form and content as a Special Topic.

Results: It was possible to present a tutorial overview of the field. The interested reader was able to learn the basic necessities and was referred to further references for details on the possible alternatives. A recipe style all-embracing plan, covering all eventualities and possibilities was not only beyond the scope of an introductory tutorial-like presentation, but was also not yet agreed upon by the scientific society.

Conclusions: It proved feasible to find a framework for integrating the interdisciplinary approaches to the challenging field of gene expression analysis with microarrays, hopefully contributing to a rapid and comprehensive introduction for novices.

Sample Selection for Microarray Gene Expression Studies

Objectives: The choice of biomedical samples for microarray gene expression studies is decisive for both validity and interpretability of results. We present a consistent, comprehensive framework to deal with the typical selection problems in microarray studies.

Methods: Microarray studies are designed either as case-control studies or as comparisons of parallel groups from cohort studies, since high levels of random variation in the experimental approach thwart absolute measurements of gene expression levels. Validity and results of gene expression studies heavily rely on the appropriate choice of these study groups. Therefore, the so-called principles of comparability, which are well known from both clinical and epidemiological studies, need to be applied to microarray experiments.

Results: The principles of comparability are the study-base principle, the principle of deconfounding and the principle of comparable accuracy in measurements. We explain each of these principles, show how they apply to microarray experiments, and illustrate them with examples. The examples are chosen as to represent typical stumbling blocks of microarray experimental design, and to exemplify the benefits of implementing the principles of comparability in the setting of micro-array experiments.

Conclusions: Microarray studies are closely related to classical study designs and therefore have to obey the same principles of comparability as these. Their validity should not be compromised by selection, confounding or information bias. The so-called study-base principle, calling for comparability and thorough definition of the compared cell populations, is the key principle for the choice of biomedical samples and controls in microarray studies.

Practical Considerations on the Use of Extreme Sib-pairs for Obesity

Objectives: The extreme sib-pair approach has been shown to be a powerful strategy to identify susceptibility loci linked to quantitative traits. The body mass index is the usually assessed trait in genetic studies on human obesity. Environmental factors clearly play an important role for this trait. We hypothesized that the low weight of most sibs who were seemingly discordant to the obese index proband was influenced by other environmental and/or genetic factors like restrained eating or psychiatric disorders.

Methods: A screening questionnaire was sent to parents of all consecutively admitted patients three weeks prior to referral of index probands for inpatient treatment of obesity. The first 320 families were further investigated. Twenty-seven seemingly extremely discordant sib-pairs (ED) were identified and examined in detail.

Results: The low weight of most sibs who were seemingly discordant to the obese index proband was influenced by factors like restrained eating or psychiatric disorders. Only 20% of the interviewed ED could be considered as genuine ED.

Conclusion: We conclude that extensive medical evaluation is necessary if the aim is to guarantee genuine ED in family studies for human obesity. Non-paternity deserves specific attention in ED studies.