Post-hoc standardisation of parametric T1 maps in cardiovascular magnetic resonance imaging: a proof-of-concept

BACKGROUND

In cardiovascular magnetic resonance imaging parametric T1 mapping lacks universally valid reference values. This limits its extensive use in the clinical routine. The aim of this work was the introduction of our self-developed Magnetic Resonance Imaging Software for Standardization (MARISSA) as a post-hoc standardisation approach.

METHODS

Our standardisation approach minimises the bias of confounding parameters (CPs) on the base of regression models. 214 healthy subjects with 814 parametric T1 maps were used for training those models on the CPs: age, gender, scanner and sequence. The training dataset included both sex, eleven different scanners and eight different sequences. The regression model type and four other adjustable standardisation parameters were optimised among 240 tested settings to achieve the lowest coefficient of variation, as measure for the inter-subject variability, in the mean T1 value across the healthy test datasets (HTE, N = 40, 156 T1 maps). The HTE were then compared to 135 patients with left ventricular hypertrophy including hypertrophic cardiomyopathy (HCM, N = 112, 121 T1 maps) and amyloidosis (AMY, N = 24, 24 T1 maps) after applying the best performing standardisation pipeline (BPSP) to evaluate the diagnostic accuracy.

FINDINGS

The BPSP reduced the COV of the HTE from 12.47% to 5.81%. Sensitivity and specificity reached 95.83% / 91.67% between HTE and AMY, 71.90% / 72.44% between HTE and HCM, and 87.50% / 98.35% between HCM and AMY.

INTERPRETATION

Regarding the BPSP, MARISSA enabled the comparability of T1 maps independently of CPs while keeping the discrimination of healthy and patient groups as found in literature.

FUNDING

This study was supported by the BMBF / DZHK.

Lazy Luna: Extendible software for multilevel reader comparison in cardiovascular magnetic resonance imaging

BACKGROUND AND OBJECTIVES

Cardiovascular Magnetic Resonance (CMR) imaging is a growing field with increasing diagnostic utility in clinical routine. Quantitative diagnostic parameters are typically calculated based on contours or points provided by readers, e.g. natural intelligences (NI) such as clinicians or researchers, and artificial intelligences (AI). As clinical applications multiply, evaluating the precision and reproducibility of quantitative parameters becomes increasingly important. Although segmentation challenges for AIs and guidelines for clinicians provide quality assessments and regulation, the methods ought to be combined and streamlined for clinical applications. The goal of the developed software, Lazy Luna (LL), is to offer a flexible evaluation tool that is readily extendible to new sequences and scientific endeavours.

METHODS

An interface was designed for LL, which allows for comparing annotated CMR images. Geometric objects ensure precise calculations of metric values and clinical results regardless of whether annotations originate from AIs or NIs. A graphical user interface (GUI) is provided to make the software available to non-programmers. The GUI allows for an interactive inspection of image datasets as well as implementing tracing procedures, which follow statistical reader differences in clinical results to their origins in individual image contours. The backend software builds on a set of meta-classes, which can be extended to new imaging sequences and clinical parameters. Following an agile development procedure with clinical feedback allows for a quick implementation of new classes, figures and tables for evaluation.

RESULTS

Two application cases present LL's extendibility to clinical evaluation and AI development contexts. The first concerns T1 parametric mapping images segmented by two expert readers. Quantitative result differences are traced to reveal typical segmentation dissimilarities from which these differences originate. The meta-classes are extended to this new application scenario. The second applies to the open source Late Gadolinium Enhancement (LGE) quantification challenge for AI developers "Emidec", which illustrates LL's usability as open source software.

CONCLUSION

The presented software Lazy Luna allows for an automated multilevel comparison of readers as well as identifying qualitative reasons for statistical reader differences. The open source software LL can be extended to new application cases in the future.

Introduction of a cascaded segmentation pipeline for parametric T1 mapping in cardiovascular magnetic resonance to improve segmentation performance

The manual and often time-consuming segmentation of the myocardium in cardiovascular magnetic resonance is increasingly automated using convolutional neural networks (CNNs). This study proposes a cascaded segmentation (CASEG) approach to improve automatic image segmentation quality. First, an object detection algorithm predicts a bounding box (BB) for the left ventricular myocardium whose 1.5 times enlargement defines the region of interest (ROI). Then, the ROI image section is fed into a U-Net based segmentation. Two CASEG variants were evaluated: one using the ROI cropped image solely (cropU) and the other using a 2-channel-image additionally containing the original BB image section (crinU). Both were compared to a classical U-Net segmentation (refU). All networks share the same hyperparameters and were tested on basal and midventricular slices of native and contrast enhanced (CE) MOLLI T1 maps. Dice Similarity Coefficient improved significantly (p < 0.05) in cropU and crinU compared to refU (81.06%, 81.22%, 72.79% for native and 80.70%, 79.18%, 71.41% for CE data), while no significant improvement (p < 0.05) was achieved in the mean absolute error of the T1 time (11.94 ms, 12.45 ms, 14.22 ms for native and 5.32 ms, 6.07 ms, 5.89 ms for CE data). In conclusion, CASEG provides an improved geometric concordance but needs further improvement in the quantitative outcome.

Multilevel comparison of deep learning models for function quantification in cardiovascular magnetic resonance: On the redundancy of architectural variations

Background

Cardiac function quantification in cardiovascular magnetic resonance requires precise contouring of the heart chambers. This time-consuming task is increasingly being addressed by a plethora of ever more complex deep learning methods. However, only a small fraction of these have made their way from academia into clinical practice. In the quality assessment and control of medical artificial intelligence, the opaque reasoning and associated distinctive errors of neural networks meet an extraordinarily low tolerance for failure.

Aim

The aim of this study is a multilevel analysis and comparison of the performance of three popular convolutional neural network (CNN) models for cardiac function quantification.

Methods

U-Net, FCN, and MultiResUNet were trained for the segmentation of the left and right ventricles on short-axis cine images of 119 patients from clinical routine. The training pipeline and hyperparameters were kept constant to isolate the influence of network architecture. CNN performance was evaluated against expert segmentations for 29 test cases on contour level and in terms of quantitative clinical parameters. Multilevel analysis included breakdown of results by slice position, as well as visualization of segmentation deviations and linkage of volume differences to segmentation metrics via correlation plots for qualitative analysis.

Results

All models showed strong correlation to the expert with respect to quantitative clinical parameters (r_{z '} = 0.978, 0.977, 0.978 for U-Net, FCN, MultiResUNet respectively). The MultiResUNet significantly underestimated ventricular volumes and left ventricular myocardial mass. Segmentation difficulties and failures clustered in basal and apical slices for all CNNs, with the largest volume differences in the basal slices (mean absolute error per slice: 4.2 ± 4.5 ml for basal, 0.9 ± 1.3 ml for midventricular, 0.9 ± 0.9 ml for apical slices). Results for the right ventricle had higher variance and more outliers compared to the left ventricle. Intraclass correlation for clinical parameters was excellent (≥0.91) among the CNNs.

Conclusion

Modifications to CNN architecture were not critical to the quality of error for our dataset. Despite good overall agreement with the expert, errors accumulated in basal and apical slices for all models.

Fast acquisition of left and right ventricular function parameters applying cardiovascular magnetic resonance in clinical routine - validation of a 2-shot compressed sensing cine sequence

Objectives. To evaluate if cine sequences accelerated by compressed sensing (CS) are feasible in clinical routine and yield equivalent cardiac morphology in less time. Design. We evaluated 155 consecutive patients with various cardiac diseases scanned during our clinical routine. LV and RV short axis (SAX) cine images were acquired by conventional and prototype 2-shot CS sequences on a 1.5 T CMR. The 2-shot prototype captures the entire heart over a period of 3 beats making the acquisition potentially even faster. Both scans were performed with identical slice parameters and positions. We compared LV and RV morphology with Bland-Altmann plots and weighted the results in relation to pre-defined tolerance intervals. Subjective and objective image quality was evaluated using a 4-point score and adapted standardized criteria. Scan times were evaluated for each sequence. Results. In total, no acquisitions were lost due to non-diagnostic image quality in the subjective image score. Objective image quality analysis showed no statistically significant differences. The scan time of the CS cines was significantly shorter (p < .001) with mean scan times of 178 ± 36 s compared to 313 ± 65 s for the conventional cine. All cardiac function parameters showed excellent correlation (r 0.978-0.996). Both sequences were considered equivalent for the assessment of LV and RV morphology. Conclusions. The 2-shot CS SAX cines can be used in clinical routine to acquire cardiac morphology in less time compared to the conventional method, with no total loss of acquisitions due to nondiagnostic quality.

TRIAL REGISTRATION

ISRCTN12344380. Registered 20 November 2020, retrospectively registered.

An agenda-setting paper on data sharing platforms: euCanSHare workshop

Various data sharing platforms are being developed to enhance the sharing of cohort data by addressing the fragmented state of data storage and access systems. However, policy challenges in several domains remain unresolved. The euCanSHare workshop was organized to identify and discuss these challenges and to set the future research agenda. Concerns over the multiplicity and long-term sustainability of platforms, lack of resources, access of commercial parties to medical data, credit and recognition mechanisms in academia and the organization of data access committees are outlined. Within these areas, solutions need to be devised to ensure an optimal functioning of platforms.

Targeting human immunodeficiency virus type 1 with antibodies derived from patients with connective tissue disease

During the budding process, human immunodeficiency virus (HIV) acquires several cellular proteins from the host. Thus, antibodies against self antigens found in sera patients with autoimmune disorders may cross react with host-derived or the HIV-specific proteins gp120 and gp41 on the viral envelope and probably neutralize HIV infection. To verify this hypothesis, 88 sera from HIV negative patients suffering from systemic lupus erythematosus (SLE) and other autoimmune disorders were analysed for cross reacting antibodies against HIV-1 by Western blot and FACS analysis indicating that antibodies cross-react with epitopes expressed on HIV infected or non-infected cells. Virus capture assays revealed that HIV-1IIIBwas directly recognized by 60% of sera from patients with autoimmune disorders. Sera were also tested in HIV neutralization assays with stimulated T cells. Reduction of the viral load by patient sera correlated with their reactivity in Western blot analysis. Complement further enhanced the reduction of viral titres, although no complement-mediated lysis was observed. These data suggest a possible protective role of auto-antibodies against HIV infection in lupus patients.

An ecosystem-scale perspective of the net land methanol flux: synthesis of micrometeorological flux measurements

Methanol is the second most abundant volatile organic compound in the troposphere and plays a significant role in atmospheric chemistry. While there is consensus about the dominant role of living plants as the major source and the reaction with OH as the major sink of methanol, global methanol budgets diverge considerably in terms of source/sink estimates reflecting uncertainties in the approaches used to model, and the empirical data used to separately constrain these terms. Here we compiled micrometeorological methanol flux data from eight different study sites and reviewed the corresponding literature in order to provide a first cross-site synthesis of the terrestrial ecosystem-scale methanol exchange and present an independent data-driven view of the land-atmosphere methanol exchange. Our study shows that the controls of plant growth on the production, and thus the methanol emission magnitude, and stomatal conductance on the hourly methanol emission variability, established at the leaf level, hold across sites at the ecosystem-level. Unequivocal evidence for bi-directional methanol exchange at the ecosystem scale is presented. Deposition, which at some sites even exceeds methanol emissions, represents an emerging feature of ecosystem-scale measurements and is likely related to environmental factors favouring the formation of surface wetness. Methanol may adsorb to or dissolve in this surface water and eventually be chemically or biologically removed from it. Management activities in agriculture and forestry are shown to increase local methanol emission by orders of magnitude; they are however neglected at present in global budgets. While contemporary net land methanol budgets are overall consistent with the grand mean of the micrometeorological methanol flux measurements, we caution that the present approach of simulating methanol emission and deposition separately is prone to opposing systematic errors and does not allow taking full advantage of the rich information content of micrometeorological flux measurements.

An ecosystem-scale perspective of the net land methanol flux: synthesis of micrometeorological flux measurements

Methanol is the second most abundant volatile organic compound in the troposphere and plays a significant role in atmospheric chemistry. While there is consensus about the dominant role of living plants as the major source and the reaction with OH as the major sink of methanol, global methanol budgets diverge considerably in terms of source/sink estimates reflecting uncertainties in the approaches used to model, and the empirical data used to separately constrain these terms. Here we compiled micrometeorological methanol flux data from eight different study sites and reviewed the corresponding literature in order to provide a first cross-site synthesis of the terrestrial ecosystem-scale methanol exchange and present an independent data-driven view of the land-atmosphere methanol exchange. Our study shows that the controls of plant growth on the production, and thus the methanol emission magnitude, and stomatal conductance on the hourly methanol emission variability, established at the leaf level, hold across sites at the ecosystem-level. Unequivocal evidence for bi-directional methanol exchange at the ecosystem scale is presented. Deposition, which at some sites even exceeds methanol emissions, represents an emerging feature of ecosystem-scale measurements and is likely related to environmental factors favouring the formation of surface wetness. Methanol may adsorb to or dissolve in this surface water and eventually be chemically or biologically removed from it. Management activities in agriculture and forestry are shown to increase local methanol emission by orders of magnitude; they are however neglected at present in global budgets. While contemporary net land methanol budgets are overall consistent with the grand mean of the micrometeorological methanol flux measurements, we caution that the present approach of simulating methanol emission and deposition separately is prone to opposing systematic errors and does not allow taking full advantage of the rich information content of micrometeorological flux measurements.

Summertime elemental mercury exchange of temperate grasslands on an ecosystem-scale

In order to estimate the air-surface mercury exchange of grasslands in temperate climate regions, fluxes of gaseous elemental mercury (GEM) were measured at two sites in Switzerland and one in Austria during summer 2006. Two classic micrometeorological methods (aerodynamic and modified Bowen ratio) have been applied to estimate net GEM exchange rates and to determine the response of the GEM flux to changes in environmental conditions (e.g. heavy rain, summer ozone) on an ecosystem-scale. Both methods proved to be appropriate to estimate fluxes on time scales of a few hours and longer. Average dry deposition rates up to 4.3 ng m^-2 h^-1 and mean deposition velocities up to 0.10 cm s^-1 were measured, which indicates that during the active vegetation period temperate grasslands are a small net sink for atmospheric mercury. With increasing ozone concentrations depletion of GEM was observed, but could not be quantified from the flux signal. Night-time deposition fluxes of GEM were measured and seem to be the result of mercury co-deposition with condensing water. Effects of grass cuts could also be observed, but were of minor magnitude.

Methanol emissions from deciduous tree species: dependence on temperature and light intensity

Methanol emissions from several deciduous tree species with predominantly mature leaves were measured under laboratory and field conditions. The emissions were modulated by temperature and light. Under constant light conditions in the laboratory, methanol emissions increased with leaf temperature, by up to 12% per degree. At constant temperatures, emissions doubled when light intensity (PAR) increased from darkness to 800 micromol x m(-2) x s(-1). A phenomenological description of light and temperature dependencies was derived from the laboratory measurements. This description was successfully applied to reproduce the diel cycle of methanol emissions from an English oak measured in the field. Labelling experiments with (13)CO(2) provided evidence that less than 10% of the emitted methanol was produced de novo by photosynthesis directly prior to emission. Hence, the light dependence of the emissions cannot be explained by instantaneous production from CO(2) fixation. Additional experiments with selective cooling of plant roots indicated that a substantial fraction of the emitted methanol may be produced in the roots or stem and transported to stomata by the transpiration stream. However, the transpiration stream cannot be considered as the main factor that determines methanol emissions by the investigated plants.

Cut-induced VOC emissions from agricultural grasslands

The introduction of proton transfer reaction mass spectrometry (PTR-MS) for fast response measurements of volatile organic compounds (VOC) has enabled the use of eddy covariance methods to investigate VOC fluxes on the ecosystem scale. In this study PTR-MS flux measurements of VOC were performed over agricultural grassland during and after a cut event. Selected masses detected by the PTR-MS showed fluxes of methanol, acetaldehyde, and acetone. They were highest directly after cutting and during the hay drying phase. Simultaneously, significant fluxes of protonated ion masses 73, 81, and 83 were observed. Due to the limited identification of compounds with the PTR-MS technique, GC-MS and GC-FID-PTR-MS techniques were additionally applied. In this way, ion mass 73 could be identified as 2-butanone, mass 81 mainly as (Z)-3-hexenal, and mass 83 mainly as the sum of (Z)-3-hexenol and hexenyl acetates. Hexenal, hexenols, and the hexenyl acetates are mostly related to plant wounding during cutting. It was found that legume plants and forbs emit a higher number of different VOC species than graminoids.

Contributions of anthropogenic and natural forcing to recent tropopause height changes

Observations indicate that the height of the tropopause-the boundary between the stratosphere and troposphere-has increased by several hundred meters since 1979. Comparable increases are evident in climate model experiments. The latter show that human-induced changes in ozone and well-mixed greenhouse gases account for approximately 80% of the simulated rise in tropopause height over 1979-1999. Their primary contributions are through cooling of the stratosphere (caused by ozone) and warming of the troposphere (caused by well-mixed greenhouse gases). A model-predicted fingerprint of tropopause height changes is statistically detectable in two different observational ("reanalysis") data sets. This positive detection result allows us to attribute overall tropopause height changes to a combination of anthropogenic and natural external forcings, with the anthropogenic component predominating.

Influence of satellite data uncertainties on the detection of externally forced climate change

Two independent analyses of the same satellite-based radiative emissions data yield tropospheric temperature trends that differ by 0.1 degrees C per decade over 1979 to 2001. The troposphere warms appreciably in one satellite data set, while the other data set shows little overall change. These satellite data uncertainties are important in studies seeking to identify human effects on climate. A model-predicted "fingerprint" of combined anthropogenic and natural effects is statistically detectable only in the satellite data set with a warming troposphere. Our findings show that claimed inconsistencies between model predictions and satellite tropospheric temperature data (and between the latter and surface data) may be an artifact of data uncertainties.

Enhancement of complement-mediated lysis by a peptide derived from SCR 13 of complement factor H

Complement factor H (fH) is an important regulator of complement activation. It contributes to protection of cells against homologous complement attack. In this study we tested the effect of fH-depletion of normal human serum (NHS) on lysis of antibody-coated sheep and human erythrocytes (EshA and EhuA). In the absence of fH, lysis of sensitised Esh and Ehu was clearly increased. Addition of fH to fH-depleted serum re-established protection of cells against complement similar to that seen with NHS. A fH-derived peptide (pepAred), covering the N-terminal half of SCR 13 in fH, was able to enhance complement-mediated lysis of EhuA significantly. However, the oxidised form of this peptide (pepAox) had no effect. Biotinylated pepAred, but not pepAox, was able to directly bind to cells. Additionally, pepAred competed with direct fH-cell interaction which was observable only after treatment of purified fH with mercaptoethanol. Only pepAred increased the amount of C3 fragments and reduced levels of fH detectable on cells as shown by FACS analysis and radio-immuno assay. Furthermore, fH and factor I (fI)-mediated cleavage of agarose bound C3b into iC3b was decreased in the presence of pepAred. These data indicate that a fH-derived peptide can enhance complement-mediated lysis. We will continue to investigate whether the use of a fH peptide can be exploited for therapeutical purposes.

[Preoperative administration of antacids to prevent Mendelson's syndrome in pregnant women receiving general anesthesia (author's transl)]

Preoperative administration of 15 mls of the antacid Andursil to women undergoing surgical procedures in GA during pregnancy does reduce the risk of aspiration pneumonitis (Mendelson-Syndrome) from 1:3,5 to 1:15,5. The data show that the number of risk patients is increased in early stages of pregnancy and--surprisingly--in patients undergoing elective surgery. Pains and costs of prophylactic antacids are negligible. Neonates of LSCS patients receiving an antacid are not at an increased risk. As a conclusion the prophylactic administration of oral antacids to all pregnant women having a GA is recommended.