| 1. |
|
|
| 2. |
- Buddenkotte, Thomas, et al.
(författare)
-
Calibrating ensembles for scalable uncertainty quantification in deep learning-based medical image segmentation
- 2023
-
Ingår i: Computers in Biology and Medicine. - : Elsevier Ltd. - 0010-4825 .- 1879-0534. ; 163
-
Tidskriftsartikel (refereegranskat)abstract
- Uncertainty quantification in automated image analysis is highly desired in many applications. Typically, machine learning models in classification or segmentation are only developed to provide binary answers; however, quantifying the uncertainty of the models can play a critical role for example in active learning or machine human interaction. Uncertainty quantification is especially difficult when using deep learning-based models, which are the state-of-the-art in many imaging applications. The current uncertainty quantification approaches do not scale well in high-dimensional real-world problems. Scalable solutions often rely on classical techniques, such as dropout, during inference or training ensembles of identical models with different random seeds to obtain a posterior distribution. In this paper, we present the following contributions. First, we show that the classical approaches fail to approximate the classification probability. Second, we propose a scalable and intuitive framework for uncertainty quantification in medical image segmentation that yields measurements that approximate the classification probability. Third, we suggest the usage of k-fold cross-validation to overcome the need for held out calibration data. Lastly, we motivate the adoption of our method in active learning, creating pseudo-labels to learn from unlabeled images and human–machine collaboration.
|
|
| 3. |
- Buddenkotte, Thomas, et al.
(författare)
-
Deep learning-based segmentation of multisite disease in ovarian cancer
- 2023
-
Ingår i: EUROPEAN RADIOLOGY EXPERIMENTAL. - : Springer Nature. - 2509-9280. ; 7:1
-
Tidskriftsartikel (refereegranskat)abstract
- Purpose: To determine if pelvic/ovarian and omental lesions of ovarian cancer can be reliably segmented on computed tomography (CT) using fully automated deep learning-based methods.Methods: A deep learning model for the two most common disease sites of high-grade serous ovarian cancer lesions (pelvis/ovaries and omentum) was developed and compared against the well-established “no-new-Net” framework and unrevised trainee radiologist segmentations. A total of 451 CT scans collected from four different institutions were used for training (n = 276), evaluation (n = 104) and testing (n = 71) of the methods. The performance was evaluated using the Dice similarity coefficient (DSC) and compared using a Wilcoxon test.Results: Our model outperformed no-new-Net for the pelvic/ovarian lesions in cross-validation, on the evaluation and test set by a significant margin (p values being 4 × 10–7, 3 × 10–4, 4 × 10–2, respectively), and for the omental lesions on the evaluation set (p = 1 × 10–3). Our model did not perform significantly differently in segmenting pelvic/ovarian lesions (p = 0.371) compared to a trainee radiologist. On an independent test set, the model achieved a DSC performance of 71 ± 20 (mean ± standard deviation) for pelvic/ovarian and 61 ± 24 for omental lesions.Conclusion: Automated ovarian cancer segmentation on CT scans using deep neural networks is feasible and achieves performance close to a trainee-level radiologist for pelvic/ovarian lesions.Relevance statement: Automated segmentation of ovarian cancer may be used by clinicians for CT-based volumetric assessments and researchers for building complex analysis pipelines.Key points:The first automated approach for pelvic/ovarian and omental ovarian cancer lesion segmentation on CT images has been presented.Automated segmentation of ovarian cancer lesions can be comparable with manual segmentation of trainee radiologists.Careful hyperparameter tuning can provide models significantly outperforming strong state-of-the-art baselines. Graphical Abstract: [Figure not available: see fulltext.]
|
|
| 4. |
- Caforio, Alida L. P., et al.
(författare)
-
Endomyocardial biopsy: safety and prognostic utility in paediatric and adult myocarditis in the European Society of Cardiology EURObservational Research Programme Cardiomyopathy and Myocarditis Long-Term Registry
- 2024
-
Ingår i: EUROPEAN HEART JOURNAL. - 0195-668X .- 1522-9645.
-
Tidskriftsartikel (refereegranskat)abstract
- Background and Aims Contemporary multicentre data on clinical and diagnostic spectrum and outcome in myocarditis are limited. Study aims were to describe baseline features, 1-year follow-up, and baseline predictors of outcome in clinically suspected or biopsy-proven myocarditis (2013 European Society of Cardiology criteria) in adult and paediatric patients from the EURObservational Research Programme Cardiomyopathy and Myocarditis Long-Term Registry. Methods Five hundred eighty-one (68.0% male) patients, 493 adults, median age 38 (27-52) years, and 88 children, aged 8 (3-13) years, were divided into 3 groups: Group 1 (n = 233), clinically suspected myocarditis with abnormal cardiac magnetic resonance; Group 2 (n = 222), biopsy-proven myocarditis; and Group 3 (n = 126) clinically suspected myocarditis with normal or inconclusive or no cardiac magnetic resonance. Baseline features were analysed overall, in adults vs. children, and among groups. One-year outcome events included death/heart transplantation, ventricular assist device (VAD) or implantable cardioverter defibrillator (ICD) implantation, and hospitalization for cardiac causes. Results Endomyocardial biopsy, mainly right ventricular, had a similarly low complication rate in children and adults (4.7% vs. 4.9%, P = NS), with no procedure-related death. A classical myocarditis pattern on cardiac magnetic resonance was found in 31.3% of children and in 57.9% of adults with biopsy-proven myocarditis (P < .001). At 1-year follow-up, 11/410 patients (2.7%) died, 7 (1.7%) received a heart transplant, 3 underwent VAD (0.7%), and 16 (3.9%) underwent ICD implantation. Independent predictors at diagnosis of death or heart transplantation or hospitalization or VAD implantation or ICD implantation at 1-year follow-up were lower left ventricular ejection fraction and the need for immunosuppressants for new myocarditis diagnosis refractory to non-aetiology-driven therapy. Conclusions Endomyocardial biopsy was safe, and cardiac magnetic resonance using Lake Louise criteria was less sensitive, particularly in children. Virus-negative lymphocytic myocarditis was predominant both in children and adults, and use of immunosuppressive treatments was low. Lower left ventricular ejection fraction and the need for immunosuppressants at diagnosis were independent predictors of unfavourable outcome events at 1 year.
|
|
| 5. |
- Kayser, Yves, et al.
(författare)
-
Core-level nonlinear spectroscopy triggered by stochastic X-ray pulses
- 2019
-
Ingår i: Nature Communications. - : Springer Nature. - 2041-1723. ; 10:1
-
Tidskriftsartikel (refereegranskat)abstract
- Stochastic processes are highly relevant in research fields as different as neuroscience, economy, ecology, chemistry, and fundamental physics. However, due to their intrinsic unpredictability, stochastic mechanisms are very challenging for any kind of investigations and practical applications. Here we report the deliberate use of stochastic X-ray pulses in two-dimensional spectroscopy to the simultaneous mapping of unoccupied and occupied electronic states of atoms in a regime where the opacity and transparency properties of matter are subject to the incident intensity and photon energy. A readily transferable matrix formalism is presented to extract the electronic states from a dataset measured with the monitored input from a stochastic excitation source. The presented formalism enables investigations of the response of the electronic structure to irradiation with intense X-ray pulses while the time structure of the incident pulses is preserved.
|
|
| 6. |
|
|
| 7. |
- Leonardo, Diego A., et al.
(författare)
-
Orientational Ambiguity in Septin Coiled Coils and its Structural Basis
- 2021
-
Ingår i: Journal of Molecular Biology. - : Elsevier BV. - 0022-2836. ; 433:9
-
Tidskriftsartikel (refereegranskat)abstract
- Septins are an example of subtle molecular recognition whereby different paralogues must correctly assemble into functional filaments important for essential cellular events such as cytokinesis. Most possess C-terminal domains capable of forming coiled coils which are believed to be involved in filament formation and bundling. Here, we report an integrated structural approach which aims to unravel their architectural diversity and in so doing provide direct structural information for the coiled-coil regions of five human septins. Unexpectedly, we encounter dimeric structures presenting both parallel and antiparallel arrangements which are in consonance with molecular modelling suggesting that both are energetically accessible. These sequences therefore code for two metastable states of different orientations which employ different but overlapping interfaces. The antiparallel structures present a mixed coiled-coil interface, one side of which is dominated by a continuous chain of core hydrophilic residues. This unusual type of coiled coil could be used to expand the toolkit currently available to the protein engineer for the design of previously unforeseen coiled-coil based assemblies. Within a physiological context, our data provide the first atomic details related to the assumption that the parallel orientation is likely formed between septin monomers from the same filament whilst antiparallelism may participate in the widely described interfilament cross bridges necessary for higher order structures and thereby septin function.
|
|
| 8. |
- Martinelli, Leonardo, et al.
(författare)
-
Fractional Spin Excitations in the Infinite-Layer Cuprate CaCuO2
- 2022
-
Ingår i: Physical Review X. - 2160-3308. ; 12:2
-
Tidskriftsartikel (refereegranskat)abstract
- We use resonant inelastic x-ray scattering (RIXS) to investigate the magnetic dynamics of the infinite-layer cuprate CaCuO2. We find that close to the (1/2,0) point, the single magnon decays into a broad continuum of excitations accounting for about 80% of the total magnetic spectral weight. Polarization-resolved RIXS spectra reveal the overwhelming dominance of the spin-flip (Delta S = 1) character of this continuum with respect to the Delta S = 0 multimagnon contributions. Moreover, its incident-energy dependence is identical to that of the magnon, supporting a common physical origin. We propose that the continuum originates from the decay of the magnon into spinon pairs, and we relate it to the exceptionally high ring exchange J(c) similar to J(1) of CaCuO2. In the infinite-layer cuprates, long-range and multisite hopping integrals are very important, and they amplify the 2D quantum magnetism effects in spite of the 3D antiferromagnetic Neel order.
|
|
| 9. |
- Milne, Christopher J., et al.
(författare)
-
Disentangling the evolution of electrons and holes in photoexcited ZnO nanoparticles
- 2023
-
Ingår i: Structural Dynamics. - : American Institute of Physics (AIP). - 2329-7778. ; 10:6
-
Tidskriftsartikel (refereegranskat)abstract
- The evolution of charge carriers in photoexcited room temperature ZnO nanoparticles in solution is investigated using ultrafast ultraviolet photoluminescence spectroscopy, ultrafast Zn K-edge absorption spectroscopy, and ab initio molecular dynamics (MD) simulations. The photoluminescence is excited at 4.66 eV, well above the band edge, and shows that electron cooling in the conduction band and exciton formation occur in <500 fs, in excellent agreement with theoretical predictions. The x-ray absorption measurements, obtained upon excitation close to the band edge at 3.49 eV, are sensitive to the migration and trapping of holes. They reveal that the 2 ps transient largely reproduces the previously reported transient obtained at 100 ps time delay in synchrotron studies. In addition, the x-ray absorption signal is found to rise in similar to 1.4 ps, which we attribute to the diffusion of holes through the lattice prior to their trapping at singly charged oxygen vacancies. Indeed, the MD simulations show that impulsive trapping of holes induces an ultrafast expansion of the cage of Zn atoms in <200 fs, followed by an oscillatory response at a frequency of similar to 100 cm-1, which corresponds to a phonon mode of the system involving the Zn sub-lattice.
|
|
| 10. |
- Mokso, Rajmund, et al.
(författare)
-
GigaFRoST : The gigabit fast readout system for tomography
- 2017
-
Ingår i: Journal of Synchrotron Radiation. - 0909-0495. ; 24:6, s. 1250-1259
-
Tidskriftsartikel (refereegranskat)abstract
- Owing to recent developments in CMOS technology, it is now possible to exploit tomographic microscopy at third-generation synchrotron facilities with unprecedented speeds. Despite this rapid technical progress, one crucial limitation for the investigation of realistic dynamic systems has remained: A generally short total acquisition time at high frame rates due to the limited internal memory of available detectors. To address and solve this shortcoming, a new detection and readout system, coined GigaFRoST, has been developed based on a commercial CMOS sensor, acquiring and streaming data continuously at 7.7GBs-1 directly to a dedicated backend server. This architecture allows for dynamic data pre-processing as well as data reduction, an increasingly indispensable step considering the vast amounts of data acquired in typical fast tomographic experiments at synchrotron beamlines (up to several tens of TByte per day of raw data).The GigaFRoST detector enables high acquisition rates and long scanning times for dynamic experiments.
|
|
