Joint [Formula: see text] and Image Reconstruction in Low-Field MRI by Physics-Informed Deep-Learning.

OBJECTIVE: We present a model-based image reconstruction approach based on unrolled neural networks which corrects for image distortion and noise in low-field ( B0  âˆ¼  50 mT) MRI. METHODS: Utilising knowledge about the underlying physics, a novel network architecture (SH-Net) is introduced which involves the estimation of spherical harmonic coefficients to guarantee a spatially smooth field map estimate. The SH-Net is integrated in an end-to-end trainable model which jointly estimates the B0-field map as well as the image. Experiments were conducted on retrospectively simulated low-field data of human knees. RESULTS: We compare our model to different model-based approaches at distinct noise levels and various B0-field distributions. Our results show that our physics-informed neural network approach outperforms the purely model-based methods by improving the PSNR up to 11.7% and the RMSE up to 86.3%. CONCLUSION: Our end-to-end trained model-based approach outperforms existing methods in reconstructing image and B0-field maps in the low-field regime. SIGNIFICANCE: low-field MRI is becoming increasingly more popular as it enables access to MR in challenging situations such as intensive care units or resource poor areas. Our method allows for fast and accurate image reconstruction in such low-field imaging with B0-inhomogeneity compensation under a wide range of various environmental conditions.

Quantitative MR Image Reconstruction Using Parameter-Specific Dictionary Learning With Adaptive Dictionary-Size and Sparsity-Level Choice.

OBJECTIVE: We propose a method for the reconstruction of parameter-maps in Quantitative Magnetic Resonance Imaging (QMRI). METHODS: Because different quantitative parameter-maps differ from each other in terms of local features, we propose a method where the employed dictionary learning (DL) and sparse coding (SC) algorithms automatically estimate the optimal dictionary-size and sparsity level separately for each parameter-map. We evaluated the method on a T1-mapping QMRI problem in the brain using the BrainWeb data as well as in-vivo brain images acquired on an ultra-high field 7 T scanner. We compared it to a model-based acceleration for parameter mapping (MAP) approach, other sparsity-based methods using total variation (TV), Wavelets (Wl), and Shearlets (Sh) to a method which uses DL and SC to reconstruct qualitative images, followed by a non-linear (DL+Fit). RESULTS: Our algorithm surpasses MAP, TV, Wl, and Sh in terms of RMSE and PSNR. It yields better or comparable results to DL+Fit by additionally significantly accelerating the reconstruction by a factor of approximately seven. CONCLUSION: The proposed method outperforms the reported methods of comparison and yields accurate T1-maps. Although presented for T1-mapping in the brain, our method's structure is general and thus most probably also applicable for the the reconstruction of other quantitative parameters in other organs. SIGNIFICANCE: From a clinical perspective, the obtained T1-maps could be utilized to differentiate between healthy subjects and patients with Alzheimer's disease. From a technical perspective, the proposed unsupervised method could be employed to obtain ground-truth data for the development of data-driven methods based on supervised learning.

Data-efficient Bayesian learning for radial dynamic MR reconstruction.

BACKGROUND: Cardiac MRI has become the gold-standard imaging technique for assessing cardiovascular morphology and function. In spite of this, its slow data acquisition process presents imaging challenges due to the motion from heartbeats, respiration, and blood flow. In recent studies, deep learning (DL) algorithms have shown promising results for the task of image reconstruction. However, there have been instances where they have introduced artifacts that may be misinterpreted as pathologies or may obscure the detection of pathologies. Therefore, it is important to obtain a metric, such as the uncertainty of the network output, that identifies such artifacts. However, this can be quite challenging for large-scale image reconstruction problems such as dynamic multi-coil non-Cartesian MRI. PURPOSE: To efficiently quantify uncertainties of a physics-informed DL-based image reconstruction method for a large-scale accelerated 2D multi-coil dynamic radial MRI reconstruction problem, and demonstrate the benefits of physics-informed DL over model-agnostic DL in reducing uncertainties while at the same time improving image quality. METHODS: We extended a recently proposed physics-informed 2D U-Net that learns spatio-temporal slices (named XT-YT U-Net), and employed it for the task of uncertainty quantification (UQ) by using Monte Carlo dropout and a Gaussian negative log-likelihood loss function. Our data comprised 2D dynamic MR images acquired with a radial balanced steady-state free precession sequence. The XT-YT U-Net, which allows for training with a limited amount of data, was trained and validated on a dataset of 15 healthy volunteers, and further tested on data from four patients. An extensive comparison between physics-informed and model-agnostic neural networks (NNs) concerning the obtained image quality and uncertainty estimates was performed. Further, we employed calibration plots to assess the quality of the UQ. RESULTS: The inclusion of the MR-physics model of data acquisition as a building block in the NN architecture led to higher image quality (NRMSE: - 33 ± 8.2 % $-33 \pm 8.2 \%$ , PSNR: 6.3 ± 1.3 % $6.3 \pm 1.3 \%$ , and SSIM: 1.9 ± 0.96 % $1.9 \pm 0.96 \%$ ), lower uncertainties ( - 46 ± 8.7 % $-46 \pm 8.7 \%$ ), and, based on the calibration plots, an improved UQ compared to its model-agnostic counterpart. Furthermore, the UQ information can be used to differentiate between anatomical structures (e.g., coronary arteries, ventricle boundaries) and artifacts. CONCLUSIONS: Using an XT-YT U-Net, we were able to quantify uncertainties of a physics-informed NN for a high-dimensional and computationally demanding 2D multi-coil dynamic MR imaging problem. In addition to improving the image quality, embedding the acquisition model in the network architecture decreased the reconstruction uncertainties as well as quantitatively improved the UQ. The UQ provides additional information to assess the performance of different network approaches.

Deep supervised dictionary learning by algorithm unrolling-Application to fast 2D dynamic MR image reconstruction.

BACKGROUND: Unrolled neural networks (NNs) have been extensively applied to different image reconstruction problems across all imaging modalities. A key component of the latter is that they allow for physics-informed learning of the regularization method, which is parametrized by the NN. However, due to the lack of understanding of deep NNs from a theoretical point of view, unrolled NNs are still black-boxes when the regularizers are given by deep NNs, for example, U-Nets. PURPOSE: Dictionarylearning (DL) is a well-established regularization method, which is based on learning a transform to sparsely approximate the signals of interest. Typically, DL-based image reconstruction either employs a dictionary, which was pretrained on a set of patches which were extracted from ground-truth images or a dictionary which is jointly trained during the reconstruction. However, in both cases, the used DL-algorithms are not designed to take into account the reconstruction problem or the underlying physical model, which describes the imaging process. In this work, we propose a DL-algorithm based on unrolled NNs to overcome these limitations. METHODS: We construct an unrolled NN, which corresponds to an unrolled DL-based reconstruction algorithm and train the unrolled NN to optimize its weights, that is, the atoms of the dictionary, by back-propagation in a supervised manner. Further, we propose a new way to employ a 2D dictionary in the spatio-temporal domain. We tested and evaluated the method on an accelerated cardiac cine MR image reconstruction problem using 216/36/36 dynamic images for training, validation, and testing and compared it to two well-known state-of-the-art approaches for cardiac cine MRI based on deep iterative CNNs. Further, we analyze the obtained dictionaries in terms of dictionary-coherence and structure of the atoms. Last, we compare the reported methods in terms of stability by applying them to an entirely different dataset consisting of 49 different test images. RESULTS: The investigated physics-informed DL-approach yields significantly more accurate reconstructions compared to the DL-method, which uses dictionaries obtained by decoupled pretraining, thereby providing an improvement of up to 4.90 dB in terms of PSNR and 5% in terms of SSIM. Further, the proposed spatio-temporal 2D dictionary outperforms the 1D and 3D dictionaries by preventing smoothing of image details while still accurately removing undersampling artifacts and noise resulting in an increase of up to 1.10 dB in terms of PSNR and 4% in terms of SSIM. Although being surpassed by the CNNs on the first dataset, the proposed NNs-based DL method is more stable compared to the latter approach and yields comparable results on the second dataset. Last, it has the advantage of being entirely interpretable in each component. CONCLUSIONS: The presented physics-informed NN can be used as training algorithm for a classical and interpretable data-driven regularization method based on a learned dictionary, which can then not only be linked to the considered data but also to the reconstruction method that the NN defines.

3D model-based super-resolution motion-corrected cardiac T1 mapping.

Objective. To provide 3D high-resolution cardiac T1 maps using model-based super-resolution reconstruction (SRR).Approach. Due to signal-to-noise ratio limitations and the motion of the heart during imaging, often 2D T1 maps with only low through-plane resolution (i.e. slice thickness of 6-8 mm) can be obtained. Here, a model-based SRR approach is presented, which combines multiple stacks of 2D acquisitions with 6-8 mm slice thickness and generates 3D high-resolution T1 maps with a slice thickness of 1.5-2 mm. Every stack was acquired in a different breath hold (BH) and any misalignment between BH was corrected retrospectively. The novelty of the proposed approach is the BH correction and the application of model-based SRR on cardiac T1 Mapping. The proposed approach was evaluated in numerical simulations and phantom experiments and demonstrated in four healthy subjects.Main results. Alignment of BH states was essential for SRR even in healthy volunteers. In simulations, respiratory motion could be estimated with an RMS error of 0.18 ± 0.28 mm. SRR improved the visualization of small structures. High accuracy and precision (average standard deviation of 69.62 ms) of the T1 values was ensured by SRR while the detectability of small structures increased by 40%.Significance. The proposed SRR approach provided T1 maps with high in-plane and high through-plane resolution (1.3 × 1.3 × 1.5-2 mm3). The approach led to improvements in the visualization of small structures and precise T1 values.

An end-to-end-trainable iterative network architecture for accelerated radial multi-coil 2D cine MR image reconstruction.

PURPOSE: Iterative convolutional neural networks (CNNs) which resemble unrolled learned iterative schemes have shown to consistently deliver state-of-the-art results for image reconstruction problems across different imaging modalities. However, because these methods include the forward model in the architecture, their applicability is often restricted to either relatively small reconstruction problems or to problems with operators which are computationally cheap to compute. As a consequence, they have not been applied to dynamic non-Cartesian multi-coil reconstruction problems so far. METHODS: In this work, we propose a CNN architecture for image reconstruction of accelerated 2D radial cine MRI with multiple receiver coils. The network is based on a computationally light CNN component and a subsequent conjugate gradient (CG) method which can be jointly trained end-to-end using an efficient training strategy. We investigate the proposed training strategy and compare our method with other well-known reconstruction techniques with learned and non-learned regularization methods. RESULTS: Our proposed method outperforms all other methods based on non-learned regularization. Further, it performs similar or better than a CNN-based method employing a 3D U-Net and a method using adaptive dictionary learning. In addition, we empirically demonstrate that even by training the network with only iteration, it is possible to increase the length of the network at test time and further improve the results. CONCLUSIONS: End-to-end training allows to highly reduce the number of trainable parameters of and stabilize the reconstruction network. Further, because it is possible to change the length of the network at the test time, the need to find a compromise between the complexity of the CNN-block and the number of iterations in each CG-block becomes irrelevant.

Adaptive sparsity level and dictionary size estimation for image reconstruction in accelerated 2D radial cine MRI.

PURPOSE: In the past, dictionary learning (DL) and sparse coding (SC) have been proposed for the regularization of image reconstruction problems. The regularization is given by a sparse approximation of all image patches using a learned dictionary, that is, an overcomplete set of basis functions learned from data. Despite its competitiveness, DL and SC require the tuning of two essential hyperparameters: the sparsity level S - the number of basis functions of the dictionary, called atoms, which are used to approximate each patch, and K - the overall number of such atoms in the dictionary. These two hyperparameters usually have to be chosen a priori and are determined by repetitive and computationally expensive experiments. Furthermore, the final reported values vary depending on the specific situation. As a result, the clinical application of the method is limited, as standardized reconstruction protocols have to be used. METHODS: In this work, we use adaptive DL and propose a novel adaptive sparse coding algorithm for two-dimensional (2D) radial cine MR image reconstruction. Using adaptive DL and adaptive SC, the optimal dictionary size K as well as the optimal sparsity level S are chosen dependent on the considered data. RESULTS: Our three main results are the following: First, adaptive DL and adaptive SC deliver results which are comparable or better than the most widely used nonadaptive version of DL and SC. Second, the time needed for the regularization is accelerated due to the fact that the sparsity level S is never overestimated. Finally, the a priori choice of S and K is no longer needed but is optimally chosen dependent on the data under consideration. CONCLUSIONS: Adaptive DL and adaptive SC can highly facilitate the application of DL- and SC-based regularization methods. While in this work we focused on 2D radial cine MR image reconstruction, we expect the method to be applicable to different imaging modalities as well.

Clinical quantitative cardiac imaging for the assessment of myocardial ischaemia.

Cardiac imaging has a pivotal role in the prevention, diagnosis and treatment of ischaemic heart disease. SPECT is most commonly used for clinical myocardial perfusion imaging, whereas PET is the clinical reference standard for the quantification of myocardial perfusion. MRI does not involve exposure to ionizing radiation, similar to echocardiography, which can be performed at the bedside. CT perfusion imaging is not frequently used but CT offers coronary angiography data, and invasive catheter-based methods can measure coronary flow and pressure. Technical improvements to the quantification of pathophysiological parameters of myocardial ischaemia can be achieved. Clinical consensus recommendations on the appropriateness of each technique were derived following a European quantitative cardiac imaging meeting and using a real-time Delphi process. SPECT using new detectors allows the quantification of myocardial blood flow and is now also suited to patients with a high BMI. PET is well suited to patients with multivessel disease to confirm or exclude balanced ischaemia. MRI allows the evaluation of patients with complex disease who would benefit from imaging of function and fibrosis in addition to perfusion. Echocardiography remains the preferred technique for assessing ischaemia in bedside situations, whereas CT has the greatest value for combined quantification of stenosis and characterization of atherosclerosis in relation to myocardial ischaemia. In patients with a high probability of needing invasive treatment, invasive coronary flow and pressure measurement is well suited to guide treatment decisions. In this Consensus Statement, we summarize the strengths and weaknesses as well as the future technological potential of each imaging modality.

Spatio-Temporal Deep Learning-Based Undersampling Artefact Reduction for 2D Radial Cine MRI With Limited Training Data.

In this work we reduce undersampling artefacts in two-dimensional (2D) golden-angle radial cine cardiac MRI by applying a modified version of the U-net. The network is trained on 2D spatio-temporal slices which are previously extracted from the image sequences. We compare our approach to two 2D and a 3D deep learning-based post processing methods, three iterative reconstruction methods and two recently proposed methods for dynamic cardiac MRI based on 2D and 3D cascaded networks. Our method outperforms the 2D spatially trained U-net and the 2D spatio-temporal U-net. Compared to the 3D spatio-temporal U-net, our method delivers comparable results, but requiring shorter training times and less training data. Compared to the compressed sensing-based methods kt-FOCUSS and a total variation regularized reconstruction approach, our method improves image quality with respect to all reported metrics. Further, it achieves competitive results when compared to the iterative reconstruction method based on adaptive regularization with dictionary learning and total variation and when compared to the methods based on cascaded networks, while only requiring a small fraction of the computational and training time. A persistent homology analysis demonstrates that the data manifold of the spatio-temporal domain has a lower complexity than the one of the spatial domain and therefore, the learning of a projection-like mapping is facilitated. Even when trained on only one single subject without data-augmentation, our approach yields results which are similar to the ones obtained on a large training dataset. This makes the method particularly suitable for training a network on limited training data. Finally, in contrast to the spatial 2D U-net, our proposed method is shown to be naturally robust with respect to image rotation in image space and almost achieves rotation-equivariance where neither data-augmentation nor a particular network design are required.

WHO 2004 criteria and CK19 are reliable prognostic markers in pancreatic endocrine tumors.

BACKGROUND: It is difficult to predict the biologic behavior of pancreatic endocrine tumors in absence of metastases or invasion into adjacent organs. The World Health Organization (WHO) has proposed in 2004 size, angioinvasion, mitotic activity, and MIB1 proliferation index as prognostic criteria. Our aim was to test retrospectively the predictive value of these 2004 WHO criteria and of CK19, CD99, COX2, and p27 immunohistochemistry in a large series of patients with long-term follow-up. DESIGN: The histology of 216 pancreatic endocrine tumor specimens was reviewed and the tumors were reclassified according to the 2004 WHO classification. The prognostic value of the WHO classification and the histopathologic criteria necrosis and nodular fibrosis was tested in 113 patients. A tissue microarray was constructed for immunohistochemical staining. The staining results were scored quantitatively for MIB1 and semiquantitatively for CK19, COX2, p27, and CD99. The prognostic value of these markers was tested in 93 patients. RESULTS: The stratification of the patients into 4 risk groups according to the 2004 WHO classification was reliable with regard to both time span to relapse and tumor-specific death. In a multivariate analysis, the CK19 status was shown to be independent of the WHO criteria. By contrast, the prognostic significance of COX2, p27, and CD99 could not be confirmed. CONCLUSIONS: The 2004 WHO classification with 4 risk groups is very reliable for predicting both disease-free survival and the time span until tumor-specific death. CK19 staining is a potential additional prognostic marker independent from the WHO criteria for pancreatic endocrine tumors.

Chromosomal translocations t(4;14), t(11;14) and proliferation rate stratify patients with mature plasma cell myelomas into groups with different survival probabilities: a molecular epidemiologic study on tissue microarrays.

Plasma cell myelomas (PMs) exhibit clinical and molecular heterogeneity. To date, morphology and immunohistochemistry on bone marrow trephines are of limited value to stratify patients into different prognostic categories. However, some chromosomal translocations are of prognostic and/or of predictive importance in PMs. In this study, the prognostic significance of morphology, CyclinD1 expression, proliferation index (Mib1) and presence of the translocations FGFR3/IgH [t(4;14)] and CCND1/IgH [t(11;14)] are compared in 119 patients with PM. Immunohistochemistry and fluorescence in situ hybridization analysis were carried out on a tissue microarray containing bone marrow trephines. Hundred and one PMs showed a mature morphology whereas 10 were immature. All but one PM carrying a translocation showed a mature morphology. Patients with a t(4;14) (12%) had a statistically significant shorter 1-year survival (P=0.004), whereas those with a t(11;14) (21%) had a trend towards a better clinical outcome. CyclinD1 protein expression was not significantly associated with survival. Besides the t(4;14), an immature morphology (P<0.001) and a proliferation index (Mib1) of more than 10% (P=0.002) were associated with a significantly worse outcome. A high occurrence of strong CyclinD1 protein expression in the tumor cells was predictive of either a t(11;14) or of a low level amplification of the CCND1 gene, suggesting that different molecular mechanisms may have lead to an over-expression of the CyclinD1 protein in PMs. These findings demonstrate that a high proliferation rate and translocations involving the IgH locus can stratify mature PMs into groups with distinct survival probabilities.

Cancer risk in the Swiss HIV Cohort Study: associations with immunodeficiency, smoking, and highly active antiretroviral therapy.

BACKGROUND: Persons infected with human immunodeficiency virus (HIV) have an increased risk for several cancers, but the influences of behavioral risk factors, such as smoking and intravenous drug use, and highly active antiretroviral therapy (HAART) on cancer risk are not clear. METHODS: Patient records were linked between the Swiss HIV Cohort Study and Swiss cantonal cancer registries. Observed and expected numbers of incident cancers were assessed in 7304 persons infected with HIV followed for 28,836 person-years. Relative risks for cancer compared with those for the general population were determined by estimating cancer registry-, sex-, age-, and period-standardized incidence ratios (SIRs). RESULTS: Highly elevated SIRs were confirmed in persons infected with HIV for Kaposi sarcoma (KS) (SIR = 192, 95% confidence interval [CI] = 170 to 217) and non-Hodgkin lymphoma (SIR = 76.4, 95% CI = 66.5 to 87.4). Statistically significantly elevated SIRs were also observed for anal cancer (SIR = 33.4, 95% CI = 10.5 to 78.6); Hodgkin lymphoma (SIR = 17.3, 95% CI = 10.2 to 27.4); cancers of the cervix (SIR = 8.0, 95% CI = 2.9 to 17.4); liver (SIR = 7.0, 95% CI = 2.2 to 16.5); lip, mouth, and pharynx (SIR = 4.1, 95% CI = 2.1 to 7.4); trachea, lung, and bronchus (SIR = 3.2, 95% CI = 1.7 to 5.4); and skin, nonmelanomatous (SIR = 3.2, 95% CI = 2.2 to 4.5). In HAART users, SIRs for KS (SIR = 25.3, 95% CI = 10.8 to 50.1) and non-Hodgkin lymphoma (SIR = 24.2, 95% CI = 15.0 to 37.1) were lower than those for nonusers (KS SIR = 239, 95% CI = 211 to 270; non-Hodgkin lymphoma SIR = 99.3, 95% CI = 85.8 to 114). Among HAART users, however, the SIR (although not absolute numbers) for Hodgkin lymphoma (SIR = 36.2, 95% CI = 16.4 to 68.9) was comparable to that for KS and non-Hodgkin lymphoma. No clear impact of HAART on SIRs emerged for cervical cancer or non-acquired immunodeficiency syndrome-defining cancers. Cancers of the lung, lip, mouth, or pharynx were not observed among nonsmokers. CONCLUSION: In persons infected with HIV, HAART use may prevent most excess risk of KS and non-Hodgkin lymphoma, but not that of Hodgkin lymphoma and other non-acquired immunodeficiency syndrome-defining cancers. No cancers of the lip, mouth, pharynx, or lung were observed in nonsmokers.

Meta-analysis of randomized controlled trials of simplified versus continued protease inhibitor-based antiretroviral therapy in HIV-1-infected patients.

OBJECTIVE: To evaluate the efficacy and safety of simplified maintenance therapy (SMT) compared with continued protease inhibitor (PI) therapy. DESIGN: Meta-analysis of nine randomized controlled trials in which 833 patients were switched to SMT (abacavir, efavirenz or nevirapine) and 616 continued PI, assessing virologic failure (primary outcome), discontinuation of therapy for reasons other than virologic failure, CD4 cell count, total plasma cholesterol and triglycerides. RESULTS: The risk ratio for virologic failure for SMT compared to continued PI was 1.06 [95% confidence interval (CI) 0.58-1.92; test for homogeneity P = 0.01] for SMT, 2.56, (95% CI, 1.17-5.64) for abacavir, 0.83 (95% CI, 0.36-1.91) for efavirenz and 0.54 (95% CI, 0.29-1.02) for nevirapine. The risk ratio for premature discontinuation of therapy with SMT was 0.61 (95% CI, 0.48-0.77; test for homogeneity P < 0.10). The difference in absolute mean cholesterol for SMT compared to continued PI was -0.15 mmol/l, (95% CI, -0.40 to 0.09; test for homogeneity P < 0.01) for SMT, -0.51 mmol/l (95% CI, -0.70 to -0.33) for abacavir, 0.22 mmol/l (95% CI, 0 to 0.43) for efavirenz and -0.19 mmol/l (95% CI, -0.48 to 0.09) for nevirapine. CONCLUSIONS: Current evidence suggests that SMT with abacavir rather than continued PI increases the risk of virologic failure, this increased risk may be confined to patients with prior mono or dual therapy with reverse transcriptase inhibitors. There is not enough evidence on whether SMT with efavirenz and nevirapine influences the risk of virologic failure. SMT with any of the three drugs reduces the risk of discontinuation of therapy, and SMT with abacavir reduces plasma cholesterol.