Your browser doesn't support javascript.
loading
Show: 20 | 50 | 100
Results 1 - 20 de 291
Filter
1.
Radiology ; 307(4): e221922, 2023 05.
Article in English | MEDLINE | ID: mdl-36975820

ABSTRACT

Background Several single-center studies found that high contralateral parenchymal enhancement (CPE) at breast MRI was associated with improved long-term survival in patients with estrogen receptor (ER)-positive and human epidermal growth factor receptor 2 (HER2)-negative breast cancer. Due to varying sample sizes, population characteristics, and follow-up times, consensus of the association is currently lacking. Purpose To confirm whether CPE is associated with long-term survival in a large multicenter retrospective cohort, and to investigate if CPE is associated with endocrine therapy effectiveness. Materials and Methods This multicenter observational cohort included women with unilateral ER-positive HER2-negative breast cancer (tumor size ≤50 mm and ≤three positive lymph nodes) who underwent MRI from January 2005 to December 2010. Overall survival (OS), recurrence-free survival (RFS), and distant RFS (DRFS) were assessed. Kaplan-Meier analysis was performed to investigate differences in absolute risk after 10 years, stratified according to CPE tertile. Multivariable Cox proportional hazards regression analysis was performed to investigate whether CPE was associated with prognosis and endocrine therapy effectiveness. Results Overall, 1432 women (median age, 54 years [IQR, 47-63 years]) were included from 10 centers. Differences in absolute OS after 10 years were stratified according to CPE tertile as follows: 88.5% (95% CI: 88.1, 89.1) in tertile 1, 85.8% (95% CI: 85.2, 86.3) in tertile 2, and 85.9% (95% CI: 85.4, 86.4) in tertile 3. CPE was independently associated with OS, with a hazard ratio (HR) of 1.17 (95% CI: 1.0, 1.36; P = .047), but was not associated with RFS (HR, 1.11; P = .16) or DRFS (HR, 1.11; P = .19). The effect of endocrine therapy on survival could not be accurately assessed; therefore, the association between endocrine therapy efficacy and CPE could not reliably be estimated. Conclusion High contralateral parenchymal enhancement was associated with a marginally decreased overall survival in patients with estrogen receptor-positive and human epidermal growth factor receptor 2-negative breast cancer, but was not associated with recurrence-free survival (RFS) or distant RFS. Published under a CC BY 4.0 license. Supplemental material is available for this article. See also the editorial by Honda and Iima in this issue.


Subject(s)
Breast Neoplasms , Triple Negative Breast Neoplasms , Humans , Female , Middle Aged , Breast Neoplasms/diagnostic imaging , Breast Neoplasms/drug therapy , Breast Neoplasms/metabolism , Receptors, Estrogen , Retrospective Studies , Triple Negative Breast Neoplasms/pathology , Breast/diagnostic imaging , Breast/metabolism , Prognosis , Receptor, ErbB-2/metabolism , Magnetic Resonance Imaging/methods , Disease-Free Survival , Neoplasm Recurrence, Local/pathology
2.
J Magn Reson Imaging ; 58(6): 1739-1749, 2023 12.
Article in English | MEDLINE | ID: mdl-36928988

ABSTRACT

BACKGROUND: While several methods have been proposed for automated assessment of breast-cancer response to neoadjuvant chemotherapy on breast MRI, limited information is available about their performance across multiple institutions. PURPOSE: To assess the value and robustness of deep learning-derived volumes of locally advanced breast cancer (LABC) on MRI to infer the presence of residual disease after neoadjuvant chemotherapy. STUDY TYPE: Retrospective. SUBJECTS: Training cohort: 102 consecutive female patients with LABC scheduled for neoadjuvant chemotherapy (NAC) from a single institution (age: 25-73 years). Independent testing cohort: 55 consecutive female patients with LABC from four institutions (age: 25-72 years). FIELD STRENGTH/SEQUENCE: Training cohort: single vendor 1.5 T or 3.0 T. Testing cohort: multivendor 3.0 T. Gradient echo dynamic contrast-enhanced sequences. ASSESSMENT: A convolutional neural network (nnU-Net) was trained to segment LABC. Based on resulting tumor volumes, an extremely randomized tree model was trained to assess residual cancer burden (RCB)-0/I vs. RCB-II/III. An independent model was developed using functional tumor volume (FTV). Models were tested on an independent testing cohort and response assessment performance and robustness across multiple institutions were assessed. STATISTICAL TESTS: The receiver operating characteristic (ROC) was used to calculate the area under the ROC curve (AUC). DeLong's method was used to compare AUCs. Correlations were calculated using Pearson's method. P values <0.05 were considered significant. RESULTS: Automated segmentation resulted in a median (interquartile range [IQR]) Dice score of 0.87 (0.62-0.93), with similar volumetric measurements (R = 0.95, P < 0.05). Automated volumetric measurements were significantly correlated with FTV (R = 0.80). Tumor volume-derived from deep learning of DCE-MRI was associated with RCB, yielding an AUC of 0.76 to discriminate between RCB-0/I and RCB-II/III, performing similar to the FTV-based model (AUC = 0.77, P = 0.66). Performance was comparable across institutions (IQR AUC: 0.71-0.84). DATA CONCLUSION: Deep learning-based segmentation estimates changes in tumor load on DCE-MRI that are associated with RCB after NAC and is robust against variations between institutions. EVIDENCE LEVEL: 2. TECHNICAL EFFICACY: Stage 4.


Subject(s)
Breast Neoplasms , Deep Learning , Humans , Adult , Middle Aged , Aged , Female , Breast Neoplasms/pathology , Retrospective Studies , Neoplasm, Residual/diagnostic imaging , Treatment Outcome , Magnetic Resonance Imaging/methods , Neoadjuvant Therapy/methods
3.
Eur Radiol ; 32(7): 4537-4546, 2022 Jul.
Article in English | MEDLINE | ID: mdl-35190891

ABSTRACT

OBJECTIVES: Visualization of the bone distribution is an important prerequisite for MRI-guided high-intensity focused ultrasound (MRI-HIFU) treatment planning of bone metastases. In this context, we evaluated MRI-based synthetic CT (sCT) imaging for the visualization of cortical bone. METHODS: MR and CT images of nine patients with pelvic and femoral metastases were retrospectively analyzed in this study. The metastatic lesions were osteolytic, osteoblastic or mixed. sCT were generated from pre-treatment or treatment MR images using a UNet-like neural network. sCT was qualitatively and quantitatively compared to CT in the bone (pelvis or femur) containing the metastasis and in a region of interest placed on the metastasis itself, through mean absolute difference (MAD), mean difference (MD), Dice similarity coefficient (DSC), and root mean square surface distance (RMSD). RESULTS: The dataset consisted of 3 osteolytic, 4 osteoblastic and 2 mixed metastases. For most patients, the general morphology of the bone was well represented in the sCT images and osteolytic, osteoblastic and mixed lesions could be discriminated. Despite an average timespan between MR and CT acquisitions of 61 days, in bone, the average (± standard deviation) MAD was 116 ± 26 HU, MD - 14 ± 66 HU, DSC 0.85 ± 0.05, and RMSD 2.05 ± 0.48 mm and, in the lesion, MAD was 132 ± 62 HU, MD - 31 ± 106 HU, DSC 0.75 ± 0.2, and RMSD 2.73 ± 2.28 mm. CONCLUSIONS: Synthetic CT images adequately depicted the cancellous and cortical bone distribution in the different lesion types, which shows its potential for MRI-HIFU treatment planning. KEY POINTS: • Synthetic computed tomography was able to depict bone distribution in metastatic lesions. • Synthetic computed tomography images intrinsically aligned with treatment MR images may have the potential to facilitate MR-HIFU treatment planning of bone metastases, by combining visualization of soft tissues and cancellous and cortical bone.


Subject(s)
Bone Neoplasms , Magnetic Resonance Imaging , Bone Neoplasms/diagnostic imaging , Bone Neoplasms/therapy , Feasibility Studies , Femur/diagnostic imaging , Humans , Magnetic Resonance Imaging/methods , Pelvis , Radiotherapy Planning, Computer-Assisted/methods , Retrospective Studies , Tomography, X-Ray Computed/methods
4.
Hum Brain Mapp ; 42(2): 367-383, 2021 02 01.
Article in English | MEDLINE | ID: mdl-33035372

ABSTRACT

Gradient nonlinearities in magnetic resonance imaging (MRI) cause spatially varying mismatches between the imposed and the effective gradients and can cause significant biases in rotationally invariant diffusion MRI measures derived from, for example, diffusion tensor imaging. The estimation of the orientational organization of fibrous tissue, which is nowadays frequently performed with spherical deconvolution techniques ideally using higher diffusion weightings, can likewise be biased by gradient nonlinearities. We explore the sensitivity of two established spherical deconvolution approaches to gradient nonlinearities, namely constrained spherical deconvolution (CSD) and damped Richardson-Lucy (dRL). Additionally, we propose an extension of dRL to take into account gradient imperfections, without the need of data interpolation. Simulations show that using the effective b-matrix can improve dRL fiber orientation estimation and reduces angular deviations, while CSD can be more robust to gradient nonlinearity depending on the implementation. Angular errors depend on a complex interplay of many factors, including the direction and magnitude of gradient deviations, underlying microstructure, SNR, anisotropy of the effective response function, and diffusion weighting. Notably, angular deviations can also be observed at lower b-values in contrast to the perhaps common assumption that only high b-value data are affected. In in vivo Human Connectome Project data and acquisitions from an ultrastrong gradient (300 mT/m) scanner, angular differences are observed between applying and not applying the effective gradients in dRL estimation. As even small angular differences can lead to error propagation during tractography and as such impact connectivity analyses, incorporating gradient deviations into the estimation of fiber orientations should make such analyses more reliable.


Subject(s)
Brain/diagnostic imaging , Databases, Factual , Diffusion Magnetic Resonance Imaging/methods , Nerve Fibers, Myelinated , Nonlinear Dynamics , White Matter/diagnostic imaging , Anisotropy , Databases, Factual/standards , Diffusion Magnetic Resonance Imaging/standards , Humans
5.
Magn Reson Med ; 86(5): 2647-2655, 2021 11.
Article in English | MEDLINE | ID: mdl-34061390

ABSTRACT

PURPOSE: To demonstrate that interleaved MR thermometry can monitor temperature in water and fat with adequate temporal resolution. This is relevant for high intensity focused uUltrasounds (HIFU) treatment of bone lesions, which are often found near aqueous tissues, as muscle, or embedded in adipose tissues, as subcutaneous fat and bone marrow. METHODS: Proton resonance frequency shift (PRFS)-based thermometry scans and T1 -based 2D variable flip angle (2D-VFA) thermometry scans were acquired alternatingly over time. Temperature in water was monitored using PRFS thermometry, and in fat by 2D-VFA thermometry with slice profile effect correction. The feasibility of interleaved water/fat temperature monitoring was studied ex vivo in porcine bone during MR-HIFU sonication. Precision and stability of measurements in vivo were evaluated in a healthy volunteer under non-heating conditions. RESULTS: The method allowed observing temperature change over time in muscle and fat, including bone marrow, during MR-HIFU sonication, with a temporal resolution of 6.1 s. In vivo, the apparent temperature change was stable on the time scale of the experiment: In 7 min the systematic drift was <0.042°C/min in muscle (PRFS after drift correction) and <0.096°C/min in bone marrow (2D-VFA). The SD of the temperature change averaged over time was 0.98°C (PRFS) and 2.7°C (2D-VFA). CONCLUSIONS: Interleaved MR thermometry allows temperature measurements in water and fat with a temporal resolution high enough for monitoring HIFU ablation. Specifically, combined fat and water thermometry provides uninterrupted information on temperature changes in tissue close to the bone cortex.


Subject(s)
High-Intensity Focused Ultrasound Ablation , Thermometry , Animals , Humans , Magnetic Resonance Imaging , Swine , Temperature , Water
6.
NMR Biomed ; 34(8): e4542, 2021 08.
Article in English | MEDLINE | ID: mdl-34031938

ABSTRACT

PURPOSE: To perform dynamic T1 mapping using a 2D variable flip angle (VFA) method, a correction for the slice profile effect is needed. In this work we investigated the impact of flip angle selection and excitation RF pulse profile on the performance of slice profile correction when applied to T1 mapping over a range of T1 values. METHODS: A correction of the slice profile effect is proposed, based on Bloch simulation of steady-state signals. With this correction, Monte Carlo simulations were performed to assess the accuracy and precision of 2D VFA T1 mapping in the presence of noise, for RF pulses with time-bandwidth products of 2, 3 and 10 and with flip angle pairs in the range [1°-90°]. To evaluate its performance over a wide range of T1 , maximum errors were calculated for six T1 values between 50 ms and 1250 ms. The method was demonstrated using in vitro and in vivo experiments. RESULTS: Without corrections, 2D VFA severely underestimates T1 . Slice profile errors were effectively reduced with the correction based on simulations, both in vitro and in vivo. The precision and accuracy of the method depend on the nominal T1 values, the FA pair, and the RF pulse shape. FA pairs leading to <5% errors in T1 can be identified for the common RF shapes, for T1 values between 50 ms and 1250 ms. CONCLUSIONS: 2D VFA T1 mapping with Bloch-simulation-based correction can deliver T1 estimates that are accurate and precise to within 5% over a wide T1 range.


Subject(s)
Algorithms , Magnetic Resonance Imaging , Humans , Phantoms, Imaging , Radio Waves , Reproducibility of Results
7.
Neuroimage ; 205: 116127, 2020 01 15.
Article in English | MEDLINE | ID: mdl-31476431

ABSTRACT

Nonlinearities of gradient magnetic fields in diffusion MRI (dMRI) can introduce systematic errors in estimates of diffusion measures. While there are correction methods that can compensate for these errors, as presented in the Human Connectome Project, such nonlinear effects are often assumed to be negligible for typical applications, and as a result, gradient nonlinearities are mostly left uncorrected. In this work, we perform a systematic analysis to investigate the effect of gradient nonlinearities on dMRI studies, from voxel-wise estimates to group study outcomes. We present a novel framework to quantify and visualize these effects by decomposing them into their magnitude and angle components. Mean magnitude deviation and fractional gradient anisotropy are introduced to quantify the distortions in the size and shape of gradient vector distributions. By means of Monte-Carlo simulations and real data from the Human Connectome Project, the errors on dMRI measures derived from the diffusion tensor imaging and diffusional kurtosis imaging are highlighted. We perform a group study to showcase the alteration in the significance and effect size due to ignoring gradient nonlinearity correction. Our results indicate that the effect of gradient field nonlinearities on dMRI studies can be significant and may complicate the interpretation of the results and conclusions.


Subject(s)
Brain/diagnostic imaging , Diffusion Magnetic Resonance Imaging/methods , Neuroimaging/methods , Adult , Computer Simulation , Connectome , Diffusion Tensor Imaging/methods , Female , Humans , Male , Phantoms, Imaging , Research Design , Young Adult
8.
Neuroimage ; 208: 116466, 2020 03.
Article in English | MEDLINE | ID: mdl-31843712

ABSTRACT

Displacement Encoding with Stimulated Echoes (DENSE) has recently shown potential for measuring cardiac-induced cerebral volumetric strain in the human brain. As such, it may provide a powerful tool for investigating the cerebral small vessels. However, further development and validation are necessary. This study aims, first, to validate a retrospectively-gated implementation of the DENSE method for assessing brain tissue pulsations as a physiological marker, and second, to use the acquired measurements to explore intracranial volume dynamics. We acquired repeated measurements of cerebral volumetric strain in 8 healthy subjects, and internally validated these measurements by comparing them to spinal CSF stroke volumes obtained in the same scan session. Peak volumetric strain was found to be highly repeatable between scan sessions. First/second measured peak volumetric strains were: (6.4 â€‹± â€‹1.7)x10-4/(6.7 â€‹± â€‹1.6)x10-4 for whole brain, (9.5 â€‹± â€‹2.5)x10-4/(9.6 â€‹± â€‹2.4)x10-4 for grey matter, and (4.4 â€‹± â€‹1.7)x10-4/(4.1 â€‹± â€‹0.8)x10-4 for white matter. Grey matter showed significantly higher peak strain (p â€‹< â€‹0.001) and earlier time-to-peak strain (p â€‹< â€‹0.02) than white matter. An approximately linear relationship was found between CSF and brain tissue volume pulsations over the cardiac cycle (mean slope and R2 of 0.88 â€‹± â€‹0.23 and 0.89 â€‹± â€‹0.07, respectively). The close similarity between CSF and brain tissue volume pulsations implies limited contributions from large intracranial vessel pulsations, providing further evidence for venous compression as an additional mechanism for maintaining stable intracranial pressures over the cardiac cycle. Cerebral pulsatility showed consistent inter-subject peak values in healthy subjects, and was strongly correlated to CSF stroke volumes. These results strengthen the potential of brain tissue volumetric strain as a means for investigating the intracranial dynamics of the ageing brain in normal or diseased states.


Subject(s)
Cerebrospinal Fluid/diagnostic imaging , Cerebrovascular Circulation/physiology , Gray Matter/diagnostic imaging , Magnetic Resonance Imaging/methods , Microvessels/diagnostic imaging , Neuroimaging/methods , Pulsatile Flow/physiology , White Matter/physiology , Adult , Cerebrospinal Fluid/physiology , Female , Gray Matter/physiology , Humans , Male , Microvessels/physiology , Young Adult
9.
Neuroimage ; 218: 116948, 2020 09.
Article in English | MEDLINE | ID: mdl-32428705

ABSTRACT

Spherical deconvolution is a widely used approach to quantify the fiber orientation distribution (FOD) from diffusion MRI data of the brain. The damped Richardson-Lucy (dRL) is an algorithm developed to perform robust spherical deconvolution on single-shell diffusion MRI data while suppressing spurious FOD peaks due to noise or partial volume effects. Due to recent progress in acquisition hardware and scanning protocols, it is becoming increasingly common to acquire multi-shell diffusion MRI data, which allows for the modelling of multiple tissue types beyond white matter. While the dRL algorithm could, in theory, be directly applied to multi-shell data, it is not optimised to exploit its information content to model the signal from multiple tissue types. In this work, we introduce a new framework based on dRL - dubbed generalized Richardson-Lucy (GRL) - that uses multi-shell data in combination with user-chosen tissue models to disentangle partial volume effects and increase the accuracy in FOD estimation. Further, GRL estimates signal fraction maps associated to each user-selected model, which can be used during fiber tractography to dissect and terminate the tracking without need for additional structural data. The optimal weighting of multi-shell data in the fit and the robustness to noise and to partial volume effects of GRL was studied with synthetic data. Subsequently, we investigated the performance of GRL in comparison to dRL and to multi-shell constrained spherical deconvolution (MSCSD) on a high-resolution diffusion MRI dataset from the Human Connectome Project and on an MRI dataset acquired at 3T on a clinical scanner. In line with previous studies, we described the signal of the cerebrospinal-fluid and of the grey matter with isotropic diffusion models, whereas four diffusion models were considered to describe the white matter. With a third dataset including small diffusion weightings, we studied the feasibility of including intra-voxel incoherent motion effects due to blood pseudo-diffusion in the modelling. Further, the reliability of GRL was demonstrated with a test-retest scan of a subject acquired at 3T. Results of simulations show that GRL can robustly disentangle different tissue types at SNR above 20 with respect to the non-weighted image, and that it improves the angular accuracy of the FOD estimation as compared to dRL. On real data, GRL provides signal fraction maps that are physiologically plausible and consistent with those obtained with MSCSD, with correlation coefficients between the two methods up to 0.96. When considering IVIM effects, a high blood pseudo-diffusion fraction is observed in the medial temporal lobe and in the sagittal sinus. In comparison to dRL and MSCSD, GRL provided sharper FODs and less spurious peaks in presence of partial volume effects, but the FOD reconstructions are also highly dependent on the chosen modelling of white matter. When performing fiber tractography, GRL allows to terminate fiber tractography using the signal fraction maps, which results in a better tract termination at the grey-white matter interface or at the outer cortical surface. In terms of inter-scan reliability, GRL performed similarly to or better than compared methods. In conclusion, GRL offers a new modular and flexible framework to perform spherical deconvolution of multi-shell data.


Subject(s)
Algorithms , Diffusion Magnetic Resonance Imaging/methods , Image Processing, Computer-Assisted/methods , Brain Mapping , Cerebrospinal Fluid , Computer Simulation , Connectome , Diffusion Magnetic Resonance Imaging/statistics & numerical data , Feasibility Studies , Humans , Reproducibility of Results , Superior Sagittal Sinus/diagnostic imaging , Temporal Lobe/diagnostic imaging , White Matter/diagnostic imaging
10.
Hum Brain Mapp ; 41(16): 4478-4499, 2020 11.
Article in English | MEDLINE | ID: mdl-32851729

ABSTRACT

Diffusion magnetic resonance imaging can indirectly infer the microstructure of tissues and provide metrics subject to normal variability in a population. Potentially abnormal values may yield essential information to support analysis of controls and patients cohorts, but subtle confounds could be mistaken for purely biologically driven variations amongst subjects. In this work, we propose a new harmonization algorithm based on adaptive dictionary learning to mitigate the unwanted variability caused by different scanner hardware while preserving the natural biological variability of the data. Our harmonization algorithm does not require paired training data sets, nor spatial registration or matching spatial resolution. Overcomplete dictionaries are learned iteratively from all data sets at the same time with an adaptive regularization criterion, removing variability attributable to the scanners in the process. The obtained mapping is applied directly in the native space of each subject toward a scanner-space. The method is evaluated with a public database which consists of two different protocols acquired on three different scanners. Results show that the effect size of the four studied diffusion metrics is preserved while removing variability attributable to the scanner. Experiments with alterations using a free water compartment, which is not simulated in the training data, shows that the modifications applied to the diffusion weighted images are preserved in the diffusion metrics after harmonization, while still reducing global variability at the same time. The algorithm could help multicenter studies pooling their data by removing scanner specific confounds, and increase statistical power in the process.


Subject(s)
Algorithms , Biological Variation, Population , Brain/diagnostic imaging , Diffusion Magnetic Resonance Imaging , Image Processing, Computer-Assisted , Multicenter Studies as Topic , Neuroimaging , Diffusion Magnetic Resonance Imaging/methods , Diffusion Magnetic Resonance Imaging/standards , Humans , Image Processing, Computer-Assisted/methods , Image Processing, Computer-Assisted/standards , Neuroimaging/methods , Neuroimaging/standards
11.
Radiology ; 296(2): 277-287, 2020 08.
Article in English | MEDLINE | ID: mdl-32452738

ABSTRACT

Background Better understanding of the molecular biology associated with MRI phenotypes may aid in the diagnosis and treatment of breast cancer. Purpose To discover the associations between MRI phenotypes of breast cancer and their underlying molecular biology derived from gene expression data. Materials and Methods This is a secondary analysis of the Multimodality Analysis and Radiologic Guidance in Breast-Conserving Therapy, or MARGINS, study. MARGINS included patients eligible for breast-conserving therapy between November 2000 and December 2008 for preoperative breast MRI. Tumor RNA was collected for sequencing from surgical specimen. Twenty-one computer-generated MRI features of tumors were condensed into seven MRI factors related to tumor size, shape, initial enhancement, late enhancement, smoothness of enhancement, sharpness, and sharpness variation. These factors were associated with gene expression levels from RNA sequencing by using gene set enrichment analysis. Statistical significance of these associations was evaluated by using a sample permutation test and the false discovery rate. Results Gene expression and MRI data were obtained for 295 patients (mean age, 56 years ± 10.3 [standard deviation]). Larger and more irregular tumors showed increased expression of cell cycle and DNA damage checkpoint genes (false discovery rate <0.25; normalized enrichment statistic [NES], 2.15). Enhancement and sharpness of the tumor margin were associated with expression of ribosomal proteins (false discovery rate <0.25; NES, 1.95). Smoothness of enhancement, tumor size, and tumor shape were associated with expression of genes involved in the extracellular matrix (false discovery rate <0.25; NES, 2.25). Conclusion Breast cancer MRI phenotypes were related to their underlying molecular biology revealed by using RNA sequencing. The association between enhancements and sharpness of the tumor margin with the ribosome suggests that these MRI features may be imaging biomarkers for drugs targeting the ribosome. © RSNA, 2020 Online supplemental material is available for this article. See also the editorial by Cho in this issue.


Subject(s)
Breast Neoplasms , Imaging Genomics/classification , Magnetic Resonance Imaging/classification , Transcriptome/genetics , Adult , Aged , Aged, 80 and over , Breast/diagnostic imaging , Breast/metabolism , Breast/pathology , Breast Neoplasms/diagnostic imaging , Breast Neoplasms/genetics , Breast Neoplasms/metabolism , Breast Neoplasms/pathology , Cohort Studies , Female , Humans , Middle Aged , Phenotype
12.
Radiology ; 295(1): 66-79, 2020 04.
Article in English | MEDLINE | ID: mdl-32043947

ABSTRACT

Background Although several deep learning (DL) calcium scoring methods have achieved excellent performance for specific CT protocols, their performance in a range of CT examination types is unknown. Purpose To evaluate the performance of a DL method for automatic calcium scoring across a wide range of CT examination types and to investigate whether the method can adapt to different types of CT examinations when representative images are added to the existing training data set. Materials and Methods The study included 7240 participants who underwent various types of nonenhanced CT examinations that included the heart: coronary artery calcium (CAC) scoring CT, diagnostic CT of the chest, PET attenuation correction CT, radiation therapy treatment planning CT, CAC screening CT, and low-dose CT of the chest. CAC and thoracic aorta calcification (TAC) were quantified using a convolutional neural network trained with (a) 1181 low-dose chest CT examinations (baseline), (b) a small set of examinations of the respective type supplemented to the baseline (data specific), and (c) a combination of examinations of all available types (combined). Supplemental training sets contained 199-568 CT images depending on the calcium burden of each population. The DL algorithm performance was evaluated with intraclass correlation coefficients (ICCs) between DL and manual (Agatston) CAC and (volume) TAC scoring and with linearly weighted κ values for cardiovascular risk categories (Agatston score; cardiovascular disease risk categories: 0, 1-10, 11-100, 101-400, >400). Results At baseline, the DL algorithm yielded ICCs of 0.79-0.97 for CAC and 0.66-0.98 for TAC across the range of different types of CT examinations. ICCs improved to 0.84-0.99 (CAC) and 0.92-0.99 (TAC) for CT protocol-specific training and to 0.85-0.99 (CAC) and 0.96-0.99 (TAC) for combined training. For assignment of cardiovascular disease risk category, the κ value for all test CT scans was 0.90 (95% confidence interval [CI]: 0.89, 0.91) for the baseline training. It increased to 0.92 (95% CI: 0.91, 0.93) for both data-specific and combined training. Conclusion A deep learning calcium scoring algorithm for quantification of coronary and thoracic calcium was robust, despite substantial differences in CT protocol and variations in subject population. Augmenting the algorithm training with CT protocol-specific images further improved algorithm performance. © RSNA, 2020 See also the editorial by Vannier in this issue.


Subject(s)
Coronary Artery Disease/diagnostic imaging , Deep Learning , Heart/diagnostic imaging , Thorax/diagnostic imaging , Tomography, X-Ray Computed/methods , Vascular Calcification/diagnostic imaging , Aged , Clinical Protocols , Female , Humans , Male , Middle Aged , Retrospective Studies
13.
Magn Reson Med ; 83(4): 1429-1441, 2020 04.
Article in English | MEDLINE | ID: mdl-31593328

ABSTRACT

PURPOSE: To study the influence of gradient echo-based contrasts as input channels to a 3D patch-based neural network trained for synthetic CT (sCT) generation in canine and human populations. METHODS: Magnetic resonance images and CT scans of human and canine pelvic regions were acquired and paired using nonrigid registration. Magnitude MR images and Dixon reconstructed water, fat, in-phase and opposed-phase images were obtained from a single T1 -weighted multi-echo gradient-echo acquisition. From this set, 6 input configurations were defined, each containing 1 to 4 MR images regarded as input channels. For each configuration, a UNet-derived deep learning model was trained for synthetic CT generation. Reconstructed Hounsfield unit maps were evaluated with peak SNR, mean absolute error, and mean error. Dice similarity coefficient and surface distance maps assessed the geometric fidelity of bones. Repeatability was estimated by replicating the training up to 10 times. RESULTS: Seventeen canines and 23 human subjects were included in the study. Performance and repeatability of single-channel models were dependent on the TE-related water-fat interference with variations of up to 17% in mean absolute error, and variations of up to 28% specifically in bones. Repeatability, Dice similarity coefficient, and mean absolute error were statistically significantly better in multichannel models with mean absolute error ranging from 33 to 40 Hounsfield units in humans and from 35 to 47 Hounsfield units in canines. CONCLUSION: Significant differences in performance and robustness of deep learning models for synthetic CT generation were observed depending on the input. In-phase images outperformed opposed-phase images, and Dixon reconstructed multichannel inputs outperformed single-channel inputs.


Subject(s)
Deep Learning , Image Processing, Computer-Assisted , Animals , Dogs , Humans , Magnetic Resonance Imaging , Neural Networks, Computer , Tomography, X-Ray Computed
14.
J Magn Reson Imaging ; 51(6): 1858-1867, 2020 06.
Article in English | MEDLINE | ID: mdl-31854487

ABSTRACT

BACKGROUND: Previous studies have shown discrepancies between index and synchronous breast cancer in histology and molecular phenotype. It is yet unknown whether this observation also applies to the MRI phenotype. PURPOSE: To investigate whether the appearance of breast cancer on MRI (i.e. phenotype) is different from that of additional breast cancer (i.e. synchronous cancer), and whether such a difference, if it exists, is associated with prognosis. STUDY TYPE: Retrospective. POPULATION: In all, 464 consecutive patients with early-stage ER+/HER2- breast cancer were included; 34/464 (7.3%) had 44 synchronous cancers in total (34 ipsilateral, 10 contralateral). SEQUENCE: 1.5T, contrast-enhanced T1 -weighted. ASSESSMENT: We assessed imaging phenotype using 50 quantitative features from each cancer and applied principal component analysis (PCA) to identify independent properties. The degree of phenotype difference was assessed. An association between phenotype differences and prognosis in terms of the Nottingham Prognostic Index (NPI) and PREDICT score were analyzed. STATISTICAL TESTS: PCA; Wilcoxon rank sum test; Benjamini-Hochberg to control the false discovery rate. RESULTS: PCA identified eight components in patients with ipsilateral synchronous cancer. Six out of eight were significantly different between index and synchronous cancer. These components represented features describing texture (three components, P < 0.001, P < 0.001, P = 0.004), size (P < 0.001), smoothness (P < 0.001), and kinetics (P = 0.004). Phenotype differences in terms of the six components were split in tertiles. Larger phenotype differences in size, kinetics, and texture were associated with significantly worse prognosis in terms of NPI (P = 0.019, P = 0.045, P = 0.014), but not for the PREDICT score (P = 0.109, P = 0.479, P = 0.109). PCA identified six components in patients with contralateral synchronous cancer. None were significantly different from the index cancer (P = 0.178, P = 0.178, P = 0.178, P = 0.326, P = 0.739, P = 0.423). DATA CONCLUSION: The MRI phenotype of ER+/HER2- breast cancer was different from that of ipsilateral synchronous cancer and a large phenotype difference was associated with worse prognosis. No significant difference was found for synchronous contralateral cancer. LEVEL OF EVIDENCE: 3 Technical Efficacy: Stage 4 J. Magn. Reson. Imaging 2020;51:1858-1867.


Subject(s)
Breast Neoplasms , Breast , Breast Neoplasms/diagnostic imaging , Humans , Magnetic Resonance Imaging , Phenotype , Prognosis , Retrospective Studies
15.
Neuroimage ; 199: 663-679, 2019 10 01.
Article in English | MEDLINE | ID: mdl-31195073

ABSTRACT

Diffusion weighted magnetic resonance imaging (dMRI) provides a non invasive virtual reconstruction of the brain's white matter structures through tractography. Analyzing dMRI measures along the trajectory of white matter bundles can provide a more specific investigation than considering a region of interest or tract-averaged measurements. However, performing group analyses with this along-tract strategy requires correspondence between points of tract pathways across subjects. This is usually achieved by creating a new common space where the representative streamlines from every subject are resampled to the same number of points. If the underlying anatomy of some subjects was altered due to, e.g., disease or developmental changes, such information might be lost by resampling to a fixed number of points. In this work, we propose to address the issue of possible misalignment, which might be present even after resampling, by realigning the representative streamline of each subject in this 1D space with a new method, coined diffusion profile realignment (DPR). Experiments on synthetic datasets show that DPR reduces the coefficient of variation for the mean diffusivity, fractional anisotropy and apparent fiber density when compared to the unaligned case. Using 100 in vivo datasets from the human connectome project, we simulated changes in mean diffusivity, fractional anisotropy and apparent fiber density. Independent Student's t-tests between these altered subjects and the original subjects indicate that regional changes are identified after realignment with the DPR algorithm, while preserving differences previously detected in the unaligned case. This new correction strategy contributes to revealing effects of interest which might be hidden by misalignment and has the potential to improve the specificity in longitudinal population studies beyond the traditional region of interest based analysis and along-tract analysis workflows.


Subject(s)
Algorithms , Diffusion Tensor Imaging/methods , Image Processing, Computer-Assisted/methods , White Matter/diagnostic imaging , Adult , Computer Simulation , Data Interpretation, Statistical , Humans
16.
Neuroimage ; 201: 116016, 2019 11 01.
Article in English | MEDLINE | ID: mdl-31310861

ABSTRACT

Childhood obesity is a rising problem caused in part by unhealthy food choices. Food choices are based on a neural value signal encoded in the ventromedial prefrontal cortex, and self-control involves modulation of this signal by the dorsolateral prefrontal cortex (dlPFC). We determined the effects of development, body mass (BMI Cole score) and body mass history on the neural correlates of healthy food choice in children. 141 children (aged 10-17y) from Germany, Hungary and Sweden were scanned with fMRI while performing a food choice task. Afterwards health and taste ratings of the foods were collected. In the food choice task children were asked to consider the healthiness or tastiness of the food or to choose naturally. Overall, children made healthier choices when asked to consider healthiness. However, children who had a higher weight gain per year chose less healthy foods when considering healthiness but not when choosing naturally. Pubertal development stage correlated positively while current body mass correlated negatively with dlPFC activation when accepting foods. Pubertal development negatively and current body mass positively influenced the effect of considering healthiness on activation of brain areas involved in salience and motivation. In conclusion, children in earlier stages of pubertal development and children with a higher body weight exhibited less activation in the dlPFC, which has been implicated in self-control during food choice. Furthermore, pubertal development and body mass influenced neural responses to a health cue in areas involved in salience and motivation. Thus, these findings suggest that children in earlier stages of pubertal development, children with a higher body mass gain and children with overweight may possibly be less susceptible to healthy eating interventions that rely on self-control or that highlight health aspects of food.


Subject(s)
Body Mass Index , Choice Behavior/physiology , Food Preferences/physiology , Prefrontal Cortex/physiology , Self-Control , Adolescent , Child , Diet, Healthy , Female , Humans , Magnetic Resonance Imaging , Male , Overweight
17.
Radiology ; 290(3): 833-838, 2019 03.
Article in English | MEDLINE | ID: mdl-30620257

ABSTRACT

Purpose To develop and evaluate a dual-layer detector capable of acquiring intrinsically registered real-time fluoroscopic and nuclear images in the interventional radiology suite. Materials and Methods The dual-layer detector consists of an x-ray flat panel detector placed in front of a γ camera with cone beam collimator focused at the x-ray focal spot. This design relies on the x-ray detector absorbing the majority of the x-rays while it is more transparent to the higher energy γ photons. A prototype was built and dynamic phantom images were acquired. In addition, spatial resolution and system sensitivity (evaluated as counts detected within the energy window per second per megabecquerel) were measured with the prototype. Monte Carlo simulations for an improved system with varying flat panel compositions were performed to assess potential spatial resolution and system sensitivity. Results Experiments with the dual-layer detector prototype showed that spatial resolution of the nuclear images was unaffected by the addition of the flat panel (full width at half maximum, 13.6 mm at 15 cm from the collimator surface). However, addition of the flat panel lowered system sensitivity by 45%-60% because of the nonoptimized transmission of the flat panel. Simulations showed that an attenuation of 27%-35% of the γ rays in the flat panel could be achieved by decreasing the crystal thickness and housing attenuation of the flat panel. Conclusion A dual-layer detector was capable of acquiring real-time intrinsically registered hybrid images, which could aid interventional procedures involving radionuclides. Published under a CC BY-NC-ND 4.0 license. Online supplemental material is available for this article.


Subject(s)
Fluoroscopy/instrumentation , Radiography, Interventional/instrumentation , Radionuclide Imaging/instrumentation , Equipment Design , Gamma Cameras , Humans , Monte Carlo Method , Phantoms, Imaging
18.
Magn Reson Med ; 81(3): 2038-2051, 2019 03.
Article in English | MEDLINE | ID: mdl-30346055

ABSTRACT

PURPOSE: To derive a generic approach for accurate localization and characterization of susceptibility markers in MRI, compatible with many common types of pulse sequences, sampling trajectories, and acceleration methods. THEORY AND METHODS: A susceptibility marker's dipolar phase evolution creates 3 saddles in the phase gradient of the spatial encoding, for each sampled data point in k-space. The signal originating from these saddles can be focused at the location of the marker to create positive contrast. The required phase shift can be calculated from the scan parameters and the marker properties, providing a marker detection algorithm generic for different scan types. The method was validated numerically and experimentally for a broad range of spherical susceptibility markers (0.3 < radius < 1.6 mm, 10 < |∆χ| < 3300 ppm), under various conditions. RESULTS: For all numerical and experimental phantoms, the average localization error was below one third of the voxel size, whereas the average error in magnetic strength quantification was 7%. The experiments included different pulse sequences (gradient echo, spin echo [SE], and free induction decay scans), sampling strategies (Cartesian, radial), and acceleration methods (echo planar imaging EPI, turbo SE). CONCLUSION: Spherical markers can be identified from their phase saddles, enabling clear visualization, precise localization, and accurate quantification of their magnetic strength, in a wide range of clinically relevant pulse sequences and sampling strategies.


Subject(s)
Brachytherapy/methods , Image Processing, Computer-Assisted/methods , Imaging, Three-Dimensional/methods , Magnetic Resonance Imaging/methods , Algorithms , Animals , Contrast Media , Models, Theoretical , Normal Distribution , Phantoms, Imaging , Software , Swine
19.
NMR Biomed ; 32(2): e4050, 2019 02.
Article in English | MEDLINE | ID: mdl-30575151

ABSTRACT

Brain tissue undergoes viscoelastic deformation and volumetric strain as it expands over the cardiac cycle due to blood volume changes within the underlying microvasculature. Volumetric strain measurements may therefore provide insights into small vessel function and tissue viscoelastic properties. Displacement encoding via stimulated echoes (DENSE) is an MRI technique that can quantify the submillimetre displacements associated with brain tissue motion. Despite previous studies reporting brain tissue displacements using DENSE and other MRI techniques, a complete picture of brain tissue volumetric strain over the cardiac cycle has not yet been obtained. To address this need we implemented 3D cine-DENSE at 7 T and 3 T to investigate the feasibility of measuring cardiac-induced volumetric strain as a marker for small vessel blood volume changes. Volumetric strain over the entire cardiac cycle was computed for the whole brain and for grey and white matter tissue separately in six healthy human subjects. Signal-to-noise ratio (SNR) measurements were used to determine the voxel-wise volumetric strain noise. Mean peak whole brain volumetric strain at 7 T (mean ± SD) was (4.5 ± 1.0) × 10-4 (corresponding to a volume expansion of 0.48 ± 0.1 mL), which is in agreement with literature values for cerebrospinal fluid that is displaced into the spinal canal to maintain a stable intracranial pressure. The peak volumetric strain ratio of grey to white matter was 4.4 ± 2.8, reflecting blood volume and tissue stiffness differences between these tissue types. The mean peak volumetric strains of grey and white matter tissue were found to be significantly different (p < 0.001). The mean SNR at 7 T and 3 T of the DENSE measurements was 22.0 ± 7.3 and 7.0 ± 2.8 respectively, which currently limits a voxel-wise strain analysis at both field strengths. We demonstrate that tissue specific quantification of volumetric strain is feasible with DENSE. This metric holds potential for studying blood volume pulsations in the ageing brain in healthy and diseased states.


Subject(s)
Algorithms , Brain/diagnostic imaging , Heart/diagnostic imaging , Magnetic Resonance Imaging , Female , Humans , Male , Motion , Signal-To-Noise Ratio , Young Adult
20.
Eur Radiol ; 29(5): 2350-2359, 2019 May.
Article in English | MEDLINE | ID: mdl-30421020

ABSTRACT

OBJECTIVES: To evaluate the added value of deep learning (DL) analysis of the left ventricular myocardium (LVM) in resting coronary CT angiography (CCTA) over determination of coronary degree of stenosis (DS), for identification of patients with functionally significant coronary artery stenosis. METHODS: Patients who underwent CCTA prior to an invasive fractional flow reserve (FFR) measurement were retrospectively selected. Highest DS from CCTA was used to classify patients as having non-significant (≤ 24% DS), intermediate (25-69% DS), or significant stenosis (≥ 70% DS). Patients with intermediate stenosis were referred for fully automatic DL analysis of the LVM. The DL algorithm characterized the LVM, and likely encoded information regarding shape, texture, contrast enhancement, and more. Based on these encodings, features were extracted and patients classified as having a non-significant or significant stenosis. Diagnostic performance of the combined method was evaluated and compared to DS evaluation only. Functionally significant stenosis was defined as FFR ≤ 0.8 or presence of angiographic high-grade stenosis (≥ 90% DS). RESULTS: The final study population consisted of 126 patients (77% male, 59 ± 9 years). Eighty-one patients (64%) had a functionally significant stenosis. The proposed method resulted in improved discrimination (AUC = 0.76) compared to classification based on DS only (AUC = 0.68). Sensitivity and specificity were 92.6% and 31.1% for DS only (≥ 50% indicating functionally significant stenosis), and 84.6% and 48.4% for the proposed method. CONCLUSION: The combination of DS with DL analysis of the LVM in intermediate-degree coronary stenosis may result in improved diagnostic performance for identification of patients with functionally significant coronary artery stenosis. KEY POINTS: • Assessment of degree of coronary stenosis on CCTA has consistently high sensitivity and negative predictive value, but has limited specificity for identifying the functional significance of a stenosis. • Deep learning algorithms are able to learn complex patterns and relationships directly from the images without prior specification of which image features represent presence of disease, and thereby may be more sensitive to subtle changes in the LVM caused by functionally significant stenosis. • Addition of deep learning analysis of the left ventricular myocardium to the evaluation of degree of coronary artery stenosis improves diagnostic performance and increases specificity of resting CCTA. This could potentially decrease the number of patients undergoing invasive coronary angiography.


Subject(s)
Algorithms , Computed Tomography Angiography/methods , Coronary Angiography/methods , Coronary Stenosis/diagnosis , Deep Learning , Fractional Flow Reserve, Myocardial/physiology , Heart Ventricles/diagnostic imaging , Coronary Stenosis/physiopathology , Female , Heart Ventricles/physiopathology , Humans , Male , Middle Aged , Multidetector Computed Tomography/methods , Reproducibility of Results , Retrospective Studies , Severity of Illness Index , Ventricular Function, Left/physiology
SELECTION OF CITATIONS
SEARCH DETAIL