Quantitative body DW-MRI biomarkers uncertainty estimation using unscented wild-bootstrap

Prospective pediatric study comparing glomerular filtration rate estimates based on motion-robust dynamic contrast-enhanced magnetic resonance imaging and serum creatinine (eGFR) to 99mTc DTPA

BACKGROUND

Current methods to estimate glomerular filtration rate (GFR) have shortcomings. Estimates based on serum creatinine are known to be inaccurate in the chronically ill and during acute changes in renal function. Gold standard methods such as inulin and ^99mTc diethylenetriamine pentaacetic acid (DTPA) require blood or urine sampling and thus can be difficult to perform in children. Motion-robust radial volumetric interpolated breath-hold examination (VIBE) dynamic contrast-enhanced MRI represents a novel tool for estimating GFR that has not been validated in children.

OBJECTIVE

The purpose of our study was to determine the feasibility and accuracy of GFR measured by motion-robust radial VIBE dynamic contrast-enhanced MRI compared to estimates by serum creatinine (eGFR) and ^99mTc DTPA in children.

MATERIALS AND METHODS

We enrolled children, 0-18 years of age, who were undergoing both a contrast-enhanced MRI and nuclear medicine ^99mTc DTPA glomerular filtration rate (NM-GFR) within 2 weeks of each other. Enrolled children consented to an additional 6-min dynamic contrast-enhanced MRI scan using the motion-robust high spatiotemporal resolution prototype dynamic radial VIBE sequence (Siemens, Erlangen, Germany) at 3 tesla (T). The images were reconstructed offline with high temporal resolution (~3 s/volume) using compressed sensing image reconstruction including regularization in temporal dimension to improve image quality and reduce streaking artifacts. Images were then automatically post-processed using in-house-developed software. Post-processing steps included automatic segmentation of kidney parenchyma and aorta using convolutional neural network techniques and tracer kinetic model fitting using the Sourbron two-compartment model to calculate the MR-based GFR (MR-GFR). The NM-GFR was compared to MR-GFR and estimated GFR based on serum creatinine (eGFR) using Pearson correlation coefficient and Bland-Altman analysis.

RESULTS

Twenty-one children (7 female, 14 male) were enrolled between February 2017 and May 2018. Data from six of these children were not further analyzed because of deviations from the MRI protocol. Fifteen patients were analyzed (5 female, 10 male; average age 5.9 years); the method was technically feasible in all children. The results showed that the MR-GFR correlated with NM-GFR with a Pearson correlation coefficient (r-value) of 0.98. Bland-Altman analysis (i.e. difference of MR-GFR and NM-GFR versus mean of NM-GFR and MR-GFR) showed a mean difference of -0.32 and reproducibility coefficient of 18 with 95% confidence interval, and the coefficient of variation of 6.7% with values between -19 (-1.96 standard deviation) and 18 (+1.96 standard deviation). In contrast, serum creatinine compared with NM-GFR yielded an r-value of 0.73. Bland-Altman analysis (i.e. difference of eGFR and NM-GFR versus mean of NM-GFR and eGFR) showed a mean difference of 2.9 and reproducibility coefficient of 70 with 95% confidence interval, and the coefficient of variation of 25% with values between -67 (-1.96 standard deviation) and 73 (+1.96 standard deviation).

CONCLUSION

MR-GFR is a technically feasible and reliable method of measuring GFR when compared to the reference standard, NM-GFR by serum ^99mTc DTPA, and MR-GFR is more reliable than estimates based on serum creatinine.

Motion-Robust Spatially Constrained Parameter Estimation in Renal Diffusion-Weighted MRI by 3D Motion Tracking and Correction of Sequential Slices

In this work, we introduce a novel motion-robust spatially constrained parameter estimation (MOSCOPE) technique for kidney diffusion-weighted MRI. The proposed motion compensation technique does not require a navigator, trigger, or breath-hold but only uses the intrinsic features of the acquired data to track and compensate for motion to reconstruct precise models of the renal diffusion signal. We have developed a technique for physiological motion tracking based on robust state estimation and sequential registration of diffusion sensitized slices acquired within 200ms. This allows a sampling rate of 5Hz for state estimation in motion tracking that is sufficiently faster than both respiratory and cardiac motion rates in children and adults, which range between 0.8 to 0.2Hz, and 2.5 to 1Hz, respectively. We then apply the estimated motion parameters to data from each slice and use motion-compensated data for 1) robust intra-voxel incoherent motion (IVIM) model estimation in the kidney using a spatially constrained model fitting approach, and 2) robust weighted least squares estimation of the diffusion tensor model. Experimental results, including precision of IVIM model parameters using bootstrap-sampling and in-vivo whole kidney tractography, showed significant improvement in precision and accuracy of these models using the proposed method compared to models based on the original data and volumetric registration.

Motion-robust parameter estimation in abdominal diffusion-weighted MRI by simultaneous image registration and model estimation

Quantitative body DW-MRI can detect abdominal abnormalities as well as monitor response-to-therapy for applications including cancer and inflammatory bowel disease with increased accuracy. Parameter estimates are obtained by fitting a forward model of DW-MRI signal decay to the observed data acquired with several b-values. The DW-MRI signal decay models typically used do not account for respiratory, cardiac and peristaltic motion, however, which may deteriorate the accuracy and robustness of parameter estimates. In this work, we introduce a new model of DW-MRI signal decay that explicitly accounts for motion. Specifically, we estimated motion-compensated model parameters by simultaneously solving image registration and model estimation (SIR-ME) problems utilizing the interdependence of acquired volumes along the diffusion-weighting dimension. To accomplish this, we applied the SIR-ME model to the in-vivo DW-MRI data sets of 26 Crohn's disease (CD) patients and achieved improved precision of the estimated parameters by reducing the coefficient of variation by 8%, 24% and 8% for slow diffusion (D), fast diffusion (D*) and fast diffusion fraction (f) parameters respectively, compared to parameters estimated with independent registration in normal-appearing bowel regions. Moreover, the parameters estimated with the SIR-ME model reduced the error rate in classifying normal and abnormal bowel loops to 12% for D and 10% for f parameter with a reduction in error rate by 13% and 11% for D and f parameters, respectively, compared to the error rate in classifying parameter estimates obtained with independent registration. The experiments in DW-MRI of liver in 20 subjects also showed that the SIR-ME model improved the precision of parameter estimation by reducing the coefficient of variation to 7% for D, 23% for D*, and 8% for the f parameter. Using the SIR-ME model, the coefficient of variation was reduced by 4%, 14% and 6% for D, D* and f parameters, respectively, compared to parameters estimated with independent registration. These results demonstrate that the proposed SIR-ME model improves the accuracy and robustness of quantitative body DW-MRI in characterizing tissue microstructure.

Fetal lung apparent diffusion coefficient measurement using diffusion-weighted MRI at 3 Tesla: Correlation with gestational age

PURPOSE

To evaluate the feasibility of using diffusion-weighted magnetic resonance imaging (DW-MRI) to assess the fetal lung apparent diffusion coefficient (ADC) at 3 Tesla (T).

MATERIALS AND METHODS

Seventy-one pregnant women (32 second trimester, 39 third trimester) were scanned with a twice-refocused Echo-planar diffusion-weighted imaging sequence with 6 different b-values in 3 orthogonal diffusion orientations at 3T. After each scan, a region-of-interest (ROI) mask was drawn to select a region in the fetal lung and an automated robust maximum likelihood estimation algorithm was used to compute the ADC parameter. The amount of motion in each scan was visually rated.

RESULTS

When scans with unacceptable levels of motion were eliminated, the lung ADC values showed a strong association with gestational age (P < 0.01), increasing dramatically between 16 and 27 weeks and then achieving a plateau around 27 weeks.

CONCLUSION

We show that to get reliable estimates of ADC values of fetal lungs, a multiple b-value acquisition, where motion is either corrected or considered, can be performed. J. Magn. Reson. Imaging 2016;44:1650-1655.

Spatially-constrained probability distribution model of incoherent motion (SPIM) for abdominal diffusion-weighted MRI

Quantitative diffusion-weighted MR imaging (DW-MRI) of the body enables characterization of the tissue microenvironment by measuring variations in the mobility of water molecules. The diffusion signal decay model parameters are increasingly used to evaluate various diseases of abdominal organs such as the liver and spleen. However, previous signal decay models (i.e., mono-exponential, bi-exponential intra-voxel incoherent motion (IVIM) and stretched exponential models) only provide insight into the average of the distribution of the signal decay rather than explicitly describe the entire range of diffusion scales. In this work, we propose a probability distribution model of incoherent motion that uses a mixture of Gamma distributions to fully characterize the multi-scale nature of diffusion within a voxel. Further, we improve the robustness of the distribution parameter estimates by integrating spatial homogeneity prior into the probability distribution model of incoherent motion (SPIM) and by using the fusion bootstrap solver (FBM) to estimate the model parameters. We evaluated the improvement in quantitative DW-MRI analysis achieved with the SPIM model in terms of accuracy, precision and reproducibility of parameter estimation in both simulated data and in 68 abdominal in-vivo DW-MRIs. Our results show that the SPIM model not only substantially reduced parameter estimation errors by up to 26%; it also significantly improved the robustness of the parameter estimates (paired Student's t-test, p < 0.0001) by reducing the coefficient of variation (CV) of estimated parameters compared to those produced by previous models. In addition, the SPIM model improves the parameter estimates reproducibility for both intra- (up to 47%) and inter-session (up to 30%) estimates compared to those generated by previous models. Thus, the SPIM model has the potential to improve accuracy, precision and robustness of quantitative abdominal DW-MRI analysis for clinical applications.

Reliable estimation of incoherent motion parametric maps from diffusion-weighted MRI using fusion bootstrap moves

Diffusion-weighted MRI has the potential to provide important new insights into physiological and microstructural properties of the body. The Intra-Voxel Incoherent Motion (IVIM) model relates the observed DW-MRI signal decay to parameters that reflect blood flow in the capillaries (D*), capillaries volume fraction (f), and diffusivity (D). However, the commonly used, independent voxel-wise fitting of the IVIM model leads to imprecise parameter estimates, which has hampered their practical usage. In this work, we improve the precision of estimates by introducing a spatially-constrained Incoherent Motion (IM) model of DW-MRI signal decay. We also introduce an efficient iterative "fusion bootstrap moves" (FBM) solver that enables precise parameter estimates with this new IM model. This solver updates parameter estimates by applying a binary graph-cut solver to fuse the current estimate of parameter values with a new proposal of the parameter values into a new estimate of parameter values that better fits the observed DW-MRI data. The proposals of parameter values are sampled from the independent voxel-wise distributions of the parameter values with a model-based bootstrap resampling of the residuals. We assessed both the improvement in the precision of the incoherent motion parameter estimates and the characterization of heterogeneous tumor environments by analyzing simulated and in vivo abdominal DW-MRI data of 30 patients, and in vivo DW-MRI data of three patients with musculoskeletal lesions. We found our IM-FBM reduces the relative root mean square error of the D* parameter estimates by 80%, and of the f and D parameter estimates by 50% compared to the IVIM model with the simulated data. Similarly, we observed that our IM-FBM method significantly reduces the coefficient of variation of parameter estimates of the D* parameter by 43%, the f parameter by 37%, and the D parameter by 17% compared to the IVIM model (paired Student's t-test, p<0.0001). In addition, we found our IM-FBM method improved the characterization of heterogeneous musculoskeletal lesions by means of increased contrast-to-noise ratio of 19.3%. The IM model and FBM solver combined, provide more precise estimate of the physiological model parameter values that describing the DW-MRI signal decay and a better mechanism for characterizing heterogeneous lesions than does the independent voxel-wise IVIM model.

In vivo assessment of optimal b-value range for perfusion-insensitive apparent diffusion coefficient imaging

PURPOSE

To assess the optimal b-values range for perfusion-insensitive apparent diffusion coefficient (ADC) imaging of abdominal organs using short-duration DW-MRI acquisitions with currently available ADC estimation methods.

METHODS

DW-MRI data of 15 subjects were acquired with eight b-values in the range of 5-800 s∕mm(2). The reference-standard, a perfusion insensitive, ADC value (ADC(IVIM)), was computed using an intravoxel incoherent motion (IVIM) model with all acquired diffusion-weighted images. Simulated DW-MRI data was generated using an IVIM model with b-values in the range of 0-1200 s∕mm(2). Monoexponential ADC estimates were calculated using: (1) Two-point estimator (ADC(2)); (2) least squares three-point (ADC(3)) estimator and; (3) Rician noise model estimator (ADC(R)). The authors found the optimal b-values for perfusion-insensitive ADC calculations by minimizing the relative root mean square error (RRMS) between the ADC(IVIM) and the monoexponential ADC values for each estimation method and organ.

RESULTS

Low b-value = 300 s∕mm(2) and high b-value = 1200 s∕mm(2) minimized the RRMS between the estimated ADC and the reference-standard ADC(IVIM) to less than 5% using the ADC(3) estimator. By considering only the in vivo DW-MRI data, the combination of low b-value = 270 s∕mm(2) and high b-value of 800 s∕mm(2) minimized the RRMS between the estimated ADC and the reference-standard ADC(IVIM) to <7% using the ADC(3) estimator. For all estimators, the RRMS between the estimated ADC and the reference standard ADC correlated strongly with the perfusion-fraction parameter of the IVIM model (r = [0.78-0.83], p ≤ 0.003).

CONCLUSIONS

The perfusion compartment in DW-MRI signal decay correlates strongly with the RRMS in ADC estimates from short-duration DW-MRI. The impact of the perfusion compartment on ADC estimations depends, however, on the choice of b-values and estimation method utilized. Likewise, perfusion-related errors can be reduced to <7% by carefully selecting the b-values used for ADC calculations and method of estimation.

Characterization of fast and slow diffusion from diffusion-weighted MRI of pediatric Crohn's disease

PURPOSE

To characterize fast and slow diffusion components in diffusion-weighted magnetic resonance imaging (DW-MRI) of pediatric Crohn's disease (CD). Overall diffusivity reduction as measured by the apparent diffusion coefficient (ADC) in patients with CD has been previously demonstrated. However, the ADC reduction may be due to changes in either fast or slow diffusion components. In this study we distinguished between the fast and slow diffusion components in the DW-MRI signal decay of pediatric CD.

MATERIALS AND METHODS

We acquired MRI from 24 patients, including MR enterography (MRE) and DW-MRI with 8 b-values (0-800 s/mm(2)). We characterized fast and slow diffusivity by intravoxel incoherent motion (IVIM) model parameters (f, D*, D), and overall diffusivity by ADC values. We determined which model best described the DW-MRI signal decay. We assessed the influence of the IVIM model parameters on the ADC. We evaluated differences in model parameter values between the enhancing and nonenhancing groups.

RESULTS

The IVIM model described the observed data significantly better than the ADC model (P = 0.0088). The ADC was correlated with f (r = 0.67, P = 0.0003), but not with D (r = 0.39, P = 0.062) and D* (r = -0.39, P = 0.057). f values were significantly lower (P < 0.003) and D* values were significantly higher (P = 0.03) in the enhancing segments, while D values were not significantly different between the groups (P = 0.14).

CONCLUSION

For this study population the IVIM model provides a better description of the DW-MRI signal decay than the ADC model. The reduced ADC is related to changes in the fast diffusion rather than to changes in the slow diffusion.

Reliable assessment of perfusivity and diffusivity from diffusion imaging of the body

Diffusion-weighted MRI of the body has the potential to provide important new insights into physiological and microstructural properties. The intra-voxel incoherent motion (IVIM) model relates the observed DW-MRI signal decay to parameters that reflect perfusivity (D*) and its volume fraction (f), and diffusivity (D). However, the commonly used voxel-wise fitting of the IVIM model leads to parameter estimates with poor precision, which has hampered their practical usage. In this work, we increase the estimates' precision by introducing a model of spatial homogeneity, through which we obtain estimates of model parameters for all of the voxels at once, instead of solving for each voxel independently. Furthermore, we introduce an efficient iterative solver which utilizes a model-based bootstrap estimate of the distribution of residuals and a binary graph cut to generate optimal model parameter updates. Simulation experiments show that our approach reduces the relative root mean square error of the estimated parameters by 80% for the D* parameter and by 50% for the f and D parameters. We demonstrated the clinical impact of our model in distinguishing between enhancing and nonenhancing ileum segments in 24 Crohn's disease patients. Our model detected the enhanced segments with 91%/92% sensitivity/specificity which is better than the 81%/85% obtained by the voxel-independent approach.