Dataset Information

Artificial neural networks for magnetic resonance elastography stiffness estimation in inhomogeneous materials.

ABSTRACT:

Purpose

To test the hypothesis that removing the assumption of material homogeneity will improve the spatial accuracy of stiffness estimates made by Magnetic Resonance Elastography (MRE).

Methods

An artificial neural network was trained using synthetic wave data computed using a coupled harmonic oscillator model. Material properties were allowed to vary in a piecewise smooth pattern. This neural network inversion (Inhomogeneous Learned Inversion (ILI)) was compared against a previous homogeneous neural network inversion (Homogeneous Learned Inversion (HLI)) and conventional direct inversion (DI) in simulation, phantom, and in-vivo experiments.

Results

In simulation experiments, ILI was more accurate than HLI and DI in predicting the stiffness of an inclusion in noise-free, low-noise, and high-noise data. In the phantom experiment, ILI delineated inclusions ≤ 2.25 cm in diameter more clearly than HLI and DI, and provided a higher contrast-to-noise ratio for all inclusions. In a series of stiff brain tumors, ILI shows sharper stiffness transitions at the edges of tumors than the other inversions evaluated.

Conclusion

ILI is an artificial neural network based framework for MRE inversion that does not assume homogeneity in material stiffness. Preliminary results suggest that it provides more accurate stiffness estimates and better contrast in small inclusions and at large stiffness gradients than existing algorithms that assume local homogeneity. These results support the need for continued exploration of learning-based approaches to MRE inversion, particularly for applications where high resolution is required.

SUBMITTER: Scott JM

PROVIDER: S-EPMC7583310 | biostudies-literature |

REPOSITORIES: biostudies-literature

ACCESS DATA

Similar Datasets

Project description:ObjectivesTo evaluate the relationship between biopsy-assessed hepatic steatosis, magnetic resonance imaging (MRI)-assessed proton density fat fraction (PDFF), and magnetic resonance elastography (MRE)-assessed liver stiffness measurement (LSM), in patients with or at risk for nonalcoholic fatty liver disease (NAFLD).MethodsA retrospective study was performed, encompassing 256 patients who had a liver biopsy and MRI/MRE examination performed within 1 year. Clinical and laboratory data were retrieved from the electronic medical record. Hepatic steatosis and fibrosis were assessed by histopathological grading/staging. First, we analyzed the diagnostic performance of PDFF for distinguishing hepatic steatosis with the receiver operating characteristic analyses. Second, variables influencing LSM were screened with univariant analyses, then identified with multivariable linear regression. Finally, the potential relationship between PDFF and LSM was assessed with linear regression after adjustment for other influencing factors, in patients with diagnosed steatosis (PDFF ≥ 5%).ResultsThe diagnostic accuracy of PDFF in distinguishing steatosis grades (S0-3) was above 0.82. No significant difference in LSM was found between patients with S1, S2, and S3 steatosis and between all steatosis grades after patients were grouped according to fibrosis stage. No statistically significant relationship was found between the LSM and PDFF (estimate = - 0.02, p = 0.065) after adjustment for fibrosis stage and age in patients with diagnosed steatosis (PDFF ≥ 5%).ConclusionsIn patients with NAFLD, the severity of hepatic steatosis has no significant influence on the liver stiffness measurement with magnetic resonance elastography.Key points• The MRI-based proton density fat fraction provides a quantitative assessment of hepatic steatosis with high accuracy. • No significant effect of hepatic steatosis on MRE-based liver stiffness measurement was found in patients with S1, S2, and S3 steatosis and between all steatosis grades after patients were grouped according to fibrosis stage. • After adjusting for fibrosis stage and age, there was no statistically significant relationship between liver stiffness and proton density fat fraction in patients with hepatic steatosis (p = 0.065).

Project description:BackgroundLarge-scale normative studies of pancreatic stiffness and potential influences have yet to be pursued via magnetic resonance elastography (MRE).PurposeTo determine normative MRE-based pancreatic stiffness values and to examine related influential factors.Study typeProspective.SubjectsIn all, 361 volunteers (men, 199; women, 162) with a median age of 54.0 years and a median body mass index (BMI) of 22.86 kg/m2 were prospectively recruited. Those with no histories of smoking, alcohol abuse, and diabetes mellitus (DM) were grouped as healthy volunteers, designating all others as positive controls.Field strength/sequenceEach volunteer underwent 3.0T pancreatic MRI at a frequency of 40 Hz.AssessmentPancreatic stiffness values, pancreatic width and volume, waist circumference, and wave distance were measured in all subjects.Statistical testsMultiple linear regression analyses were performed to determine variables that influence MRE-determined stiffness.ResultsThe mean pancreatic stiffness in all volunteers was 1.20 ± 0.16 kPa. Stiffness levels in positive control volunteers proved significantly greater than levels in healthy volunteers (1.29 ± 0.17 kPa vs. 1.14 ± 0.13 kPa; P < 0.001). In multiple linear regression analysis, sex (P = 0.004), BMI (P < 0.001), pancreatic width (P = 0.005), smoking (P < 0.001), alcohol abuse (P < 0.001), and DM (P = 0.001) emerged as significant independent factors impacting pancreatic stiffness. Smoking, alcohol abuse, DM, and wide pancreas were associated with greater pancreatic stiffness (coefficients = 0.202, 0.183, 0.149, and 0.160, respectively), while reduced pancreatic stiffness corresponded with female sex and larger BMI (coefficient = -0.155 and -0.192, respectively).Data conclusionMRE-based pancreatic stiffness values are impacted by sex, BMI, pancreatic width, smoking, alcohol abuse, and DM. Reference values are essential for future clinical studies.Level of evidence1 Technical Efficacy: Stage 2 J. Magn. Reson. Imaging 2020;52:448-458.