Accelerated 3D whole-brain T1, T2, and proton density mapping: feasibility for clinical glioma MR imaging

Abstract

Purpose: Advanced MRI-based biomarkers offer comprehensive and quantitative information for the evaluation and characterization of brain tumors. In this study, we report initial clinical experience in routine glioma imaging with a novel, fully 3D multiparametric quantitative transient-state imaging (QTI) method for tissue characterization based on T1 and T2 values.

Methods: To demonstrate the viability of the proposed 3D QTI technique, nine glioma patients (grade II-IV), with a variety of disease states and treatment histories, were included in this study. First, we investigated the feasibility of 3D QTI (6:25 min scan time) for its use in clinical routine imaging, focusing on image reconstruction, parameter estimation, and contrast-weighted image synthesis. Second, for an initial assessment of 3D QTI-based quantitative MR biomarkers, we performed a ROI-based analysis to characterize T1 and T2 components in tumor and peritumoral tissue.

Results: The 3D acquisition combined with a compressed sensing reconstruction and neural network-based parameter inference produced parametric maps with high isotropic resolution (1.125 × 1.125 × 1.125 mm³ voxel size) and whole-brain coverage (22.5 × 22.5 × 22.5 cm³ FOV), enabling the synthesis of clinically relevant T1-weighted, T2-weighted, and FLAIR contrasts without any extra scan time. Our study revealed increased T1 and T2 values in tumor and peritumoral regions compared to contralateral white matter, good agreement with healthy volunteer data, and high inter-subject consistency.

Conclusion: 3D QTI demonstrated comprehensive tissue assessment of tumor substructures captured in T1 and T2 parameters. Aiming for fast acquisition of quantitative MR biomarkers, 3D QTI has potential to improve disease characterization in brain tumor patients under tight clinical time-constraints.

Keywords: Glioma imaging; Image-based biomarkers; MRI; Multiparametric imaging; Neural networks.

Fig. 1

3D QTI data processing. a Reconstruction and processing. After acquisition (➀), raw k-space data is processed via naïve zero-filling (dotted line), k-space weighted view-sharing (dashed line), or a compressed sensing LRTV technique (solid line). All methods in ➁ are followed by dimensionality reduction via SVD subspace projection in the time domain (➂), gridding onto a Cartesian grid followed by a 3D IFFT (➃), and coil sensitivity estimation and combination (➄). The reconstructed image series are then fed into a neural network or are matched to a precomputed dictionary to output parametric maps of T1, T2 and PD (➅). We then synthesize clinical image contrasts using the parametric maps (➆). b Neural network architecture for parameter inference. The model receives the complex, voxel-wise signal in SVD subspace x and infers the underlying tissue parameter vector θ with T1, T2, and a PD-related scaling factor. The input signal x is phase-aligned (green lines) to transfer the complex into real-valued signal, followed by a normalization layer (purple lines). The model then divides into separate pathways, each with three ReLU-activated hidden layers and 200, 100, and one node, to eventually yield the concatenated parametric output vector θ

Fig. 2

Application to quantitative characterization of tumor substructures. For quantitative analysis of tumor substructures, intra-tumoral structures, i.e., peritumoral edema, necrotic/non-enhancing tumor core, and enhancing tumor, were annotated by a trained expert based on the clinical contrast-weighted MR data. Voxel-wise T1-T2 distributions were then derived for the individual ROIs. Using a Gaussian mixture model, we explored whether we can identify the two voxel classes that are apparent in the T1-T2 space in necrotic/non-enhancing tumor areas, which were then mapped back to the image space

Fig. 3

Contrast-weighted image synthesis for a representative patient case. From the T1, T2, and PD maps, we produce clinically relevant, fully 3D qualitative image information with high isotropic resolution and without additional scan time. As seen from the axial views and the histogram-based comparison considering the whole image volumes, synthetic T1-weighted, T2-weighted, and FLAIR MRI contrasts correspond to the clinical reference acquisitions. Corresponding sagittal and coronal views are shown in Fig. 11

Fig. 4

Sensitivity to rigid head motion. Profuse head motion can affect image acquisition in the transient-state, which leads to image degradation in the parametric maps (a) and the synthetic image contrasts (b) compared to the clinical contrast-weighted acquisitions. The post-contrast T1-weighted MRI indicates that state-of-the-art conventional MRI cannot fully recoup the pronounced head motion in this case. Corresponding sagittal and coronal views are shown in Fig. 12

Fig. 5

Sensitivity to physiological motion. Pulsating blood flow and thereby induced pulsation of the cerebrospinal fluid (CSF) can impact the T2 estimation (a) and subsequent synthesis of T2-weighted image contrasts (b) as observed in large vessels and in regions with high CSF pulsation, e.g., along the brainstem (white arrows)

Fig. 6

Qualitative comparison of tumor patient cases. Expert ROIs (green: peritumoral edema, red: necrotic core/non-enhancing tumor, yellow: enhancing tumor) are shown together with clinical T1-weighted FSPGR, T2-weighted, FLAIR, Gd-enhanced T1-weighted FSPGR images and quantitative T1 and T2 maps

Fig. 7

Qualitative T1-T2-analysis based on manual ROI annotations together with additional explorative parameter-driven tumor subclassification for two representative patient cases. Classification of necrotic/non-enhancing tissue voxels based on quantitative T1 and T2 values can give more insights into the heterogenous structure, which we attribute to fluidic (necrotic/non-enhancing tumor core II, magenta) and solid (necrotic/non-enhancing tumor core I, red) components, within the gross tumor regions (right). Expected spatial correlations of the two subcomponents are maintained as the back-projection of the T1-T2-based classification of necrotic and solid tissue results in connected annotations (left)

Fig. 8

Quantitative ROI-based parameter analysis. Boxplots of the patient-wise T1 and T2 parameter spaces (a) and a scatter plot of the respective mean T1 and T2 values (b) indicate increased T1 and T2 values in diseased, tumorous regions compared to healthy, contralateral WM and GM regions with high inter-subject consistency and small variance. Outliers in the boxplots are omitted for clarity

Fig. 9

Parametric maps of T1, T2, and PD obtained with zero-filling, k-space weighted view-sharing, and LRTV reconstructions and subsequent dictionary matching for a representative patient case (a) and a healthy volunteer (b). Parameter estimation is consistent across all reconstruction approaches. View-sharing reconstruction provides better image quality in the quantitative maps than zero-filling with best reduction of undersampling artifacts obtained with LRTV reconstruction

Fig. 10

Neural network-based parameter estimation for a representative patient case (top) and a healthy volunteer (bottom). T1, T2, and PD maps obtained from voxel-wise neural network inference are consistent with dictionary matching results. The parameter maps are computed from SVD-compressed image-series after LRTV reconstruction

Fig. 11

Contrast-weighted image synthesis for a representative patient case. Sagittal and coronal views in correspondence to the axial views in Fig. 3 are shown

Fig. 12

Sensitivity to rigid head motion. Sagittal and coronal views in correspondence to the axial views in Fig. 4 are shown