Accelerating Whole-Body Diffusion-weighted MRI with Deep Learning-based Denoising Image Filters.

PURPOSE: To use deep learning to improve the image quality of subsampled images (number of acquisitions = 1 [NOA1]) to reduce whole-body diffusion-weighted MRI (WBDWI) acquisition times. MATERIALS AND METHODS: Both retrospective and prospective patient groups were used to develop a deep learning-based denoising image filter (DNIF) model. For initial model training and validation, 17 patients with metastatic prostate cancer with acquired WBDWI NOA1 and NOA9 images (acquisition period, 2015-2017) were retrospectively included. An additional 22 prospective patients with advanced prostate cancer, myeloma, and advanced breast cancer were used for model testing (2019), and the radiologic quality of DNIF-processed NOA1 (NOA1-DNIF) images were compared with NOA1 images and clinical NOA16 images by using a three-point Likert scale (good, average, or poor; statistical significance was calculated by using a Wilcoxon signed ranked test). The model was also retrained and tested in 28 patients with malignant pleural mesothelioma (MPM) who underwent lung MRI (2015-2017) to demonstrate feasibility in other body regions. RESULTS: The model visually improved the quality of NOA1 images in all test patients, with the majority of NOA1-DNIF and NOA16 images being graded as either "average" or "good" across all image-quality criteria. From validation data, the mean apparent diffusion coefficient (ADC) values within NOA1-DNIF images of bone disease deviated from those within NOA9 images by an average of 1.9% (range, 1.1%-2.6%). The model was also successfully applied in the context of MPM; the mean ADCs from NOA1-DNIF images of MPM deviated from those measured by using clinical-standard images (NOA12) by 3.7% (range, 0.2%-10.6%). CONCLUSION: Clinical-standard images were generated from subsampled images by using a DNIF.Keywords: Image Postprocessing, MR-Diffusion-weighted Imaging, Neural Networks, Oncology, Whole-Body Imaging, Supervised Learning, MR-Functional Imaging, Metastases, Prostate, Lung Supplemental material is available for this article. Published under a CC BY 4.0 license.

Noise-Corrected, Exponentially Weighted, Diffusion-Weighted MRI (niceDWI) Improves Image Signal Uniformity in Whole-Body Imaging of Metastatic Prostate Cancer.

Purpose: To characterize the voxel-wise uncertainties of Apparent Diffusion Coefficient (ADC) estimation from whole-body diffusion-weighted imaging (WBDWI). This enables the calculation of a new parametric map based on estimates of ADC and ADC uncertainty to improve WBDWI imaging standardization and interpretation: NoIse-Corrected Exponentially-weighted diffusion-weighted MRI (niceDWI). Methods: Three approaches to the joint modeling of voxel-wise ADC and ADC uncertainty (σADC) are evaluated: (i) direct weighted least squares (DWLS), (ii) iterative linear-weighted least-squares (IWLS), and (iii) smoothed IWLS (SIWLS). The statistical properties of these approaches in terms of ADC/σADC accuracy and precision is compared using Monte Carlo simulations. Our proposed post-processing methodology (niceDWI) is evaluated using an ice-water phantom, by comparing the contrast-to-noise ratio (CNR) with conventional exponentially-weighted DWI. We present the clinical feasibility of niceDWI in a pilot cohort of 16 patients with metastatic prostate cancer. Results: The statistical properties of ADC and σADC conformed closely to the theoretical predictions for DWLS, IWLS, and SIWLS fitting routines (a minor bias in parameter estimation is observed with DWLS). Ice-water phantom experiments demonstrated that a range of CNR could be generated using the niceDWI approach, and could improve CNR compared to conventional methods. We successfully implemented the niceDWI technique in our patient cohort, which visually improved the in-plane bias field compared with conventional WBDWI. Conclusions: Measurement of the statistical uncertainty in ADC estimation provides a practical way to standardize WBDWI across different scanners, by providing quantitative image signals that improve its reliability. Our proposed method can overcome inter-scanner and intra-scanner WBDWI signal variations that can confound image interpretation.

Magnetic Resonance Imaging-Based Assessment of Breast Cancer-Related Lymphoedema Tissue Composition.

OBJECTIVES: The aim of this study was to propose a magnetic resonance imaging acquisition and analysis protocol that uses image segmentation to measure and depict fluid, fat, and muscle volumes in breast cancer-related lymphoedema (BCRL). This study also aims to compare affected and control (unaffected) arms of patients with diagnosed BCRL, providing an analysis of both the volume and the distribution of the different tissue components. MATERIALS AND METHODS: The entire arm was imaged with a fluid-sensitive STIR and a 2-point 3-dimensional T1W gradient-echo-based Dixon sequences, acquired in sagittal orientation and covering the same imaging volume. An automated image postprocessing procedure was developed to simultaneously (1) contour the external volume of the arm and the muscle fascia, allowing separation of the epifacial and subfascial volumes; and to (2) separate the voxels belonging to the muscle, fat, and fluid components. The total, subfascial, epifascial, muscle (subfascial), fluid (epifascial), and fat (epifascial) volumes were measured in 13 patients with unilateral BCRL. Affected versus unaffected volumes were compared using a 2-tailed paired t test; a value of P < 0.05 was considered to be significant. Pearson correlation was used to investigate the linear relationship between fat and fluid excess volumes. The distribution of fluid, fat, and epifascial excess volumes (affected minus unaffected) along the arm was also evaluated using dedicated tissue composition maps. RESULTS: Total arm, epifascial, epifascial fluid, and epifascial fat volumes were significantly different (P < 0.0005), with greater volume in the affected arms. The increase in epifascial volume (globally, 94% of the excess volume) constituted the bulk of the lymphoedematous swelling, with fat comprising the main component. The total fat excess volume summed over all patients was 2.1 times that of fluid. Furthermore, fat and fluid excess volumes were linearly correlated (Pearson r = 0.75), with the fat excess volume being greater than the fluid in 11 subjects. Differences in muscle compartment volume between affected and unaffected arms were not statistically significant, and contributed only 6% to the total excess volume. Considering the distribution of the different tissue excess volumes, fluid accumulated prevalently around the elbow, with substantial involvement of the upper arm in only 3 cases. Fat excess volume was generally greater in the upper arm; however, the relative increase in epifascial volume, which considers the total swelling relative to the original size of the arm, was in 9 cases maximal within the forearm. CONCLUSIONS: Our measurements indicate that excess of fat within the epifascial layer was the main contributor to the swelling, even when a substantial accumulation of fluid was present. The proposed approach could be used to monitor how the internal components of BCRL evolve after presentation, to stratify patients for treatment, and to objectively assess treatment response. This methodology provides quantitative metrics not currently available during the standard clinical assessment of BCRL and shows potential for implementation in clinical practice.

Quantitative Contrast-Enhanced Magnetic Resonance Lymphangiography of the Upper Limbs in Breast Cancer Related Lymphedema: An Exploratory Study.

BACKGROUND: Contrast-Enhanced Magnetic Resonance Lymphangiography (CE-MRL) presents some limitations: (i) it does not quantify lymphatic functionality; and (ii) enhancement of vascular structures may confound image interpretation. Furthermore, although CE-MRL is well described in the published literature for the lower limbs, there is a paucity of data with regards to its use in the upper limbs. In this proof-of-principle study, we propose a new protocol to perform CE-MRL in the upper limbs of patients with breast cancer-related lymphedema (BCRL) which addresses these limitations. METHODS AND RESULTS: CE-MRL was performed using a previously published (morphological) protocol and the proposed protocol (quantitative) on both the ipsilateral (abnormal) and contralateral (normal) arms of patients with BCRL. The quantitative protocol employs contrast agent (CA) intradermal injections at a lower concentration to prevent T2*-related signal decay. Both protocols provided high-resolution three-dimensional images of upper limb lymphatic vessels. CA uptake curves were utilized to distinguish between lymphatic vessels and vascular structures. The quantitative protocol minimized venous enhancement and avoided spurious delays in lymphatic enhancement due to short T2* values, enabling correct CA uptake characterization. The quantitative protocol was therefore employed to measure the lymphatic fluid velocity, which demonstrated functional differences between abnormal and normal arms. The velocity values were in agreement with previously reported lymphoscintigraphy and near infra-red lymphangiography measurements. CONCLUSIONS: This work demonstrated the feasibility of CE-MRL of the upper limbs in patients with BRCL, introducing an advanced imaging and analysis protocol suitable for anatomical and functional study of the lymphatic system.