Prospective image registration for automated scan prescription of follow-up knee images in quantitative studies

Machine learning-based automated scan prescription of lumbar spine MRI acquisitions

PURPOSE

High quality scan prescription that optimally covers the area of interest with scan planes aligned to relevant anatomical structures is crucial for error-free radiologic interpretation. The goal of this project was to develop a machine learning pipeline for oblique scan prescription that could be trained on localizer images and metadata from previously acquired MR exams.

METHODS

A novel Multislice Rotational Region-based Convolutional Neural Network (MS-R2CNN) architecture was developed. Based on this architecture, models for automated prescription sagittal lumbar spine acquisitions from axial, sagittal, and coronal localizer slices were trained. The automated prescription pipeline was integrated with the scanner console software and evaluated in experiments with healthy volunteers (N = 3) and patients with lower-back pain (N = 20).

RESULTS

Experiments in healthy volunteers demonstrated high accuracy of automated prescription in all subjects. There was good agreement between alignment and coverage of manual and automated prescriptions, as well as consistent views of the lumbar spine at different positions of the subjects within the scanner bore. In patients with lower-back pain, the generated prescription was applied in 18 cases (90% of the total number). None of the cases required major adjustment, while in 11 cases (55%) there were minor manual adjustments to the generated prescription.

CONCLUSIONS

This study demonstrates the ability of oriented object detection-based models to be trained to prescribe oblique lumbar spine MRI acquisitions without the need of manual annotation or feature engineering and the feasibility of using machine learning-based pipelines on the scanner for automated prescription of MRI acquisitions.

Automated MRI Field of View Prescription from Region of Interest Prediction by Intra-Stack Attention Neural Network

Manual prescription of the field of view (FOV) by MRI technologists is variable and prolongs the scanning process. Often, the FOV is too large or crops critical anatomy. We propose a deep learning framework, trained by radiologists' supervision, for automating FOV prescription. An intra-stack shared feature extraction network and an attention network are used to process a stack of 2D image inputs to generate scalars defining the location of a rectangular region of interest (ROI). The attention mechanism is used to make the model focus on a small number of informative slices in a stack. Then, the smallest FOV that makes the neural network predicted ROI free of aliasing is calculated by an algebraic operation derived from MR sampling theory. The framework's performance is examined quantitatively with intersection over union (IoU) and pixel error on position and qualitatively with a reader study. The proposed model achieves an average IoU of 0.867 and an average ROI position error of 9.06 out of 512 pixels on 80 test cases, significantly better than two baseline models and not significantly different from a radiologist. Finally, the FOV given by the proposed framework achieves an acceptance rate of 92% from an experienced radiologist.

Zonal differences in meniscus MR relaxation times in response to in vivo static loading in knee osteoarthritis

This study assessed the effects of static loading on MRI relaxation times of menisci in individuals with and without radiographic knee OA. High-resolution fast spin-echo (FSE) and T(1ρ)/T(2) relaxation time MR sequences were obtained with and without loading at 50% body weight in 124 subjects. T(1ρ)/T(2) relaxation times were calculated in menisci, and meniscus lesions were assessed through clinical grading. Student's t-test compared OA and control unloaded relaxation times as well as within-group changes with loading, Generalized Linear Models evaluated zonal variation, and ANCOVA compared loading response between groups. Unloaded T(1ρ) and T(2) in the middle and inner zones of the lateral anterior horn and outer zone of the medial posterior horn were significantly higher in OA and suggest that meniscal OA change occurs unevenly. Zonal T(1ρ) and T(2) showed differing patterns between anterior and posterior horns, suggesting differences in macromolecular organization. Significant increases with loading were seen largely in the T(2) of controls and less frequently in subjects with OA. In the medial posterior horn, T(1ρ) and T(2) decreased with loading in OA but changed negligibly in controls; these significantly different loading responses between groups may indicate load transmission failure in OA menisci.

Response of knee cartilage T1rho and T2 relaxation times to in vivo mechanical loading in individuals with and without knee osteoarthritis

OBJECTIVE

The objective of this study was to evaluate the effects of mechanical loading on knee articular cartilage T1ρ and T2 relaxation times in patients with and without osteoarthritis (OA).

DESIGN

Magnetic resonance (MR) images were acquired from 137 subjects with and without knee OA under two conditions: unloaded and loaded at 50% body weight. Three sequences were acquired: a high-resolution 3D-CUBE, a T1ρ relaxation time, and a T2 relaxation time sequences. Cartilage regions of interest included: medial and lateral femur (MF, LF); medial and lateral tibia (MT, LT), laminar analysis (superficial and deep layers), and subcompartments. Changes in relaxation times in response to loading were evaluated.

RESULTS

In response to loading, we observed significant reductions in T1ρ relaxation times in the MT and LT. In both the MF and LF, loading resulted in significant decreases in the superficial layer and significant increases in the deep layer of the cartilage for T1ρ and T2. All subcompartments of the MT and LT showed significant reduction in T1ρ relaxation times. Reductions were larger for subjects with OA (range: 13-19% change) when compared to healthy controls (range: 3-13% change).

CONCLUSIONS

Loading of the cartilage resulted in significant changes in relaxation times in the femur and tibia, with novel findings regarding laminar and subcompartmental variations. In general, changes in relaxation times with loading were larger in the OA group suggesting that the collagen-proteoglycan matrix of subjects with OA is less capable of retaining water, and may reflect a reduced ability to dissipate loads.