Background: Modic changes (MCs) are the most prevalent classification system for describing magnetic resonance imaging (MRI) signal intensity changes in the vertebrae. However, there is a growing need for novel quantitative and standardized methods of characterizing these anomalies, particularly for lesions of transitional or mixed nature, due to the lack of conclusive evidence of their associations with low back pain. This retrospective imaging study aims to develop an interpretable deep learning-based detection tool for voxel-wise mapping of MCs.Methods: Seventy-five lumbar spine MRI exams that presented with acute-tochronic low back pain, radiculopathy, and other symptoms of the lumbar spine were enrolled. The pipeline consists of two deep convolutional neural networks to generate an interpretable voxel-wise Modic map. First, an autoencoder was trained to segment vertebral bodies from T 1 -weighted sagittal lumbar spine images. Next, two radiologists segmented and labeled MCs from a combined T 1 -and T 2 -weighted assessment to serve as ground truth for training a second autoencoder that performs segmentation of MCs. The voxels in the detected regions were then categorized to the appropriate Modic type using a rule-based signal intensity algorithm. Post hoc, three radiologists independently graded a second dataset with the aid of the model predictions in an artificial (AI)-assisted experiment. Results:The model successfully identified the presence of changes in 85.7% of samples in the unseen test set with a sensitivity of 0.71 (±0.072), specificity of 0.95 (±0.022), and Cohen's kappa score of 0.63. In the AI-assisted experiment, the agreement between the junior radiologist and the senior neuroradiologist significantly improved from Cohen's kappa score of 0.52 to 0.58 (p < 0.05).Conclusions: This deep learning-based approach demonstrates substantial agreement with radiologists and may serve as a tool to improve inter-rater reliability in the assessment of MCs.Kenneth T. Gao and Radhika Tibrewala contributed equally to this study.

Structured Abstract Study Design In vivo retrospective study of fully automatic quantitative imaging feature extraction from clinically acquired lumbar spine magnetic resonance imaging (MRI). Objective To demonstrate the feasibility of substituting automatic for human-demarcated segmentation of major anatomical structures in clinical lumbar spine MRI to generate quantitative image-based features and biomechanical models. Setting Previous studies have demonstrated the viability of automatic segmentation applied to medical images; however, the feasibility of these networks to segment clinically acquired images has not yet been demonstrated, as they largely rely on specialized sequences or strict quality of imaging data to achieve good performance. Methods Convolutional neural networks were trained to demarcate vertebral bodies, intervertebral disc, and paraspinous muscles from sagittal and axial T1-weighted MRIs. Intervertebral disc height, muscle cross sectional area, and subject-specific musculoskeletal models of tissue loading in the lumbar spine were then computed from these segmentations and compared against those computed from human-demarcated masks. Results Segmentation masks, as well as the morphological metrics and biomechanical models computed from those masks, were highly similar between human- and computer-generated methods. Segmentations were similar with Dice Similarity Coefficients 0.77 or greater across networks, morphological metrics and biomechanical models were similar with Pearson R correlation coefficients 0.69 or greater when significant. Conclusions This study demonstrates the feasibility of substituting computer-generated for human-generated segmentations of major anatomical structures in lumbar spine MRI to compute quantitative image-based morphological metrics and subject-specific musculoskeletal models of tissue loading quickly, efficiently, and at scale without interrupting routine clinical care.

We demonstrate that fast acquisition images reconstructed robust to segmentation with convolutional neural networks, which are known to generate chaotic errors in response to minimal image perturbations. This finding indicates that low-NEX images with DL-based noise reduction are not only robust to human readers, but machine readers as well.

Objectives To evaluate whether combining fast acquisitions with deep-learning reconstruction can provide diagnostically useful images and quantitative assessment comparable to standard-of-care acquisitions for lumbar spine MRI. Methods Eighteen patients were imaged with both standard protocol and fast protocol using reduced signal averages, each protocol including sagittal fat-suppressed T2-weighted, sagittal T1-weighted, and axial T2-weighted 2D fast spin-echo sequences. Fast-acquisition data was additionally reconstructed using vendor-supplied deep-learning reconstruction with three different noise reduction factors. For qualitative analysis, standard images as well as fast images with and without deep-learning reconstruction were graded by three radiologists on five different categories. For quantitative analysis, convolutional neural networks were applied to sagittal T1-weighted images to segment intervertebral discs and vertebral bodies, and disc heights and vertebral body volumes were derived. Results Based on noninferiority testing on qualitative scores, fast images without deep-learning reconstruction were inferior to standard images for most categories. However, deep-learning reconstruction improved the average scores, and noninferiority was observed over 24 out of 45 comparisons (all with sagittal T2-weighted images while 4/5 comparisons with sagittal T1-weighted and axial T2-weighted images). Interobserver variability increased with 50 and 75% noise reduction factors. Deep-learning reconstructed fast images with 50% and 75% noise reduction factors had comparable disc heights and vertebral body volumes to standard images (r2 ≥ 0.86 for disc heights and r2 ≥ 0.98 for vertebral body volumes). Conclusions This study demonstrated that deep-learning-reconstructed fast-acquisition images have the potential to provide noninferior image quality and comparable quantitative assessment to standard clinical images.