Democratizing Artificial Intelligence in Anatomic Pathology.

CONTEXT.­: Artificial intelligence is a transforming technology for anatomic pathology. Involvement within the workforce will foster support for algorithm development and implementation. OBJECTIVE.­: To develop a supportive ecosystem that enables pathologists with variable expertise in artificial intelligence to create algorithms in a development environment with seamless transition to a production environment. RESULTS.­: The development team considered internal development and vended solutions. Because of the extended timeline and resource requirements for internal development, a decision was made to use a vended solution. Vendor proposals were solicited and reviewed by pathologists, IT, and security groups. A vendor was selected and pipelines for development and production were established. Proposals for development were solicited from the pathology department. Eighty-four investigators were selected for the initial cohort, receiving training and access to dedicated subject matter experts. A total of 30 of 31 projects progressed through the model development process of annotating, training, and validation. Based on these projects, 15 abstracts were submitted to national meetings. CONCLUSIONS.­: Democratizing artificial intelligence by creating an ecosystem to support pathologists with varying levels of expertise can break down entry barriers, reduce overall cost of algorithm development, improve algorithm quality, and enhance the speed of adoption.

Core Needle Biopsy Guidance Based on Tissue Morphology Assessment with AI-OCT Imaging.

This paper presents a combined optical imaging/artificial intelligence (OI/AI) technique for the real-time analysis of tissue morphology at the tip of the biopsy needle, prior to collecting a biopsy specimen. This is an important clinical problem as up to 40% of collected biopsy cores provide low diagnostic value due to high adipose or necrotic content. Micron-scale-resolution optical coherence tomography (OCT) images can be collected with a minimally invasive needle probe and automatically analyzed using a computer neural network (CNN)-based AI software. The results can be conveyed to the clinician in real time and used to select the biopsy location more adequately. This technology was evaluated on a rabbit model of cancer. OCT images were collected with a hand-held custom-made OCT probe. Annotated OCT images were used as ground truth for AI algorithm training. The overall performance of the AI model was very close to that of the humans performing the same classification tasks. Specifically, tissue segmentation was excellent (~99% accuracy) and provided segmentation that closely mimicked the ground truth provided by the human annotations, while over 84% correlation accuracy was obtained for tumor and non-tumor classification.

Differential Trajectory of Diffusion and Perfusion Magnetic Resonance Imaging of Rat Spinal Cord Injury.

Traumatic spinal cord injury causes rapid neuronal and vascular injury, and predictive biomarkers are needed to facilitate acute patient management. This study examined the progression of magnetic resonance imaging (MRI) biomarkers after spinal cord injury and their ability to predict long-term neurological outcomes in a rodent model, with an emphasis on diffusion-weighted imaging (DWI) markers of axonal injury and perfusion-weighted imaging of spinal cord blood flow (SCBF). Adult Sprague-Dawley rats received a cervical contusion injury of varying severity (injured = 30, sham = 9). MRI at 4 h, 48-h, and 12-weeks post-injury included T1, T2, perfusion, and DWI. Locomotor outcome was assessed up to 12 weeks post-injury. At 4 h, the deficit in SCBF was larger than the DWI lesion, and although SCBF partially recovered by 48 h, the DWI lesion expanded. At 4 h, the volume of the SCBF deficit (R2 = 0.56, padj < 0.01) was significantly correlated with 12-week locomotor outcome, whereas DWI (R2 = 0.30, padj < 0.01) was less predictive of outcome. At 48 h, SCBF (R2 = 0.41, padj < 0.01) became less associated with outcome, and DWI (R2 = 0.38, padj < 0.01) lesion volume became more closely related to outcome. Spinal cord perfusion has unique spatiotemporal dynamics compared with diffusion measures of axonal damage and highlights the importance of acute perfusion abnormalities. Perfusion and diffusion offer complementary and clinically relevant insight into physiological and structural abnormalities following spinal cord injury beyond those afforded by T1 or T2 contrasts.

Acute Magnetic Resonance Imaging Predictors of Chronic Motor Function and Tissue Sparing in Rat Cervical Spinal Cord Injury.

Predicting functional outcomes from spinal cord injury (SCI) at the acute setting is important for patient management. This work investigated the relationship of early magnetic resonance imaging (MRI) biomarkers in a rat model of cervical contusion SCI with long-term functional outcome and tissue sparing. Forty rats with contusion injury at C5 at either the spinal cord midline (bilateral) or over the lateral cord (unilateral) were examined using in vivo multi-modal quantitative MRI at 1 day post-injury. The extent of T2-weighted hyperintensity reflecting edema was greater in the bilateral model compared with the unilateral injury. Diffusion tensor imaging (DTI) exhibited microscopic damage in similar regions of the cord as reductions in fractional anisotropy (FA) and mean diffusivity (MD), but DTI parameter maps were also confounded by the presence of vasogenic edema that locally increased FA and MD. In comparison, filtered diffusion-weighted imaging (fDWI) more clearly delineated the location of acute axonal damage without effects of vasogenic edema. Pairwise correlation analysis revealed that 28-day motor functional outcomes were most strongly associated with the extent of edema (R = -0.69). Principal component analysis identified close associations of motor functional score with tissue sparing, the extent of edema, lesion area, and injury type (unilateral or bilateral). Among the diffusion MRI parameters, lesion areas measured with fDWI had the strongest association with functional outcome (R = -0.41). Voxelwise correlation analysis identified a locus of white matter damage associated with function in the dorsal white matter, although this was likely driven by variance across the two injury patterns (unilateral and bilateral injury). Nonetheless, correlation with motor function within the damaged region found in the voxelwise analysis outperformed morphological lesion area measurement as a predictor of chronic function. Collectively, this study characterized anatomical and diffusion MRI signatures of acute SCI at cervical spine and their association with chronic functional outcomes and histological results.

Diffusion-prepared fast spin echo for artifact-free spinal cord imaging.

PURPOSE: Diffusion MRI provides unique contrast important for the detection and examination of pathophysiology after acute neurologic insults, including spinal cord injury. Diffusion weighted imaging of the rodent spinal cord has typically been evaluated with axial EPI readout. However, Diffusion weighted imaging is prone to motion artifacts, whereas EPI is prone to susceptibility artifacts. In the context of acute spinal cord injury, diffusion filtering has previously been shown to improve detection of injury by minimizing the confounding effects of edema. We propose a diffusion-preparation module combined with a rapid acquisition with relaxation enhancement readout to minimize artifacts for sagittal imaging. METHODS: Sprague-Dawley rats with cervical contusion spinal cord injury were scanned at 9.4 Tesla. The sequence optimization included the evaluation of motion-compensated encoding diffusion gradients, gating strategy, and different spinal cord-specific diffusion-weighting schemes. RESULTS: A diffusion-prepared rapid acquisition with relaxation enhancement achieved high-quality images free from susceptibility artifacts with both second-order motion-compensated encoding and gating necessary for reduction of motion artifacts. Axial diffusivity obtained from the filtered diffusion-encoding scheme had greater lesion-to-healthy tissue contrast (52%) compared to the similar metric from DTI (25%). CONCLUSION: This work demonstrated the feasibility of high-quality diffusion sagittal imaging in the rodent cervical cord with diffusion-prepared relaxation enhancement. The sequence and results are expected to improve injury detection and evaluation in acute spinal cord injury.

Filter-probe diffusion imaging improves spinal cord injury outcome prediction.

OBJECTIVE: Diffusion-weighted imaging (DWI) is a powerful tool for investigating spinal cord injury (SCI), but has limited specificity for axonal damage, which is the most predictive feature of long-term functional outcome. In this study, a technique designed to detect acute axonal injury, filter-probe double diffusion encoding (FP-DDE), is compared with standard DWI for predicting long-term functional and cellular outcomes. METHODS: This study extends FP-DDE to predict long-term functional and histological outcomes in a rat SCI model of varying severities (n = 58). Using a 9.4T magnetic resonance imaging (MRI) system, a whole-cord FP-DDE spectroscopic voxel was acquired in 3 minutes at the lesion site and compared to DWI at 48 hours postinjury. Relationships with chronic (30-day) locomotor and histological outcomes were evaluated with linear regression. RESULTS: The FP-DDE measure of parallel diffusivity (ADC|| ) was significantly related to chronic hind limb locomotor functional outcome (R2 = 0.63, p < 0.0001), and combining this measurement with acute functional scores demonstrated prognostic benefit versus functional testing alone (p = 0.0007). Acute ADC|| measurements were also more closely related to the number of injured axons measured 30 days after the injury than standard DWI. Furthermore, acute FP-DDE images showed a clear and easily interpretable pattern of injury that closely corresponded with chronic MRI and histology observations. INTERPRETATION: Collectively, these results demonstrate FP-DDE benefits from greater specificity for acute axonal damage in predicting functional and histological outcomes with rapid acquisition and fully automated analysis, improving over standard DWI. FP-DDE is a promising technique compatible with clinical settings, with potential research and clinical applications for evaluation of spinal cord pathology. Ann Neurol 2018;83:37-50.