Clinical Evaluation of a 2-Minute Ultrafast Brain MR Protocol for Evaluation of Acute Pathology in the Emergency and Inpatient Settings.

BACKGROUND AND PURPOSE: The use of MR imaging in emergency settings has been limited by availability, long scan times, and sensitivity to motion. This study assessed the diagnostic performance of an ultrafast brain MR imaging protocol for evaluation of acute intracranial pathology in the emergency department and inpatient settings. MATERIALS AND METHODS: Sixty-six adult patients who underwent brain MR imaging in the emergency department and inpatient settings were included in the study. All patients underwent both the reference and the ultrafast brain MR protocols. Both brain MR imaging protocols consisted of T1-weighted, T2/T2*-weighted, FLAIR, and DWI sequences. The ultrafast MR images were reconstructed by using a machine-learning assisted framework. All images were reviewed by 2 blinded neuroradiologists. RESULTS: The average acquisition time was 2.1 minutes for the ultrafast brain MR protocol and 10 minutes for the reference brain MR protocol. There was 98.5% agreement on the main clinical diagnosis between the 2 protocols. In head-to-head comparison, the reference protocol was preferred in terms of image noise and geometric distortion (P < .05 for both). The ultrafast ms-EPI protocol was preferred over the reference protocol in terms of reduced motion artifacts (P < .01). Overall diagnostic quality was not significantly different between the 2 protocols (P > .05). CONCLUSIONS: The ultrafast brain MR imaging protocol provides high accuracy for evaluating acute pathology while only requiring a fraction of the scan time. Although there was greater image noise and geometric distortion on the ultrafast brain MR protocol images, there was significant reduction in motion artifacts with similar overall diagnostic quality between the 2 protocols.

Accurate auto-labeling of chest X-ray images based on quantitative similarity to an explainable AI model.

The inability to accurately, efficiently label large, open-access medical imaging datasets limits the widespread implementation of artificial intelligence models in healthcare. There have been few attempts, however, to automate the annotation of such public databases; one approach, for example, focused on labor-intensive, manual labeling of subsets of these datasets to be used to train new models. In this study, we describe a method for standardized, automated labeling based on similarity to a previously validated, explainable AI (xAI) model-derived-atlas, for which the user can specify a quantitative threshold for a desired level of accuracy (the probability-of-similarity, pSim metric). We show that our xAI model, by calculating the pSim values for each clinical output label based on comparison to its training-set derived reference atlas, can automatically label the external datasets to a user-selected, high level of accuracy, equaling or exceeding that of human experts. We additionally show that, by fine-tuning the original model using the automatically labelled exams for retraining, performance can be preserved or improved, resulting in a highly accurate, more generalized model.

Evaluation of highly accelerated wave controlled aliasing in parallel imaging (Wave-CAIPI) susceptibility-weighted imaging in the non-sedated pediatric setting: a pilot study.

BACKGROUND: Susceptibility-weighted imaging (SWI) is highly sensitive for intracranial hemorrhagic and mineralized lesions but is associated with long scan times. Wave controlled aliasing in parallel imaging (Wave-CAIPI) enables greater acceleration factors and might facilitate broader application of SWI, especially in motion-prone populations. OBJECTIVE: To compare highly accelerated Wave-CAIPI SWI to standard SWI in the non-sedated pediatric outpatient setting, with respect to the following variables: estimated scan time, image noise, artifacts, visualization of normal anatomy and visualization of pathology. MATERIALS AND METHODS: Twenty-eight children (11 girls, 17 boys; mean age ± standard deviation [SD] = 128.3±62 months) underwent 3-tesla (T) brain MRI, including standard three-dimensional (3-D) SWI sequence followed by a highly accelerated Wave-CAIPI SWI sequence for each subject. We rated all studies using a predefined 5-point scale and used the Wilcoxon signed rank test to assess the difference for each variable between sequences. RESULTS: Wave-CAIPI SWI provided a 78% and 67% reduction in estimated scan time using the 32- and 20-channel coils, respectively, corresponding to estimated scan time reductions of 3.5 min and 3 min, respectively. All 28 children were imaged without anesthesia. Inter-reader agreement ranged from fair to substantial (k=0.67 for evaluation of pathology, 0.55 for anatomical contrast, 0.3 for central noise, and 0.71 for artifacts). Image noise was rated higher in the central brain with wave SWI (P<0.01), but not in the peripheral brain. There was no significant difference in the visualization of normal anatomical structures and visualization of pathology between the standard and wave SWI sequences (P=0.77 and P=0.79, respectively). CONCLUSION: Highly accelerated Wave-CAIPI SWI of the brain can provide similar image quality to standard SWI, with estimated scan time reduction of 3-3.5 min depending on the radiofrequency coil used, with fewer motion artifacts, at a cost of mild but perceptibly increased noise in the central brain.

An artificial intelligence-accelerated 2-minute multi-shot echo planar imaging protocol for comprehensive high-quality clinical brain imaging.

PURPOSE: We introduce and validate an artificial intelligence (AI)-accelerated multi-shot echo-planar imaging (msEPI)-based method that provides T1w, T2w, T2∗ , T2-FLAIR, and DWI images with high SNR, high tissue contrast, low specific absorption rates (SAR), and minimal distortion in 2 minutes. METHODS: The rapid imaging technique combines a novel machine learning (ML) scheme to limit g-factor noise amplification and improve SNR, a magnetization transfer preparation module to provide clinically desirable contrast, and high per-shot EPI undersampling factors to reduce distortion. The ML training and image reconstruction incorporates a tunable parameter for controlling the level of denoising/smoothness. The performance of the reconstruction method is evaluated across various acceleration factors, contrasts, and SNR conditions. The 2-minute protocol is directly compared to a 10-minute clinical reference protocol through deployment in a clinical setting, where five representative cases with pathology are examined. RESULTS: Optimization of custom msEPI sequences and protocols was performed to balance acquisition efficiency and image quality compared to the five-fold longer clinical reference. Training data from 16 healthy subjects across multiple contrasts and orientations were used to produce ML networks at various acceleration levels. The flexibility of the ML reconstruction was demonstrated across SNR levels, and an optimized regularization was determined through radiological review. Network generalization toward novel pathology, unobserved during training, was illustrated in five clinical case studies with clinical reference images provided for comparison. CONCLUSION: The rapid 2-minute msEPI-based protocol with tunable ML reconstruction allows for advantageous trade-offs between acquisition speed, SNR, and tissue contrast when compared to the five-fold slower standard clinical reference exam.

MarkIt: A Collaborative Artificial Intelligence Annotation Platform Leveraging Blockchain For Medical Imaging Research.

Current research on medical image processing relies heavily on the amount and quality of input data. Specifically, supervised machine learning methods require well-annotated datasets. A lack of annotation tools limits the potential to achieve high-volume processing and scaled systems with a proper reward mechanism. We developed MarkIt, a web-based tool, for collaborative annotation of medical imaging data with artificial intelligence and blockchain technologies. Our platform handles both Digital Imaging and Communications in Medicine (DICOM) and non-DICOM images, and allows users to annotate them for classification and object detection tasks in an efficient manner. MarkIt can accelerate the annotation process and keep track of user activities to calculate a fair reward. A proof-of-concept experiment was conducted with three fellowship-trained radiologists, each of whom annotated 1,000 chest X-ray studies for multi-label classification. We calculated the inter-rater agreement and estimated the value of the dataset to distribute the reward for annotators using a crypto currency. We hypothesize that MarkIt allows the typically arduous annotation task to become more efficient. In addition, MarkIt can serve as a platform to evaluate the value of data and trade the annotation results in a more scalable manner in the future. The platform is publicly available for testing on https://markit.mgh.harvard.edu.