Diagnostic performance of deep learning-based reconstruction algorithm in 3D MR neurography.

OBJECTIVE: The study aims to evaluate the diagnostic performance of deep learning-based reconstruction method (DLRecon) in 3D MR neurography for assessment of the brachial and lumbosacral plexus. MATERIALS AND METHODS: Thirty-five exams (18 brachial and 17 lumbosacral plexus) of 34 patients undergoing routine clinical MR neurography at 1.5 T were retrospectively included (mean age: 49 ± 12 years, 15 female). Coronal 3D T2-weighted short tau inversion recovery fast spin echo with variable flip angle sequences covering plexial nerves on both sides were obtained as part of the standard protocol. In addition to standard-of-care (SOC) reconstruction, k-space was reconstructed with a 3D DLRecon algorithm. Two blinded readers evaluated images for image quality and diagnostic confidence in assessing nerves, muscles, and pathology using a 4-point scale. Additionally, signal-to-noise ratio (SNR) and contrast-to-noise ratios (CNR) between nerve, muscle, and fat were measured. For comparison of visual scoring result non-parametric paired sample Wilcoxon signed-rank testing and for quantitative analysis paired sample Student's t-testing was performed. RESULTS: DLRecon scored significantly higher than SOC in all categories of image quality (p < 0.05) and diagnostic confidence (p < 0.05), including conspicuity of nerve branches and pathology. With regard to artifacts there was no significant difference between the reconstruction methods. Quantitatively, DLRecon achieved significantly higher CNR and SNR than SOC (p < 0.05). CONCLUSION: DLRecon enhanced overall image quality, leading to improved conspicuity of nerve branches and pathology, and allowing for increased diagnostic confidence in evaluation of the brachial and lumbosacral plexus.

Improving workflow in prostate MRI: AI-based decision-making on biparametric or multiparametric MRI.

OBJECTIVES: To develop and validate an artificial intelligence algorithm to decide on the necessity of dynamic contrast-enhanced sequences (DCE) in prostate MRI. METHODS: This study was approved by the institutional review board and requirement for study-specific informed consent was waived. A convolutional neural network (CNN) was developed on 300 prostate MRI examinations. Consensus of two expert readers on the necessity of DCE acted as reference standard. The CNN was validated in a separate cohort of 100 prostate MRI examinations from the same vendor and 31 examinations from a different vendor. Sensitivity/specificity were calculated using ROC curve analysis and results were compared to decisions made by a radiology technician. RESULTS: The CNN reached a sensitivity of 94.4% and specificity of 68.8% (AUC: 0.88) for the necessity of DCE, correctly assigning 44%/34% of patients to a biparametric/multiparametric protocol. In 2% of all patients, the CNN incorrectly decided on omitting DCE. With a technician reaching a sensitivity of 63.9% and specificity of 89.1%, the use of the CNN would allow for an increase in sensitivity of 30.5%. The CNN achieved an AUC of 0.73 in a set of examinations from a different vendor. CONCLUSIONS: The CNN would have correctly assigned 78% of patients to a biparametric or multiparametric protocol, with only 2% of all patients requiring re-examination to add DCE sequences. Integrating this CNN in clinical routine could render the requirement for on-table monitoring obsolete by performing contrast-enhanced MRI only when needed.