Deep Learning-Generated Synthetic MR Imaging STIR Spine Images Are Superior in Image Quality and Diagnostically Equivalent to Conventional STIR: A Multicenter, Multireader Trial.

BACKGROUND AND PURPOSE: Deep learning image reconstruction allows faster MR imaging acquisitions while matching or exceeding the standard of care and can create synthetic images from existing data sets. This multicenter, multireader spine study evaluated the performance of synthetically created STIR compared with acquired STIR. MATERIALS AND METHODS: From a multicenter, multiscanner data base of 328 clinical cases, a nonreader neuroradiologist randomly selected 110 spine MR imaging studies in 93 patients (sagittal T1, T2, and STIR) and classified them into 5 categories of disease and healthy. A DICOM-based deep learning application generated a synthetically created STIR series from the sagittal T1 and T2 images. Five radiologists (3 neuroradiologists, 1 musculoskeletal radiologist, and 1 general radiologist) rated the STIR quality and classified disease pathology (study 1, n = 80). They then assessed the presence or absence of findings typically evaluated with STIR in patients with trauma (study 2, n = 30). The readers evaluated studies with either acquired STIR or synthetically created STIR in a blinded and randomized fashion with a 1-month washout period. The interchangeability of acquired STIR and synthetically created STIR was assessed using a noninferiority threshold of 10%. RESULTS: For classification, there was a decrease in interreader agreement expected by randomly introducing synthetically created STIR of 3.23%. For trauma, there was an overall increase in interreader agreement by +1.9%. The lower bound of confidence for both exceeded the noninferiority threshold, indicating interchangeability of synthetically created STIR with acquired STIR. Both the Wilcoxon signed-rank and t tests showed higher image-quality scores for synthetically created STIR over acquired STIR (P < .0001). CONCLUSIONS: Synthetically created STIR spine MR images were diagnostically interchangeable with acquired STIR, while providing significantly higher image quality, suggesting routine clinical practice potential.

Deep Learning Image Processing Enables 40% Faster Spinal MR Scans Which Match or Exceed Quality of Standard of Care : A Prospective Multicenter Multireader Study.

OBJECTIVE: This prospective multicenter multireader study evaluated the performance of 40% scan-time reduced spinal magnetic resonance imaging (MRI) reconstructed with deep learning (DL). METHODS: A total of 61 patients underwent standard of care (SOC) and accelerated (FAST) spine MRI. DL was used to enhance the accelerated set (FAST-DL). Three neuroradiologists were presented with paired side-by-side datasets (666 series). Datasets were blinded and randomized in sequence and left-right display order. Image features were preference rated. Structural similarity index (SSIM) and per pixel L1 was assessed for the image sets pre and post DL-enhancement as a quantitative assessment of image integrity impact. RESULTS: FAST-DL was qualitatively better than SOC for perceived signal-to-noise ratio (SNR) and artifacts and equivalent for other features. Quantitative SSIM was high, supporting the absence of image corruption by DL processing. CONCLUSION: DL enables 40% spine MRI scan time reduction while maintaining diagnostic integrity and image quality with perceived benefits in SNR and artifact reduction, suggesting potential for clinical practice utility.

Deep Learning Enables 60% Accelerated Volumetric Brain MRI While Preserving Quantitative Performance: A Prospective, Multicenter, Multireader Trial.

BACKGROUND AND PURPOSE: In this prospective, multicenter, multireader study, we evaluated the impact on both image quality and quantitative image-analysis consistency of 60% accelerated volumetric MR imaging sequences processed with a commercially available, vendor-agnostic, DICOM-based, deep learning tool (SubtleMR) compared with that of standard of care. MATERIALS AND METHODS: Forty subjects underwent brain MR imaging examinations on 6 scanners from 5 institutions. Standard of care and accelerated datasets were acquired for each subject, and the accelerated scans were enhanced with deep learning processing. Standard of care, accelerated scans, and accelerated-deep learning were subjected to NeuroQuant quantitative analysis and classified by a neuroradiologist into clinical disease categories. Concordance of standard of care and accelerated-deep learning biomarker measurements were assessed. Randomized, side-by-side, multiplanar datasets (360 series) were presented blinded to 2 neuroradiologists and rated for apparent SNR, image sharpness, artifacts, anatomic/lesion conspicuity, image contrast, and gray-white differentiation to evaluate image quality. RESULTS: Accelerated-deep learning was statistically superior to standard of care for perceived quality across imaging features despite a 60% sequence scan-time reduction. Both accelerated-deep learning and standard of care were superior to accelerated scans for all features. There was no difference in quantitative volumetric biomarkers or clinical classification for standard of care and accelerated-deep learning datasets. CONCLUSIONS: Deep learning reconstruction allows 60% sequence scan-time reduction while maintaining high volumetric quantification accuracy, consistent clinical classification, and what radiologists perceive as superior image quality compared with standard of care. This trial supports the reliability, efficiency, and utility of deep learning-based enhancement for quantitative imaging. Shorter scan times may heighten the use of volumetric quantitative MR imaging in routine clinical settings.

Synthetic MRI for Clinical Neuroimaging: Results of the Magnetic Resonance Image Compilation (MAGiC) Prospective, Multicenter, Multireader Trial.

BACKGROUND AND PURPOSE: Synthetic MR imaging enables reconstruction of various image contrasts from 1 scan, reducing scan times and potentially providing novel information. This study is the first large, prospective comparison of synthetic-versus-conventional MR imaging for routine neuroimaging. MATERIALS AND METHODS: A prospective multireader, multicase noninferiority trial of 1526 images read by 7 blinded neuroradiologists was performed with prospectively acquired synthetic and conventional brain MR imaging case-control pairs from 109 subjects (mean, 53.0 ± 18.5 years of age; range, 19-89 years of age) with neuroimaging indications. Each case included conventional T1- and T2-weighted, T1 and T2 FLAIR, and STIR and/or proton density and synthetic reconstructions from multiple-dynamic multiple-echo imaging. Images were randomized and independently assessed for diagnostic quality, morphologic legibility, radiologic findings indicative of diagnosis, and artifacts. RESULTS: Clinical MR imaging studies revealed 46 healthy and 63 pathologic cases. Overall diagnostic quality of synthetic MR images was noninferior to conventional imaging on a 5-level Likert scale (P < .001; mean synthetic-conventional, -0.335 ± 0.352; Δ = 0.5; lower limit of the 95% CI, -0.402). Legibility of synthetic and conventional morphology agreed in >95%, except in the posterior limb of the internal capsule for T1, T1 FLAIR, and proton-density views (all, >80%). Synthetic T2 FLAIR had more pronounced artifacts, including +24.1% of cases with flow artifacts and +17.6% cases with white noise artifacts. CONCLUSIONS: Overall synthetic MR imaging quality was similar to that of conventional proton-density, STIR, and T1- and T2-weighted contrast views across neurologic conditions. While artifacts were more common in synthetic T2 FLAIR, these were readily recognizable and did not mimic pathology but could necessitate additional conventional T2 FLAIR to confirm the diagnosis.

Quiet PROPELLER MRI techniques match the quality of conventional PROPELLER brain imaging techniques.

BACKGROUND AND PURPOSE: Switching of magnetic field gradients is the primary source of acoustic noise in MR imaging. Sound pressure levels can run as high as 120 dB, capable of producing physical discomfort and at least temporary hearing loss, mandating hearing protection. New technology has made quieter techniques feasible, which range from as low as 80 dB to nearly silent. The purpose of this study was to evaluate the image quality of new commercially available quiet T2 and quiet FLAIR fast spin-echo PROPELLER acquisitions in comparison with equivalent conventional PROPELLER techniques in current day-to-day practice in imaging of the brain. MATERIALS AND METHODS: Thirty-four consecutive patients were prospectively scanned with quiet T2 and quiet T2 FLAIR PROPELLER, in addition to spatial resolution-matched conventional T2 and T2 FLAIR PROPELLER imaging sequences on a clinical 1.5T MR imaging scanner. Measurement of sound pressure levels and qualitative evaluation of relative image quality was performed. RESULTS: Quiet T2 and quiet T2 FLAIR were comparable in image quality with conventional acquisitions, with sound levels of approximately 75 dB, a reduction in average sound pressure levels of up to 28.5 dB, with no significant trade-offs aside from longer scan times. CONCLUSIONS: Quiet FSE provides equivalent image quality at comfortable sound pressure levels at the cost of slightly longer scan times. The significant reduction in potentially injurious noise is particularly important in vulnerable populations such as children, the elderly, and the debilitated. Quiet techniques should be considered in these special situations for routine use in clinical practice.

Radiation dose reduction in CT-guided spine biopsies does not reduce diagnostic yield.

BACKGROUND AND PURPOSE: CT-guided biopsy is the most commonly used method to obtain tissue for diagnosis in suspected cases of malignancy involving the spine. The purpose of this study was to demonstrate that a low-dose CT-guided spine biopsy protocol is as effective in tissue sampling as a regular-dose protocol, without adversely affecting procedural time or complication rates. MATERIALS AND METHODS: We retrospectively reviewed all patients who underwent CT-guided spine procedures at our institution between May 2010 and October 2013. Biopsy duration, total number of scans, total volume CT dose index, total dose-length product, and diagnostic tissue yield of low-dose and regular-dose groups were compared. RESULTS: Sixty-four patients were included, of whom 31 underwent low-dose and 33 regular-dose spine biopsies. There was a statistically significant difference in total volume CT dose index and total dose-length product between the low-dose and regular-dose groups (P < .0001). There was no significant difference in the total number of scans obtained (P = .3385), duration of procedure (P = .149), or diagnostic tissue yield (P = .6017). CONCLUSIONS: Use of a low-dose CT-guided spine biopsy protocol is a practical alternative to regular-dose approaches, maintaining overall quality and efficiency at reduced ionizing radiation dose.

Whole-brain adaptive 70-kVp perfusion imaging with variable and extended sampling improves quality and consistency while reducing dose.

BACKGROUND AND PURPOSE: Despite common use of CTP to assess cerebral hemodynamics in the setting of ischemia, concerns over radiation exposure remain. Our aim was to evaluate the efficacy of an adaptive 70-kVp (peak) whole-brain CTP protocol with variable sampling intervals and extended duration against an established fixed-sampling, limited-period protocol at 80 kVp. MATERIALS AND METHODS: A retrospective analysis of 37 patients with stroke scanned with conventional (n = 17) and variant-protocol (n = 20) whole-brain CTP was performed. We compared radiation dose, parametric map quality, and consistency of full-contrast circulation capture between a modified 70-kVp protocol, with 20 whole-brain passes at variable sampling intervals over an extended sampling period, and a conventional 80-kVp CTP examination with 24 passes at fixed-sampling intervals and a more limited scanning window. Mann-Whitney U test analysis was used to compare both protocols. RESULTS: The 70-kVp CTP scan provided superior image quality at a 45% lower CT dose index volume and 13% lower dose-length product/effective dose compared with the conventional 80-kVp scan. With respect to the consistency of contrast-passage capture, 95% of the adaptive, extended protocol continued through the venous return to baseline, compared with only 47% by using the conventional limited-length protocol. Rapid sampling during the critical arterial arrival and washout period was accomplished in nearly 95% with both the variable and fixed-sampling-interval protocols. CONCLUSIONS: Seventy-kilovolt (peak) CTP with variable and extended sampling produces improved image quality at lower radiation doses with greater consistency of full contrast passage capture.

Diffusion-weighted MRI "claw sign" improves differentiation of infectious from degenerative modic type 1 signal changes of the spine.

BACKGROUND AND PURPOSE: Modic type 1 degenerative signal changes can mimic/suggest infection, leading to additional costly and sometimes invasive investigations. This retrospective study analyzes the utility and accuracy of a novel, diffusion-weighted "claw sign" for distinguishing symptomatic type 1 degeneration from vertebral diskitis/osteomyelitis. MATERIALS AND METHODS: Seventy-three patients with imaging features resembling type 1 degeneration were classified clinically into 3 groups: true degenerative type 1 changes (n = 33), confirmed diskitis/osteomyelitis (n = 20), and radiologically suspected infection later disproved clinically (n = 20). A claw sign was defined on DWI as well-marginated, linear, regions of high signal situated within the adjacent vertebral bodies at the interface of normal with abnormal marrow. Two blinded neuroradiologists independently rated the presence of the claw sign, along with T2 disk signal and disk and endplate enhancement to determine the utility of each for identifying degeneration versus infection. RESULTS: When the 2 neuroradiologists identified a definite claw, 38 of 39 patients (97%) and 29 of 29 patients (100%) proved to be infection-free. When the readers identified a probable claw, 14 of 14 patients (100%) and 16 of 19 patients (84%) proved to be infection-free. Conversely, when the readers identified the absence of claw sign (diffuse DWI pattern), there was proved infection in 17 of 17 cases (100%) and 13 of 14 cases (93%). CONCLUSIONS: In patients with type 1 signal changes of the vertebral disk space, a claw sign is highly suggestive of degeneration and its absence strongly suggests diskitis/osteomyelitis.

Repeated head CT in the neurosurgical intensive care unit: feasibility of sinogram-affirmed iterative reconstruction-based ultra-low-dose CT for surveillance.

BACKGROUND AND PURPOSE: Patients in the neurosurgical intensive care unit undergo multiple head CT scans, resulting in high cumulative radiation exposures. Our aim was to assess the acceptability of a dedicated, special-purpose sinogram-affirmed iterative reconstruction-based ultra-low-dose CT protocol for neurosurgical intensive care unit surveillance head CT examinations, comparing image quality with studies performed with our standard-of-care sinogram-affirmed iterative reconstruction low-dose CT and legacy filtered back-projection standard-dose CT protocols. MATERIAL AND METHODS: A retrospective analysis was performed of 54 head CT examinations: ultra-low-dose CT (n = 22), low-dose CT (n = 12), and standard-dose CT (n = 20) in 22 patients in the neurosurgical intensive care unit. Standard-dose CT was reconstructed by using filtered back-projection on a Somatom Sensation 64 scanner. Ultra-low-dose CT and ultra-low-dose CT examinations were performed on a Siemens AS+128 scanner with commercially available sinogram-affirmed iterative reconstruction. Qualitative and quantitative parameters, including image quality and dose, were evaluated. RESULTS: Sinogram-affirmed iterative reconstruction ultra-low-dose CT represented a 68% lower dose index volume compared with filtered back-projection standard-dose CT techniques in the same patients while maintaining similar quality and SNR levels. Sinogram-affirmed iterative reconstruction low-dose CT offered higher image quality than filtered back-projection standard-dose CT (P < .05) with no differences in SNR at a 24% lower dose index volume. Compared with low-dose CT, ultra-low-dose CT had significantly lower SNR (P = .001) but demonstrated clinically satisfactory measures of image quality. CONCLUSIONS: In this cohort of patients in the neurosurgical intensive care unit, dedicated ultra-low-dose CT for surveillance head CT imaging led to a significant dose reduction while maintaining adequate image quality.

A clinical comparison of the safety and efficacy of MultiHance (gadobenate dimeglumine) and Omniscan (Gadodiamide) in magnetic resonance imaging in patients with central nervous system pathology.

RATIONALE AND OBJECTIVES: The safety and diagnostic efficacy of MultiHance (gadobenate dimeglumine) in the central nervous system (CNS) were evaluated in a double-blind, multicenter, phase III clinical trial. METHODS: Two hundred five patients highly suspected of having a CNS lesion (by previous imaging exam) were enrolled at 16 sites in the United States. Patients were randomized to one of three incremental dosing regimens. Magnetic resonance imaging with Omniscan (gadodiamide) at doses of 0.1 and 0.3 mmol/kg was compared with MultiHance (gadobenate dimeglumine) at doses of 0.05 and 0.15 mmol/kg and at 0.1 and 0.2 mmol/kg. RESULTS: Compared with predose images alone, efficacy was demonstrated in each of the gadobenate dimeglumine and gadodiamide groups (single and cumulative doses) as indicated by the level of diagnostic information, number of lesions detected, and contrast-to-noise ratio measurements. The level of diagnostic information from gadobenate dimeglumine at 0.1 mmol/kg was equivalent to that with gadodiamide at the same dose. One of the two blinded reviewers found equivalence between the gadobenate dimeglumine 0.05 mmol/kg dose and gadodiamide at 0.1 mmol/kg. Both reviewers found the level of diagnostic information to be equivalent after the second dose of contrast for all three dosing regimens. The cumulative doses of gadobenate dimeglumine were well tolerated and as safe as gadodiamide. CONCLUSIONS: Gadobenate dimeglumine is comparable to gadodiamide in terms of safety and efficacy for imaging of CNS lesions, with a possible advantage in imaging applications owing to enhanced T1 relaxivity.