Objective
Test a practical realignment approach to compensate the technical variability of MR radiomic features.
Methods
T1 phantom images acquired on 2 scanners, FLAIR and contrast-enhanced ...T1-weighted (CE-T1w) images of 18 brain tumor patients scanned on both 1.5-T and 3-T scanners, and 36 T2-weighted (T2w) images of prostate cancer patients scanned in one of two centers were investigated. The ComBat procedure was used for harmonizing radiomic features. Differences in statistical distributions in feature values between 1.5- and 3-T images were tested before and after harmonization. The prostate studies were used to determine the impact of harmonization to distinguish between Gleason grades (GGs).
Results
In the phantom data, 40 out of 42 radiomic feature values were significantly different between the 2 scanners before harmonization and none after. In white matter regions, the statistical distributions of features were significantly different (
p
< 0.05) between the 1.5- and 3-T images for 37 out of 42 features in both FLAIR and CE-T1w images. After harmonization, no statistically significant differences were observed. In brain tumors, 41 (FLAIR) or 36 (CE-T1w) out of 42 features were significantly different between the 1.5- and 3-T images without harmonization, against 1 (FLAIR) or none (CE-T1w) with harmonization. In prostate studies, 636 radiomic features were significantly different between GGs after harmonization against 461 before. The ability to distinguish between GGs using radiomic features was increased after harmonization.
Conclusion
ComBat harmonization efficiently removes inter-center technical inconsistencies in radiomic feature values and increases the sensitivity of studies using data from several scanners.
Key Points
• Radiomic feature values obtained using different MR scanners or imaging protocols can be harmonized by combining off-the-shelf image standardization and feature realignment procedures.
• Harmonized radiomic features enable one to pool data from different scanners and centers without a substantial loss of statistical power caused by intra- and inter-center variability.
• The proposed realignment method is applicable to radiomic features from different MR sequences and tumor types and does not rely on any phantom acquisition.
To assess the influence of gray-level discretization on inter- and intra-observer reproducibility of texture radiomics features on clinical MR images.
We studied two independent MRI datasets of 74 ...lacrymal gland tumors and 30 breast lesions from two different centers. Two pairs of readers performed three two-dimensional delineations for each dataset. Texture features were extracted using two radiomics softwares (Pyradiomics and an in-house software). Reproducible features were selected using a combination of intra-class correlation coefficient (ICC) and concordance and coherence coefficient (CCC) with 0.8 and 0.9 as thresholds, respectively. We tested six absolute and eight relative gray-level discretization methods and analyzed the distribution and highest number of reproducible features obtained for each discretization. We also analyzed the number of reproducible features extracted from computer simulated delineations representative of inter-observer variability.
The gray-level discretization method had a direct impact on texture feature reproducibility, independent of observers, software or method of delineation (simulated vs. human). The absolute discretization consistently provided statistically significantly more reproducible features than the relative discretization. Varying the bin number of relative discretization led to statistically significantly more variable results than varying the bin size of absolute discretization.
When considering inter-observer reproducible results of MRI texture radiomics features, an absolute discretization should be favored to allow the extraction of the highest number of potential candidates for new imaging biomarkers. Whichever the chosen method, it should be systematically documented to allow replicability of results.
Celotno besedilo
Dostopno za:
DOBA, IZUM, KILJ, NUK, PILJ, PNG, SAZU, SIK, UILJ, UKNU, UL, UM, UPUK
Objectives
To determine the diagnostic performances of a single Dixon-T2-weighted imaging (WI) sequence compared to a conventional protocol including T1-, T2-, and fat-suppressed T2-weighted MRI at ...3 T when assessing thyroid eye disease (TED).
Materials and methods
This IRB-approved prospective single-center study enrolled participants presenting with confirmed TED from April 2015 to October 2019. They underwent an MRI, including a conventional protocol and a Dixon-T2WI sequence. Two neuroradiologists, blinded to all data, read both datasets independently and randomly. They assessed the presence of extraocular muscle (EOM) inflammation, enlargement, fatty degeneration, or fibrosis as well as the presence of artifacts. The Wilcoxon signed-rank test was used.
Results
Two hundred six participants were enrolled (135/206 66% women, 71/206 34% men, age 52.3 ± 13.2 years). Dixon-T2WI was significantly more likely to detect at least one inflamed EOM as compared to the conventional set (248/412 60% versus 228/412 55% eyes; (
p
= 0.02). Dixon-T2WI was more sensitive and specific than the conventional set for assessing muscular inflammation (100% versus 94.7% and 71.2% versus 68.5%, respectively). Dixon-T2WI was significantly less likely to show major or minor artifacts as compared to fat-suppressed T2WI (20/412 5% versus 109/412 27% eyes,
p
< 0.001, and 175/412 42% versus 257/412 62% eyes,
p
< 0.001). Confidence was significantly higher with Dixon-T2WI than with the conventional set (2.35 versus 2.24,
p
= 0.003).
Conclusion
Dixon-T2WI showed higher sensitivity and specificity and showed fewer artifacts than a conventional protocol when assessing thyroid eye disease, in addition to higher self-reported confidence.
Key Points
•
Dixon-T2WI has better sensitivity and specificity than a conventional protocol for assessing inflamed extraocular muscles in patients with thyroid eye disease.
•
Dixon-T2WI shows significantly fewer artifacts than fat-suppressed T2WI.
•
Dixon-T2WI is faster and is associated with significantly higher self-reported reader confidence as compared to a conventional protocol when assessing inflammatory extraocular muscles.
OBJECTIVESDistinguishing benign from malignant orbital lesions remains challenging both clinically and with imaging, leading to risky biopsies. The objective was to differentiate benign from ...malignant orbital lesions using radiomics on 3 T magnetic resonance imaging (MRI) examinations.
MATERIALS AND METHODSThis institutional review board–approved prospective single-center study enrolled consecutive patients presenting with an orbital lesion undergoing a 3 T MRI prior to surgery from December 2015 to July 2019. Radiomics features were extracted from 6 MRI sequences (T1-weighted images WIs, DIXON-T2-WI, diffusion-WI, postcontrast DIXON-T1-WI) using the Pyradiomics software. Features were selected based on their intraobserver and interobserver reproducibility, nonredundancy, and with a sequential step forward feature selection method. Selected features were used to train and optimize a Random Forest algorithm on the training set (75%) with 5-fold cross-validation. Performance metrics were computed on a held-out test set (25%) with bootstrap 95% confidence intervals (95% CIs). Five residents, 4 general radiologists, and 3 expert neuroradiologists were evaluated on their ability to visually distinguish benign from malignant lesions on the test set. Performance comparisons between reader groups and the model were performed using McNemar test. The impact of clinical and categorizable imaging data on algorithm performance was also assessed.
RESULTSA total of 200 patients (116 58% women and 84 42% men; mean age, 53.0 ± 17.9 years) with 126 of 200 (63%) benign and 74 of 200 (37%) malignant orbital lesions were included in the study. A total of 606 radiomics features were extracted. The best performing model on the training set was composed of 8 features including apparent diffusion coefficient mean value, maximum diameter on T1-WIs, and texture features. Area under the receiver operating characteristic curve, accuracy, sensitivity, and specificity on the test set were respectively 0.869 (95% CI, 0.834–0.898), 0.840 (95% CI, 0.806–0.874), 0.684 (95% CI, 0.615–0.751), and 0.935 (95% CI, 0.905–0.961). The radiomics model outperformed all reader groups, including expert neuroradiologists (P < 0.01). Adding clinical and categorizable imaging data did not significantly impact the algorithm performance (P = 0.49).
CONCLUSIONSAn MRI radiomics signature is helpful in differentiating benign from malignant orbital lesions and may outperform expert radiologists.
Objective
To determine the sensitivity and specificity of high-resolution (HR) MRI for detecting signal abnormalities of cranial nerves (CN) in giant cell arteritis (GCA) patients presenting with ...diplopia.
Methods
This IRB-approved retrospective single-center study included GCA patients who underwent 3-T HR MRI from December 2014 to January 2020. Two radiologists, blinded to all data, individually assessed for the presence of enhancement of the 3rd, 4th, and/or 6th CN on post-contrast HR imaging and high signal intensity on HR T2-WI, for signal abnormalities of extraocular muscles and the brainstem, and for inflammatory changes of the ophthalmic and extracranial arteries. A Fisher’s exact test was used to compare patients with or without diplopia.
Results
In total, 64 patients (42/64 (66%) women and 22/64 (34%) men, mean age 76.3 ± 8 years) were included. Of the 64 patients, 14 (21.9%) presented with diplopia. Third CN enhancement was detected in 7/8 (87.5%) patients with 3rd CN impairment, as compared to no patients with 4th or 6th CN impairment or to patients without diplopia (
p
< 0.001). Third CN abnormal high signal intensity on HR T2-WI was detected in 4/5 patients (80%) with 3rd CN impairment versus none of other patients (
p
< 0.001). Sensitivity, specificity, positive predictive value, and negative predictive value for detecting 3rd CN signal abnormalities were of 0.88, 1, 1, and 0.99 and 0.8, 1, 1, and 0.98 for post-contrast HR imaging and HR T2-WI, respectively.
Conclusions
HR MRI had excellent diagnostic sensitivity and specificity when detecting signal abnormalities of the 3rd CN in GCA patients presenting with 3rd CN impairment.
Key Points
•
Third cranial nerve enhancement was detected in all patients with 3rd cranial nerve impairment except for one with transient diplopia
.
•
The “check mark sign” might be useful to identify 3rd cranial nerve signal abnormalities in the orbital apex
.
•
No signal abnormalities of the 4th or 6th cranial nerves could be detected on high-resolution MRI
.
Objectives
To assess the impact of timing from visual symptoms’ onset to diffusion-weighted (DW) 3 T MRI completion to detect ischemic changes of the optic disc and optic nerve in AION patients.
...Methods
This IRB-approved retrospective single-center study included 3 T MRI data from 126 patients with AION and 111 controls with optic neuritis treated between January 2015 and May 2020. Two radiologists blinded to all data individually analyzed imaging. A senior neuroradiologist resolved any discrepancies by consensus. The primary judgment criterion was the restricted diffusion of the optic disc and/or the optic nerve assessed subjectively on the ADC maps. ADC values were also measured. Spearman rank correlations were used to examine the relationships between timing from visual symptoms’ onset to MRI completion and both the restricted diffusion and the ADC values.
Results
One hundred twenty-six patients (47/126 37.3% women and 79/126 62.7% men, mean age 69.1 ± 13.7 years) with AION were included. Restricted diffusion of the optic disc in AION eyes was more frequent in the early MRI group than in the late MRI group: 35/49 (71.4%) eyes versus 3/83 (3.6%) eyes,
p
< 0.001. ADC values of the pathological optic discs and optic nerves were lower in the early MRI group than in the late MRI group: 0.61 0.52–0.94 × 10
−3
mm
2
/s versus 1.28 1.01–1.44 × 10
−3
mm
2
/s,
p
< 0.001, and 0.74 0.61–0.88 × 10
−3
mm
2
/s versus 0.89 0.72–1.10 × 10
−3
mm
2
/s,
p
< 0.001, respectively.
Conclusions
DWI MRI showed good diagnostic performance to detect AION when performed early after the onset of visual symptoms.
Key Points
• Restricted diffusion of the optic disc in eyes affected by AION was significantly more likely to be observed in patients who had undergone MRI within 5 days after onset of visual symptoms.
•
ADC values of the pathological optic discs and optic nerves were significantly lower in patients who had undergone MRI within 5 days after onset of visual symptoms of AION: 0.61 × 10
−3
mm
2
/
s versus 1.28 × 10
−3
mm
2
/s,
p < 0.001, and 0.74 × 10
−3
mm
2
/
s versus 0.89 × 10
−3
mm
2
/s,
p < 0.001, respectively.
•
The optimal threshold for timing from visual symptoms’ onset to MRI completion to detect restricted diffusion of the optic disc and/or optic nerve was 5 days, with an AUC of 0.88
(
CI
95%
: 0.82–0.94
).
Objectives
To determine the diagnostic accuracy of MRI intravoxel incoherent motion (IVIM) when characterizing orbital lesions, which is challenging due to a wide range of locations and histologic ...types.
Methods
This IRB-approved prospective single-center study enrolled participants presenting with an orbital lesion undergoing a 3-T MRI prior to surgery from December 2015 to July 2019. An IVIM sequence with 15
b
values ranging from 0 to 2000 s/mm
2
was performed. Two neuroradiologists, blinded to clinical data, individually analyzed morphological MRIs. They drew one region of interest inside each orbital lesion, providing apparent diffusion coefficient (ADC), true diffusion coefficient (
D
), perfusion fraction (
f
), and pseudodiffusion coefficient (
D
*) values.
T
test, Mann-Whitney
U
test, and receiver operating characteristic curve analyses were performed to discriminate between orbital lesions and to determine the diagnostic accuracy of the IVIM parameters.
Results
One hundred fifty-six participants (84 women and 72 men, mean age 54.4 ± 17.5 years) with 167 orbital lesions (98/167 59% benign lesions including 54 orbital inflammations and 69/167 41% malignant lesions including 32 lymphomas) were included in the study.
ADC and
D
were significantly lower in malignant than in benign lesions: 0.8 × 10
−3
mm
2
/s 0.45 versus 1.04 × 10
−3
mm
2
/s 0.33,
p
< 0.001, and 0.75 × 10
−3
mm
2
/s 0.40 versus 0.98 × 10
−3
mm
2
/s 0.42,
p
< 0.001, respectively.
D
* was significantly higher in malignant lesions than in benign ones: 12.8 × 10
−3
mm
2
/s 20.17 versus 7.52 × 10
−3
mm
2
/s 7.57,
p
= 0.005. Area under curve was of 0.73, 0.74, 0.72, and 0.81 for ADC,
D
,
D
*, and a combination of
D
,
f
, and
D
*, respectively.
Conclusions
Our study showed that IVIM might help better characterize orbital lesions.
Key Points
• Intravoxel incoherent motion (IVIM) helps clinicians to assess patients with orbital lesions.
• Intravoxel incoherent motion (IVIM) helps clinicians to characterize orbital lymphoma versus orbital inflammation.
• Management of patients becomes more appropriate.
To assess the efficacy of endovascular treatment (EVT) of intracranial aneurysms for recurrence, bleeding, and de novo aneurysm formation at long-term follow-up (> 10 years after treatment) with ...magnetic resonance (MR) angiography and to identify risk factors for recurrence through a prospective study and a systematic review of the literature.
Clinical examinations and 3-T MR angiography were performed prospectively 10 years after EVT of intracranial aneurysms in a single institution. Ethics committee approval and informed consent were obtained. PubMed, EMBASE, and Cochrane databases were searched to identify studies in which authors reported bleeding and/or aneurysm recurrence rates in patients who received follow-up more than 10 years after EVT. Univariate and multivariate subgroup analyses were performed to identify risk factors (midterm MR angiographic results, aneurysm characteristics, retreatment within 5 years).
In the prospective study, sac recanalization occurred between midterm and long-term MR angiography in 16 of 129 (12.4%) aneurysms. Grade 2 classification on the Raymond scale at midterm MR angiography (relative risk RR, 4.16; 99% confidence interval CI: 2.12, 8.14) and retreatment within 5 years (RR, 4.67; 99% CI: 1.55, 14.03) were risk factors for late recurrence. In the systematic review (15 cohorts, 2773 patients, 2902 aneurysms), bleeding, aneurysm recurrence, and de novo lesion formation rates were, respectively, 0.7% (99% CI: 0.2%, 2.7%; I(2), 0%; one of 694 patients), 11.4% (99% CI: 7.0%, 18.0%; I(2), 21.6%), and 4.1% (99% CI: 1.7, 9.4%; I(2), 54.1%). Raymond grade 2 initial result (RR, 7.08; 99% CI: 1.24, 40.37; I(2), 82.6%) and aneurysm size greater than 10 mm (RR, 4.37; 99% CI: 1.83, 10.44; I(2), 0%) were risk factors for late recurrence.
EVT of intracranial aneurysm is effective for prevention of long-term bleeding, but recurrences occur in a clinically relevant percentage of patients, a finding that may justify follow-up of selected patients for 10 years or more, such as patients with aneurysms larger than 10 mm or classified as Raymond grade 2 at midterm MR angiography.
Objectives
To compare the diagnostic accuracy of 3D versus 2D contrast-enhanced vessel-wall (CE-VW) MRI of extracranial and intracranial arteries in the diagnosis of GCA.
Methods
This prospective ...two-center study was approved by a national research ethics board and enrolled participants from December 2014 to October 2017. A protocol including both a 2D and a 3D CE-VW MRI at 3 T was performed in all patients. Two neuroradiologists, blinded to clinical data, individually analyzed separately and in random order 2D and 3D sequences in the axial plane only or with reformatting. The primary judgment criterion was the presence of GCA-related inflammatory changes of extracranial arteries. Secondary judgment criteria included inflammatory changes of intracranial arteries and the presence of artifacts. A McNemar’s test was used to compare 2D to 3D CE-VW MRIs.
Results
Seventy-nine participants were included in the study (42 men and 37 women, mean age 75 (± 9.5 years)). Fifty-one had a final diagnosis of GCA. Reformatted 3D CE-VW was significantly more sensitive than axial-only 3D CE-VW or 2D CE-VW when showing inflammatory change of extracranial arteries: 41/51(80%) versus 37/51 (73%) (
p
= 0.046) and 35/50 (70%) (
p
= 0.03). Reformatted 3D CE-VW was significantly more specific than 2D CE-VW: 27/27 (100%) versus 22/26 (85%) (
p
= 0.04). 3D CE-VW showed higher sensitivity than 2D CE-VW when detecting inflammatory changes of intracranial arteries: 10/51(20%) versus 4/50(8%),
p
= 0.01. Interobserver agreement was excellent for both 2D and 3D CE-VW MRI: κ = 0.84 and 0.82 respectively.
Conclusions
3D CE-VW MRI supported more accurate diagnoses of GCA than 2D CE-VW.
Key Points
• 3D contrast-enhanced vessel-wall magnetic resonance imaging is a high accuracy, non-invasive diagnostic tool used to diagnose giant cell arteritis.
• 3D contrast-enhanced vessel-wall imaging is feasible for clinicians to complete within a relatively short time, allowing immediate assessment of extra and intracranial arteries.
• 3D contrast-enhanced vessel-wall magnetic resonance imaging might be considered a diagnostic tool when intracranial manifestation of GCA is suspected.
PDF
