Prognostic value of global deep white matter DTI metrics for 1-year outcome prediction in ICU traumatic brain injury patients: an MRI-COMA and CENTER-TBI combined study.

PURPOSE: A reliable tool for outcome prognostication in severe traumatic brain injury (TBI) would improve intensive care unit (ICU) decision-making process by providing objective information to caregivers and family. This study aimed at designing a new classification score based on magnetic resonance (MR) diffusion metrics measured in the deep white matter between day 7 and day 35 after TBI to predict 1-year clinical outcome. METHODS: Two multicenter cohorts (29 centers) were used. MRI-COMA cohort (NCT00577954) was split into MRI-COMA-Train (50 patients enrolled between 2006 and mid-2014) and MRI-COMA-Test (140 patients followed up in clinical routine from 2014) sub-cohorts. These latter patients were pooled with 56 ICU patients (enrolled from 2014 to 2020) from CENTER-TBI cohort (NCT02210221). Patients were dichotomised depending on their 1-year Glasgow outcome scale extended (GOSE) score: GOSE 1-3, unfavorable outcome (UFO); GOSE 4-8, favorable outcome (FO). A support vector classifier incorporating fractional anisotropy and mean diffusivity measured in deep white matter, and age at the time of injury was developed to predict whether the patients would be either UFO or FO. RESULTS: The model achieved an area under the ROC curve of 0.93 on MRI-COMA-Train training dataset, and 49% sensitivity for 96.8% specificity in predicting UFO and 58.5% sensitivity for 97.1% specificity in predicting FO on the pooled MRI-COMA-Test and CENTER-TBI validation datasets. CONCLUSION: The model successfully identified, with a specificity compatible with a personalized decision-making process in ICU, one in two patients who had an unfavorable outcome at 1 year after the injury, and two-thirds of the patients who experienced a favorable outcome.

Global mean diffusivity: A radiomarker discriminating good outcome long term after traumatic brain injury.

BACKGROUND: Traumatic brain injury (TBI) is a chronic pathology responsible for cognitive disorders impacting outcome. Global clinical outcome several years after TBI may be associated with anatomical sequelae. Anatomical lesions are not well described because characterizing diffuse axonal injury and brain atrophy require using specific MRI sequences with quantitative measures. The best radiologic parameter to describe the lesions long term after TBI is not known. OBJECTIVE: We aimed to first, assess the global volumetric and diffusion parameters related to long-term outcome after TBI and second, define the most discriminating parameter. METHODS: In this observational study, we included 96 patients with severe TBI and 22 healthy volunteers. The mean delay after TBI was 63.2 months [range 31-119]. The Glasgow Outcome Scale Extended (GOS-E) was used to assess the global long-term clinical outcome. All patients underwent multimodal MRI with measures of brain volume, ventricle volume, global fractional anisotropy (FA) and global mean diffusivity (MD). RESULTS: All 96 participants had significant impairment in global FA, global MD, brain volume and ventricle volume as compared with the 22 controls (P<0.01). Only global MD significantly differed between the "good recovery" group (GOS-E score 7-8) and the other two groups: GOS-E scores 3-4 and 5-6. Brain volume significantly differed between the GOS-E 7-8 and 3-4 groups. Global MD was the most discriminating radiological parameter for the "good recovery" group versus other patients, long term after TBI. FA appeared less relevant at this time. Global atrophy was higher in patients than controls but lacked reliability to discriminate groups of patients. CONCLUSION: Global mean diffusivity seems a more promising radiomarker than global FA for discriminating good outcome long term after TBI. Further work is needed to understand the evolution of these long-term radiological parameters after TBI.

Long-term cognitive disability after traumatic brain injury: Contribution of the DEX relative questionnaires.

Executive functions are high-level cognitive processes commonly impaired after severe traumatic brain injury (sTBI), which may be associated with persistent anosognosia. The dysexecutive questionnaire (DEX) was designed to assess different domains of executive functioning in daily life. Two versions of the DEX exist (DEX-S completed by the patient, DEX-O completed by a relative) to compare cognitive complaints and patient's awareness. This work was aimed at studying the relevance of DEX-O for assessing daily-life limitations, the persistence of anosognosia and its association with global disability (GOSE) and magnetic resonance imaging (MRI) markers of brain alterations. Sixty-three patients (and relatives) were included within 63.4 months (±20.7) after sTBI. DEX-S and DEX-O scores were significantly positively correlated. We obtained significant correlations between DEX-S and episodic memory and phasic alert but not with executive assessment, GOSE and diffusion MRI markers. DEX-O was significantly correlated with executive function, episodic memory, attention (phasic alert sustained and divided attention), with the GOSE and the volume of the body of the corpus callosum (MRI marker). Anosognosia score (DEX-O minus DEX-S) correlated with mean diffusivity measure. These results highlight the clinical interest of DEX-O in assessing long-term disability.

Transcutaneous Laryngeal Ultrasonography for Laryngeal Immobility Diagnosis in Patients with Voice Disorders After Thyroid/Parathyroid Surgery.

BACKGROUND: Transcutaneous laryngeal ultrasonography (TLUS) was recently developed to assess recurrent nerve palsy after thyroid/parathyroid surgery, with variable rates of efficiency. The aim of the current study was to evaluate this technique using subjective estimation and post-processing quantitative data. METHODS: Fifty subjects presenting with a recurrent nerve palsy and 50 "controls" presenting with voice, swallowing, or breathing disorders following thyroid/parathyroid surgery were prospectively included. All of them underwent a flexible laryngoscopy, considered the gold standard, and a ten-second TLUS clip within the 10 days following surgery. In addition to the subjective interpretation of vocal fold motion, two quantitative criteria taking into account motion symmetry (symmetry index, SI) and amplitude (mobility index) of the two hemi-larynges were defined on TLUS acquisitions in adduction and abduction. RESULTS: The subjective interpretation provided a sensitivity of 100% and a specificity of 96%, compared to the gold standard. The quantitative criteria provided a sensitivity and specificity of both 82%, when based on SI solely. When combining SI and mobility index, the sensitivity reached 94%, but the specificity fell to 66%. CONCLUSIONS: Visual assessment of recurrent nerve palsy using TLUS after thyroid/parathyroid surgery appeared a high sensitive and specific test compared to flexible laryngoscopy. Quantitative criteria are promising and need to be refined to better describe the whole TLUS video clip.

Shear Wave Elastography in Thyroid Nodules with Indeterminate Cytology: Results of a Prospective Bicentric Study.

BACKGROUND: The clinical management of thyroid nodules with indeterminate cytology (IC) remains challenging. The role of shear wave elastography (SWE) in this setting is controversial. The aim of the study was to assess the performances of SWE in terms of prediction of malignancy, reproducibility, and combined analysis with ultrasound (US) examination in thyroid nodules with IC. METHODS: This prospective study was conducted in two referral centers. Eligible patients had a thyroid nodule ≥15 mm with IC (Bethesda class III-V) for which surgery had been recommended. Patients underwent a standardized US evaluation combined with a SWE exam followed by surgery. SWE parameters included mean (meanEI; kPa) and max (maxEI) elasticity values, and ratio (meanEI nodule/parenchyma). RESULTS: One hundred and thirty-one nodules (median size 30 mm) in 131 patients were studied. IC was class III in 28%, class IV in 64%, and class V in 8% of cases. After surgery, 21 (16%) nodules were malignant, including nine papillary thyroid cancers (PTC), six follicular thyroid cancers, five poorly differentiated carcinomas, and one large B-cell lymphoma. SWE parameters were similar in benign and malignant nodules, including meanEI (20.2 vs. 19.6 kPa), maxEI (34.3 vs. 32.5 kPa), and ratio (1.57 vs. 1.38). In malignant nodules, meanEI, maxEI, and ratio were higher in the classic PTC variants (n = 4) than in the other PTC variants (n = 5; p < 0.02) and in non-PTC tumors (n = 12; p < 0.005). Intra- and inter-observer coefficients of variations for meanEI in nodules were 23% and 26%, respectively. The French Thyroid Imaging Reporting and Data System score, the American Thyroid Association US classification, and the EU-Thyroid Imaging Reporting and Data System were not associated with malignancy. CONCLUSIONS: Despite high elasticity values in classic PTC variants, conventional SWE indexes failed to discriminate between benign and malignant tumors in thyroid nodules with IC.

Sonographic Dynamic Description of the Laryngeal Tract: Definition of Quantitative Measures to Characterize Vocal Fold Motion and Estimation of Their Normal Values.

Vocal fold motion was analyzed during free breathing using two-dimensional dynamic ultrasound imaging. Two cadavers were first analyzed to define easily identifiable landmarks. Motion of the laryngeal tract was then analyzed in an axial plane. Left and right arytenoids and thyroid cartilage were defined on images corresponding to abduction and adduction of the laryngeal tract. Associated area measurements were established for 50 healthy subjects. All area indices were significantly larger during abduction than adduction. Symmetry of motion was established by comparing each hemi-larynx, and mobility fractions were defined. Normal values of laryngeal motion during free breathing were thus established.

Does motion affect liver stiffness estimates in shear wave elastography? Phantom and clinical study.

This study was undertaken to evaluate the impact of free-breathing (FB) vs. Apnea on Shear-wave elastography (SWE) measurements. Quantitative liver-stiffness measurements were obtained during FB and Apnea for 97 patients with various body-morphologies and liver textures. Quality indexes of FB and Apnea elasticity maps (percentage of non-filling (PNF), temporal (TV) and spatial (SV) variabilities) were computed. SWE measurements were also obtained from an homogeneous phantom at rest and during a mechanically-induced motion. Liver-stiffness values estimated from FB and Apnea acquisitions were correlated, particularly for homogeneous livers (r=0.76, P<0.001) and favorable body-morphologies (r=0.68, P<0.001). However FB values were consistently 20-25% lower than Apnea ones (P<0.001). FB also systematically resulted in degradation of TV (P<0.005) and PNF (P<0.001) compared to Apnea but had no impact on SV. With the phantom, no differences between SWE measurements at rest and during motion were observed. Apnea and FB measurements are highly correlated, although FB data quality is degraded compared to Apnea and estimated stiffness in FB is systematically lower than in Apnea. These discrepancies between rest and motion states were observed for patients but not for phantom data, suggesting that patient breath-holding impacts liver stiffness.

Improved Estimation of Cardiac Function Parameters Using a Combination of Independent Automated Segmentation Results in Cardiovascular Magnetic Resonance Imaging.

This work aimed at combining different segmentation approaches to produce a robust and accurate segmentation result. Three to five segmentation results of the left ventricle were combined using the STAPLE algorithm and the reliability of the resulting segmentation was evaluated in comparison with the result of each individual segmentation method. This comparison was performed using a supervised approach based on a reference method. Then, we used an unsupervised statistical evaluation, the extended Regression Without Truth (eRWT) that ranks different methods according to their accuracy in estimating a specific biomarker in a population. The segmentation accuracy was evaluated by estimating six cardiac function parameters resulting from the left ventricle contour delineation using a public cardiac cine MRI database. Eight different segmentation methods, including three expert delineations and five automated methods, were considered, and sixteen combinations of the automated methods using STAPLE were investigated. The supervised and unsupervised evaluations demonstrated that in most cases, STAPLE results provided better estimates than individual automated segmentation methods. Overall, combining different automated segmentation methods improved the reliability of the segmentation result compared to that obtained using an individual method and could achieve the accuracy of an expert.

Automatic assessment of shear wave elastography quality and measurement reliability in the liver.

A strategy is proposed that accesses the quality of individual shear wave elastography (SWE) exams and the reliability of elasticity measurements in clinical practice. For that purpose, a confidence index based on temporal stability and SWE filling was defined to provide an automatic estimation of each scan quality: high (HG) or low (LG) grade. With this index, the intra-observer acquisition variability assessed by comparing consecutive scans of the same patient was 17% and 32% for HG and LG clips, respectively. The measurement quantification variability assessed by comparing the measurements of a radiologist with those of a trained operator and of two automatic measurements on a same clip averaged 13% and 22% for HG and LG exams, respectively. It was found that SWE measurements depend greatly on the quality of the acquired data. The proposed quality index (HG or LG) provides objective input on the accuracy and diagnostic reliability of SWE measurements.

Hearts and minds: linking vascular rigidity and aerobic fitness with cognitive aging.

Human aging is accompanied by both vascular and cognitive changes. Although arteries throughout the body are known to become stiffer with age, this vessel hardening is believed to start at the level of the aorta and progress to other organs, including the brain. Progression of this vascular impairment may contribute to cognitive changes that arise with a similar time course during aging. Conversely, it has been proposed that regular exercise plays a protective role, attenuating the impact of age on vascular and metabolic physiology. Here, the impact of vascular degradation in the absence of disease was investigated within 2 groups of healthy younger and older adults. Age-related changes in executive function, elasticity of the aortic arch, cardiorespiratory fitness, and cerebrovascular reactivity were quantified, as well as the association between these parameters within the older group. In the cohort studied, older adults exhibited a decline in executive functions, measured as a slower performance in a modified Stroop task (1247.90 ± 204.50 vs. 898.20 ± 211.10 ms on the inhibition and/or switching component, respectively) than younger adults. Older participants also showed higher aortic pulse wave velocity (8.98 ± 3.56 vs. 3.95 ± 0.82 m/s, respectively) and lower VO2 max (29.04 ± 6.92 vs. 42.32 ± 7.31 mL O2/kg/min, respectively) than younger adults. Within the older group, faster performance of the modified Stroop task was associated with preserved aortic elasticity (lower aortic pulse wave velocity; p = 0.046) and higher cardiorespiratory fitness (VO2 max; p = 0.036). Furthermore, VO2 max was found to be negatively associated with blood oxygenation level dependent cerebrovascular reactivity to CO2 in frontal regions involved in the task (p = 0.038) but positively associated with cerebrovascular reactivity in periventricular watershed regions and within the postcentral gyrus. Overall, the results of this study support the hypothesis that cognitive status in aging is linked to vascular health, and that preservation of vessel elasticity may be one of the key mechanisms by which physical exercise helps to alleviate cognitive aging.

An efficient strategy based on an individualized selection of registration methods. Application to the coregistration of MR and SPECT images in neuro-oncology.

An efficient registration strategy is described that aims to help solve delicate medical imaging registration problems. It consists of running several registration methods for each dataset and selecting the best one for each specific dataset, according to an evaluation criterion. Finally, the quality of the registration results, obtained with the best method, is visually scored by an expert as excellent, correct or poor. The strategy was applied to coregister Technetium-99m Sestamibi SPECT and MRI data in the framework of a follow-up protocol in patients with high grade gliomas receiving antiangiogenic therapy. To adapt the strategy to this clinical context, a robust semi-automatic evaluation criterion based on the physiological uptake of the Sestamibi tracer was defined. A panel of eighteen multimodal registration algorithms issued from BrainVisa, SPM or AIR software environments was systematically applied to the clinical database composed of sixty-two datasets. According to the expert visual validation, this new strategy provides 85% excellent registrations, 12% correct ones and only 3% poor ones. These results compare favorably to the ones obtained by the globally most efficient registration method over the whole database, for which only 61% of excellent registration results have been reported. Thus the registration strategy in its current implementation proves to be suitable for clinical application.

Geometry is a major determinant of flow reversal in proximal aorta.

The aim of this study is to quantify aortic backward flow (BF) using phase-contrast cardiovascular magnetic resonance (PC-CMR) and to study its associations with age, indexes of arterial stiffness, and geometry. Although PC-CMR blood flow studies showed a simultaneous presence of BF and forward flow (FF) in the ascending aorta (AA), the relationship between aortic flows and aging as well as arterial stiffness and geometry in healthy volunteers has never been reported. We studied 96 healthy subjects [47 women, 39 ± 15 yr old (19-79 yr)]. Aortic stiffness [arch pulse wave velocity (PWVAO), AA distensibility], geometry (AA diameter and arch length), and parameters related to AA BF and FF (volumes, peaks, and onset times) were estimated from CMR. Applanation tonometry carotid-femoral pulse-wave velocity (PWVCF), carotid augmentation index, and time to return of the reflected pressure wave were assessed. Whereas FF parameters remained unchanged, BF onset time shortened significantly (R(2) = 0.18, P < 0.0001) and BF volume and BF-to-FF peaks ratio increased significantly (R(2) = 0.38 and R(2) = 0.44, respectively, P < 0.0001) with aging. These two latter BF indexes were also related to stiffness indexes (PWVCF, R(2) > 0.30; PWVAO, R(2) > 0.24; and distensibility, R(2) > 0.20, P < 0.001), augmentation index (R(2) > 0.20, P < 0.001), and aortic geometry (AA diameter, R(2) > 0.58; and arch length, R(2) > 0.31, P < 0.001). In multivariate analysis, aortic diameter was the strongest independent correlate of BF beyond age effect. In conclusion, AA BF estimated using PC-CMR increased significantly in terms of magnitude and volume and appeared earlier with aging and was mostly determined by aortic geometry. Thus BF indexes could be relevant markers of subclinical arterial wall alterations.

A framework using multimodal imaging for longitudinal monitoring of patients in neuro-oncology. Application to a SPECT/MRI study.

This paper proposes a framework to assess the potential value of 99mTc Sestamibi SPECT in addition to Gadolinium-enhanced MRI for the monitoring of patients with high grade gliomas under antiangiogenic treatment. It includes: 1) multimodal and monomodal high precision registration steps achieved thanks to a registration strategy which selects the best method among several ones for each dataset, 2) tumor segmentation steps dedicated to each modality and 3) a tumor comparison step which consists in the computation of some global (volume, intensity) and local (matching and mismatching) quantitative indices to analyze the tumor using different imaging modalities and at different times during the treatment. Each step is checked via 2D and 3D visualization. This framework was applied to a database of fifteen patients. For all patients, except one, the tumor volumes decrease globally and locally. Furthermore, a high correlation (r=0.77) was observed between MRI and Sestamibi tumor volumes. Finally, local indices show some possible mismatches between MRI Gadolinium uptake and Sestamibi uptake, which need to be further investigated.

A new strategy to improve coregistration of SPECT and MR images in patients with high grade glioma.

This paper proposes a new strategy to optimize the coregistration of Technetium-99m Sestamibi SPECT and MRI data in case of patients with high grade glioma. It consists in a personalized approach which selects, for each data set, the best registration method among several ones. To achieve this selection, a quantitative dedicated evaluation criterion based on the average intensities within specific anatomical structures corresponding to physiological areas of uptake of Sestamibi was defined. The strategy was applied to sixty-two data sets using nine registration methods based on mutual information and chamfer distance registration approaches, with different settings. It was implemented within the Anatomist/Brainvisa environment, using its basic registration functions. The visual evaluation by experts indicated that this strategy provides 60% good quality registrations, and 26% intermediate quality ones. Compared to the single use of the best global registration method, the number of registrations of good quality was multiplied by 1.4 when using the data specific strategy.

Structure and function of the ascending aorta in palliated transposition of the great arteries.

BACKGROUND: In transposition of the great arteries (TGA), the right ventricle (RV) is subaortic and abnormal aortic structure or function could adversely affect the capacity of the RV to supply the systemic circulation. Our aim was to assess aortic dimensions and distensibility and RV function in patients with palliated TGA using cardiovascular magnetic resonance imaging (CMR). METHODS: We studied 29 patients (22 males; age 29±4 years) with simple TGA, who underwent an atrial switch procedure, and 29 age and sex matched controls. All subjects had cine and phase contrast CMR to evaluate aortic function and global RV function. RESULTS: TGA patients had significant dilatation of the aortic annulus (21.0±3.6 mm vs. 17.6±4.1 mm, p=0.002) and the sinus of Valsalva (30.0±4 mm vs. 26.8±4.2 mm, p=0.005), compared to controls. These findings were associated with reduced distensibility of the ascending aorta in patients with TGA (3.5±1.6 vs. 5.3±2.4 mmHg(-1) x 10(-3), p=0.0009). We could not show a significant correlation between aortic stiffness indices and RV size, function, mass or presence of fibrosis. CONCLUSION: The aortic root dilates and the ascending aorta stiffens in TGA, during young adult life. Although these proximal aortic changes did not show adverse effects on the RV in our young TGA sample, they might have important long-term physiopathological consequences in these patients.

Nonsupervised ranking of different segmentation approaches: application to the estimation of the left ventricular ejection fraction from cardiac cine MRI sequences.

A statistical methodology is proposed to rank several estimation methods of a relevant clinical parameter when no gold standard is available. Based on a regression without truth method, the proposed approach was applied to rank eight methods without using any a priori information regarding the reliability of each method and its degree of automation. It was only based on a prior concerning the statistical distribution of the parameter of interest in the database. The ranking of the methods relies on figures of merit derived from the regression and computed using a bootstrap process. The methodology was applied to the estimation of the left ventricular ejection fraction derived from cardiac magnetic resonance images segmented using eight approaches with different degrees of automation: three segmentations were entirely manually performed and the others were variously automated. The ranking of methods was consistent with the expected performance of the estimation methods: the most accurate estimates of the ejection fraction were obtained using manual segmentations. The robustness of the ranking was demonstrated when at least three methods were compared. These results suggest that the proposed statistical approach might be helpful to assess the performance of estimation methods on clinical data for which no gold standard is available.

Fully automated segmentation of the left ventricle applied to cine MR images: description and results on a database of 45 subjects.

A fully automated segmentation method of the left ventricle from short-axis cardiac MR images is proposed and evaluated. The segmentation is based on morphological filtering and gradient vector flow snake for which an automatic setting of parameters has already been proposed. The present work focuses on the automatic detection of a region of interest (ROI) surrounding the left ventricle, prior to the segmentation step. The whole process was applied to the MICCAI 2009 Left Ventricle Challenge database containing 45 subjects (9 healthy subjects and 36 with pathology). The automatic detection of the ROI was judged accurate in 86% of the cases. It failed in 2% of the slices and provided an overestimation in 9% of the slices. Furthermore, the endocardial segmentation was accurate in 80% of the slices and the epicardial was judged satisfactory in 71% of the slices. This fully automated procedure can thus be used as a first step in a user controlled approach, in order to reduce the total number of interactions.

Measurement of aortic arch pulse wave velocity in cardiovascular MR: comparison of transit time estimators and description of a new approach.

PURPOSE: To investigate the efficiency of a new method (TT-Upslope) for transit time (Δt) estimation from cardiovascular MR (CMR) velocity curves. MATERIALS AND METHODS: Fifty healthy volunteers (40 ± 15 years) underwent applanation tonometry to estimate carotid-femoral pulse wave velocity (cf-PWV) and carotid pressure measurements, and CMR to estimate aortic arch-PWV and ascending aorta distensibility (AAD). The Δt was calculated with TT-Upslope by minimizing the area delimited by two sigmoid curves fitted to the systolic upslope of the ascending (AAC) and descending (DAC) aorta velocity curves, and compared with previously described methods: TT-Point using the half maximum of AAC and DAC, TT-Foot using AAC and DAC feet, and TT-Wave by minimizing the area between AAC and DAC curves using cross correlation. RESULTS: All the Δt methods provided a high reproducibility of arch-PWV. However, TT-Upslope and TT-Wave resulted in better correlations with aging (r = 0.83/r = 0.83 versus r = 0.47/r = 0.72), cf-PWV (r = 0.69/r = 0.70 versus r = 0.34/r = 0.59), and AAD (r = 0.81/r = 0.71 versus r = 0.61/r = 0.60). Furthermore, TT-Upslope resulted in stronger relationship between arch-PWV and AAD according to a theoretical model and provided better characterization of older subjects compared with TT-Wave. CONCLUSION: Arch-PWV estimated with CMR using the TT-Upslope method was found to be reproducible and accurate, providing strong correlations with age and aortic stiffness indices.

Automated estimation of aortic strain from steady-state free-precession and phase contrast MR images.

The strain values extracted from steady-state free-precession (SSFP) and phase contrast (PC) images acquired with a 1.5T scanner on a compliant flow phantom and within the thoracic aorta of 52 healthy subjects were compared. Aortic data were acquired perpendicular to the aorta at the level of the pulmonary artery bifurcation. Cross sectional areas were obtained by using an automatic and robust segmentation method. While a good correlation (r = 0.99) was found between the aortic areas extracted from SSFP and PC sequences, a lower correlation (r = 0.71) was found between the corresponding aortic strain values. Strain values estimated using SSFP and PC sequences were equally correlated with age. Interobserver reproducibility was better for SSFP than for PC. Strain values in the ascending and descending aorta were better correlated for SSFP (r = 0.8) than for PC (r = 0.65) and fitted with the expectation of a larger strain in the ascending aorta when using SSFP. The spatial and temporal resolutions of the acquisitions had a minor influence upon the estimated strain values. Thus, if PC acquisitions can be used to estimate both pulse wave velocity and aortic strain, an additional SSFP sequence may be useful to improve the accuracy in estimating the aortic strain.