Research Presentation Sessions: Neuro

RPS 1811 - Advanced neuroimaging in various diseases

July 16, 10:30 - 11:30 CEST

RPS 1811-2
8 min
Distinct brain structural-functional network topological coupling explains different outcomes in tinnitus patients treated with sound therapy
Qian Chen, Beijing / China
Author Block: Q. Chen, Z. Wang, H. Lv; Beijing/CN
Purpose or Learning Objective: Topological properties, which serve as the core of the neural network, and their couplings can reflect different therapeutic effects in tinnitus patients. We hypothesised that tinnitus patients with different outcomes after sound therapy (narrowband noise) would have distinct brain network topological alterations.
Methods or Background: Diffusion tensor imaging and resting-state functional magnetic resonance imaging were prospectively performed in 60 patients with idiopathic tinnitus and 57 healthy controls (HCs). Graph-theoretical network analyses of structural connectivity (SC), functional connectivity (FC) and SC and FC coupling were performed. Associations between clinical performance and graph-theoretical features were also analysed.
Results or Findings: Treatment was effective (effective group; EG) in 28 patients and ineffective (ineffective group; IG) in 32 patients. For FC, the patients in the EG showed higher local efficiency than patients in the IG. For SC, patients in both the EG and IG displayed lower normalised characteristic path length, characteristic path length and global efficiency than the HCs. More importantly, patients in the IG had higher coupling than the HCs, whereas there was no difference in coupling between patients in the EG and HCs. Additionally, there were significant associations between the SC features and clinical performance in patients in the EG.
Conclusion: Our findings demonstrate that tinnitus patients exhibited significant brain network topological alterations, especially in the structural brain network. More importantly, patients who demonstrated different curative effects showed distinct SC-FC topological coupling properties. SC-FC coupling could be an indicator that could be used to predict prognoses in patients with idiopathic tinnitus before sound therapy.
Limitations: We didn't conduct it as a longitudinal study.
Ethics committee approval: This study was approved by the ethics committees of Beijing Friendship Hospital, Capital Medical University.
Funding for this study: No. [2015] 160 from the Beijing Scholars Program.
RPS 1811-3
8 min
Periventricular radiomic features in differentiating neuromyelitis optica spectrum disorder (NMOSD) and multiple sclerosis (MS)
Viknesh Balasingam, klang / Malaysia
Author Block: V. Balasingam, N. Mohd Ramli, N. Gowdh, T. Li Kuo; Klang/MY
Purpose or Learning Objective: To (1) extract periventricular radiomics features in MRI brain of NMOSD and MS patients and (2) design a radiomic score and validated nomogram that is useful in differentiating NMOSD and MS.
Methods or Background: The two major idiopathic inflammatory demyelinating diseases (IIDD) of the brain are NMOSD and MS. The clinical features and MRI findings overlap with both, causing periventricular changes on MRI brain. Immunotherapies for MS appear to exacerbate NMOSD, while immunoregulatory is ineffective or even harmful in NMOSD. Radiomics is a relatively new field that aims to convert and analyse medical imaging quantitatively.
Results or Findings: We analysed the MRI of 78 patients who were confirmed IIDD from June 2012 until March 2019. The axial T2W sequence was segmented at three regions of interest at the periventricular region of the lateral, third and fourth ventricles. 117 radiomics features were extracted, and, using multivariate logistic regression, we found five significant radiomics features, which were used to develop the radiomics score. The mean values of the radiomics score for MS and NMOSD were 0.457 and ? 0.729. The highest AUC (0.93) was obtained when the prediction model combined the radiomics and clinical features at the lateral ventricle level. We then developed an individualised prediction model represented by nomograms that were validated using the Harell?s C-index with 1000 bootstrap resampling.
Conclusion: Radiomics provides a good method of discriminating NMOSD and MS. The developed individualised radiomics nomogram has good discriminating potential and can lead to better diagnostic accuracy.
Limitations: Sample size limitation of 78 patients, as this is a rare disease.
Ethics committee approval: Not applicable.
Funding for this study: No funding was received for this work.
RPS 1811-4
8 min
Intraspinal stem cell infusion in amyotrophic lateral sclerosis patients: longitudinal analysis of MR spectroscopy ratios, cortical thickness, clinical parameters and survival
Carmen Trejo Gallego, Murcia / Spain
Author Block: C. Trejo Gallego1, N. Rodríguez Albacete1, F. Iniesta Martinez1, J. M. Plasencia Martinez1, C. Vázquez Olmos1, M. Blanquer1, S. Martínez Pérez2, J. M. Garcia Santos1; 1Murcia/ES, 2Alicante/ES
Purpose or Learning Objective: Evaluate metabolic changes in the precentral gyrus and cortical thickness (CT) after autologous bone marrow mononuclear cell (ABMMC) intraspinal infusion in amyotrophic lateral sclerosis (ALS) patients and its clinical correlation.
Methods or Background: In this phase II clinical trial, 63 patients were randomized in three groups with different interventions each: intraspinal (G1) or intrathecal (G2) injection of ABMMC, or intrathecal (G3) injection of saline. Patients were followed with 1.5 T MR, performed before treatment and 12 months later. The protocol included univoxel MR spectroscopy at both hand knobs and Brain Volume BRAVO, to extract region-wise CT through surface-based morphometry analysis. Clinical surveillance included scales to evaluate functional status, muscular strength and respiratory function, as well as duration and disease progression. For statistical analysis we used SPSS 20.0.0 for Windows and JASP 0.14.
Results or Findings: NAA/Cho and NAA/Cho Cr in the dominant hemisphere increased 12 months after intraspinal infusion. NAA/Cho Cr in the dominant hemisphere before treatment was lower in G1 than in G2 patients. Symptoms progression before treatment was greater in the G1 patients as compared with the other groups. Besides, NAA/Cr and NAA/Cho Cr in the non-dominant hemisphere and rate of progression in the G1 patients demonstrated a negative correlation. Regarding CT in the G1 patients, 3-month MRIs showed a reduction in some points of the frontal and temporo-occipital lobes in the dominant hemisphere, as well as in the parietal and frontal lobes of the non-dominant hemisphere, which, however, increased after 12 months.
Conclusion: Our results go along with those of previous investigations regarding the potential trophic effect of ABMMC intraspinal injection in ALS patients .
Limitations: Small number of patients in the 12 month follow-up and heterogeneous distribution of the disease within each group.
Ethics committee approval: Obtained.
Funding for this study: Absent.
RPS 1811-5
8 min
Validation of automated MS lesion detection in two independent cohorts
Peter Ngum, Turku / Finland
Author Block: P. Ngum1, Y. Forslin2, C. Cananau2, J. Koikkalainen1, R. Ouellette2, E. Sandberg2, F. Barkhof3, F. Piehl2, T. Granberg2; 1Tampere/FI, 2Stockholm/SE, 3London/UK
Purpose or Learning Objective: To examine the performance of the cNeuro® cMRI software in detecting MS lesions from T2-FLAIR images using two independent cohorts.
Methods or Background: The Karolinska Institutet (KI) cohort contained data from 39 relapsing-remitting MS patients on anti-CD-20 therapy with manual segmentation (19 males, 44.7±8.2 years). The MICCAI-2016 cohort contained data from 38 MS patients (8 males, 46.9±10.2 years, four sites, four scanners) with manual segmentations by seven experts used for generating a consensus ground truth. Lesions were detected using the cNeuro® cMRI software (Combinostics, Finland). Several metrics were used to evaluate performance against the ground truth: F1 score = 2*(SENS*PPV)/(SENS+PPV) for lesion detection (lesions > 3 mm3), a correlation coefficient for lesion counts and lesion volumes, and an absolute value of difference in the lesion counts (median and 95th percentile) (lesions > 10 mm3).
Results or Findings: The F1 score was 0.61 (KI) and 0.64 (MICCAI-2016) when compared with the ground truth. When each expert was compared with all other experts, the experts had F1 scores 0.545-0.706, while the corresponding value for automatic detection was 0.552. The correlation coefficient was 0.88 (0.92 without one outlier) (KI), 0.91 (MICCAI-2016) for lesion counts and R=0.97 (KI) and R=0.96 (MICCAI-2016) for lesion volumes. The median (95th percentile) for the absolute difference in lesion counts was 8 (28.6) (KI) and 5.5 (26.6) (MICCAI-2016).
Conclusion: The results show comparable overall performance, F1 score 0.61 and 0.64, in two independent cohorts. As F1 score for automated detection was within the range of seven experts (0.552 vs. 0.545-0.706), the results suggest that image quantification tools are reaching the detection level comparable to human experts.
Limitations: A limited number of cases was used.
Ethics committee approval: The study was approved by the Swedish Ethical Review Authority.
Funding for this study: Supported by grants provided by Medical Diagnostics Karolinska.
RPS 1811-6
8 min
Comparison of quick-brain and routine-brain image quality measured via MRQy, an open-source quality-control tool for imaging
Bronson Ciavarra, Richmond / United States
Author Block: B. Ciavarra, D. Abeysekera, J. Wood, B. Reed, J. Yung, N. Swanston, K-P. Hwang, J. M. Johnson; Houston, TX/US
Purpose or Learning Objective: To evaluate quantitative performance measures between "quick" brain (QB) protocol versus standard-of-care (SOC) protocol in patients at high risk of motion.
Methods or Background: QB sequences reduce scanning time, motion and patient sedation needs. Quantitative imaging measures were provided by MRQy, an open-source quality-control tool for imaging data. The SOC imaging group consisted of ER and inpatients receiving brain MR. QB patients were those suspected not to tolerate the SOC examination. Both studies included Ax T1, T2, T2 FLAIR, T2*, DWI, T1 post and 3D T1 post. Imaging metrics between the groups were analysed using a student's t-test.
Results or Findings: Mean imaging time for the SOC study (420 exams) was 25 min versus 12 min for the QB study (185 scans). Both studies included Ax T1, T2, T2 FLAIR, T2*, DWI, T1 post and 3D T1 post. The QB version utilises GRE sequences and added acceleration. Comparison of mean, range, and variance for the two groups showed no statistical difference for 3D T1 post, T2 and T2 FLAIR imaging (P > .01). Variance within measures of SNR were greater for quick sequences (T2: 0.07, 3D T1: 0.05, T1: 0.08, FLAIR 0.06) versus traditional sequences (T2: 0.06, 3D T1: 0.05, T1: 0.06, FLAIR 0.06).
Conclusion: The QB study reduced imaging time from ~25 to ~12 minutes without significant change in mean, range and variance of foreground intensity values, though variability within measures of SNR were slightly higher in the quick scans compared to SOC studies. Thus, the gain in speed is a trade-off for potential decrease in quantitative imaging metrics.
Limitations: Study relies on quantitative measures, not qualitative, which may not correlate with radiologists' quality impression.
Ethics committee approval: IRB of UT MDACC approved this study.
Funding for this study: Not applicable.
RPS 1811-7
8 min
Radiomics in MRI in the study of patients with multiple sclerosis
Melania Stubos, Trieste / Italy
Author Block: M. Stubos, L. Calderoni, I. Gandin, A. Sartori, M. Ukmar, M. A. A. Cova; Trieste/IT
Purpose or Learning Objective: The aim of this study was to evaluate the contribution of radiomics through the texture analysis in magnetic resonance imaging (MRI) in patients with MS and compare the apparently normal white matter of patients with MS with the white matter of healthy controls.
Methods or Background: A retrospective study on 25 patients with MS (24 with a relapsing-remitting type and 1 with a primary-progressive type) and 25 healthy controls who underwent a 3T MRI was performed. Through a dedicated software, ROIs were placed in the following areas: bilateral frontal, parietal and temporal lobes, thalami and cerebellar peduncles, genu and splenium of the corpus callosum and the lateral ventricles. Finally, features on which to develop a logistic model were chosen.
Results or Findings: Statistically significant differences (p<0.05/nfeature) were found between the two groups, especially in the frontal and parietal lobes and in the splenium of the corpus callosum. In the temporal lobe no statistically significant difference was found. As far as the thalami, the middle cerebellar peduncles and the genu are concerned, the model did not show an optimal performance.
Conclusion: This study allowed the identification of the statistically significative differences between the apparently normal white matter of patients with MS and the white matter of healthy controls through radiomics. The chosen features can become a classification model and be further used to implement machine learning in order to achieve the possibility of an early diagnoses between several different white matter diseases.
Limitations: Not applicable.
Ethics committee approval: This study was approved by an ethics committee.
Funding for this study: Not applicable.
RPS 1811-8
8 min
Left inferior frontal gyrus neuroplastisity during novel speech-sound learning: an fMRI and DTI study
Sahal Saad AL Otaibi, Liverpool / UK
Author Block: S. S. AL Otaibi1, G. Meyer2, S. Wuerger2; 1Taif/SA, 2Liverpool/UK
Purpose or Learning Objective: While the functional and structural changes that occur when we learn new language skills are well documented, relatively little is known about the time course of these changes, especially in the short term. Here we present a combined functional magnetic resonance (fMRI) and diffusion tensor imaging (DTI) study that tracks functional and microstructural change over three days of learning phonetic categorisation in a new language.
Methods or Background: Twenty adult native English-speaking participants were scanned before and after attending a new language (Arabic) training course of one hour for three consecutive days. During the course, the participants were trained to do two tasks: phonetic discrimination and a pronunciation task.
Results or Findings: Behavioural results show significant performance improvements on both tasks after training. Functional imaging analysis provides evidence for (i) significant pre-training blood-oxygenation-level-dependent (BOLD) signal differences between native and non-native languages in the left inferior frontal gyrus (IFG) but not in any other language-related areas; and (ii) significant post-training BOLD signal increases in the left IFG. This increase is correlated with the participants' behavioural performance change. Microstructural (DTI) analysis in the left IFG shows a significant post-training increase in fractional anisotropy (FA) after just three days of training. In functional connectivity, positive correlation was observed between the left inferior frontal gyrus and two areas (the right cerebellum and the left middle occipital gyrus) and a negative correlation in the right lingual gyrus.
Conclusion: These findings indicate that short-term speech-sound learning recruits the same brain area and the same functional network that is being used for processing the L1. The microstructural changes could indicate an early stage of myelination?process in the left IFG.
Limitations: No follow-up scan.
Ethics committee approval: University of Liverpool Ethics [ref# 3384].
Funding for this study: University of Taif.

CME Information

This session is accredited with 1 CME credit.