Machine learning may predict individual hand motor activation from resting-state fMRI in patients with brain tumors in perirolandic cortex

OBJECTIVE

The study aimed to evaluate the predictive validity of the neural network (NN) method for presurgical mapping of motor areas using resting-state functional MRI (rs-fMRI) data of patients with brain tumor located in the perirolandic cortex (PRC).

METHODS

A total of 109 patients with brain tumors occupying PRC underwent rs-fMRI and hand movement task-based fMRI (tb-fMRI) scans. Using a NN model trained on fMRI data of 47 healthy controls, individual task activation maps were predicted from their rs-fMRI data. NN-predicted maps were compared with task activation and independent component analysis (ICA)-derived maps. Spatial Pearson's correlation coefficients (CC) matrices and Dice coefficients (DC) between task activation and predicted activation using NN (DC_{NN_Act}) and ICA (DC_{ICA_Act}) were calculated and compared using non-parametric tests. The effects of tumor types and head motion on predicted maps were demonstrated.

RESULTS

The CC matrix of NN-predicted maps showed higher diagonal values compared with ICA-derived maps (p < 0.001). DC_{NN_Act} were higher than DC_{ICA_Act} (p < 0.001) for patients with or without motor deficits. Lower DCs were found in subjects with head motion greater than one voxel. DCs were higher on the nontumor side than on the tumor side (p < 0.001), especially in the glioma group compared with meningioma and metastatic groups.

CONCLUSIONS

This study indicated that the NN approach could predict individual motor activation using rs-fMRI data and could have promising clinical applications in brain tumor patients with anatomical and functional reorganizations.

KEY POINTS

• The neural network machine learning approach successfully predicted hand motor activation in patients with a tumor in the perirolandic cortex, despite space-occupying effects and possible functional reorganization. • Compared to the conventional independent component analysis, the neural network approach utilizing resting-state fMRI data yielded a higher correlation to the active task hand activation data. • The Dice coefficient of machine learning-predicted activation vs. task fMRI activation was different between tumor and nontumor side, also between tumor types, which might indicate different effects of possible neurovascular uncoupling on resting-state and task fMRI.

Neutrophil CD64 as a diagnostic marker for neonatal sepsis: Meta-analysis

BACKGROUND

Neutrophil CD64 (nCD64) is a promising marker for diagnosing bacterial infections. Several studies have investigated the performance of nCD64 for diagnosing neonatal sepsis and the results are variable. Interest in nCD64 for detecting serious bacterial infections is increasing rapidly.

OBJECTIVES

The aim of the present study was to carry out a meta-analysis to systematically evaluate the diagnostic accuracy of nCD64 in neonatal sepsis. As far as the authors know, no previous studies have undertaken this.

MATERIAL AND METHODS

A review of studies from Pubmed, Embase and the Cochrane Library, from inception through June 2015, found 7 studies (involving 2213 neonates) fulfilling the inclusion criteria. These 7 studies were subjected to a bivariate meta-analysis of sensitivity and specificity and a summary receiveroperating characteristic (SROC) curve; I2 was used to test heterogeneity, and the source of heterogeneity was investigated by influence analysis and meta-regression.

RESULTS

The pooled sensitivity and specificity were 80% (95%CI, 69-88%) and 83% (95%CI, 71-90%), respectively. The area under the SROC curve (AUC) was 0.88 (95%CI, 0.85-0.91). The studies had substantial heterogeneity (I2 = 87.1%).

CONCLUSIONS

The results showed that nCD64 is a reliable biomarker for diagnosing neonatal sepsis (AUC = 0.88).

Assessing Region of Interest Schemes for the Corticospinal Tract in Patients With Brain Tumors

Diffusion tensor imaging (DTI) and diffusion tensor tractography (DTT) techniques are widely used for identifying the corticospinal tract (CST) white matter pathways as part of presurgical planning. However, mass effects in patients with brain tumors tend to cause anatomical distortions and compensatory functional reorganization of the cortex, which may lead to inaccurate mapping of white matter tracts. To overcome these problems, we compared different region-of-interest (ROI) selection schemes to track CST fibers in patients with brain tumors. Our study investigated the CSTs of 16 patients with intracranial tumors. The patients were classified into 3 subgroups according to the spatial relationships of the lesion and the primary motor cortex (PMC)/internal capsule. Specifically, we investigated the key factors that cause distorted tractography in patients with tumors. We compared 3 CST tractography methods that used different ROI selection schemes. The results indicate that CST fiber tracking methods based only on anatomical ROIs could possibly lead to distortions near the PMC region and may be unable to effectively localize the PMC. In contrast, the dual ROI method, which uses ROIs that have been selected from both blood oxygen level-dependent functional MRI (BOLD-fMRI) activation and anatomical landmarks, enabled the tracking of fibers to the motor cortex. The results demonstrate that the dual ROI method can localize the entire CST fiber pathway and can accurately describe the spatial relationships of CST fibers relative to the tumor. These results illustrate the reliability of using fMRI-guided DTT in patients with tumors. The combination of fMRI and anatomical information enhances the identification of tracts of interest in brains with anatomical deformations, which provides neurosurgeons with a more accurate approach for visualizing and localizing white matter fiber tracts in patients with brain tumors. This approach enhances surgical performance and perserves brain function.

Motor network plasticity and low-frequency oscillations abnormalities in patients with brain gliomas: a functional MRI study

Brain plasticity is often associated with the process of slow-growing tumor formation, which remodels neural organization and optimizes brain network function. In this study, we aimed to investigate whether motor function plasticity would display deficits in patients with slow-growing brain tumors located in or near motor areas, but who were without motor neurological deficits. We used resting-state functional magnetic resonance imaging to probe motor networks in 15 patients with histopathologically confirmed brain gliomas and 15 age-matched healthy controls. All subjects performed a motor task to help identify individual motor activity in the bilateral primary motor cortex (PMC) and supplementary motor area (SMA). Frequency-based analysis at three different frequencies was then used to investigate possible alterations in the power spectral density (PSD) of low-frequency oscillations. For each group, the average PSD was determined for each brain region and a nonparametric test was performed to determine the difference in power between the two groups. Significantly reduced inter-hemispheric functional connectivity between the left and right PMC was observed in patients compared with controls (P<0.05). We also found significantly decreased PSD in patients compared to that in controls, in all three frequency bands (low: 0.01-0.02 Hz; middle: 0.02-0.06 Hz; and high: 0.06-0.1 Hz), at three key motor regions. These findings suggest that in asymptomatic patients with brain tumors located in eloquent regions, inter-hemispheric connection may be more vulnerable. A comparison of the two approaches indicated that power spectral analysis is more sensitive than functional connectivity analysis for identifying the neurological abnormalities underlying motor function plasticity induced by slow-growing tumors.

Does diffusion tensor tractography of the corticospinal tract correctly reflect motor function?

OBJECTIVE

To investigate the consistency of diffusion tensor tractography of the corticospinal tract on motor function. CLINICAL PRESENTATIONS AND INTERVENTION: Three patients with brain tumor were admitted to our hospital with impaired motor function. Diffusion tensor imaging (DTI) and tractography were performed in these patients to assess their affected corticospinal tract. The corticospinal tract showed interruption with moderately impaired motor function in 2 patients. The third case had significantly weakened muscle strength on the left upper limb but an intact right corticospinal tract.

CONCLUSION

These cases showed that the corticospinal tracts obtained by DTI with tractography were inconsistent with motor function. Hence, DTI should be interpreted with caution.

[Transforming growth factor β1 enhances the invasiveness of human glioma cell line via ERK/MAPK pathway]

OBJECTIVE

To investigate the effect of transforming growth factor β1 (TGF-β1) on the invasiveness of human glioma SF767 cell line in vitro and explore the possible molecular mechanism.

METHODS

Human glioma SF767 cell line treated with TGF-β1 for 12, 24, or 48 h were examined for the expressions of epithelial-mesenchymal transition- and ERK/MAPK pathway-associated proteins using Western blotting. MTT assay and Transwell assay were employed to assess the changes in the cell proliferation and invasiveness after TGF-β1 treatment, respectively.

RESULTS

Western blotting showed that TGF-β1 treatment up-regulated the expressions of vimentin, MMP-2, P-ERK, and Zeb-1 and down-regulated E-cadherin expression in SF767 cells. TGF-β1 treatment of the cells resulted in significantly shortened doubling time of cells and obviously increased cell invasiveness.

CONCLUSION

TGF-β1 induces epithelial-mesenchymal transition of SF767 cell line to increase its invasiveness possibly via ERK/MAPK pathway.

Association of Helicobacter pylori with elevated blood ammonia levels in cirrhotic patients: a meta-analysis

PURPOSE

The association between Helicobacter pylori (H. pylori) and blood ammonia levels in cirrhotic patients is controversial. We aimed to clarify this controvercy by performing a meta-analysis of published studies.

MATERIALS AND METHODS

We searched PubMed, EMBASE and Cochrane library for studies which explored the association between H. pylori and blood ammonia levels in cirrhotic patients before May 2012. Six cohort studies involved in 632 H. pylori positive and 396 H. pylori negative cirrhotic patients were eligible for our analysis. The summary estimates were presented as standard means differences (SMD) and 95% confidence intervals (CI) from individual studies.

RESULTS

Overall, there was significant association between H. pylori infection and the elevated blood ammonia levels in cirrhotic patients (SMD=0.34, 95% CI=0.21-0.47, I²=42.1%). Sensitivity analysis further confirmed this association. Subgroup analysis showed that the association was found only in Asian ethnicity, but not in Caucasian ethnicity.

CONCLUSION

H. pylori infection is associated with elevated blood ammonia levels in cirrhotic patients, and more large scale studies and stratify analysis are warranted in order to further evaluate this association.

Diffusion tensor imaging and proton magnetic resonance spectroscopy in brain tumor: Correlation between structure and metabolism

Proton magnetic resonance spectroscopy and diffusion tensor imaging are non-invasive techniques used to detect metabolites and water diffusion in vivo. Previous studies have confirmed a positive correlation of individual fractional anisotropy values with N-acetylaspartate/creatine and N-acetylaspartate/choline ratios in tumors, edema, and normal white matter. This study divided the brain parenchyma into tumor, peritumoral edema, and normal-appearing white matter according to MRI data, and analyzed the correlation of metabolites with water molecular diffusion. Results demonstrated that in normal-appearing white matter, N-acetylaspartate/creatine ratios were positively correlated with fractional anisotropy values, negatively correlated with radial diffusivities, and positively correlated with maximum eigenvalues. Maximum eigenvalues and radial diffusivities in peritumoral edema showed a negative correlation with choline, N-acetylaspartate, and creatine. Radial diffusivities in tumor demonstrated a negative correlation with choline. These data suggest that the relationship between metabolism and structure is markedly changed from normal white matter to peritumoral edema and tumor. Neural metabolism in the peritumoral edema area decreased with expanding extracellular space. The normal relationship of neural function and microstructure disappeared in the tumor region.

Differentiation of pure vasogenic edema and tumor-infiltrated edema in patients with peritumoral edema by analyzing the relationship of axial and radial diffusivities on 3.0T MRI

OBJECT

The purpose of this study was to analyze the relationship of axial diffusivity (AD) and radial diffusivity (RD) in peritumoral edema of high grade glioma, metastasis, and meningioma, and to differentiate tumor-infiltrated edema from pure vasogenic edema.

MATERIALS AND METHODS

Twenty patients with high-grade glioma and 16 with meningioma or metastatic tumor were enrolled in our study. All cases were confirmed by histopathological study. Diffusion tensor imaging (DTI) was performed in all patients. Peritumoral edema of high-grade glioma was considered tumor-infiltrated edema, and edema of meningioma or metastasis was considered pure vasogenic edema. Fractional anisotropy (FA), mean diffusivity (MD), AD, RD, regression coefficient of RD to AD (RCRD-AD), and tumor infiltration index (TII) in tumor-infiltrated edema and pure vasogenic edema were analyzed and compared. Receiver operating characteristic (ROC) curve analysis was conducted to demonstrate their differential effectiveness.

RESULTS

RCRD-AD in tumor-infiltrated edema (0.724±0.125) was significantly higher than that in pure vasogenic edema (0.571±0.111) (P=0.001). FA of tumor-infiltrated edema (0.175±0.025) was significantly lower than that of pure vasogenic edema (0.203±0.035) (P=0.007). Other parameters showed no significant difference between the 2 types of edema. ROC curve analysis showed RCRD-AD was the most effective parameter in distinguishing tumor-infiltrated edema from pure vasogenic edema. Using a threshold of 0.6, a sensitivity of 0.85 and specificity of 0.69 can be achieved with RCRD-AD.

CONCLUSION

Analysis of the AD and RD relationship may reflect differences in diffusion characteristics of edema surrounding high-grade glioma and meningioma or metastasis, and may be helpful in detecting peritumoral infiltration in high-grade glioma.

[Accuracy of MR perfusion weighted imaging for cerebral glioma grading: a meta-analysis]

OBJECTIVE

To evaluate values of MR perfusion weighted imaging(PWI) in the grading of intracranial gliomas by a meta-analysis.

METHODS

All English and Chinese literatures published before November 2009 in PUBMED and CNKI (China National Knowledge Infrastructure) were searched. Literature searching requirements: (1) MR PWI with region of interest (ROI) must have been applied in the grading of gliomas; (2) histopathological results must have been used as the reference standards. In addition, after data extraction, a bivariate random-effect model and hierarchical weighted symmetric summary receiver-operating curve must have been utilized to pool the data. Furthermore, a meta regression method must have been applied to detect and analyze the factors that affected the diagnostic accuracy.

RESULTS

Overall, fourteen studies were included with a total amount of 1021 patients. The pooled sensitivity, specificity and diagnostic odds ratio with 95% of CI were 93% (89% - 96%), 81% (73% - 87%) and 55 (28 - 107), respectively. The factors that affected the diagnostic accuracy were the sample size, the ratio of malignant glioma, the injection rate of contrast agents, the repetition time and the cutoff value. Heterogeneity did exist among the results obtained from different studies.

CONCLUSION

the relative cerebral blood volume (rCBV) of MR PWI can be referred to differentiate malignant cerebral gliomas from benign ones with sound sensitivity and specificity. However, there is not a unified threshold applied in the researches that have been conducted so far, and techniques used differ among the studies. In summary, clinical application values of rCBV of MR PWI are to be corroborated by further studies with a larger sample size.

[Analysis on the relationship between tensor veli palatini and secretory otitis media in patients with nasopharyngeal carcinoma after radiotherapy]

OBJECTIVE

The analyze the relationship between secretory otitis media (SOM) and injury of tensor veli palatini (TVP) muscle injury after radiotherapy, then to explore the pathogenesis of SOM in patients with nasopharyngeal carcinoma (NPC) after radiotherapy.

METHOD

The cross section area (CSA) of TVP and medial pterygoid (MP) muscle were measured in MRI of 32 patients with NPC before and after radiotherapy, meanwhile the incidence of SOM were surveyed after radiotherapy, then to analyze the relationship between the change of TVP and the incidence of SOM after radiotherapy.

RESULT

Of 48 ears without SOM before radiotherapy, 27 ears developed post-irradiation SOM, including 24 ears with TVP atrophy and 3 ears without TVP atrophy, and 21 ears had no post-irradiation SOM, including 8 ears with TVP atrophy and 13 ears without TVP atrophy. chi2 test showed significant difference (P < 0.01). It indicated that post-irradiation SOM have correlation with TVP atrophy. The more possibility of TVP atrophy occurred after long time radiotherapy.

CONCLUSION

The atrophy of TVP in patients with NPC usually occurs 6 months after radiotherapy, and this may result in the post-irradiation SOM. The pathogenesis of post-irradiation SOM need further study functionally.