Risk Factors for Unilateral Trigeminal Neuralgia Based on Machine Learning

Application research on the diagnosis of classic trigeminal neuralgia based on VB-Net technology and radiomics

BACKGROUND

This study aims to utilize the deep learning method of VB-Net to locate and segment the trigeminal nerve, and employ radiomics methods to distinguish between CTN patients and healthy individuals.

METHODS

A total of 165 CTN patients and 175 healthy controls, matched for gender and age, were recruited. All subjects underwent magnetic resonance scans. VB-Net was used to locate and segment the bilateral trigeminal nerve of all subjects, followed by the application of radiomics methods for feature extraction, dimensionality reduction, feature selection, model construction, and model evaluation.

RESULTS

On the test set for trigeminal nerve segmentation, our segmentation parameters are as follows: the mean Dice Similarity Coefficient (mDCS) is 0.74, the Average Symmetric Surface Distance (ASSD) is 0.64 mm, and the Hausdorff Distance (HD) is 3.34 mm, which are within the acceptable range. Analysis of CTN patients and healthy controls identified 12 features with larger weights, and there was a statistically significant difference in Rad_score between the two groups (p < 0.05). The Area Under the Curve (AUC) values for the three models (Gradient Boosting Decision Tree, Gaussian Process, and Random Forest) are 0.90, 0.87, and 0.86, respectively. After testing with DeLong and McNemar methods, these three models all exhibit good performance in distinguishing CTN from normal individuals.

CONCLUSIONS

Radiomics can aid in the clinical diagnosis of CTN, and it is a more objective approach. It serves as a reliable neurobiological indicator for the clinical diagnosis of CTN and the assessment of changes in the trigeminal nerve in patients with CTN.

The Challenges in Clinical Diagnosis of Trigeminal Neuralgia: A Review

The lack of established laboratory tests or biomarkers for trigeminal neuralgia (TN) makes diagnosing this relatively rare condition extremely challenging. Trigeminal nerve compression observable on magnetic resonance imaging may indicate TN, but many patients do not have visible lesions or compression. In particular, TN may be confused with migraine, cluster headache, temporomandibular disorder, and other types of headache. An accurate diagnosis is imperative for proper treatment since these conditions do not respond to the same treatment. Many symptoms of these headaches can be vague or overlap, and clinicians depend in large measure on the subjective reports of their patients. Nevertheless, it is imperative to diagnose TN better, which can cause excruciating pain, reduce the quality of life, and even result in disability. It is possible that TN is underestimated.

Analyzing the risk factors of unilateral trigeminal neuralgia under neurovascular compression

Background

This study aimed to explore the risk factors and potential causes of unilateral classical or idiopathic trigeminal neuralgia (C-ITN) by comparing patients and healthy controls (HCs) with neurovascular compression (NVC) using machine learning (ML).

Methods

A total of 84 C-ITN patients and 78 age- and sex-matched HCs were enrolled. We assessed the trigeminal pons angle and identified the compressing vessels and their location and severity. Machine learning was employed to analyze the cisternal segment of the trigeminal nerve (CN V).

Results

Among the C-ITN patients, 53 had NVC on the unaffected side, while 25 HCs exhibited bilateral NVC, and 24 HCs showed unilateral NVC. By comparing the cisternal segment of CN V between C-ITN patients on the affected side and HCs with NVC, we identified the side of NVC, the compressing vessel, and certain texture features as risk factors for C-ITN. Additionally, four texture features differed in the structure of the cisternal segment of CN V between C-ITN patients on the unaffected side and HCs with NVC.

Conclusion

Our findings suggest that the side of NVC, the compressing vessel, and the microstructure of the cisternal segment of CN V are associated with the risk of C-ITN. Furthermore, microstructural changes observed in the cisternal segment of CN V on the unaffected side of C-ITN patients with NVC indicate possible indirect effects on the CN V to some extent.

Altered trends of local brain function in classical trigeminal neuralgia patients after a single trigger pain

OBJECTIVE

To investigate the altered trends of regional homogeneity (ReHo) based on time and frequency, and clarify the time-frequency characteristics of ReHo in 48 classical trigeminal neuralgia (CTN) patients after a single pain stimulate.

METHODS

All patients underwent three times resting-state functional MRI (before stimulation (baseline), after stimulation within 5 s (triggering-5 s), and in the 30th min of stimulation (triggering-30 min)). The spontaneous brain activity was investigated by static ReHo (sReHo) in five different frequency bands and dynamic ReHo (dReHo) methods.

RESULTS

In the five frequency bands, the number of brain regions which the sReHo value changed in classical frequency band were most, followed by slow 4 frequency band. The left superior occipital gyrus was only found in slow 2 frequency band and the left superior parietal gyrus was only found in slow 3 frequency band. The dReHo values were changed in midbrain, left thalamus, right putamen, and anterior cingulate cortex, which were all different from the brain regions that the sReHo value altered. There were four altered trends of the sReHo and dReHo, which dominated by decreased at triggering-5 s and increased at triggering-30 min.

CONCLUSIONS

The duration of brain function changed was more than 30 min after a single pain stimulate, although the pain of CTN was transient. The localized functional homogeneity has time-frequency characteristic in CTN patients after a single pain stimulate, and the changed brain regions of the sReHo in five frequency bands and dReHo complemented to each other. Which provided a certain theoretical basis for exploring the pathophysiology of CTN.

Independent association of the Meckel's cave with trigeminal neuralgia and development of a screening tool

PURPOSE

To 1) investigate the association of the properties of the Meckel's cave (MC) with TN occurrence (i.e., affected vs. unaffected nerves) and whether such association was independent of neurovascular contact (NVC); and 2) develop an objective screening tool for TN.

MATERIALS AND METHODS

Two hundred and nineteen trigeminal nerves were included. (The severity of) NVC was identified for individual nerve, and a set of 107 radiomic features were extracted to characterize various properties of each MC. Both procedures were primarily based on magnetic resonance imaging sequences. A radiomic score (Rad-score) was constructed for each MC to integrate the features associated with TN occurrence. Independent t-test and logistic regression were conducted to assess the association and develop the screening tool mentioned above.

RESULTS

Twelve features were selected to build the Rad-score, with the Inverse Difference Moment Normalized (IDMN) having the greatest weight. The Rad-score was significantly (p ≤ 0.05) higher in the affected compared to the unaffected nerves, irrespective of NVC. The Rad-score and NVC were incorporated in the regression model/screening tool, which demonstrated an acceptable discriminating ability (C-statistic = 0.84).

CONCLUSION

This study has identified a potential association of the properties/features of the MC with TN occurrence, probably involving the demyelination and axonal injury of the trigeminal ganglion within the MC as suggested by the IDMN. Such association may be independent of NVC. This finding may provide new insight into the etiology and/or pathophysiology of TN. The screening tool, which demonstrated an acceptable discriminating ability, may contribute to an improvement in its diagnosis.

Static and dynamic functional connectivity combined with the triple network model in amnestic mild cognitive impairment and Alzheimer's disease

Background

Alzheimer's disease (AD) and amnestic mild cognitive impairment (aMCI) are characterized by abnormal functional connectivity (FC) of default-mode network (DMN), salience network (SN), and central executive network (CEN). Static FC (sFC) and dynamic FC (dFC) combined with triple network model can better study the dynamic and static changes of brain networks, and improve its potential diagnostic value in the diagnosis of AD spectrum disorders.

Methods

Differences in sFC values and dFC variability patterns among the three brain networks of the three groups (53 AD patients, 40 aMCI patients, and 40 NCs) were computed by ANOVA using Gaussian Random Field theory (GRF) correction. The correlation between FC values (sFC values and dFC variability) in the three networks and cognitive scores (MMSE and MoCA) in AD and aMCI groups was analyzed separately.

Results

Within the DMN network, there were significant differences of sFC values in right/left medial superior frontal gyrus and dFC variability in left opercular part inferior frontal gyrus and right dorsolateral superior frontal gyrus among the three groups. Within the CEN network, there were significant differences of sFC values in left superior parietal gyrus. Within the SN network, there were significant differences of dFC variability in right Cerebelum_7b and left opercular part inferior frontal gyrus. In addition, there was a significant negative correlation between FC values (sFC values of CEN and dFC variability of SN) and MMSE and MoCA scores.

Conclusion

It suggests that sFC, dFC combined with triple network model can be considered as potential biomarkers for AD and aMCI.

The role of artificial intelligence in the management of trigeminal neuralgia

Trigeminal neuralgia (TN) is the most frequent facial pain. It is difficult to treat pharmacologically and a significant amount of patients can become drug-resistant requiring surgical intervention. From an etiologically point of view TN can be distinguished in a classic form, usually due to a neurovascular conflict, a secondary form (for example related to multiple sclerosis or a cerebello-pontine angle tumor) and an idiopathic form in which no anatomical cause is identifiable. Despite numerous efforts to treat TN, many patients experience recurrence after multiple operations. This fact reflects our incomplete understanding of TN pathogenesis. Artificial intelligence (AI) uses computer technology to develop systems for extension of human intelligence. In the last few years, it has been a widespread of AI in different areas of medicine to implement diagnostic accuracy, treatment selection and even drug production. The aim of this mini-review is to provide an up to date of the state-of-art of AI applications in TN diagnosis and management.

Development and validation of radiomics models for the prediction of diagnosis of classic trigeminal neuralgia

The study aims to develop a magnetic resonance imaging (MRI)-based radiomics model for the diagnosis of classic trigeminal neuralgia (cTN). This study involved 350 patients with cTN and 100 control participants. MRI data were collected retrospectively for all the enrolled subjects. The symptomatic side trigeminal nerve regions of patients and both sides of the trigeminal nerve regions of control participants were manually labeled on MRI images. Radiomics features of the areas labeled were extracted. Principle component analysis (PCA) and least absolute shrinkage and selection operator (LASSO) regression were utilized as the preliminary feature reduction methods to decrease the high dimensionality of radiomics features. Machine learning methods were established, including LASSO logistic regression, support vector machine (SVM), and Adaboost methods, evaluating each model's diagnostic abilities using 10-fold cross-validation. All the models showed excellent diagnostic ability in predicting trigeminal neuralgia. A prospective study was conducted, 20 cTN patients and 20 control subjects were enrolled to validate the clinical utility of all models. Results showed that the radiomics models based on MRI can predict trigeminal neuralgia with high accuracy, which could be used as a diagnostic tool for this disorder.

Alteration of brain network centrality in CTN patients after a single triggering pain

Objective

The central nervous system may also be involved in the pathogenesis of classical trigeminal neuralgia (CTN). The present study aimed to explore the characteristics of static degree centrality (sDC) and dynamic degree centrality (dDC) at multiple time points after a single triggering pain in CTN patients.

Materials and methods

A total of 43 CTN patients underwent resting-state function magnetic resonance imaging (rs-fMRI) before triggering pain (baseline), within 5 s after triggering pain (triggering-5 s), and 30 min after triggering pain (triggering-30 min). Voxel-based degree centrality (DC) was used to assess the alteration of functional connection at different time points.

Results

The sDC values of the right caudate nucleus, fusiform gyrus, middle temporal gyrus, middle frontal gyrus, and orbital part were decreased in triggering-5 s and increased in triggering-30 min. The sDC value of the bilateral superior frontal gyrus were increased in triggering-5 s and decreased in triggering-30 min. The dDC value of the right lingual gyrus was gradually increased in triggering-5 s and triggering-30 min.

Conclusion

Both the sDC and dDC values were changed after triggering pain, and the brain regions were different between the two parameters, which supplemented each other. The brain regions which the sDC and dDC values were changing reflect the global brain function of CTN patients, and provides a basis for further exploration of the central mechanism of CTN.

Radiomics analysis of unaffected side changes in classic trigeminal neuralgia

OBJECTIVE

To investigate the subtle differences in the structure of the unaffected trigeminal nerve between patients with classic trigeminal neuralgia (CTN) and healthy controls (HCs) by means of radiomics, so as to further explore the etiological mechanism of trigeminal neuralgia (TN).

METHODS

The imagine data of 95 CTN patients and 89 matched HCs were collected and retrospectively analyzed. They were assigned to four groups according to the presence or absence of neurovascular compression (NVC) of the unaffected trigeminal nerve (HCs with and without NVC; CTN patients with and without NVC on the unaffected side). All patients underwent magnetic resonance imaging (MRI) scans. Bilateral trigeminal cisternal segments were manually delineated, followed by feature extraction, dimensionality reduction, feature selection, model construction and model evaluation.

RESULTS

Six weighted textural signatures (sphericity, maximum 2D diameter, skewness, robust mean absolute deviation, large dependence low gray level emphasis, and surface-to-volume ratio) were found in HCs with and without NVC, while 7 were found in CTN patients without NVC on the unaffected side and HCs without NVC. The Rad_score was statistically different between the two groups (P < 0.05). The AUC of the training set was consistent with that of the validation set. The calibration curves of the training and validation sets demonstrated the high accuracy of the model.

CONCLUSIONS

NVC can alter trigeminal nerve structure and cause alterations in related characteristics; but NVC is not a necessary condition for the formation of CTN, and its incidence is also high in asymptomatic healthy people, and thus it is necessary to grade the severity of NVC. In addition, there are differences in the characteristics of the unaffected side between CTN patients and HCs, which may be due to congenital or secondary factors.

Amplitude of low-frequency fluctuation after a single-trigger pain in patients with classical trigeminal neuralgia

Objective

This study aimed to explore the central mechanism of classical trigeminal neuralgia (CTN) by analyzing the static amplitude of low-frequency fluctuation (sALFF) and dynamic amplitude of low-frequency fluctuation (dALFF) in patients with CTN before and after a single-trigger pain.

Methods

This study included 48 patients (37 women and 11 men, age 55.65 ± 11.41 years) with CTN. All participants underwent 3D-T1WI and three times resting-state functional magnetic resonance imaging. The images were taken before stimulating the trigger zone (baseline), within 5 s after stimulating the trigger zone (triggering-5 s), and in the 30th minute after stimulating the trigger zone (triggering-30 min). The differences between the three measurements were analyzed using a repeated-measures analysis of variance.

Results

The sALFF values of the bilateral middle occipital gyrus and right cuneus gradually increased, and the values of the left posterior cingulum gyrus and bilateral superior frontal gyrus gradually decreased in triggering-5 s and triggering-30 min. The values of the right middle temporal gyrus and right thalamus decreased in triggering-5 s and subsequently increased in triggering-30 min. The sALFF values of the left superior temporal gyrus increased in triggering-5 s and then decreased in triggering-30 min. The dALFF values of the right fusiform gyrus, bilateral lingual gyrus, left middle temporal gyrus, and right cuneus gyrus gradually increased in both triggering-5 s and triggering-30 min.

Conclusions

The sALFF and dALFF values changed differently in multiple brain regions in triggering-5 s and triggering-30 min of CTN patients after a single trigger of pain, and dALFF is complementary to sALFF. The results might help explore the therapeutic targets for relieving pain and improving the quality of life of patients with CTN.

Supplementary Information

The online version contains supplementary material available at 10.1186/s10194-022-01488-8.

Textural features reflecting local activity of the hippocampus improve the diagnosis of Alzheimer's disease and amnestic mild cognitive impairment: A radiomics study based on functional magnetic resonance imaging

Background

Textural features of the hippocampus in structural magnetic resonance imaging (sMRI) images can serve as potential diagnostic biomarkers for Alzheimer's disease (AD), while exhibiting a relatively poor discriminant performance in detecting early AD, such as amnestic mild cognitive impairment (aMCI). In contrast to sMRI, functional magnetic resonance imaging (fMRI) can identify brain functional abnormalities in the early stages of cerebral disorders. However, whether the textural features reflecting local functional activity in the hippocampus can improve the diagnostic performance for AD and aMCI remains unclear. In this study, we combined the textural features of the amplitude of low frequency fluctuation (ALFF) in the slow-5 frequency band and structural images in the hippocampus to investigate their diagnostic performance for AD and aMCI using multimodal radiomics technique.

Methods

Totally, 84 AD, 50 aMCI, and 44 normal controls (NCs) were included in the current study. After feature extraction and feature selection, the radiomics models incorporating sMRI images, ALFF values and their combinations in the bilateral hippocampus were established for the diagnosis of AD and aMCI. The effectiveness of these models was evaluated by receiver operating characteristic (ROC) analysis. The radiomics models were further validated using the external data from the Alzheimer's Disease Neuroimaging Initiative (ADNI) database.

Results

The results of ROC analysis showed that the radiomics models based on structural images in the hippocampus had a better diagnostic performance for AD compared with the models using ALFF, while the ALFF-based model exhibited better discriminant performance for aMCI than the models with structural images. The radiomics models based on the combinations of structural images and ALFF were found to exhibit the highest accuracy for distinguishing AD from NCs and aMCI from NCs.

Conclusion

In this study, we found that the textural features reflecting local functional activity could improve the diagnostic performance of traditional structural models for both AD and aMCI. These findings may deepen our understanding of the pathogenesis of AD, contributing to the early diagnosis of AD.