MRI features of responsible contacts in vascular compressive trigeminal neuralgia and prediction modeling

Abstract

Background

Multiple neurovascular contacts in patients with vascular compressive trigeminal neuralgia often challenge the diagnosis of responsible contacts.

Purpose

To analyze the magnetic resonance imaging (MRI) features of responsible contacts and establish a predictive model to accurately pinpoint the responsible contacts.

Material and Methods

Sixty-seven patients with unilateral trigeminal neuralgia were enrolled. A total of 153 definite contacts (45 responsible, 108 non-responsible) were analyzed for their MRI characteristics, including neurovascular compression (NVC) grading, distance from pons to contact (D_pons-contact), vascular origin of compressing vessels, diameter of vessel (D_vessel) and trigeminal nerve (D_{trigeminal nerve}) at contact. The MRI characteristics of the responsible and non-responsible contacts were compared, and their diagnostic efficiencies were further evaluated using a receiver operating characteristic (ROC) curve. The significant MRI features were incorporated into the logistics regression analysis to build a predictive model for responsible contacts.

Results

Compared with non-responsible contacts, NVC grading and arterial compression ratio (84.44%) were significantly higher, D_pons-contact was significantly lower at responsible contacts (P < 0.001, 0.002, and 0.033, respectively). NVC grading had a highest diagnostic area under the ROC curve (AUC) of 0.742, with a sensitivity of 64.44% and specificity of 75.00%. The logistic regression model showed a higher diagnostic efficiency, with an AUC of 0.808, sensitivity of 88.89%, and specificity of 62.04%.

Conclusion

Contact degree and position are important MRI features in identifying the responsible contacts of the trigeminal neuralgia. The logistic predictive model based on D_pons-contact, NVC grading, and vascular origin can qualitatively improve the prediction of responsible contacts for radiologists.

Keywords

Trigeminal neuralgia magnetic resonance tomographic angiography responsible contact neurovascular compression grading transition zone

Introduction

Primary trigeminal neuralgia is a paroxysmal needle-like, knife-like, or electric shock-like pain with an incidence rate of about 4–20/100,000 (1). It often occurs in the distribution area of the trigeminal nerve and can erupt through a trigger point, usually lasting for several seconds to 2 min. Though not a fatal disease, it seriously compromises the quality of life of those affected (2).

Generally, the condition is caused by neurovascular contact in the cisternal segment of the trigeminal nerve (3). The current mainstream treatments for the trigeminal neuralgia rely heavily on drugs, microvascular decompression (MVD) surgery, and stereotactic radiosurgery (4 –8). Before treatment, routine magnetic resonance tomographic angiography (MRTA) is often prescribed to rule out secondary trigeminal neuralgia caused by conditions such as neoplasm or multiple sclerosis. This routine examination can reveal potential vascular compressive lesions by clearly indicating the contact position, degree, and origin of blood vessels (arteries and veins) between the nerve and blood vessel in the cisternal segment (9 –12). However, neurovascular contacts are very common in the afflicted and even in healthy people, and merely a small part of them are responsible contacts, where vascular compression-induced nerve demyelination causes symptoms. Additionally, multiple neurovascular contacts can be present in both symptomatic and asymptomatic sides (9). These conditions often pose challenges in making an accurate diagnosis for vascular compressive trigeminal neuralgia. Accurate diagnosis of responsible contacts will help radiologists and surgeons better understand the mechanism of the trigeminal neuralgia and improve the effectiveness of the treatments. In contrast, equivocal contacts may delay treatment or potentially lead to unnecessary treatment for patients. Consequently, it is of great significance to develop a set of specific criteria that may help define the characteristic radiological features for neurovascular contact(s) responsible for a patient’s symptoms.

Previous studies have found that the incidence of vascular compressive trigeminal neuralgia is highly correlated with the degree of neurovascular contact (13). The latter, usually assessed by neurovascular compression (NVC) grading, is often significantly higher on the symptomatic side than on the asymptomatic side (14). However, only about 50% of the patients have serious neurovascular contact on the affected side (13); hence, the identification of responsible contacts by contact degree alone is not accurate enough.

Previous research focuses more on the affected trigeminal nerve of the patients than on the neurovascular contacts. In fact, neurovascular contact is a better choice. When multiple neurovascular contacts are present, it may be difficult to discern which one is responsible for causing symptoms. In addition, the efficacy of the MR features in diagnosing responsible contacts has not been well evaluated, let alone the multivariate analysis based on these features.

Therefore, the aim of the present study was to investigate the role of several MR features in diagnosing responsible neurovascular contacts, evaluate their predictive value, and further develop a model incorporating useful MRI features for an accurate and prospective prediction of responsible contacts.

Material and Methods

Inclusion and exclusion of patients

A total of 75 patients, admitted to Fujian Medical University Union Hospital from July 2013 to May 2018, were diagnosed with unilateral vascular compressive trigeminal neuralgia and enrolled in the study. Their clinical records were retrospectively sorted and reviewed. The inclusion criteria were: (i) no obvious pain alleviation with the medicinal administration such as carbamazepine; (ii) reception of the first MVD surgical treatment; and (iii) the adoption of the same pre-surgery MRTA scanning technique. The exclusion criteria were: (i) trigeminal neuralgia due to intracranial tumor, multiple sclerosis, or other diseases; (ii) incomplete MRTA image or poor MRTA image quality; (iii) no significant symptom improvement after MVD; and (iv) absence of operation records. The study protocol was approved by the Institutional Ethical Committee of Fujian Medical University Union Hospital.

Acquisition and processing of magnetic resonance images

The trigeminal nerve was examined by a GE MR 750 system (GE Healthcare, Waukesha, WI, USA), a 3.0-T superconducting MR scanner, and an eight-channel head and neck combined phased array coil. The scanning sequence included 3D-FIESTA and 3D-TOF-MRA. The parameters of the former were set as follows: TR/TE = 6.7/2.6 ms; field of view (FOV) = 180 × 180 mm; slice thickness = 0.5 mm; slice spacing = 0 mm; matrix = 256 × 256; NEX = 4. The parameters of the latter were as follows: TR/TE = 25/6.8 ms; FOV = 220 × 220 mm; slice thickness = 0.6 mm; slice spacing = 0 mm; matrix = 256 × 256; NEX = 1. The scanning baseline was parallel to the auditory-orbital line and extended from the frontal base to the bridging sulcus, including the trigeminal nerve, facial auditory nerve, and the main branches of the vertebrobasilar artery.

All transverse image data were imported into GE Advanced Workstation (Version 4.6, GE Healthcare, Waukesha, WI, USA), and the 3D-FIESTA and 3D-TOF-MRA sequences were simultaneously opened in the 3D Synchro View for synchronous analysis. The 3D-FIESTA sequence was reconstructed in the oblique sagittal and coronal positions. The neurovascular relationship was observed from the axial, oblique coronal, and sagittal perspectives. Three-dimensional MIP reconstruction of vertebrobasilar artery and their major branches was performed using the 3D-TOF-MRA data.

Inclusion and exclusion of neurovascular contacts

The MRI features of all neurovascular contacts of the bilateral trigeminal nerves in all enrolled patients were analyzed by two senior radiologists, who were blinded to the symptomatic sides. The inclusion and exclusion criteria of the neurovascular contacts were as follows: (i) neurovascular contact was observed in at least one of the transverse, oblique sagittal, or coronal images; (ii) the contacts of the trigeminal nerve on the symptomatic side were all excluded if the contacts were not identified in the surgical record. The included contacts were considered responsible contacts if: (i) the neurovascular contact was located on the symptomatic side; (ii) the neurovascular contacts were clearly identified in the surgical record of MVD; and (iii) the patient experienced pain relief after the MVD surgery, which indicated that the responsible contacts were effectively treated. The rest of the contacts were designated as non-responsible contacts.

Qualitative and quantitative evaluation of contacts

NVC grading

The relationship between blood vessel and trigeminal nerve was analyzed from the transverse, coronal, and sagittal planes according to the evaluation criteria recommended by previous literature: 0 points = no contact; 1 point = nerve and blood vessel contact with no indentation and nerve shift; 2 points = nerve contact with blood vessel with indentation; 3 points = compression of nerves by blood vessels, with nerve displacement or distortion, as shown in Fig. 1. The degree of nerve compression by blood vessels (NVC grading) was evaluated using a 9-point scale, which sums up the NVC points of transverse, oblique coronal, and sagittal planes (14).

Fig. 1.

Examples of neurovascular compression subscores in transverse and sagittal views. 1 point (A1, A2): nerve and blood vessel contact with no indentation and nerve shift; 2 points (B1, B2): nerve contact with blood vessel with indentation; 3 points (C1, C2): compression of nerves by blood vessels, with nerve displacement or distortion. A3, B3, and C3 were the corresponding schematic diagrams in oblique sagittal view for patients A, B, and C, respectively. The trigeminal nerves are colored in yellow and the blood vessels in red.

Distance from pons to contact

The distance from pons to contact (D_pons-contact) was defined as the shortest distance from pons to the neurovascular contact position. In order to measure D_pons-contact, the best plane of the oblique sagittal 3D-FIESTA images parallel to the trigeminal nerve was selected and magnified properly to display the whole cistern segment of the trigeminal nerve. Afterwards, the distance from the neurovascular contact point to the pons was measured in the best plane, as shown in Fig. 2a.

Fig. 2.

Illustrations of evaluation of MRI features of neurovascular contact. (a) D_pons-contact was measured from the neurovascular contact point to the pons in the best plane of the oblique sagittal 3D-FIESTA images parallel to the trigeminal nerve. (b) D_vessel and D_{trigeminal nerve} were measured perpendicular to the long axis of the trigeminal nerve and blood vessels. (c) An example of multiple neurovascular contacts on the same side. (d, e) Two cases demonstrating the arterial and venous compressing vessel in vascular compressive trigeminal neuralgia. Arrows in the figure indicate a compressive artery with high intensity on 3D-TOF-MRA (in d) or a vein with low intensity on 3D-TOF-MRA (in e) next to the trigeminal nerve.

Fig. 3.

Screening process for enrolled patients and contacts.

Fig. 4.

ROC curves of the MRI features in diagnosing responsible contacts. The AUC of the NVC grading, D_pons-contact, and logistic regression predictive model for diagnosing responsible contacts were 0.742, 0.656, and 0.808, respectively. AUC, area under the ROC curve; D_pons-contact: distance from pons to contact; MRI, magnetic resonance imaging; NVC, neurovascular compression; ROC, receiver operating characteristic.

Diameter of vessel and trigeminal nerve at contacts

The diameter of vessel (D_vessel) and diameter of trigeminal nerve (D_{trigeminal nerve}) were measured on a GE Advanced workstation. The neurovascular contact was targeted and magnified properly in the transverse, oblique sagittal, and coronal positions. The diameters at contact were measured perpendicular to the long axis of the trigeminal nerve and blood vessels with the measurement tools provided by the workstation, as shown in Fig. 2b.

Tracing the vascular origin of the compressing vessels

The vascular origin of the compressing vessel was determined according to the vascular intensity on 3D-TOF-MRA. A high intensity on 3D-TOF-MRA indicates the vessel is of arterial origin; otherwise, the vessel is of venous origin. Therefore, the vascular origin was divided into arterial compression (with a high intensity on 3D-TOF-MRA) and venous compression (without a high intensity on 3D-TOF-MRA). Two cases demonstrating the determination of the vascular origin were illustrated in Fig. 2d and 2e.

Statistical analysis

SPSS software (Version 18.0.0, IBM, Armonk, NY, USA) was used for statistical analysis. A chi-square test was used to compare the statistical differences of vascular origin between the responsible and non-responsible contacts; rank sum tests were used to compare the statistical differences of other MR features between the responsible and non-responsible contacts. The logistic regression analysis was adopted to establish a predictive model by incorporating the significant MRI features. A receiver operating characteristic (ROC) curve was further used to evaluate the diagnostic efficiency of these MR features and logistic regression predictive model for responsible contacts. The statistical significance was set at P < 0.05.

Results

Descriptive statistics of patients and contacts

The study retrospectively analyzed 75 patients with unilateral vascular compressive trigeminal neuralgia. One case was excluded due to cerebellopontine angle tumor, three cases due to incomplete MRTA data, and four cases for lack of surgical records, leaving 67 eligible cases. A total of 201 contacts from all the enrolled patients were identified and 48 points were excluded due to unclear surgical records, and 153 qualified contacts were obtained for the analysis, which were further classified into responsible contacts (n = 45) and non-responsible contacts (n = 108). The process of inclusion and exclusion was shown in Fig. 3.

Descriptive statistics of basic characteristics of patients and contacts are shown in Table 1. As indicated in Table 1, all enrolled patients had multiple neurovascular contacts, with a high incidence rate on the symptomatic side (up to 43.28%), indicating the difficulty in diagnosing the responsible contacts and the urgency and significance of identifying the MRI characteristics of responsible contacts. An example of multiple contacts on the same side is shown in Fig. 2c and 2d.

Table 1.

Descriptive statistics of enrolled patients and neurovascular contacts.

Feature
The feature of enrolled patients
Age (years)	61.52 ± 11.78
Sex
Male	31/67 (46.27)
Female	36/67 (53.73)
The ratio of patients with multiple contacts	67/67 (100.00)
The ratio of patients with multiple contacts on the symptomatic side	29/67 (43.28)
The feature of enrolled contacts
Contact on the right side	82/153 (53.59)
Contact on the symptomatic side	55/153 (35.95)

Values are given as n (%) or mean ± SD.

Table 2.

Comparison of TZ contact, NVC grading, and vascular origin between the responsible and non-responsible contacts.

MRI feature	Responsible contacts (n = 45)	Non-responsible contacts (n = 108)	Statistics	P*
NVC grading	6.00 (5.00–8.00)	5.00 (4.00–5.75)	Z = –4.817	<0.001
D_pons-contact (mm)	3.10 (2.85–3.65)	4.55 (2.60–6.58)	Z = –3.033	0.002
Vascular origin (artery)	38 (84.44)	73 (67.59)	X² = 4.530	0.033
D_vessel (mm)	1.50 (1.30–1.80)	1.40 (1.20–1.60)	Z = –1.828	0.068
D_{trigeminal nerve} (mm)	1.80 (1.40–2.00)	1.80 (1.60–2.10)	Z = –0.866	0.386

*P < 0.05 is statistically significant.

NVC, neurovascular compression.

Table 3.

Logistic regression analysis of indicators of responsible contacts judgment.

Independent indicators	β	OR (95% CI)	P
NVC grading	0.549	1.732 (1.333–2.249)	<0.001
D_pons-contact	–0.316	0.729 (0.595–0.892)	0.002
Vascular origin (vein as reference)	–0.669	0.512 (0.188–1.393)	0.190

*P < 0.05 is statistically significant.

CI, confidence interval; NVC, neurovascular compression; OR, odds ratio.

Table 4.

Diagnostic statistics of the MRI features in diagnosing the responsible contacts.

Feature	AUC	Threshold value	Sensitivity (%)	Specificity (%)	PLR	NLR
NVC grading	0.742	5	64.44	75.00	2.58	0.47
D_pons-contact (mm)	0.656	4.2	97.78	53.70	2.11	0.041
D_vessel (mm)	0.594	1.6	40.00	79.63	1.96	0.75
D_{trigeminal nerve} (mm)	0.544	1.4	28.89	81.48	1.56	0.87
Logistic regression predictive model	0.808	0.233	88.89	62.04	2.34	0.18

AUC, area under the receiver operating characteristic curve; MRI, magnetic resonance imaging; NLR, negative likelihood ratio; NVC: neurovascular compression; PLR, positive likelihood ratio.

Tracing the origin of blood vessels

The vascular origin of each neurovascular contact was traced on 3D-TOF-MRA. Of the responsible vessels, 38 of 45 (84.44%) were arteries and 7 of 45 (15.56%) were veins. The superior cerebellar artery (SCA) and anterior inferior cerebellar artery (AICA) were the two most common responsible sources of neurovascular contacts, accounting for 53.33% (24/45) and 26.67% (12/45), respectively; in contrast, the basilar artery (BA) and posterior inferior cerebellar artery (PICA) rarely compressed the trigeminal nerve, both accounting for 2.22% (1/45).

Comparison of MRI characteristics between responsible and non-responsible contacts

The comparison of MRI characteristics revealed that NVC grading of the responsible contacts were significantly higher than those of the non-responsible contacts (P < 0.001), while the D_pons-contact in the responsible contacts was significantly lower than that of the non-responsible contacts (P = 0.002). Compared with the non-responsible contacts, the incidence of arterial vascular compression in the responsible contacts was also significantly higher (84.44% vs. 67.59%, P = 0.0175). In contrast, no significant differences in diameters of both D_vessel (P = 0.068) and D_{trigeminal nerve} (P = 0.386) at contact position were found between responsible and non-responsible contacts (Table 2).

Establishment of the predictive model for responsible contacts by multivariate logistic regression

Multivariate logistic regression analysis was conducted with D_pons-contact, NVC grading, and vascular origin as independent variables and responsible contacts as dependent variable (Table 3).

The logistic regression predictive model was formulated as the following equation: ln((p(y = 1)/(1–p(y = 1))) = −2.565–0.316*X₁ + 0.549*X₂ – 0.669*X₃, where X₁ represents the D_pons-contact, X₂ represents the NVC grading, X₃ the vascular origin, and P value prediction probability; the closer P value approximates to 1, the more likely it is a responsible contact. Both D_pons-contact (odds ratio [OR] = 0.729, P = 0.002) and NVC grading (OR = 1.732, P < 0.001) are independent predictors for responsible contacts, while vascular origin (OR = 0.512, P = 0.190) is not an independent predictor.

ROC analysis

The ROC curve was used to evaluate the diagnostic efficiency of NVC grading, D_pons-contact, D_vessel, D_{trigeminal nerve}, and the established logistic predictive model in distinguishing responsible and non-responsible contacts (Fig. 4 and Table 4). The analysis showed that NVC grading had the highest diagnostic efficiency among the three MRI features, with an area under the ROC curve (AUC) of 0.742, sensitivity of 64.44%, and specificity of 75.00%; D_pons-contact came second, with an AUC of 0.656, sensitivity of 97.78%, and specificity of 53.70%. In contrast, both D_vessel and D_{trigeminal nerve} had a lower diagnostic efficiency with AUCs of 0.594 and 0.544, sensitivity of 40.00% and 28.89%, and specificity of 79.63% and 81.48%, respectively.

As for the diagnostic efficiency of the predictive model, the AUC of the model in judging responsible contacts was 0.808 (95% confidence interval [CI] = 0.735–0.882). With the prediction probability of 0.233 as the optimal cut-off value, the sensitivity and specificity of the model in judging responsible contacts were 88.89% and 62.04%, respectively. The AUC of the predictive model was significantly higher than that of NVC grading, D_pons-contact, D_vessel, and D_{trigeminal nerve}, with a P value of 0.007, <0.001, <0.001, and <0.001, respectively.

Discussion

In clinical settings, radiologists are often pressed to make an accurate diagnosis for vascular compressive trigeminal neuralgia. According to the available literature, neurovascular contact appears in the affected side of about 56%–95% of patients with primary trigeminal neuralgia, in the asymptomatic side of about 10%–71% of patients with primary trigeminal neuralgia, and may even occur in individuals without symptoms (9,21,22). The present study also found a high incidence of multiple neurovascular contacts in all the 67 enrolled patients and on the symptomatic side (43.28%). All these conditions make it difficult for radiologists to accurately identify the responsible neurovascular contacts.

In contrast to prior research, this study focuses mainly on the neurovascular contacts instead of the individual patients, because this will help to eliminate the effect of multiple neurovascular contacts (an additional variable) on the results of the study. By comparing the MRI features between responsible and non-responsible contacts, we found that NVC grading, D_pons-contact, and vascular origin were of value in the diagnosis of responsible contacts. Specifically, significant differences in NVC grading and D_pons-contact were found between responsible contacts and non-responsible contacts; in addition, the arterial compression were more significantly involved in the responsible contacts than in the non-responsible contacts (84.44% vs. 67.59%). In contrast, no significant differences were found in the diameters of both D_vessel and D_{trigeminal nerve} at the contact position between responsible and non-responsible contacts. Further ROC curve analysis showed that NVC grading and D_pons-contact had a relatively higher diagnostic efficiency in diagnosing the responsible contacts, while the D_vessel and D_{trigeminal nerve} had a relatively lower diagnostic efficiency. These results signify that the contact degree and the contact position are two important diagnostic candidates in the diagnosis of responsible contacts.

Previous studies have found that the incidence of vascular compressive trigeminal neuralgia is highly correlated with the degree of neurovascular contact (13), which is usually assessed by NVC grading. NVC grading is often significantly higher on the symptomatic side than on the non-symptomatic side (14). Consistently, this study showed that NVC grading in the responsible contacts was significantly higher than that of the non-responsible contacts.

The contact position is obviously another important feature to pinpoint the responsible contacts. Previous studies have found that trigeminal neuralgia is closely associated with the compression sites of blood vessels (15) and can be easily induced by demyelination (12,16 –18) when the root entry zone or proximal trigeminal nerve (D_pons-contact < 2.5 mm or <3 mm) is compressed by blood vessels (16,19,20). However, such grouping for contact position is far from satisfactory. The current study used D_pons-contact as a quantitative feature to diagnose responsible contacts and found that the cut-off value of 4.2 mm produced optimal sensitivity and specificity. This distance is well consistent with the transition zone (TZ) of the trigeminal nerve, which refers to an area where the oligodendrocyte-composed central nerve myelin sheath meets the Schwann cell-composed peripheral nerve myelin sheath of the cranial nerve. The transition region is extremely sensitive to mechanical traction stimulation and is prone to the abnormal discharge of the nerve due to demyelination caused by vascular compression and arterial pulsation, thus causing corresponding symptoms (1,16,23). Until recently, the notion of TZ in trigeminal nerve is proposed in the literature, and according to Guclu et al. (23), the TZ of trigeminal nerve is a segment of about 2 mm in length and approximately 4.2 mm away from the brain stem. Therefore, if the shortest distance from pons to the neurovascular contact position falls within the range (2.2–4.2 mm), the transition region may have a greater chance of involvement. Accordingly, when the TZ of the trigeminal nerve is involved, responsible contact should be considered as a candidate cause for trigeminal neuralgia.

The origins of the responsible vessels were traced in this study. Arteries were the main cause for vascular compressive trigeminal neuralgia, accounting for 84.44%, and SCA and AICA were the two most common responsible arteries, accounting for 53.33% and 26.67%, respectively. Studies (1) have shown that vascular compressive trigeminal neuralgia is usually caused by a neighboring elongated SCA coming from above or by an AICA coming from below, with the SCA being more common (88% alone or in association) than the AICA (25%). Although the proportion of SCA was relatively low in this study, we consider that the results of this study are essentially consistent with the findings of previous studies.

In addition, NVC grading, D_pons-contact, and vascular origin were incorporated in the logistic regression analysis to establish a predictive model for identifying the responsible contacts. The results showed that the diagnostic efficiency of the regression model was significantly higher than that of single factor analysis, with an optimized sensitivity and specificity and an AUC reaching 0.808, significantly higher than that of any other single factor. Therefore, the predictive model signifies that in clinical practice, NVC grading, D_pons-contact, and vascular origin should be closely combined to provide a more reliable identification of the responsible contacts for radiologists. The predictive model in the present study has made a significant improvement in diagnosing responsible contacts, and, in particular, its relatively high sensitivity (88.89%) indicates that the predictive model can effectively reduce the missed diagnoses. However, the specificity of the model was relatively low (62.04%), which may be due to the low diagnostic specificity of the D_pons-contact (53.70%). In the current study, the threshold value of the D_pons-contact for diagnosing vascular compressive trigeminal neuralgia was 4.2 mm, meaning that, theoretically, vascular compressive trigeminal neuralgia tends to occur if the D_pons-contact is <4.2 mm. In fact, vascular compressive trigeminal neuralgia easily occurs when the D_pons-contact falls within the range of about 2.2–4.2 mm, rather than 0–2.2 mm, indicating that the transition region is probably involved by the vessel, which may explain the relatively low specificity of the D_pons-contact. To the best of our knowledge, most studies used diagnostic parameters alone rather than combined parameters to build a model to predict the symptomatic side of vascular compressive trigeminal neuralgia. A prediction model for the symptomatic side of trigeminal neuralgia by Blitz et al. (24) used contrast-enhanced CISS imaging to evaluate the neurovascular compression in trigeminal neuralgia, reporting a sensitivity of 87.7% and specificity of 35.1% for patients with Neurovascular Conflict Grades 1–3 and a sensitivity of 51.9% and specificity of 94.6% for patients with Neurovascular Conflict Grades 2 and 3 (24). By comparison, an improvement has been achieved in the prediction model proposed in the present study and we consider that the model’s specificity can be further improved if a larger sample size and more diagnostic parameters are included in the predictive model.

The present study has some limitations. First, the study targets at a small sample size. Many cases were excluded from the study due to inconsistent scanning conditions or unreliable MVD surgical records, resulting in a small sample size. Therefore, future design may enroll a large sample pool to verify the role of MRI characteristics in identifying the responsible contacts. Second, the trigeminal nerves and vessels are very small in size, but the acquisition intraplane resolution was limited to about 0.7 mm, which means that some of the target vessel/nerve included just 1–2 pixels. Therefore, errors may be present in measuring the D_{trigeminal nerve} and D_vessel. Meanwhile, D_pons-contact was defined on the subjective best plane parallel to the trigeminal nerve. It would affect the results without anatomical index as a reference. Finally, the parameters used in the present study were possible parameters of responsible contacts that are defined on the basis of previous literature rather than on the assessment standard for culprit vessel compression. Therefore, some other parameters may not be considered.

In conclusion, contact degree and position can serve as important MRI features in identifying the responsible contacts of the trigeminal neuralgia. The logistic predictive model based on D_pons-contact, NVC grading, and vascular origin can qualitatively improve the prediction of responsible contacts for radiologists and can act as a promising tool for an accurate diagnosis of responsible contacts.

Footnotes

Acknowledgements

The authors thank Professor HongZhi Huang for proofreading the manuscript.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Rifeng Jiang

References

Haller

Etienne

Ko vari

, et al. Imaging of neurovascular compression syndromes: trigeminal neuralgia, hemifacial spasm, vestibular paroxysmia, and glossopharyngeal neuralgia. AJN Am J Neuroradiol 2016; 37:1384–1392.

Cruccu

Finnerup

Jensen

, et al. Trigeminal neuralgia: New classification and diagnostic grading for practice and research. Neurology 2016; 87:220–228.

Gambeta

Chichorro

Zamponi

GW.

Trigeminal neuralgia: An overview from pathophysiology to pharmacological treatments. Mol Pain 2020; 16:1744806920901890.

Bick

Huie

Sneh

, et al. Older patients have better pain outcomes following microvascular decompression for trigeminal neuralgia. Neurosurgery 2019; 84:116–122.

Bendtsen

Zakrzewska

Abbott

, et al. European Academy of Neurology guideline on trigeminal neuralgia. Eur J Neurol 2019; 26:831–849.

Tuleasca

Regis

Sahgal

, et al. Stereotactic radiosurgery for trigeminal neuralgia: a systematic review. J Neurosurg 2018; 130:733–757.

Wolf

Kondziolka

Gamma knife surgery in trigeminal neuralgia. Neurosurg Clin N Am 2016; 27:297–304.

Fariselli

Ghielmetti

Bosetti

, et al. The role of radiosurgery in trigeminal neuralgia. Neurol Sci 2017; 38:63–65.

Brînzeu

Drogba

Sindou

Reliability of MRI for predicting characteristics of neurovascular conflicts in trigeminal neuralgia: implications for surgical decision making. J Neurosurg 2018; 1:1–11.

10.

Tanrikulu

Scholz

Nikoubashman

, et al. Preoperative MRI in neurovascular compression syndromes and its role for microsurgical considerations. Clin Neurol Neurosurg 2015; 129:17–20.

11.

Ruiz-Juretschke

Guzmán-de-Villoria

García-Leal

, et al. Predictive value of magnetic resonance for identifying neurovascular compressions in trigeminal neuralgia. Neurologia 2019; 34:510–519.

12.

Akimoto

Nagaoka

Nariai

, et al. Preoperative evaluation of neurovascular compression in patients with trigeminal neuralgia by use of three-dimensional reconstruction from two types of high-resolution magnetic resonance imaging. Neurosurgery 2002; 51:956–961; discussion 961–952.

13.

Maarbjerg

Wolfram

Gozalov

, et al. Significance of neurovascular contact in classical trigeminal neuralgia. Brain 2015; 138:311–319.

14.

Chen

Zhang

Guo

, et al. Preoperative evaluation of the neurovascular compression using magnetic resonance tomographic angiography: our radiologic indications for microvascular decompression to treat trigeminal neuralgia. J Craniofac Surg 2014; 25:e384–388.

15.

Hao

Zhang

Chen

, et al. Sensitivity of magnetic resonance tomographic angiography for detecting the degree of neurovascular compression in trigeminal neuralgia. Neurol Sci 2020; 41:2947–2951.

16.

Peker

Kurtkaya

Uzun

, et al. Microanatomy of the central myelin-peripheral myelin transition zone of the trigeminal nerve. Neurosurgery 2006; 59:354–359; discussion 354–359.

17.

Devor

Govrin-Lippmann

Rappaport

ZH.

Mechanism of trigeminal neuralgia: an ultrastructural analysis of trigeminal root specimens obtained during microvascular decompression surgery. J Neurosurg 2002; 96:532–543.

18.

Yang

Shen

Xia

, et al. Preoperative evaluation of neurovascular relationship in trigeminal neuralgia by three-dimensional fast low angle shot (3D-FLASH) and three-dimensional constructive interference in steady-state (3D-CISS) MRI sequence. Br J Radiol 2018; 91:20170557.

19.

Suzuki

Yoshino

Shimada

, et al. Trigeminal neuralgia: differences in magnetic resonance imaging characteristics of neurovascular compression between symptomatic and asymptomatic nerves. Oral Surg Oral Med Oral Pathol Oral Radiol 2015; 119:113–118.

20.

De Ridder

Moller

Verlooy

, et al. Is the root entry/exit zone important in microvascular compression syndromes? Neurosurgery 2002; 51:427–433; discussion 433–424.

21.

Hughes

Frederickson

Branstetter

, et al. MRI of the trigeminal nerve in patients with trigeminal neuralgia secondary to vascular compression. AJR Am J Roentgenol 2016; 206:595–600.

22.

Kuncz

Vörös

Barzó

, et al. Comparison of clinical symptoms and magnetic resonance angiographic (MRA) results in patients with trigeminal neuralgia and persistent idiopathic facial pain. Medium-term outcome after microvascular decompression of cases with positive MRA findings. Cephalalgia 2016; 26:266–276.

23.

Guclu

Sindou

Meyronet

, et al. Cranial nerve vascular compression syndromes of the trigeminal, facial and vago-glossopharyngeal nerves: comparative anatomical study of the central myelin portion and transitional zone; correlations with incidences of corresponding hyperactive dysfunctional syndromes. Acta Neurochirurgica 2011; 153:2365–2375.

24.

Blitz

Northcutt

Shin

, et al. Contrast-enhanced CISS imaging for evaluation of neurovascular compression in trigeminal neuralgia: improved correlation with symptoms and prediction of surgical outcomes. AJN Am J Neuroradiol 2018; 39:1724–1732.