ETM Experimental and Therapeutic Medicine 1792-0981 1792-1015 D.A. Spandidos ETM-25-4-11870 10.3892/etm.2023.11870 Articles 3D multimodal image fusion based on MRI in the preoperative evaluation of microvascular decompression: A meta‑analysis LiangChen 1 2 YangLing 3 ZhangBinbin 1 LiRuichun 1 GuoShiwen 1 Department of Neurosurgery, The First Affiliated Hospital of Xi'an Jiaotong University, Xi'an, Shaanxi 710061, P.R. China Department of Diagnostic and Interventional Radiology, Division of Medical Physics, University Medical Center Freiburg, Faculty of Medicine, University of Freiburg, D-79106 Freiburg im Breisgau, Germany Department of Aviation Psychology Research, Xi'an Civil Aviation Hospital, Xi'an, Shaanxi 710082, P.R. China Correspondence to: Professor Shiwen Guo, Department of Neurosurgery, The First Affiliated Hospital of Xi'an Jiaotong University, 277 Yanta West Road, Xi'an, Shaanxi 710061, P.R. China guoshiwen1962@126.com jwxu1001@163.com 04 2023 06 03 2023 25 4 171 28 09 2022 14 02 2023 Copyright: © Liang et al. 2020 This is an open access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made.

Neurovascular compression (NVC) is the main cause of hemifacial spasm (HFS) or trigeminal neuralgia (TN), and frequently occurs at the root entry zone of cranial nerves. Microvascular decompression (MVD) is an effective surgical treatment for TN and HFS caused by NVC. The accurate preoperative diagnosis of NVC is crucial to the evaluation of MVD as an appropriate treatment for TN and HFS. Three-dimensional (3D) time-of-flight magnetic resonance angiography (3D TOF MRA) and high resolution T2-weighted imaging (HR T2WI) are used to detect NVC prior to MVD; however, this combination alone has certain disadvantages. Multimodal image fusion (MIF) may combine two or more images from the same or different modalities, allowing neurosurgeons to use the reconstructed 3D model to observe anatomical details more clearly from different perspectives. The aim of the present meta-analysis was to evaluate the effect of 3D MIF based on 3D TOF MRA combined with HR T2WI in the preoperative diagnosis of NVC, and thus to evaluate its clinical application value in the preoperative evaluation of MVD. Relevant studies available on PubMed, Embase, Web of Science, Scopus, China National Knowledge Infrastructure and the Cochrane Library, and published from the inception of each database to September 2022, were retrieved. Studies using 3D MIF based on 3D TOF MRA combined with HR T2WI to diagnose NVC in patients with TN or HFS were included. The Quality Assessment of Diagnostic Accuracy Studies checklist was used to evaluate the quality of the included studies. The statistical software Stata 16.0 was used to perform the meta-analysis. Data extraction was performed by two independent investigators and discrepancies were resolved by discussion. Pooled sensitivities, specificities, positive likelihood ratio (PLR), negative likelihood ratio (NLR), diagnostic odds ratio (DOR) and the area under the receiver operating characteristic curve (AUROC) were calculated as the main summary effect size. The I² and Q-test were used to assess heterogeneity. The present search identified 702 articles, of which 7 (comprising 390 patients) fulfilled the inclusion criteria. Bivariate analysis indicated that the pooled sensitivity and specificity of 3D MIF based on 3D TOF MRA combined with HR T2WI for detecting NVC were 0.97 (95% CI, 0.95-0.99) and 0.89 (95% CI, 0.77-0.95), respectively. The pooled PLR was 8.8 (95% CI, 4.1-18.6), the pooled NLR was 0.03 (95% CI, 0.02-0.06) and the pooled DOR was 291 (95% CI, 99-853). The AUROC was 0.98 (95% CI, 0.97-0.99). The studies had no substantial heterogeneity (I2=0; Q=0.000; P=0.50). The present results suggested that 3D MIF based on 3D TOF MRA combined with HR T2WI had excellent sensitivity and specificity for diagnosing NVC in patients with TN or HFS. Therefore, this method should serve a key role in MVD preoperative evaluation.

 3D multimodal image fusion magnetic resonance imaging 3D time-of-flight magnetic resonance angiography high resolution T2-weighted imaging neurovascular compression microvascular decompression meta-analysis Funding: The present study was supported by a grant from the Key Research and Development Plan of Shaanxi Province, China (grant no. 2021SF-298). Introduction

Neurovascular compression (NVC) is the main cause of primary trigeminal neuralgia (TN) or hemifacial spasm (HFS), and frequently occurs at the root entry zone (REZ) of cranial nerves (1). Microvascular decompression (MVD) is an effective surgical treatment for TN and HFS caused by NVC (2,3). However, preoperative detection of NVC is occasionally difficult. Although MVD is not challenging for skilled and experienced neurosurgeons, it is not effective for patients with TN or HFS not caused by NVC (4). Therefore, an accurate preoperative diagnosis of NVC is crucial in deciding whether to perform MVD.

Magnetic resonance imaging (MRI) has been used to detect NVC prior to MVD for numerous years; however, it does have certain disadvantages. Routine MRI sequences cannot clearly and accurately display the relationship between nerves and blood vessels at the REZ (4). Since the 1990s, 3D time-of-flight magnetic resonance angiography (3D TOF MRA) has gradually become a common MRI sequence for detecting NVC. 3D TOF MRA is able to selectively image fast-flowing blood and clearly display nerves and blood vessels (5). However, the 3D TOF MRA sequence demonstrates a poor ability to visualize blood vessels with a slow blood flow, such as veins and arterioles (6). With the development of MRI technology, the appearance of high resolution T2-weighted imaging (HR T2WI) brings new options for NVC detection. HR T2WI may adequately demonstrate the anatomical structure in the cerebellopontine angle (CPA) against the background of cerebrospinal fluid (CSF) signal and aid in justifying a diagnosis of NVC, which may be obtained by different technologies, including 3D balanced steady state gradient echo and 3D fast turbo spin echo (5). Different manufacturers use different names for these technologies, such as constructive interference steady state (CISS), fast imaging employing steady-state acquisition (FIESTA), balanced fast field echo (bFFE), sampling perfection with application-optimized contrasts using different flip angle evolutions (SPACE), balanced steady-state free precession and turbo spin echo driven equilibrium. However, it is difficult for HR T2WI to distinguish the relationship between nerves and blood vessels when they are in close contact or there is a lack of CSF signal contrast around them. In that case, another MRI sequence is needed to assist with the diagnosis (6). Due to their advantages and disadvantages, they are frequently used together to detect NVC. Although the combination of the two sequences may make up for certain shortcomings, both display two-dimensional images, and the spatial structure of CPA cannot be displayed intuitively and dynamically. Therefore, it is necessary to identify more accurate imaging or post-processing technology to improve imaging quality.

Multimodal image fusion (MIF) may combine two or more images from the same or different modalities and reconstruct a 3D model, which allows the operator to observe anatomical details more clearly from different angles (7). Therefore, the MIF technique combined with these MRI sequences is considered to be an accurate method for the preoperative diagnosis of NVC. 3D MIF based on 3D TOF MRA combined with HR T2WI may accurately display the precise anatomical structures at the REZ and indicate the relationship between cranial nerves and blood vessels (8,9). However, there remains a lack of large-scale clinical trials to analyze its clinical application value in the preoperative diagnosis of NVC.

The present meta-analysis was designed to evaluate the value of 3D MIF based on 3D TOF MRA combined with HR T2WI in the preoperative judgment of NVC in patients with TN and HFS, and thus to evaluate its clinical application value in the preoperative evaluation of MVD.

 Materials and methods <sec> <title>Manuscript preparation

The whole study was conducted according to the PRISMA 2020 statement (10) and the manuscript was prepared and revised according to the PRISMA 2020 Checklist.

 Search strategy and selection criteria

PubMed (https://pubmed.ncbi.nlm.nih.gov/), Embase (https://www.embase.com), Web of Science (https://www.webofscience.com), Scopus (https://www.scopus.com), China National Knowledge Infrastructure (https://www.cnki.net) and the Cochrane Library (https://www.cochranelibrary.com/) were systematically searched. The medical subject heading terms or Emtree terms were ‘Magnetic Resonance Angiography’ and ‘Microvascular Decompression Surgery’. The search query was ‘((Microvascular Decompression Surgery) OR (Decompression Surgeries, Microvascular) OR (Decompression Surgery, Microvascular) OR (Microvascular Decompression Surgeries) OR (Surgeries, Microvascular Decompression) OR (Surgery, Microvascular Decompression) OR (Microvascular Decompression) OR (Decompression, Microvascular) OR (Decompressions, Microvascular) OR (Microvascular Decompressions)) AND ((Magnetic Resonance Angiography) OR (MRI Angiography) OR (Angiographies, MRI) OR (Angiography, MRI) OR (MRI Angiographies) OR (MRI Angiographies) OR (Angiographies, Magnetic Resonance) OR (Magnetic Resonance Angiographies) OR (Perfusion Magnetic Resonance Imaging) OR (Perfusion Weighted MRI) OR (MRI, Perfusion Weighted) OR (time-of-flight))’. Articles published from the inception of each database to September 2022 were retrieved.

After deleting duplicate publications, reasonable inclusion and exclusion criteria were developed to review the remaining studies. The inclusion criteria were as follows: i) The study used 3D MIF technology based on 3D TOF MRA combined with HR T2WI to judge NVC in patients with TN or HFS; ii) the study design was prospective or retrospective; and iii) the intraoperative findings were used as the reference standard for NVC diagnosis. The exclusion criteria were as follows: i) Reviews, case reports, editorials, correspondences, comments or meeting abstracts/meeting minutes; ii) 3D MIF technology was not used in the study, or 3D TOF MRA combined with HR T2WI was not used to fuse the multimodal image; iii) the patients in the study received only preoperative evaluation but not MVD; and iv) studies without sufficient data to construct the 2x2 contingency table.

 Data extraction

A total of two investigators (BZ and RL) independently extracted the data, including the quality assessment of the retrieved studies. Discrepancies were resolved in a consensus meeting, or if no agreement could be reached, such as regarding the methodological quality of the included studies, they were resolved by referral to a third investigator (SG). The extracted data included the following: i) Basic research information, including the name of the first author, publication year, country of the first author, sample size and type of research design; ii) the characteristics of the participants, including the age of the patients and their diagnosis; iii) the MRI sequences used in the study; iv) the 3D fusion software used in the study; and v) the research results, including the number of true positives, false positives, false negatives, true negatives, sensitivity and specificity.

 Literature quality assessment

The methodological quality of the included studies was assessed by two researchers on the basis of the Quality Assessment of Diagnostic Accuracy Studies (QUADAS) checklist (11). Review Manager 5.4.1 software (The Cochrane Collaboration) was used to generate a methodological quality graph and methodological quality graph summary.

 Statistical analysis

An exact binomial rendition of the bivariate mixed-effects regression model was used to synthesize data. The data analysis was performed using the meta-analytical integration of diagnostic test accuracy studies (MIDAS) module of Stata 16.0 software (StataCorp LP). The pooled sensitivity, specificity, positive likelihood ratio (PLR), negative likelihood ratio (NLR), diagnostic odds ratio (DOR), area under the receiver operating characteristic curve (AUROC) and 95% CIs were calculated. The I² and Q-test were used to assess heterogeneity. The publication bias of the included literature was examined by Deeks' test using Stata 16.0. The clinical application value of 3D MIF based on 3D TOF MRA combined with HR T2WI for the diagnosis of NVC was evaluated by Fagan nomogram and likelihood ratio (LR) scatter plot. Graphs were produced using the MIDAS module for Stata 16.0 and Review Manager 5.4.1 software.

 Results <sec> <title>Included articles

In the search, 702 articles were identified, with 435 articles remaining after discarding duplicate records. A full-text analysis was performed on the 56 articles that remained following screening by titles and abstracts, of which 6 articles did not have complete data and 43 articles were not related to 3D MIF or 3D TOF MRA combined with HR T2WI, and were therefore excluded. Finally, 7 articles were included in the present analysis (Fig. 1).

 Basic characteristics of the included studies

The basic characteristics of the studies included in the present meta-analysis are presented in Table I. A total of 390 patients were included in the seven studies. Among them, three studies (12-14) focused on TN, one (8) on HFS and three (9,15,16) on both. A total of two (8,12) of the seven studies performed 3D TOF MRA combined with the 3D FIESTA sequence, two (9,15) performed 3D TOF MRA combined with the 3D SPACE sequence, one (13) performed 3D TOF MRA combined with bFFE sequence, one (14) performed 3D TOF MRA combined with bFFE and 3D Gd-enhanced spoiled gradient recalled sequence, and one (16) performed 3D TOF MRA combined with the 3D CISS sequence. A total of three (8,9,12) of the seven studies used 3D-slicer software (17) for 3D MIF, two (13,14) used OsiriX 2.5.1 software (Pixmeo SARL), and the remaining two used iPlan Net (Brainlab AG) (15) software and Leonardo™ (Siemens AG) software (16), respectively. A total of four studies (9,14-16) were performed as prospective studies and the remaining three as retrospective studies. According to the methodological quality graph (Fig. 2A) and the methodological quality summary (Fig. 2B), the quality of the literature included in the present study was acceptable.

 Meta-analysis results. Heterogeneity of the meta-analysis

The I2 and Q-test were used to assess the heterogeneity of the studies. The results indicated that the studies in the present meta-analysis had no substantial heterogeneity (I2=0; Q=0.000; P=0.50). The same result was observed from the Galbraith plot (Fig. 3A).

 Summary effect size

Bivariate analysis yielded that the pooled sensitivity and specificity of 3D MIF based on 3D TOF MRA combined with HR T2WI for detecting NVC were 0.97 (95% CI, 0.95-0.99) and 0.89 (95% CI, 0.77-0.95), respectively (Fig. 4A). The pooled PLR was 8.8 (95% CI, 4.1-18.6), the pooled NLR was 0.03 (95% CI, 0.02-0.06; Fig. 4B) and the pooled DOR was 291 (95% CI, 99-853; Fig. 4C). The AUROC was 0.98 (95% CI, 0.97-0.99; Fig. 4D).

 Publication bias

The Deeks' funnel plot asymmetry test was used to evaluate the publication bias of the studies in the current meta-analysis, with P=0.59 indicating no significant publication bias in the included studies (Fig. 3B).

 Sensitivity analysis

Sensitivity analysis was carried out by deleting each included study one by one and calculating the respective summary effect size. The pooled sensitivity and AUROC were used as evaluation indicators. The results suggested that excluding any study did not significantly change the pooled sensitivity and AUROC (Table II), indicating that the results of the present meta-analysis were robust and reliable.

 Evaluation of clinical application value

The LR scatter plot based on summary PLR and NLR was in the lower left quadrant (Fig. 5A). According to the data of the present meta-analysis, the incidence rate of NVC in patients with TN and HFS was estimated to be 86.15%, which was consistent with the reported results (18). Therefore, with a pretest probability of 0.86, the Fagan nomogram indicated that the positive post-test probability was 98% and the negative post-test probability 16% (Fig. 5B).

 Discussion

In the present meta-analysis, the sensitivity, specificity, PLR, NLR, DOR and AUROC of 3D MIF based on 3D TOF MRA combined with HR T2WI for detecting NVC were examined. The results demonstrated a high diagnostic ability of this technology with high sensitivity and specificity. As mentioned above, 3D TOF MRA is a widely used MRI sequence for detecting NVC, which is frequently used as a standard MRI-based method to compare with other methods. In a previous meta-analysis, the authors reported that 3D TOF MRA exhibited a sensitivity of 0.95 (95% CI, 0.93-0.96) and specificity of 0.77 (95% CI, 0.66-0.86) in correctly identifying NVC in patients with TN (19). According to the present results, when using 3D MIF technology, the sensitivity and specificity in diagnosing NVC was able to be improved. In a previous report, the DOR and AUROC of 3D TOF MRA in diagnosing NVC were 52.92 (95% CI, 26.39-106.11) and 0.97 (95% CI, 0.95-0.99), respectively (19). Compared with 3D TOF MRA, the DOR of 3D MIF technology based on 3D TOF MRA combined with HR T2WI reached 291 (95% CI, 99-853) and the AUROC was 0.98 (95% CI, 0.97-0.99). These results suggested that compared with 3D TOF MRA, 3D MIF technology based on 3D TOF MRA combined with HR T2WI appears to have more advantages in diagnosing NVC in terms of diagnostic accuracy. In addition, whether 3D MIF is able to actually improve the effectiveness of 3D TOF MRA combined with HR T2 in the diagnosis of NVC is also of great concern. In another meta-analysis, the sensitivity and specificity of 3D TOF MRA combined with HR T2WI in diagnosing NVC were 0.96 (95% CI, 0.92-0.98) and 0.92 (95% CI, 0.74-0.98), respectively, and the DOR and AUROC were 283 (95% CI, 50-1,620) and 0.98 (95% CI 0.97-0.99), respectively (20). Compared with the results of the present study, the sensitivity and DOR of the diagnosis of NVC appear to be improved after the combination of 3D MIF technology. However, it is not accurate to evaluate the diagnostic effectiveness of different methods by directly comparing the values of different meta-analysis results. In a network meta-analysis, different NVC diagnostic methods were evaluated by constructing statistical models, and the results indicated that 3D MIF based on 3D TOF MRA combined with HR T2WI did have a higher superiority index in diagnosing NVC compared with the simple combination of the aforementioned MRI sequences without 3D MIF (21). These results suggested that 3D MIF technology may indeed improve the efficiency of NVC diagnosis.

To date, to the best of our knowledge, no meta-analysis has been published on HR T2WI as another type of MRI sequence commonly used in the detection of NVC. Therefore, there are no indicators that may accurately reflect the ability of HR T2WI to detect NVC, such as the pooled sensitivity, specificity, PLR, NLR, DOR and AUROC. According to existing clinical studies, the range of sensitivity and specificity of HR T2 WI in the detection of NVC are 86.17-100 and 71.47-100%, respectively (22-25). The results of different studies vary widely, which may be due to the use of different HR T2WI techniques and study designs. The value of HR T2WI in detecting NVC requires to be evaluated by further meta-analyses or large-scale clinical studies. Current data cannot be used to directly compare the ability of HR T2WI and 3D MIF based on 3D TOF MRA combined with HR T2WI to detect NVC.

Furthermore, the post-test probabilities are also associated with the clinical diagnostic ability of the diagnostic test (26). In the present study, the post-test probability for a positive test result was 98%, indicating the high clinical application value of 3D MIF based on 3D TOF MRA combined with HR T2WI in diagnosing NVC. The LR scatter plot is a qualitative effect size rating approach for diagnostic test accuracy, which divides the effect rating of diagnostic tests by quadrant (27). The LR scatter plot based on summary PLR and NLR was in the lower left quadrant. This meant that 3D MIF based on 3D TOF MRA combined with HR T2WI has the ability to exclude the diagnosis of NVC and that the diagnostic accuracy of this test has a ‘moderate’ effect rating. In combination, the above two results suggested that 3D MIF based on 3D TOF MRA combined with HR T2WI had a good clinical application value in diagnosing NVC.

For 3D MIF, the selection of MR sequences for fusion also affects the final diagnostic effect. Since the combination of 3D TOF MRA and HR T2WI has several advantages in detecting NVC (28,29), it may add to the advantages of 3D MIF technology. In fact, this is also the most commonly used combination of MRI sequences in 3D MIF clinical studies to detect NVC (8,9,12-16). There are currently few studies on the use of other MRI sequence combinations for 3D MIF in the detection of NVC (30). However, with the development of MR technology, other more suitable MRI sequence combinations may appear.

The present meta-analysis had certain limitations: i) NVC may also cause glossopharyngeal neuralgia (GN), but due to the lack of relevant clinical research data, patients with GN were not included in the current meta-analysis; ii) due to the limited number of cases, subgroup analyses were not performed for different diseases, T2 sequences or 3D fusion software; iii) of the seven included studies, three were retrospective studies, which have more potential sources of bias and confounding than prospective studies.

In conclusion, the present results suggested that 3D MIF based on 3D TOF MRA combined with HR T2WI has excellent sensitivity and specificity for detecting NVC in patients with TN or HFS. This method should serve a key role in the preoperative evaluation of MVD.

 Acknowledgements

Not applicable.

 Availability of data and materials

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

 Authors' contributions

CL was responsible for writing the main manuscript and data analysis. CL and LY contributed to the study design. BZ, SG and RL contributed to the literature search, data extraction and quality assessment. CL and SG confirm the authenticity of all the raw data. All authors have read and approved the final manuscript.

 Ethics approval and consent to participate

Not applicable.

 Patient consent for publication

Not applicable.

 Competing interests

The authors declare that they have no competing interests.

 References BroggiMAcerbiFFerroliPTringaliGSchiaritiMBroggiGMicrovascular decompression for neurovascular conflicts in the cerebello-pontine angle: which role for endoscopy?Acta Neurochir (Wien)1551709171620132388461110.1007/s00701-013-1824-8 CuiZLingZAdvances in microvascular decompression for hemifacial spasmJ Otol101620152993777410.1016/j.joto.2015.06.002 SadeBLeeJHMicrovascular decompression for trigeminal neuralgiaNeurosurg Clin N Am2574374920142524066110.1016/j.nec.2014.06.007 MontanoNConfortiGDi BonaventuraRMeglioMFernandezEPapacciFAdvances in diagnosis and treatment of trigeminal neuralgiaTher Clin Risk Manag1128929920152575053310.2147/TCRM.S37592 ChenSRNeurological Imaging for Hemifacial SpasmInt Ophthalmol Clin589710920182923988210.1097/IIO.0000000000000212 DonahueJHOrnanDAMukherjeeSImaging of Vascular Compression SyndromesRadiol Clin North Am5512313820172789018110.1016/j.rcl.2016.08.001 HuangBYangFYinMMoXZhongCA Review of multimodal medical image fusion techniquesComput Math Methods Med2020827934220203237722610.1155/2020/8279342 ShiHLiYWangYGuoWZhangKDuYShiHQianTThe preoperative evaluation value of 3D-slicer program before microsurgical vascular decompression in patients with hemifacial spasmClin Neurol Neurosurg21710724120223542985310.1016/j.clineuro.2022.107241 YaoSZhangJZhaoYHouYXuXZhangZKikinisRChenXMultimodal image-based virtual reality presurgical simulation and evaluation for trigeminal neuralgia and hemifacial spasmWorld Neurosurg113e499e50720182947699310.1016/j.wneu.2018.02.069 PageMJMcKenzieJEBossuytPMBoutronIHoffmannTCMulrowCDShamseerLTetzlaffJMAklEABrennanSEThe PRISMA 2020 statement: An updated guideline for reporting systematic reviewsBMJ372n7120213378205710.1136/bmj.n71 WhitingPFRutjesAWWestwoodMEMallettSDeeksJJReitsmaJBLeeflangMMSterneJABossuytPMQUADAS-2GroupQUADAS-2: A revised tool for the quality assessment of diagnostic accuracy studiesAnn Intern Med15552953620112200704610.7326/0003-4819-155-8-201110180-00009 JiaoYDuanFYanZMengQFengYThe value of 3D-TOF-MRA and 3D-FIESTA fusion three-dimensional images in judgment of offending vessel in primary trigeminal neuralgiaChin J Neurosurg359289322019 LeeAMcCartneySBurbidgeCRaslanAMBurchielKJTrigeminal neuralgia occurs and recurs in the absence of neurovascular compression Clinical articleJ Neurosurg1201048105420142450624110.3171/2014.1.JNS131410 MillerJAcarFHamiltonBBurchielKPreoperative visualization of neurovascular anatomy in trigeminal neuralgiaJ Neurosurg10847748220081831209410.3171/JNS/2008/108/3/0477 DolatiPGolbyAEichbergDAbolfotohMDunnIFMukundanSHulouMMAl-MeftyOPre-operative image-based segmentation of the cranial nerves and blood vessels in microvascular decompression: Can we prevent unnecessary explorations?Clin Neurol Neurosurg13915916520152647670010.1016/j.clineuro.2015.10.006 GranataFVinciSLLongoMBernavaGCaffoMCutugnoMMorabitoRSalamoneITomaselloFAlafaciCAdvanced virtual magnetic resonance imaging (MRI) techniques in neurovascular conflict: Bidimensional image fusion and virtual cisternographyRadiol Med1181045105420132371628410.1007/s11547-013-0928-9 FedorovABeichelRKalpathy-CramerJFinetJFillion-RobinJCPujolSBauerCJenningsDFennessyFSonkaM3D Slicer as an image computing platform for the quantitative imaging networkMagn Reson Imaging301323134120122277069010.1016/j.mri.2012.05.001 Bašić KesVZadro MatovinaLAccommodation to diagnosis of trigeminal neuralgiaActa Clin Croat5615716120172912055410.20471/acc.2017.56.01.21 CaiJXinZXZhangYQSunJLuJLXieFDiagnostic value of 3D time-of-flight MRA in trigeminal neuralgiaJ Clin Neurosci221343134820152607793810.1016/j.jocn.2015.02.023 LiangCYangLZhangBBGuoSWLiRCThree-dimensional time-of-flight magnetic resonance angiography combined with high resolution T2-weighted imaging in preoperative evaluation of microvascular decompressionWorld J Clin Cases10125941260420223657912110.12998/wjcc.v10.i34.12594 LiangCYangLReichardtWZhangBLiRDifferent MRI-based methods for the diagnosis of neurovascular compression in trigeminal neuralgia or hemifacial spasm: A network meta-analysisJ Clin Neurosci108192420223657731910.1016/j.jocn.2022.12.016 XiaHKongMLuHLuoSXuDZhangJValue of MR 3D FLASH-WE combined with 3D SPACE sequences applied in cranial nerve diseasesChin Comput Med Imaging215055092015(In Chinese) JiaJMGuoHHuoWJHuSWHeFSunXDLinGJPreoperative evaluation of patients with hemifacial spasm by three-dimensional time-of-flight (3D-TOF) and three-dimensional constructive interference in steady state (3D-CISS) sequenceClin Neuroradiol2643143820162579546610.1007/s00062-015-0382-2 Ruiz-JuretschkeFGuzmán-de-VilloriaJGGarcía-LealRSañudoJRPredictive value of magnetic resonance for identifying neurovascular compressions in trigeminal neuralgiaNeurologia3451051920192854975410.1016/j.nrl.2017.03.007(In English, Spanish) YangDShenJXiaXLinYYangTLinHJinYZhouKLiYPreoperative 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 sequenceBr J Radiol912017055720182938879810.1259/bjr.20170557 GogtayNJStatistical evaluation of diagnostic tests-part 2 (pre-test and post-test probability and odds, likelihood ratios, receiver operating characteristic curve, Youden's index and diagnostic test biases)J Assoc Physicians India658691201728792173 RubinsteinMLKraftCSParrottJSDetermining qualitative effect size ratings using a likelihood ratio scatter matrix in diagnostic test accuracy systematic reviewsDiagnosis520521420183024301510.1515/dx-2018-0061 SinghalSDanksRARadiologic and neurosurgical diagnosis of arterial neurovascular conflict on magnetic resonance imaging for trigeminal neuralgia in routine clinical practiceWorld Neurosurg157e166e17220223461940010.1016/j.wneu.2021.09.115 Wei ShengCYuRMengQQuCEfficacy of microvascular decompression in patients with trigeminal neuralgia with negative neurovascular relationship shown by magnetic resonance tomographyClin Neurol Neurosurg19710606320203267951310.1016/j.clineuro.2020.106063 OgiwaraMShimizuTSurface rendered three-dimensional MR imaging for the evaluation of trigeminal neuralgia and hemifacial spasmJ Clin Neurosci1184084420041551985910.1016/j.jocn.2003.06.010

Flow chart of the literature search and results of the present meta-analysis.

Methodological quality of the included studies was assessed using the Quality Assessment of Diagnostic Accuracy Studies checklist. (A) Methodological quality graph. (B) Methodological quality summary.

Heterogeneity and publication bias of the present meta-analysis. (A) Galbraith plot. (B) Deeks' funnel plot.

Summary effect size of the present meta-analysis. (A) Pooled sensitivity and specificity. (B) Pooled positive likelihood ratio and pooled negative likelihood ratio. (C) Pooled diagnostic odds ratio. (D) SROC curve. DLR, diagnostic likelihood ratio; SROC, summary receiver operating characteristic; SENS, sensitivity; SPEC, specificity; AUC, area under the curve.

Assessment of the clinical application value of 3D multimodal image fusion based on time-of-flight magnetic resonance angiography combined with high-resolution T2-weighted imaging. (A) Likelihood ratio scatter plot. (B) Fagan nomogram showing the post-test probability. LUQ, left upper quadrant; LRP, likelihood ratio positive; LRN, likelihood ratio negative; RUQ, right upper quadrant; LLQ, left lower quadrant; RLQ, right lower quadrant; LR, likelihood ratio; Prob, probability; Pos, positive; Neg, negative.

Basic characteristics of the included studies.

First author, year Country Sample size Age, years Diagnosis Research type MRI sequences 3D fusion software True-pos False-pos False-neg True-neg Sensitivity Specificity (Refs.)
Shi, 2022 China 40 49.6 (24-66) HFS Retrospective 3D TOF MRA + 3D FIESTA 3D-slicer 38 1 0 1 1.00 0.50 (8)
Jiao, 2019 China 48 ND TGN Retrospective 3D TOF MRA + 3D FIESTA 3D-slicer 46 0 1 1 0.98 1.00 (12)
Yao, 2018 China 42 51.2±11.6 (22-76) TGN/HFS Prospective 3D TOF + 3D SPACE 3D-slicer 40 0 1 1 0.98 1.00 (9)
Dolati, 2015 USA 14 65±10 TGN Prospective 3D TOF + 3D SPACE iPlan Net 19 0 0 1 1.00 1.00 (15)
Dolati, 2015 USA 6 61±7 HFS                   (15)
Lee, 2014 USA 190 56±14.2 TGN Retrospective 3D TOF + bFFE OsiriX 148 5 6 31 0.96 0.86 (13)
Miller, 2008 USA 18 52.9 (26-80) TGN Prospective 3D TOF + bFFE + 3D Gd-enhanced SPGR OsiriX 15 0 1 2 0.94 1.00 (14)
Granata, 2013 Italy 32 42±7 (23-68) TGN/HFS Prospective 3D TOF + 3D CISS Leonardo 21 0 0 11 1.00 1.00 (16)

Values are expressed as the median (range) and/or mean ± standard deviation. ND, no data; TOF MRA, time-of-flight magnetic resonance angiography; FIESTA, fast imaging employing steady-state acquisition; SPACE, sampling perfection with application-optimized contrasts using different flip angle evolutions; bFFE, balanced fast field echo; SPGR, spoiled gradient recalled; CISS, constructive interference steady state; TN, trigeminal neuralgia; HFS, hemifacial spasm; Pos, positive; Neg, negative.

Sensitivity analysis.

First author of excluded study, year Pooled sensitivity (95% CI) Pooled AUROC (95% CI) (Refs.)
Shi, 2022 0.97 (0.93-0.99) 0.98 (0.97-0.99) (8)
Jiao, 2019 0.97 (0.95-0.99) 0.98 (0.97-0.99) (12)
Yao, 2018 0.97 (0.95-0.99) 0.98 (0.97-0.99) (9)
Dolati, 2015 0.97 (0.95-0.99) 0.98 (0.97-0.99) (15)
Lee, 2014 0.98 (0.95-1.00) 0.99 (0.98-1.00) (13)
Miller, 2008 0.97 (0.95-0.99) 0.98 (0.97-0.99) (14)
Granata, 2013 0.97 (0.95-0.99) 0.98 (0.96-0.99) (16)

AUROC, area under the receiver operating characteristic curve.