Multicenter reproducibility study of diffusion MRI and fiber tractography of the lumbosacral nerves

Clinical Value and Reliability of Quantitative Assessments of Lumbosacral Nerve Root Using Diffusion Tensor and Diffusion Weighted MR Imaging: A Systematic Review

BACKGROUND

Lumbosacral radicular pain diagnosis remains challenging. Diffusion tensor imaging (DTI) and diffusion weighted imaging (DWI) have potential to quantitatively evaluate symptomatic nerve root, which may facilitate diagnosis.

PURPOSE

To determine the ability of DTI and DWI metrics, namely fractional anisotropy (FA) and apparent diffusion coefficient (ADC), to discriminate between healthy and symptomatic lumbosacral nerve roots, to evaluate the association between FA and ADC values and patient symptoms, and to determine FA and ADC reliability.

STUDY TYPE

Systematic review.

SUBJECTS

Eight hundred twelve patients with radicular pain with or without radiculopathy caused by musculoskeletal-related compression or inflammation of a single, unilateral lumbosacral nerve root and 244 healthy controls from 29 studies.

FIELD STRENGTH/SEQUENCE

Diffusion weighted echo planar imaging sequence at 1.5 T or 3 T.

ASSESSMENT

An extensive systematic review of the literature was conducted in Embase, Scopus, and Medline databases. FA and ADC values in symptomatic and contralateral lumbosacral nerve roots were extracted and summarized, together with intra- and inter-rater agreements. Where available, associations between DWI or DTI parameters and patient symptoms or symptom duration were extracted.

STATISTICAL TESTS

The main results of the included studies are summarized. No additional statistical analyses were performed.

RESULTS

The DTI studies systematically found significant differences in FA values between the symptomatic and contralateral lumbosacral nerve root of patients suffering from radicular pain with or without radiculopathy. In contrast, identification of the symptomatic nerve root with ADC values was inconsistent for both DTI and DWI studies. FA values were moderately to strongly correlated with several symptoms (eg, disability, nerve dysfunction, and symptom duration). The inter- and intra-rater reliability of DTI parameters were moderate to excellent. The methodological quality of included studies was very heterogeneous.

DATA CONCLUSION

This systematic review showed that DTI was a reliable and discriminative imaging technique for the assessment of symptomatic lumbosacral nerve root, which more consistently identified the symptomatic nerve root than DWI. Further studies of high quality are needed to confirm these results.

EVIDENCE LEVEL

N/A TECHNICAL EFFICACY: Stage 2.

Lumbosacral plexus and pudendal nerve magnetic resonance tractography: A systematic review of the clinical applications for pudendal neuralgia

Diffusion tensor imaging (DTI) is commonly used to establish three-dimensional mapping of white-matter bundles in the supraspinal central nervous system. DTI has also been the subject of many studies on cranial and peripheral nerves. This non-invasive imaging technique enables virtual dissection of nerves in vivo and provides specific measurements of microstructural integrity. Adverse effects on the lumbosacral plexus may be traumatic, compressive, tumoral, or malformative and thus require dedicated treatment. DTI could lead to new perspectives in pudendal neuralgia diagnosis and management. We performed a systematic review of all articles or posters reporting results and protocols for lumbosacral plexus mapping using the DTI technique between January 2011 and December 2023. Twenty-nine articles published were included. Ten studies with a total of 351 participants were able to track the lumbosacral plexus in a physiological context and 19 studies with a total of 402 subjects tracked lumbosacral plexus in a pathological context. Tractography was performed on a 1.5T or 3T MRI system. DTI applied to the lumbosacral plexus and pudendal nerve is feasible but no microstructural normative value has been proposed for the pudendal nerve. The most frequently tracking parameters used in our review are: 3T MRI, b-value of 800 s/mm2, 33 directions, 3 × 3 × 3 mm3, AF threshold of 0.1, minimum fiber length of 10 mm, bending angle of 30°, and 3DT2 TSE anatomical resolution. Increased use of DTI could lead to new perspectives in the management of pudendal neuralgia due to entrapment syndrome, whether at the diagnostic, prognostic, or preoperative planning level. Prospective studies of healthy subjects and patients with the optimal acquisition parameters described above are needed to establish the accuracy of MR tractography for diagnosing pudendal neuralgia and other intrapelvic nerve entrapments.

Entrapped by pain: The diagnosis and management of endometriosis affecting somatic nerves

Somatic nerve entrapment caused by endometriosis is an underrecognized and often misdiagnosed issue that leads to many women suffering unnecessarily. While the classic symptoms of endometriosis are well-known to the gynaecologic surgeon, the dermatomal-type pain caused by endometriosis impacting neural structures is not within gynecologic day-to-day practice, which often complicates diagnosis and delays treatment. A thorough understanding of pelvic neuroanatomy and a neuropelveologic approach is required for accurate assessments of patients with endometriosis and nerve entrapment. Magnetic resonance imaging is the preferred imaging modality for this presentation of endometriosis. Surgical management with laparoscopic or robotic-assisted techniques is the preferred approach to treatment, with excellent long-term results reported after nerve detrapment and endometriosis excision. The review calls for increased awareness and education on the links between endometriosis and the nervous system, advocating for patient-centered care and further research to refine the diagnosis and treatment of this challenging condition.

Intra-parotid facial nerve path by MRI tractography: radio-clinical comparison in parotid tumors

PURPOSE

The objective of our study was to evaluate the ability of preoperative MRI tractography to visualize and predict the path of the facial nerve with respect to an intra-parotid mass.

METHODS

We performed an observational bicentric study from June 2019 to August 2020. All patients older than 18 years old, treated for a parotid mass with surgical indication, without MRI contraindication and who agreed to participate in the study were enrolled prospectively. All patients underwent a cervico-facial MRI with tractographic analysis. Postprocessed tractography images of the intra-parotid facial nerve were analyzed by two expert radiologists in head and neck imaging. The intraoperative anatomical description of the facial nerve path and its relationship to the mass was performed by the surgeon during the operation, with no visibility on MRI examination results. A statistical study allowed for the description of the data collected as well as the measurement of inter-observer agreement and agreement between tractography and surgery using kappa coefficients.

RESULTS

Fifty-two patients were included. The facial nerve trunk and its first two divisional branches were visualized via tractography in 93.5% of cases (n = 43). The upper distal branches were visualized in 51.1% of cases (n = 23), and the lower branches were visualized in 73.3% of cases (n = 33). Agreement with the location described per-operatively was on average 82.9% for the trunk, 74.15% for the temporal branch, and 75.21% for the cervico-facial branch.

CONCLUSION

Fiber tractography analysis by MRI of the intra-parotid facial nerve appears to be a good test for predicting the path of the nerve over the parotid mass and could be an additional tool to guide the surgeon in the operative procedure.

Magnetic resonance neurography and diffusion tensor imaging of the sciatic nerve in hereditary transthyretin amyloidosis polyneuropathy

The therapeutic advance in hereditary transthyretin amyloidosis (ATTRv amyloidosis) requires quantitative biomarkers of nerve involvement in order to foster early diagnosis and monitor therapy response. We aimed at quantitatively assessing Magnetic Resonance Neurography (MRN) and Diffusion Tensor Imaging (DTI) properties of the sciatic nerve in subjects with ATTRv-amyloidosis-polyneuropathy (ATTRv-PN) and pre-symptomatic carriers (ATTRv-C). Twenty subjects with pathogenic variants of the TTR gene (mean age 62.20 ± 12.04 years), 13 ATTRv-PN, and 7 ATTRv-C were evaluated and compared with 20 healthy subjects (mean age 60.1 ± 8.27 years). MRN and DTI sequences were performed at the right thigh from the gluteal region to the popliteal fossa. Cross-sectional-area (CSA), normalized signal intensity (NSI), and DTI metrics, including fractional anisotropy (FA), mean (MD), axial (AD), and radial diffusivity (RD) of the right sciatic nerve were measured. Increased CSA, NSI, RD, and reduced FA of sciatic nerve differentiated ATTRv-PN from ATTRv-C and healthy subjects at all levels (p < 0.01). NSI differentiated ATTRv-C from controls at all levels (p < 0.05), RD at proximal and mid-thigh (1.04 ± 0.1 vs 0.86 ± 0.11 p < 0.01), FA at mid-thigh (0.51 ± 0.02 vs 0.58 ± 0.04 p < 0.01). According to receiver operating characteristic (ROC) curve analysis, cutoff values differentiating ATTRv-C from controls (and therefore identifying subclinical sciatic involvement) were defined for FA, RD, and NSI. Significant correlations between MRI measures, clinical involvement and neurophysiology were found. In conclusion, the combination of quantitative MRN and DTI of the sciatic nerve can reliably differentiate ATTRv-PN, ATTRv-C, and healthy controls. More important, MRN and DTI were able to non-invasively identify early subclinical microstructural changes in pre-symptomatic carriers, thus representing a potential tool for early diagnosis and disease monitoring.

Lumbosacral plexus MR tractography: A novel diagnostic tool for extraspinal sciatica and pudendal neuralgia?

BACKGROUND

Diagnosing extraspinal sciatica and pudendal neuralgia remains a clinical challenge. MRI and MR Neurography (MRN) are currently the standard techniques used to support the diagnosis of extraspinal lumbosacral plexus entrapments; however, for the intrapelvic portions of the lumbosacral plexus their accuracy is still limited. MR Tractography (MRT) feasibility to image the lumbosacral plexus has been demosntrated, but its clinical applications have yet to be determined.

PURPOSE

To correlate MRT with intraoperative findings in patients undergoing laparoscopic treatment of intrapelvic entrapments of the lumbosacral plexus and compare its accuracy with Neuropelveological clinical assessment and MRN.

MATERIALS AND METHODS

This is a retrospective analysis of MRT reconstructions of diffusion tensor imaging (DTI) sequences acquired for the MRN collected from a cohort of 13 patients undergoing laparoscopic detrapment of the lumbosacral plexus. The primary outcome of this study was the correlation of MRT reconstruction with intraoperative findings. Secondary outcomes included the correlation of MRN, preoperative Neuropelveological clinical diagnoses and the diffusion-weighted imaging (DWI) fractional anisotropy (FA) and Apparent Diffusion Coefficient (ADC) in patients undergoing pelvic MRI and MRN for the investigation of intrapelvic nerve entrapments.

RESULTS

MRT correlated with intraoperative findings in 11 of 13 patients (85%). Neuropelveological clinical assessment was able to accurately diagnose a pelvic nerve entrapment in 12/13 patients (92%) and MRN agreed with surgical findings in only 2/13 (15%) patients. MRT was significantly superior to MRN (p < 0.001). FA and ADC did not correlate with the identification of a nerve entrapment, likely due to limitations regarding the placement of the seedpoints.

CONCLUSIONS

This initial, retrospective analysis, suggests that MRT is superior to MRN at diagnosing intrapelvic entrapments of the lumbosacral plexus. A prospective, double-blinded study is underway to validate this data, but these initial findings show great potential for MRT as a diagnostic tool for extraspinal sciatica and pudendal neuralgia. Clinical Trials Registry: U1111-1261-4910 (REBEC - Brazilian Registry for Clinical Trials).

Magnetic resonance tractography of the lumbosacral plexus: Step-by-step

ABSTRACT

MR tractography of the lumbosacral plexus (LSP) is challenging due to the difficulty of acquiring high quality data and accurately estimating the neuronal tracts. We proposed an algorithm for an accurate visualization and assessment of the major LSP bundles using the segmentation of the cauda equina as seed points for the initial starting area for the fiber tracking algorithm.Twenty-six healthy volunteers underwent MRI examinations on a 3T MR scanner using the phased array coils with optimized measurement protocols for diffusion-weighted images and coronal T2 weighted 3D short-term inversion recovery sampling perfection with application optimized contrast using varying flip angle evaluation sequences used for LSP fiber reconstruction and MR neurography (MRN).The fiber bundles reconstruction was optimized in terms of eliminating the muscle fibers contamination using the segmentation of cauda equina, the effects of the normalized quantitative anisotropy (NQA) and angular threshold on reconstruction of the LSP. In this study, the NQA parameter has been used for fiber tracking instead of fractional anisotropy (FA) and the regions of interest positioning was precisely adjusted bilaterally and symmetrically in each individual subject.The diffusion data were processed in individual L3-S2 nerve fibers using the generalized Q-sampling imaging algorithm. Data (mean FA, mean diffusivity, axial diffusivity and radial diffusivity, and normalized quantitative anisotropy) were statistically analyzed using the linear mixed-effects model. The MR neurography was performed in MedINRIA and post-processed using the maximum intensity projection method to demonstrate LSP tracts in multiple planes.FA values significantly decreased towards the sacral region (P < .001); by contrast, mean diffusivity, axial diffusivity, radial diffusivity and NQA values significantly increased towards the sacral region (P < .001).Fiber tractography of the LSP was feasible in all examined subjects and closely corresponded with the nerves visible in the maximum intensity projection images of MR neurography. Usage of NQA instead of FA in the proposed algorithm enabled better separation of muscle and nerve fibers.The presented algorithm yields a high quality reconstruction of the LSP bundles that may be helpful both in research and clinical practice.

Comparing multiband and singleband EPI in NODDI at 3 T: what are the implications for reproducibility and study sample sizes?

Objective

The reproducibility of Neurite orientation dispersion and density imaging (NODDI) metrics from time-saving multiband (MB) EPI compared with singleband (SB) has not been considered. This study aims to evaluate the reproducibility of NODDI parameters from SB and MB acquisitions, determine the agreement between acquisitions and estimate the sample sizes required to detect between-group change.

Methods

Brain diffusion MRI data were acquired using SB and MB (acceleration factors 2 (MB2) and 3 (MB3)) on 8 healthy subjects on 2 separate visits. NODDI maps of isotropic volume fraction (FISO), neurite density (NDI) and orientation dispersion index (ODI) were estimated. Region-of-interest analysis was performed; variability across subjects and visits was measured using coefficients of variation (CoV). Intraclass correlation coefficient and Bland–Altman analysis were performed to assess reproducibility and detect any systematic bias between SB, MB2 and MB3. Power calculations were used to determine sample sizes required to detect group differences.

Results

Both NDI and ODI were reproducible between visits; however, FISO was variable. All parameters were not reproducible across methods; a systematic bias was observed with the derived values decreasing as the MB factor increases. The number of subjects needed to detect a between-group change is not significantly different between methods; however, ODI needs considerably higher sample sizes than NDI.

Conclusions

Both SB and MB yield highly reproducible NDI and ODI measures, but direct comparison of these parameters between methods is complicated by systematic differences that exist between the two approaches.

Supplementary Information

The online version contains supplementary material available at 10.1007/s10334-020-00897-7.

Diffusion tensor imaging and tractography for preoperative assessment of benign peripheral nerve sheath tumors

PURPOSE

To evaluate the diagnostic value of fiber tractography and diffusivity analysis generated from 3D diffusion-weighted (DW) sequences for preoperative assessment of benign peripheral nerve sheath tumors.

METHOD

MR imaging at 3 T was performed in 22 patients (mean age 41.9 ± 17.1y, 13 women) with histologically confirmed schwannomas (N = 18) and histologically confirmed neurofibromas (N = 11), including a 3D DW turbo spin echo sequence with fat suppression. Diffusion tensor parameters were computed and fiber tracks were determined. Evaluation was performed by two radiologists and one orthopedic surgeon blinded for final diagnosis. Mean diffusivity was computed to allow further assessment of tumor microstructure. Preoperative fascicle visualization was graded, fascicles were categorized regarding anatomical location and amount of fascicles surrounding the tumor. The agreement of imaging findings with intraoperative findings was assessed.

RESULTS

On 78.3 % of the DTI images, the fascicle visualization was rated as good or very good. Tractography differences were observed in schwannomas and neurofibromas, showing schwannomas to be significantly more often located eccentrically to the nerve (94.8 %) than neurofibromas (0 %, P < 0.01). Fascicles were significantly more often continuous (87.5 %) in schwannomas, while in neurofibromas, none of the tracks was graded to be continuous (0 %, P = 0.014). A substantial agreement between fiber tracking and surgical anatomy was found regarding the fascicle courses surrounding the tumor (κ = 0.78). Mean diffusivity of schwannomas (1.5 ± 0.2 × 10^-3 mm²/s) was significantly lower than in neurofibromas (1.8 ± 0.2 × 10^-3 mm²/s; P < 0.001). The Youden index showed an optimal cutoff at 1.7 × 10^-3 mm²/s (sensitivity, 0.91; specificity, 0.78; J = 0.69).

CONCLUSIONS

Preoperative diffusion tensor imaging allowed to accurately differentiate between schwannomas and neurofibromas and to describe their location in relation to the nerve fascicles for preoperative planning.

A quantitative and clinical evaluation of nerve roots in lumbosacral radiculopathy using diffusion tensor imaging

PURPOSE

This study aimed to investigate the relationship between the fractional anisotropy (FA) values of compressed nerves derived in diffusion tensor imaging (DTI) and the corresponding clinical symptoms for quantitative and clinical evaluation in patients with lumbosacral radiculopathy.

METHODS

Thirty-six patients and ten volunteers participated in the study and measured with DTI. The resultant FA values for L5-S1 lumbar nerve roots were calculated at three sub-regions. Additionally, the DTI relevant tractography was also performed on L4-S1 nerve roots. Clinical symptoms were performed by Japanese Orthopedic Association (JOA) scoring for each patient and volunteer.

RESULTS

The FA values of the nerves at the symptomatic side were significantly lower than those at the asymptomatic side (p < 0.001). Diffusion tensor tractography distinctly showed abnormalities in the symptomatic nerve tracts. There was a significant correlation between JOA scores and the FA values of the compressed nerves at middle and distal sub-regions (p < 0.005).

CONCLUSION

The clinical symptoms associated robustly with the DTI derived FA values of the compressed nerves in patients with lumbosacral radiculopathy. Therefore, the FA values can be a potential clinical tool to evaluate the nerve roots in lumbosacral radiculopathy quantitatively.

Peripheral Nerve Diffusion Tensor Imaging : Interreader and Test-retest Reliability as Quantified by the Standard Error of Measurement

PURPOSE

Diffusion tensor imaging (DTI) is increasingly being used in magnetic resonance neurography (MRN). The purpose of this study was to determine the interreader and test-retest reliability of peripheral nerve DTI in MRN with focus on the sciatic nerve.

METHODS

In this prospective study 27 healthy volunteers each underwent 3 scans of a short DTI protocol on separate days consisting of a T2-weighted turbo spin-echo and single-shot DTI sequence of the sciatic nerve of the dominant leg. The DTI parameters fractional anisotropy (FA), mean diffusivity (MD), axial diffusivity (AD), and radial diffusivity (RD) were obtained after manual nerve segmentation by two independent readers. Intraclass correlation coefficients (ICC), standard error of measurement (SEM), and Bland-Altman plots were calculated as measures for both interreader and test-retest agreement for all readout parameters.

RESULTS

The mean ± standard deviation was 0.507 ± 0.05 for FA, 1308.5 ± 162.4 × 10^-6 mm²/s for MD, 905.6 ± 145.4 ×10^-6 mm²/s for RD and 2114.1 ± 219.2 × 10^-6 mm²/s for AD. The SEM for FA was 0.02 for interreader and test-retest agreement, the SEM for MD, RD, and AD ranged between 46.2 × 10^-6 mm²/s (RD) and 70.1 × 10^-6 mm²/s (AD) for interreader reliability and between 45.9 × 10^-6 mm²/s (RD) and 70.1 × 10^-6 mm²/s (AD) for test-retest reliability. The ICC for interreader reliability of DTI parameters ranged between 0.81 and 0.92 and ICC for test-retest reliability between 0.76 and 0.91.

CONCLUSION

Peripheral nerve DTI of the sciatic nerve is reliable and reproducible. The measures presented here may serve as first orientation values of measurement accuracy when interpreting parameters of sciatic nerve DTI.

Denoising of diffusion MRI improves peripheral nerve conspicuity and reproducibility

BACKGROUND

Quantitative diffusion MRI is a promising technique for evaluating peripheral nerve integrity but low signal-to-noise ratio (SNR) can impede measurement accuracy.

PURPOSE

To evaluate principal component analysis (PCA) and generalized spherical deconvolution (genSD) denoising techniques to improve within-subject reproducibility and peripheral nerve conspicuity.

STUDY TYPE

Prospective.

SUBJECTS

Seven healthy volunteers and three peripheral neuropathy patients.

FIELD STRENGTH/SEQUENCE

3T/multiband single-shot echo planar diffusion sequence using multishell 55-direction scheme.

ASSESSMENT

Images were processed using four methods: "original" (no denoising), "average" (10 repetitions), "PCA-only," and "PCA + genSD." Tibial and common peroneal nerve segmentations and masks were generated from volunteer diffusion data. Quantitative (SNR and contrast-to-noise ratio [CNR]) values were calculated. Three radiologists qualitatively evaluated nerve conspicuity for each method. The two denoising methods were also performed in three patients with peripheral neuropathies.

STATISTICAL TESTS

For healthy volunteers, calculations included SNR and CNR_FA (computed using FA values). Coefficient of variation (CV%) of CNR_FA quantified within-subject reproducibility. Groups were compared with two-sample t-tests (significance P < 0.05; two-tailed, Bonferroni-corrected). Odds ratios (ORs) quantified the relative rates of each of three radiologists confidently identifying a nerve, per slice, for the four methods.

RESULTS

"PCA + genSD" yielded the highest SNR (mean_overall = 14.83 ± 1.99) and tibial and common peroneal nerve CNR_FA (mean_tibial = 3.45, mean_peroneal = 2.34) compared to "original" (P _SNR < 0.001; P _CNR = 0.011) and "PCA-only" (P _SNR < 0.001, P _CNR < 0.001). "PCA + genSD" had higher within-subject reproducibility (low CV%) for tibial (6.04 ± 1.98) and common peroneal nerves (8.27 ± 2.75) compared to "original" and "PCA-only." The mean FA was higher for "original" than "average" (P < 0.001), but did not differ significantly between "average" and "PCA + genSD" (P = 0.14). "PCA + genSD" had higher tibial and common peroneal nerve conspicuity than "PCA-only" (OR_tibial = 2.50, P < 0.001; OR_peroneal = 1.86, P < 0.001) and "original" (OR_tibial = 2.73, P < 0.001; OR_peroneal = 2.43, P < 0.001).

DATA CONCLUSION

PCA + genSD denoising method improved SNR, CNR_FA , and within-subject reproducibility (CV%) without biasing FA and nerve conspicuity. This technique holds promise for facilitating more reliable, unbiased diffusion measurements of peripheral nerves.

LEVEL OF EVIDENCE

2 Technical Efficacy Stage: 1 J. Magn. Reson. Imaging 2020;51:1128-1137.

Diffusion-weighted imaging and diffusion tensor imaging as adjuncts to conventional MRI for the diagnosis and management of peripheral nerve sheath tumors: current perspectives and future directions

Peripheral nerve sheath tumors (PNSTs) account for ~ 5% of soft tissue neoplasms and are responsible for a wide spectrum of morbidities ranging from localized neuropathy to fulminant metastatic spread and death. MR imaging represents the gold standard for identification of these neoplasms, however, current anatomic MR imaging markers do not reliably detect or differentiate benign and malignant lesions, and therefore, biopsy or excision is required for definitive diagnosis. Diffusion-weighted MR imaging (DWI) serves as a useful tool in the evaluation and management of PNSTs by providing functional information regarding the degree of diffusion, while diffusion tensor imaging (DTI) aids in determining the directional information of predominant diffusion and has been shown to be particularly useful for pre-operative planning of these tumors by delineating healthy and pathologic fascicles. The article focuses on these important neurogenic lesions, highlighting the current utility of diffusion MR imaging and future directions including computerized radiomic analysis. KEY POINTS: • Anatomic MRI is moderately accurate in differentiating benign from malignant PNST. • Diffusion tensor imaging facilitates pre-operative planning of PNSTs by depicting neuropathy and tractography. • Radiomics will likely augment current observer-based diagnostic criteria for PNSTs.