1 Introduction

Magnetic resonance neurography (MRN) provides visualization of the brachial plexus and brachial plexus lesions and assists with classification of brachial plexus injury [1-3]. It plays a crucial role in injury diagnosis and localization and has become the preferred diagnostic imaging modality for clinical assessment of non-traumatic brachial plexus injury [4-6].

However, MRN is associated with several technical challenges [7-11]. First, the presence of a large quantity of air in the lung and on the surface of the neck and shoulder causes a significant difference in magnetic susceptibility at the air–soft tissue interface, which may cause conspicuous magnetic susceptibility artifact and failure in fat suppression. Second, obtaining three-dimensional (3D) T2 short tau inversion recovery (STIR) imaging requires a long scanning time, and this sequence is associated with a high specific absorption rate value, which can lead to patient discomfort and consequent motion artifacts. Third, the lymph nodes and slow-flowing veins surrounding the cervical nerve roots exhibit high signal intensity on the 3D T2 STIR sequence, which affects visualization of the brachial plexus (Figure 1).

To enhance brachial plexus visualization and quality of imaging, contrast-enhanced MRN (ceMRN) may be used, in which an injected contrast agent reduces signal contamination from the lymph nodes and veins around the nerve roots [12-15]. However, the interval between contrast injection and scanning that provides the highest-quality images is not known.

Therefore, we used continuous multiphase scanning to determine the optimal time to initiate brachial plexus ceMRN scanning after contrast injection.

2 Materials and Methods

2.1 Patients

Adult patients (age > 18 years) who were clinically evaluated with suspicion of or diagnosed with brachial plexus pathology at our hospital from November, 2019 to April, 2020 were eligible to enter the study. Those with abnormal liver or kidney function or severe mental illness were excluded. We also excluded patients whose scanning was interrupted because of discomfort. A detailed study flowchart is shown in Figure 2. The study was approved by the Institutional Review Board of Union Hospital, Tongji Medical College, Huazhong University of Science and Technology (No. TJ2015-04-02). All participants provided written informed consent.

2.3 Objective Evaluation

The signal of the C7 nerve trunk (Signal_nerve), lymph node (Signal_{lymph node}), and sternocleidomastoid muscle (Signal_muscle) were measured at different coronal levels on the thin original brachial plexus images (Figure 3). The standard deviation (SD) of the region of interest (ROI) within the sternocleidomastoid muscle served as background noise (BN). The Signal_nerve, Signal_{lymph node}, Signal_muscle, BN, signal-to-noise ratio (SNR), and contrast-to-noise ratio (CNR) were compared across groups. SNR was calculated as Signal_nerve/BN. CNR was calculated as (Signal_nerve − Signal_{lymph node})/BN.

3 Results

Fifteen patients (10 men and five women) were included for analysis after applying criteria. Mean patient age was 47.3 ± 12.5 years (range, 26–63 years). All successfully completed scanning with no significant adverse effects and the images from the six groups met the diagnostic imaging criteria.

3.1 Objective Evaluation

As shown in Table 1 and Figure 4, signal_nerve, signal_muscle, BN, and SNR did not significantly differ among the six groups. However, significant differences were observed in signal_{lymph node} (F = 16.067) and CNR (F = 9.495). With increasing delay in scanning time, the number of visualized lymph nodes and signal_{lymph node} intensity gradually increased, whereas the CNR gradually decreased. This suggests that signal_{lymph node} is impacted as the contrast agent is metabolized over time after injection.

TABLE 1. The comparison of objective evaluation for image quality.

Groups	Signalnerve	Signalmuscle	Signallymph node	BN	SNR	CNR
A	512.67 ± 113.55	109.59 ± 21.41	135.37 ± 65.61c	13.70 ± 5.22	41.67 ± 15.11	29.89 ± 2.31a
B	509.75 ± 130.29	107.73 ± 16.32	240.12 ± 96.92b	13.19 ± 4.64	44.40 ± 21.25	22.25 ± 2.32a,b
C	533.87 ± 129.93	116.63 ± 22.95	306.18 ± 104.25a,b	14.10 ± 4.80	41.50 ± 16.03	17.31 ± 2.28b,c
D	509.31 ± 130.53	118.13 ± 17.25	329.31 ± 91.61a,b	14.03 ± 3.64	39.11 ± 14.73	13.46 ± 2.22b,c
E	510.91 ± 132.67	115.51 ± 18.29	354.11 ± 78.34a	12.16 ± 2.65	42.80 ± 10.91	13.43 ± 2.68b,c
F	510.21 ± 145.04	116.64 ± 21.96	387.54 ± 87.94a	12.72 ± 3.30	42.47 ± 15.98	9.51 ± 2.54c
F	0.081	0.700	16.067	0.519	0.181	9.495
p	0.995	0.625	< 0.05	0.762	0.969	< 0.05

• Note: Pairwise comparisons are indicated by letter notations (a, b, c). There is no statistical difference among any groups with the same letters. Statistical differences exist only when the letters are completely different.
• Abbreviations: BN, background noise; CNR, contrast-to-noise ratio; SNR, signal-to-noise ratio.

3.2 Subjective Evaluation

The overall subjective scores exhibited good interobserver agreement (κ = 0.868). As shown in Table 2 and Figure 5, the final subjective score for each of the six groups was ≥ 4 points. The subjective score did not significantly differ among groups B, C, D, E, and F; however, the score was significantly lower for group A (4.47 ± 0.30) than groups B (4.83 ± 0.24), C, (4.90 ± 0.21), D (4.87 ± 0.30), E (4.83 ± 0.31), and F (4.83 ± 0.31). Presumably, the contrast agent had not fully entered the small veins surrounding the nerve roots in the first 5 min after injection and suppression of venous signal (signal_vein) was insufficient, causing a lower score for group A.

TABLE 2. Subjective scoring results (cases).

Groups	4.0 points	4.5 points	5.0 points	Average score
A	3	10	2	4.47 ± 0.30b
B	0	5	10	4.83 ± 0.24a
C	0	3	12	4.90 ± 0.21a
D	1	2	12	4.87 ± 0.30a
E	1	3	11	4.83 ± 0.31a
F	1	3	11	4.83 ± 0.31a
χ2	25.476
p	< 0.05

• Note: Pairwise comparisons are indicated by letter notations (a, b, c). There is no statistical difference among any groups with the same letters. Statistical differences exist only when the letters are completely different.

4 Discussion

This study demonstrated that the time interval between injection of contrast agent and image acquisition affects the suppression of signal_{lymph node}. As time elapsed, the suppression of signal_vein was also affected (Figure 6). Although the subjective scores of all six groups were relatively high, we still recommend performing brachial plexus ceMRN 5 min after contrast injection for two reasons: First, to achieve excellent image quality, good background suppression is required, namely the suppression of high signal_{lymph node} and signal_vein. The signal_{lymph node} and signal_vein were relatively low only in group B. Second, in clinical practice, patients have some other sequences for auxiliary diagnosis that need to be scanned. If a longer waiting time is required for ceMRN, it is easy to cause discomfort to the patient and lead to examination interruption or failure.

The 3D T2-NerveView sequence is based on the STIR technique and uses high resolution and isotropic voxels, avoiding geometric distortion in the reconstructed image. The STIR technique is employed for fat suppression, using a broader bandwidth (2500 Hz) of the 180° flip insulated pulse, which addresses the issue of uneven fat suppression in the neck. Additionally, the sequence incorporates motion-sensitized driven equilibrium, which reduces the signal of moving tissues, effectively suppressing the signal of veins and subclavian arteries [16-19]. Moreover, we combined the newly developed compressed sensing technique with the 3D T2-NerveView sequence, which greatly saves scanning time by “sparse sampling”, reduces motion artifacts, and improves the success rate of the examination. These advanced techniques allowed us to perform multiphase scans of the brachial plexus after contrast injection and obtain high-quality images.

Signal_nerve and signal_muscle were largely unchanged over time owing to the presence of the blood–nerve barrier and the relative sparseness of blood vessels in muscle tissue [20]. Contrast agent shortens the T2 value of tissue. The higher the contrast agent concentration, the more obvious the T2 shortening and the faster the tissue signal attenuates. As a result, the suppression of background signals is better [20-22]. After a contrast agent is injected, its tissue concentration will go from low to high and then to low owing to the blood circulation and renal metabolism. This change in concentration causes the signal_{lymph node} and signal_vein to change over time. However, the concentration change in lymph nodes is rapid, whereas that in the perineural plexus veins is relatively slow. Understanding the dynamic changes of signal_{lymph node} and signal_vein can assist radiologists in selecting the appropriate delay time or designing new scanning protocols to achieve superior image quality.

In this study, signal_{lymph node} was lowest on the first-phase scan, whereas signal_vein around the nerve roots were completely suppressed starting from the second-phase scan (phases 2–6). Based on these findings, we hypothesize that in order to obtain high-quality brachial plexus imaging with minimal background signal interference, two contrast injections could be administered. The basic method is as follows: first, suppress the signal_vein with the first contrast injection; 5 minutes later, immediately initiate the scan upon the second contrast injection, which will also suppress the signal_{lymph node}. However, this hypothesis has not yet been tested. If proven effective, this approach could further improve the image quality of the brachial plexus.

We acknowledge that this study has several limitations. First, the sample size is small and the statistical results may lack persuasiveness. Second, although the gadodiamide contrast agent has a half-life of approximately 1.5 h [23], we only monitored signal changes during the first 30 min after injection. Third, our recommendation to perform brachial plexus ceMRN 5 min after contrast injection is relative and a function of the fact that our sequence time was 5 min. If our sequence time had been 3 min, we might have recommended a 3-min time period.

5 Conclusion

In conclusion, the timing of imaging after contrast agent injection in brachial plexus ceMRN affects image quality. We recommend performing ceMRN 5 min after contrast injection to obtain the best visualization of the plexus, which should improve diagnostic accuracy.

Author Contributions

Jun Xu: writing–original draft (equal). Xiaoli Hu: writing–original draft (equal). Xiaoyun Su: software (equal). Shen Gui: validation (equal). Ziqiao Lei: project administration (equal). Xiaoming Liu: data curation (equal). Xiangzhi Zhou: validation (equal). Lixia Wang: conceptualization (equal). Wenjun Wu: writing–review & editing (equal). Xiangchuang Kong: writing–review & editing (equal).

Ethics Statement

We have read the statement of the journal's Ethics and Integrity Policy and have strictly adhered to the provisions. The study was approved by the Institutional Review Board of Union Hospital, Tongji Medical College, Huazhong University of Science and Technology (Approval No. TJ2015-04-02).

Consent

All participants provided written informed consent.

Conflicts of Interest

The authors declare no conflicts of interest.