Introduction

The presence of structural compression of the spinal cord and resultant neurological impairment often characterizes cervical spondylotic myelopathy (CSM) [1]. Age-related degeneration of the spine serves as the primary instigator of CSM, while spinal cord trauma and arthritis may also contribute to its development. Notably, the rate of CSM-related hospitalizations was reported at 4.04 per 100,000 person-years. Males and older persons have a higher incidence rate of CSM [2]. Nevertheless, variations in clinical evaluations, particularly neurological examination results, pose challenges in effectively assessing CSM patients. There is often a certain discrepancy between the degree of spinal cord compression and the symptoms experienced by the patient. Neurological function discrepancies may occur despite similar levels of spinal cord compression. In certain cases, patients with seemingly mild spinal cord compression on MRI exhibit pronounced limb numbness or even experience a loss of strength. The identification of responsible segments for neurological dysfunction proves to be complex in cases involving multi-stage compression, hindering accurate determination of the precise cervical spine segments necessitating surgical intervention. Previous studies have suggested that invasive spinal cord electrophysiological assessments offer an objective means of quantitatively evaluating spinal cord function and even pinpointing specific spinal cord segments [3, 4]. However, non-invasive assessment methods capable of precisely identifying neurological function at distinct spinal cord segments are currently unavailable.

Diffusion tensor imaging (DTI) is an advanced MR imaging method that quantitatively evaluates water molecule diffusion in the microenvironment of biological tissues, providing indirect insights into neurological function. This technique relies on measuring the thermal Brownian motion of water molecules. DTI finds extensive application in estimating brain axonal organization and assessing spinal cord function [5,6,7]. Commonly used DTI metrics for quantitative assessment include axial diffusivity (AD), mean diffusivity (MD), radial diffusivity (RD), and fractional anisotropy (FA). AD and FA are particularly crucial indicators for evaluating spinal cord function. AD represents the diffusion coefficient along the long axis of the diffusion tensor, typically aligned with the axon, and its value generally increases in compressed segments [8]. FA characterizes the water's anisotropic movement ability, typically displaying lower values in the compressed segments than in the normal segments [9].

Previous research has highlighted the significance of DTI metrics in diagnosing CSM. For instance, several studies have demonstrated a strong correlation between DTI and the modified Japanese Orthopedic Association scale (mJOA) [10,11,12,13]. Suetomi et al. specifically examined the diagnostic performance of maximum or minimum diffusion metrics (AD or FA) in specific CSM segments [14]. These studies could qualitatively determine CSM, while they have had limited success in pinpointing the exact cervical compression segment responsible for sensory neurological dysfunction. Moreover, most previous studies have relied on the relatively subjective mJOA scale for evaluating spinal cord function. The neurological function determination efficacy of mJOA is inferior to that of invasive electrophysiological [15,16,17,18].

Consequently, there is a need to validate the diagnostic efficacy of DTI metrics against electrophysiological neurological function examination. Validating the diagnostic efficacy of DTI metrics by assessing neurological function using electrophysiological neurological examinations is necessary. Therefore, our study aimed to validate the effectiveness of DTI metrics (AD, MD, RD, and FA) in diagnosing the specific segment responsible for somatic neurological dysfunction in CSM, employing electrophysiology as a tool to validate neurological function.

Material and methods

Basic information

Patients were enrolled based on the following criteria. Inclusion criteria included: (1) Patients with comprehensive examination data, including physical examination, imaging findings, and mJOA scale; (2) Patients who underwent intraoperative electrophysiological examination during posterior cervical spine surgery. Exclusion criteria included: (1) Patients under 18 years old; (2) Patients with a history of prior spinal surgery; (3) Patients with a history of tumors or tuberculosis; (4) Patients with a history of multiple traumas; (5) Patients with claustrophobia.

Following these criteria, 19 patients with CSM were included, along with 10 healthy individuals serving as controls with no symptoms or compressions observed on MRI examination between July 2020 and July 2021. The control group comprised 5 males and 5 females, while the CSM group included 10 males and 9 females (Table 1). The individuals in the control group were confirmed to be healthy through physical examination and MRI. In contrast, the CSM patients exhibited clinical neurological deficits and cervical spinal cord compression, as indicated by MRI findings.

Table 1 Demographic data of included cases
Full size table

The average age of all subjects was 56.14 ± 12.41 years, with an average BMI of 25.99 ± 3.87 kg/ m2. The control group had an average age of 43.50 ± 7.21 years, while the CSM group had an average age of 62.79 ± 8.85 years. The BMI for the control group was 24.31 ± 2.91 kg/ m2, compared to 26.79 ± 4.08 kg/ m2 for the CSM patients. There were no statistically significant differences in gender distribution and BMI [(24.31 ± 2.91) kg/m2 vs (26.79 ± 4.08) kg/m2, (P = 0.115)] between the two groups. For blinded methodology, we blinded the healthy population and data measures. As the surgery was an invasive procedure and there was a clear difference between the control group and the patient group, the patients, the chief physician of the physical examination, and the operator were not blinded. This study was approved by the local Ethics Committee. Written informed consent was obtained from all subjects involved.

Clinical examination

Clinical routine physical examinations, including sensory nerve assessments, muscle strength evaluations, tendon reflex evaluations, and pathological reflex examinations, were conducted as evaluation references in both the control and CSM groups. These examinations were performed by a chief physician [X.H.] and a senior attending physician [JC.W.], both of whom specialize in the field of cervical disease. The mJOA scoring system was employed to quantify neurological function (Appendix Table 1).

Anatomical MRI and DTI data acquisition

Cervical spine anatomical MRI and DTI examination were conducted. For both the control group and CSM group, MRI scans were performed using a Philips 3.0 T Ingenia MRI scanner (Philips, Best, The Netherlands) equipped with a 16-channel head/neck coil. T1-weighted and T2-weighted axial and sagittal images were acquired with the following parameters. T1-weighted images were obtained with a repetition time of 400 ms, and an echo time (TE) of 6.85 ms. The field-of-view (FOV) is 250 × 250 mm2, with an image resolution of 0.7 × 0.8 mm2 and a slice thickness of 3-mm. Sagittal T2-weighted images were obtained with TE/TR = 100/3,000 ms, maintaining the same FOV and resolution as sagittal T1-weighted images. For axial T2-weighted images, the in-plane resolution is 0.9 × 0.9 mm2 with a slice thickness of 4 mm, FOV = 190 × 190 mm2, TE/ TR = 70/5286 ms. The hyperintense signal on T2W MRI was assessed and identified by a radiologist and a chief physician [X.H.] from the spine department. The aspect ratio (defined as the anteroposterior diameter / the transverse diameter of the corresponding segment's spinal cord) and transverse ratio (defined as the transverse diameter of the corresponding segment's spinal cord / the transverse diameter of the spinal cord at C2/3) were calculated as the average of the results independently measured by two residents. Both residents conducted three repeated measurements to minimize measurement error (Fig. 1).

Fig. 1
figure 1

Basic information and definitions of the parameters. a and b Conventional MRI of the cervical spine. c The aspect ratio was defined as the anteroposterior diameter (m) / the transverse diameter of the corresponding segment's spinal cord (n). d The transverse ratio was defined as the transverse diameter of the corresponding segment's spinal cord (n) / the transverse diameter of the spinal cord at C2/3 (n’). e The SCT spinal cord zoning diagram. f Demonstration of DTI spinal tractography

Full size image

The DTI datasets were acquired using zonally-magnified oblique multislice echo planar imaging (ZOOM EPI) with a restricted FOV on the axial view. A total of twenty-three slices were scanned, covering spatial areas from C2 to C7. The specific imaging parameters were as follows: FOV = 56 × 42 mm2, in-plane resolution = 0.88 × 0.88 mm3, slice thickness = 4 mm, number of signal averaging (NSA) = 1, b = 0/800 s/mm2, 32 diffusion directions, and TR = 4 heartbeats with peripheral pulse unit (PPU) triggering (trigger delay = 300 ms). The total scan time was approximately 5 min depending on the heart rate (Figs 2, 3, 4 and 5)

Fig. 2
figure 2

a The automatic registration and segmentation results of the spinal cord for each slice from a particular patient. The co-registered white matter (depicted in blue) and gray matter (represented in yellow/red) were overlaid on the b0 image. The first slice was located at the C7 cervical vertebra, and the last slice was at C1. Note that the results were reliable for both the severely compressed slice b and the non-compressed slice c

Full size image
Fig. 3
figure 3

The T2W anatomical MR image along with calculated DTI metrics and color encoded FA maps (cFA) for a non-compressed slice (Slice No. 1) and a compressed slice (Slice No. 2)

Full size image
Fig. 4
figure 4

Procedures of spinal cord electrophysiological examination. a and b showed radiographs after placing the deep spinal cord electrode on dura. c Placement of the spinal cord deep electrode. d Electrophysiological data of the patient. The greatest decrease of SSEP amplitude% could be seen at the 11th electrode points. The corresponding spinal cord segment was considered as the segment with the most serious impact on spinal cord function

Full size image
Fig. 5
figure 5

ROC curves for the diagnosis of the responsible segment for CSM. a DTI metrics of whole spinal cord for C4/5. b DTI metrics of dorsal column for C4/5. c DTI metrics of whole spinal cord for C5/6. d DTI metrics of dorsal column for C5/6. ROC: receiver operating characteristic

Full size image

Data were processed using the Spinal Cord Toolbox [19] following a specific procedure. Initially, the diffusion dataset was automatically segmented and motion-corrected. Next, it was registered to the PAM50 template [20]. Subsequently, the DTI metrics were computed and tract-specific DTI metrics were extracted with various predefined regions of interest (ROIs) based on the atlas in the toolbox. Four ROIs were selected, including the whole spinal cord, dorsal columns, white matter, and gray matter. For quality control, the ROI definition was subsequently evaluated visually by overlaying the co-registered atlas template on the mean DWI image. Finally, morphologic features from anatomical MRI (e.g., aspect ratio, transverse ratio) and DTI metrics (AD, MD, RD, and FA) were then analyzed using the receiver operating characteristic curve (ROC) to evaluate their diagnostic efficacy in determining the responsible segment of CSM.

Electrophysiological data acquisition and analysis

In the patients of the CSM group, intraoperative spinal cord electrophysiological examination was conducted for research purposes to assess spinal cord function during surgery. All CSM patients underwent spinous process splitting laminoplasty [21]. The operation procedures were as follows: A CSM patient was positioned in a prone posture under general anesthesia. A midline incision was made in the back of the neck to expose the spinous processes of segments C3–C7, while preserving the muscle insertion points of the C2 and C7 spinous processes. Subsequently, a single spinal cord deep electrode (PMT/2102-16-091, Chanhassen, MN, USA) was carefully positioned in the midline of the dural surface before decompression, allowing for the acquisition of electrophysiological signals in closer proximity to the spinal cord, thus enhancing result accuracy. The intraoperative monitoring system (Cascade 32 channels, North Kellogg St. Kennewick, WA, US) was used to collect the somatosensory evoked potentials (SSEP). Stimulation electrodes were placed on the median nerves of both wrists, using a stimulation intensity of 20 mA, a stimulation frequency of 4.13 Hz, 200 superimposition times, and a band-pass filter of 10–300 Hz. The electrode position was determined using C-arm radiographs. The responsible segment was determined as the spinal cord segment where the recorded potential exhibited the greatest decrease amplitude% or the largest prolonged latency% [22]. This electrophysiological examination served as an objective measure of neurological function.

Statistical analysis

The data were analyzed using Stata/MP 14.0 (College Station, TX, USA). One missing value in the control group’s BMI was deleted. The Shapiro–Wilk normal distribution test was performed for the measurement data. Measurement data (age, BMI, mJOA scores) that conformed the normal distribution were expressed as mean ± standard deviation (SD). Independent samples t‐tests were used for comparisons between the two groups. Data that did not meet the normal distribution were presented as the median and interquartile range (p25–p75) and were compared using the nonparametric Mann–Whitney U test between the two groups. Numerical data were expressed as percentages. Comparisons between the two groups were made using the Chi-square test or Fisher’s exact test. Intraclass correlation coefficient (ICC) and Kappa were calculated for the assessment of intra/interobserver reliability. Additionally, the area under the curve (AUC), sensitivity, specificity, and Youden index% were computed to evaluate the diagnostic performance. All tests were two-sided, and P values less than 0.05 were considered significant.

Results

The responsible segment determined by electrophysiological examination.

According to the results of the electrophysiological examination, neurological dysfunction was found in 2 patients due to C3/4 compression, in 10 patients due to C4/5 compression, in 6 patients due to C5/6 compression, and in 1 patient due to C6/7 compression (Table 2).

Table 2 Responsible segment for the neurological dysfunction determined by different methods of CSM group
Full size table

The responsible segment determined by physical examination.

According to the results of the physical examination, sensory nerve dysfunction was found in 13 patients due to C3/4 compression, in 3 patients due to C4/5 compression, and in 2 patients due to C5/6 compression. One patient did not exhibit any sensory nerve dysfunction. (Table 2). Motor nerve dysfunction was found in 6 patients due to C3/4 compression, in 6 patients due to C4/5 compression, and in 2 patients due to C5/6 compression. Five patients did not exhibit any motor nerve dysfunction. The overall mJOA score in the CSM group was 12.71 ± 1.55. Specifically, the mJOA score for sensory nerve function was 4.87 ± 0.72, and the mJOA score for motor nerve function was 5.05 ± 1.35. In contrast, none of the volunteers in the control group exhibited any neurological dysfunction. The overall mJOA score was 17 in the control group. The average mJOA score of sensory nerve function was 6 and the average mJOA score of motor nerve function was 8.

Diagnostic efficacy of MRI parameters for the the responsible segment

According to anatomical MRI examination, the most stenotic segment was C3/4 in 7 patients, C4/5 in 7 patients, C5/6 in 4 patients, and C6/7 in one patient. T2 high signal was found in 13 patients at the most stenotic segment (Table 2). None of the volunteers in the control group had compression. Upon a comprehensive evaluation of all cervical segments, the aspect ratio (AUC = 0.823, P = 0.001, Sensitivity = 68.42%, Specificity = 82.47%) proves more effective than the transverse ratio (AUC = 0.661, P = 0.027, Sensitivity = 68.42%, Specificity = 67.01%) in identifying the specific responsible segment (Table 3). Further subgroup analysis revealed that the diagnostic efficacy of MRI parameters was influenced by cervical spine segment. Since there were fewer C3/4 and C6/7 responsible segments in this study, the subgroup analysis focused on C4/5 and C5/6 segments, which had a higher probability of compression according to clinical work. As for C4/5, AUC was 0.837 for the aspect ratio (P = 0.003, Sensitivity = 90.00%, Specificity = 73.68%) and 0.7 for the transverse ratio (P = 0.081, Sensitivity = 60.00%, Specificity = 94.74%). As for C5/6, AUC was 0.674 for the aspect ratio (P = 0.196, Sensitivity = 50.00%, Specificity = 89.96%) and 0.565 for the transverse ratio (P = 0.628, Sensitivity = 66.67%, Specificity = 56.52%), respectively.

Table 3 Diagnosis efficacy of anatomical MRI parameters in determining responsible segment
Full size table

Diagnostic efficacy of the DTI metrics for the responsible segment

Among all DTI (Whole spinal cord) parameters, when taking all cervical segments into consider, AD (AUC = 0.822, P = 0.001, Sensitivity = 89.47%, Specificity = 77.32%) performed the best diagnostic efficacy, better than MD (AUC = 0.806, P = 0.001, Sensitivity = 78.95%, Specificity = 76.29%), RD (AUC = 0.783, P = 0.001, Sensitivity = 89.47%, Specificity = 60.82%) and FA (AUC = 0.750, P = 0.001, Sensitivity = 84.21%, Specificity = 54.64%) (Table 4). In subgroup analysis, for C4/5 subgroup, the AUC of the DTI metrics across the whole spinal cord were 0.732 for AD (P = 0.044), 0.768 for MD (P = 0. 019), 0.742 for RD (P = 0.035), 0.726 for FA (P = 0.049), respectively. Similarly, for C5/6 subgroup, the AUC of the AD, MD, RD, and FA across the whole spinal cord was 0.949 (P = 0.001), 0.884 (P = 0.004), 0.848 (P = 0.010), 0.884 (P = 0.004), respectively.

Table 4 Diagnosis efficacy of DTI (Whole spinal cord) parameters in determining responsible segment
Full size table

Among all DTI (Dorsal column) parameters, when taking all cervical segments into consider, AD (AUC = 0.823, P = 0.001, Sensitivity = 84.21%, Specificity = 77.32%) performed the best diagnostic efficacy, better than MD (AUC = 0.816, P = 0.001, Sensitivity = 89.47%, Specificity = 71.13%), RD (AUC = 0.791, P = 0.001, Sensitivity = 89.47%, Specificity = 58.76%) and FA (AUC = 0.772, P = 0.001, Sensitivity = 73.68%, Specificity = 69.07%) (Table 5). In subgroup analysis, for the C4/5 subgroup, the AUC of the DTI measures in the dorsal column region are 0.732 for AD (P = 0.044), 0.763 for MD (P = 0.022), 0.747 for RD (P = 0.031), 0.737 for FA (P = 0.039), respectively. Similarly, for the C5/6 subgroup, the AUC of the DTI metrics for the dorsal column was 0.949 for AD (P = 0.001), 0.928 for MD (P = 0.002), 0.920 for RD (P = 0.002), 0.913 for FA (P = 0.002), respectively.

Table 5 Diagnosis efficacy of DTI (Dorsal column) parameters in determining responsible segment
Full size table

5. Summary of the effectiveness of non-invasive examinations (physical examination, anatomical MRI, DTI) in identifying the responsible cervical spine segment.

The segments identified as responsible through electrophysiological examination showed agreement with sensory nerve examination in 5 (26.32%) patients and with motor nerve examination in 7 (36.84%) patients.

Additionally, 11 (57.89%) patients with neurological dysfunction, based on the most stenotic segment on MRI, were consistent with the segments identified through electrophysiological examination. Similarly, ten (52.63%) segments exhibiting high signal on MRI were in concordance with the segments determined responsible by electrophysiology.

Comparing the responsible and irresponsible segments, it was found that AD, MD, and RD were significantly higher, whereas FA was significantly lower in the responsible segments compared to the irresponsible segments (P < 0.05). Overall, the AUC of DTI (Dorsal Column) parameters (AD, MD, RD, FA) was higher than the corresponding parameters of DTI (Whole spinal cord). Notably, axial diffusivity of DTI (Dorsal Column) exhibited the largest AUC (0.823), along with superior sensitivity (84.21%) and specificity (77.32%) among the parameters considered (Table 6). These results suggest that DTI (Dorsal Column) shows promising potential for determining responsible segments in non-invasive examinations.

Table 6 Summary of Diagnostic efficacy MRI and DTI for the responsible segment
Full size table

Discussion

In the context of CSM with multiple segments of compression, the use of invasive intraoperative neurophysiological examination has proven valuable as a reliable and objective measure of spinal cord function. This method is essential as physical neurological examinations often struggle to precisely identify the specific responsible segment of compression that leads to neurological dysfunction. A study highlighted that SSEP revealed relatively fewer spinal segment functional abnormalities compared to the findings from traditional MRI, which primarily provided information on the anatomical structure of the spinal cord [23]. This suggested that not all compressed segments contribute to the observed neurological dysfunction. In scenarios where multiple segments are compressed, it is possible that only one segment primarily contributes to the spinal cord dysfunction. Hence, identifying the precise responsible spinal cord segment affecting nerve function becomes crucial for making well-informed and appropriate surgical decisions.

Several studies have focused on assessing the diagnostic performance of DTI met for CSM by reporting sensitivity and specificity. The most extensively studied FA and AD were discussed here [14, 24]. Wang's study reported a sensitivity of 30.77% and specificity of 91.89% for the FA to identify myelopathic segments in CSM [25]. Our results were much better in some ways (DTI-dorsal column, sensitivity = 73.68%, specificity = 69.07%, Youden index = 42.75%). Furthermore, it has been reported that FA exhibits the largest AUC of 0.750 [24], which was lower than our results (AUC: AD = 0.823, FA = 0.772). Few studies computed the AUC to assess the efficacy of DTI in diagnosing CSM. To some extent, CSM can be considered as a chronic neurological injury [26]. To provide further context, we also considered DTI findings from cases of cervical spine injuries for comparison. Previous studies have indicated that in cases of spinal cord injury, FA (AUC = 0.770) at the lesion epicenter could effectively discriminate injury severity [27]. Moreover, when combining FA, AD, and RD, the AUC was reported to be 0.92 for diagnosing cervical spine injury [28]. Notably, the results of series or parallel diagnostic examinations using FA and AD were less effective compared to those using FA or AD alone. However, it is important to acknowledge the differences that exist between acute spinal cord injury and CSM.

In addition, our study involved the utilization of the Spinal Cord Toolbox (SCT) to partition the nerve conduction tracts of the spinal cord and measure the corresponding DTI metrics. Interestingly, we discovered that, in comparison between the whole spinal cord and the dorsal column tract (includes the fasciculus cuneatus and fasciculus gracilis), the dorsal tract exhibited superior diagnostic performance to the whole spinal cord. This finding is in line with the results of a study conducted by Jan Valošek et al.[18]. Different results between spinal cord tracts may be related to the pathogenesis of the disease, as in clinical practice, the sensory function is usually the initial problem. This might also be related to the use of SSEP which was generated from the dorsal sensory functional area as an evaluation function index in our electrophysiological studies. In our study, when considering all cervical segments, AD from the DTI-Dorsal Column exhibited the most significant potential in identifying responsible segments, superior to parameters of DTI-Whole spinal cord, aspect ratio, the most stenotic segment, T2 high signals, transverse ratio, motor nerve dysfunction, and sensory nerve dysfunction. These results suggested that AD from DTI (Dorsal Column) showed promising potential for determining responsible segments in non-invasive examinations (Table 6).

The discrepancies observed between our study and other existing research may be attributed to the use of different standards for assessing spinal cord function. While previous DTI studies primarily relied on the mJOA score as the standard for evaluating spinal cord function [24]. Our study employed intraoperative neurophysiological monitoring as an invasive examination method to detect neurological function, rather than relying solely on the mJOA score [14]. Notably, the direct attachment of the probe to the surface of the spinal cord during our intraoperative neurophysiological monitoring resulted in reduced signal interference. This approach enhanced the objectivity of neurological function evaluation compared to the use of the mJOA score alone. During cervical spine surgery, we recorded SSEP generated from the dorsal sensory functional area of the spinal cord at the corresponding cervical spine segment. The electrodes we used were positioned closer to the source of the generated neurological signals compared to electrodes placed on the scalp, thereby reducing noise signals from specific structures along the pathway [29, 30]. Consequently, changes in the amplitude or latency of the potential may be more sensitive in reflecting the degree of neuronal damage in the spinal cord caused by mechanical compression [31]. These methodological differences may have contributed to the variations observed between our study and previous research.

Additionally, our study highlighted that the diagnostic efficacy of both DTI and MRI parameters was significantly influenced by the specific segment of the cervical spine. Notably, our clinical observations consistently indicated that patients with disc herniation at the C4/5 or C5/6 level were more prevalent. Interestingly, based on our findings, the diagnostic efficacy of MRI and DTI in assessing neurological deficits at the higher segment (C4/5) were similar. Conversely, when diagnosing neurological dysfunction resulting from the lower segment (C5/6), DTI exhibited a generally superior diagnostic effect compared to anatomical MRI. Several factors potentially accounted for this discrepancy. Firstly, it is commonly observed that CSM patients often suffer from compression in multiple segments. The impact of the upper segment's compression might have an additive effect on the DTI metrics of the lower segment, leading to more statistically significant results in the lower segment. Secondly, the degree of compression at the C4/5 level was notably higher than that at the C5/6 level in our study (aspect ratio: 0.238 (0.204–0.286) versus 0.304 (0.211–0.359); transverse ratio: 1.228 (1.121–1.263) versus 1.111 (1.029–1.146)). This discrepancy might indicate that DTI metrics possess a higher diagnostic utility in confirming responsible segments compared to anatomical MRI parameters, particularly in patients with milder compression. This finding could assist surgeons in choosing appropriate surgical segments and minimizing surgical injury. In cases of more severe compression, the responsible segment might be multiple, necessitating more extensive multi-segment surgery.

There are several limitations in this study. Firstly, a notable statistical difference in age (t = − 5.92, P < 0.001) was observed between the two groups. While some studies have reported that DTI values are not significantly influenced by age in subjects older than 40 years (P < 0.05) [32,33,34], it is still imperative to consider an age-matched cohort for more conclusive findings in future studies. Secondly, to further explore the diagnostic efficacy under various degrees of compression, it is crucial to grade the severity of compression. Lastly, it should be noted that the members of the healthy control group were confirmed to be in good health and without cervical disorders only based on physical examination and MRI, electrophysiological testing was not conducted due to ethical considerations.

Conclusion

When considering all cervical segments, AD from the DTI-Dorsal Column exhibited the most significant potential in identifying responsible segments. This potential was found to be superior to that of DTI-Whole spinal cord, aspect ratio, the most stenotic segment, T2 high signals, transverse ratio, motor nerve dysfunction, and sensory nerve dysfunction. The diagnostic effectiveness of both DTI and MRI parameters was notably influenced by the specific cervical spine segment.