In our current study, three major findings were observed: (1) Relative to healthy participants, Degenerative Cervical Myelopathy (DCM) patients exhibited lower threshold for pain; (2) and altered Middle Cingulate Cortex (MCC) function was associated with pain threshold which is also tightly correlated with the postoperative neck pain intensity; (3) Further, the ALFF of MCC provided additional value for predicting the occurrence of postoperative axial pain via machine learning analysis in DCM patients.
In comparison to healthy participants, DCM patients exhibited lower pain threshold; and patients with postoperative axial pain showed lower pain threshold preoperatively than those without.
In analysis 1, we found that the pain thresholds of DCM patients were significantly lower than healthy participants, and the pain thresholds were also lower in DCM patients with PAP than those without. This finding is in line with previous study conducted by Zhang et.al. in which they conducted quantitative sensory testing and revealed that patients with PAP have a lower pressure pain threshold and temporal summation (i.e., higher sensitivity to pain perception) than patients without PAP. Their findings indicated that preoperative endogenous pain modulation deficiency might be associated with axial pain after posterior cervical decompression [9]. It is not surprising that DCM patients developed abnormal pain modulation system, considering most of the patients experienced chronic pain that is associated with modifications of the central nervous system, such as central sensitization, which is responsible for alterations in pain sensitivity in acute and chronic pain situations [26,27,28]. We also found that preoperative pain threshold was negatively correlated with pre/post-operative pain intensity and preoperative pain intensity was positively correlated with postoperative pain intensity in DCM patients. These findings also supported the idea that DCM patients developed central sensitization following long-term axial pain which further aggravates or induce the postoperative axial pain.
Altered MCC function was associated with preoperative pain threshold and PAP intensity in DCM patients
In analysis 2, we found that relative to healthy participants, DCM patients exhibited significantly higher ALFF within Middle Cingulate Cortex (MCC) and Superior Frontal Gyrus (SFG), and the MCC ALFF were correlated with both preoperative pain threshold and PAP intensity. MCC, which is frequently activated during acute pain, has been shown to be responded specifically to nociceptive input from subcortical brain regions. Additionally, chronic pain also causes grey matter changes in MCC, and such changes overlaps in various chronic pain condition indicating the structural alterations of MCC could well be the biological marker for chronic pain per se [29]. Further, Davis et.al. found that greater heat pain sensitivity (i.e., lower heat pain threshold) correlated with thickening in the mid-cingulate cortex, which indicated MCC is responsible for detecting and processing nociceptive input [30]. From the functional aspect, in addition, the MCC is an important component of the cingulate-insular pathway which gates and maintains nociceptive hypersensitivity in the absence of conditioned noxious stimuli and affects the impact of pain [31]. Taken together, our observed association between pain sensitivity and MCC ALFF support the hypothesis that continuous nociceptive input causes MCC cortical reorganisation which further induces hypersensitivity in chronic pain patients.
Furthermore, we also observed significant altered ALFF within SFG, M1 and V1. These results were in line with previous reports. Kaito et.al. conducted rs-fMRI and found that the ALFF within SFG and V1 were altered in DCM patients [32]. They concluded that these brain alterations were considered as the functional reorganisation following long-term chronic spinal cord injury. We also found that relative to healthy participants, DCM patients exhibited significantly lower ALFF within primary motor cortex (M1). M1, a key region in the sensorimotor network, is involved in a various of motor functions, such as motor planning, inhibition, coordination, movement, and so on [33, 34]. It has been shown that the ALFF within M1 was significantly higher in DCM patients than healthy controls, and was also tightly correlated with the fractional anisotropy value of C2 segment which reflects the severity of myelopathy [33]. Our previous study also illustrated the potential utility of M1 ALFF for predicting the prognosis of DCM patients following decompression surgery [35]. Our current finding was in line with the previous reports, indicating potential cortical reorganisation occurs in DCM [36,37,38,39].
MCC ALFF provides additional value for PAP vs. nPAP classification
Posterior cervical decompression surgery is one of the most widely used surgical approaches, and increasing frequency of PAP after posterior decompression approach seriously affects the daily life of patients. Till now, there is still controversy about the causes of PAP and its related factors. Atsushi et al. showed that anterolithesis, current smoking, preoperative neck pain, etc. are influencing factors of axial pain after laminoplasty [7]. A systematic review summarises possible factors influencing axial pain after posterior surgery, including age, preoperative axial pain, different surgical techniques, and postoperative management [8]. It has been shown that about 40% of patients experienced axial pain after laminoplasty, but it occurred mostly in those who had preoperative axial pain [7, 8, 40]. Although multiple factors have been identified as causal factors in PAP, preoperative neck pain severity is the most commonly reported PAP marker in DCM patients, and our current results also confirmed this. Further, our univariate and multivariate analysis also identified the neural correlates of the PAP, which is associated with the pain sensitivity in DCM patients. It is also worth mentioning that despite we optimally matched the preoperative pain intensity between PAP and nPAP patients, PAP still exhibited significantly higher MCC ALFF than nPNP patients. This result indicated the independent contribution of MCC function in altered pain modulation pathway which related to hypersensitivity in PAP patients. As to the clinical implications of our findings, identifying patients with PAP could aid the clinicians to develop novel perioperative management to reduce or avoid such complication based on hypersensitivity in these patients. Preoperative analgetic use has been proved to be effective in reducing postoperative pain intensity for many other orthopedic surgeries [41,42,43]. Such perioperative preparation could reduce the central sensitization and thus relieve the pain following large trauma.
Limitation
First, the main limitation is that our patients have all received medication treatment including non-steroid-anti-inflammatory drug, etc. This may affect our results to some extent. Therefore, future studies with DCM patients who are not on medication or who have a washout period from medication are needed to confirm our findings. Postoperative fMRI data was not collected due to the possibility of artifact and heating due to surgical implants. Even though it appears to be safe and other studies have collected data on postoperative fMRI data, the majority of our patients declined to cooperate after we informed them of potential harm (e.g., loss of surgical implants) associated with postoperative fMRI. Our current study only analysed ALFF alterations between patients and healthy controls, other resting-state fMRI metrics such as functional connectivity (FC), regional homogeneity (ReHo), functional connectivity strength (FCS), need further study. Socioeconomic status is a crucial factor affecting pain process between individuals, but was not collected and thus its possible association with pain perception could not be investigated in the present study.