1 Introduction

Multilevel pincer-type cervical spondylotic myelopathy (CSM) involves simultaneous ventral and dorsal compression of the spinal cord and is commonly identified in middle-aged and older patients with cervical degeneration. The condition develops through the combined action of various age-related structural changes, including disc protrusion, posterior vertebral osteophytes, hypertrophy of the ligamentum flavum, ossification of the posterior longitudinal ligament, and cervical kyphotic deformity.¹ These factors generate bidirectional compression, producing a “pincer-like” configuration that broadens the region of spinal cord compromise, reduces the sagittal diameter of the canal, disrupts cord perfusion, and ultimately leads to progressive neurological impairment.² With the aging of the population and the widespread use of spinal imaging, multilevel pincer-type CSM is being identified with increasing frequency, and the choice of an appropriate operative strategy has become a prominent issue in clinical practice.

Patients with single-level pincer-type disease often experience meaningful neurological improvement after anterior decompression and fusion. The procedure eliminates ventral compression directly, and restoration of disc height together with cervical lordosis can lessen dorsal buckling or hypertrophy of the ligamentum flavum, which in turn reduces part of the posterior compressive load.³ In patients with extensive disease or posterior vertebral lesions, anterior cervical corpectomy and fusion (ACCF), or hybrid constructs combining ACCF and anterior cervical discectomy and fusion (ACDF), can widen the decompression corridor while maintaining segmental stability.^4,5 When three or more segments are affected and marked ventral and dorsal compression coexist, substantial debate persists regarding the choice among anterior-only, posterior-only, or combined approaches. Clinical observations indicate that residual dorsal compression may remain after anterior-only procedures in a subset of patients, resulting in suboptimal neurological recovery or the need for secondary posterior decompression.⁶ In contrast, combined approaches enable more comprehensive circumferential decompression but introduce higher operative trauma and complication risk.^7,8 Striking an appropriate balance between decompression adequacy and procedural safety remains a central problem, and the absence of quantitative and reproducible preoperative indicators forces decision-making to rely heavily on surgeon experience.

Identifying patients unlikely to benefit from an anterior-only approach is therefore a major focus in cervical spine surgery. Several radiological parameters—such as the Torg–Pavlov ratio, the K-line, and cervical sagittal alignment—have been investigated for decision-making. Among these, the K-line reflects the spatial relationship between canal diameter and vertebral alignment and is widely referenced when assessing candidacy for posterior laminoplasty; however, its diagnostic performance diminishes in multilevel disease with concurrent ventral and dorsal compression, limiting its role as a standalone criterion. Other studies have attempted to incorporate cervical curvature, disc height, and intramedullary signal alterations into composite assessments, yet these methods remain difficult to standardize and reproduce in daily clinical practice. As a result, consensus on the optimal operative strategy is still lacking. An anterior approach directly removes ventral pathology and restores cervical alignment but becomes insufficient when dorsal compression is substantial⁹; a posterior approach decompresses dorsally and allows spinal cord drift-back yet cannot adequately address prominent ventral osteophytes or disc protrusions; combined anterior–posterior procedures achieve robust decompression but increase surgical trauma and complication risk.¹⁰ Evidence from both clinical and randomized studies indicates that operative approach alone does not fully determine outcomes, reinforcing the need for a concise, objective, and quantifiable tool to help select patients who are appropriate candidates for anterior-only treatment.

Against this background, we introduced the posterior compression score (PCS) as a quantitative tool designed to characterize the extent of dorsal cord compression in segments affected by the pincer configuration. PCS classifies the severity of posterior compromise on preoperative magnetic resonance imaging and provides a direct representation of the dorsal pathological burden, with favorable reproducibility and practical applicability in routine clinical evaluation. In contrast to earlier indicators that emphasized cervical alignment or sagittal canal dimensions, PCS concentrates specifically on dorsal compression—a determinant that strongly influences the effectiveness of an anterior-only approach—and may therefore supply an objective reference for preoperative decision-making. We hypothesized that a higher PCS reflects substantial posterior compression, indicating that anterior decompression alone may be inadequate and that neurological recovery could remain limited or require supplementary posterior intervention. Conversely, patients with a lower PCS were expected to be more suitable candidates for anterior-only surgery and more likely to achieve meaningful postoperative neurological improvement.

Guided by this framework, the present study retrospectively examined clinical and imaging data from patients with multilevel pincer-type CSM to investigate whether PCS can help identify individuals appropriate for an anterior-only operation and to explore its potential usefulness in preoperative planning.

2 Methods

2.5

2.5 PCS

PCS was applied to quantify dorsal compression of the spinal cord based on preoperative cervical MRI across the C2/3 to C6/7 segments. Grading criteria were defined as follows: 0 points: absence of dorsal compression with preserved dorsal cerebrospinal fluid signal; 1 point: soft-tissue compression at the interlaminar level, typically resulting from hypertrophy of the ligamentum flavum; 2 points: bony compression at the laminar level, related to laminar hypertrophy, osteophytes, or posterior extension of OPLL. The segmental grades were summed to derive the total PCS (Fig. 1). Two spine surgeons assessed each case independently, and discrepancies were adjudicated by a third reviewer. Interobserver reliability was examined using Cohen's κ.

Close

Fig. 1 Presents the sagittal configuration of pincer-type CSM together with the grading scheme for PCS.

Export to PPT

3 Results

3.1

3.1 General findings

In this study, 148 patients with multilevel pincer-type CSM were included, comprising 85 men and 63 women, with a mean age of 58.6 ± 9.7 years and a mean follow-up of 12.8 ± 3.3 months. The mean preoperative PCS was 2.7 ± 1.5 (range 0–8). ROC analysis indicated a strong predictive capacity of PCS for postoperative neurological recovery, and a score of 2 yielded the highest accuracy for identifying patients who achieved a JOA recovery rate ≥50% (AUC = 0.85) (Fig. 2). Based on this threshold, patients were assigned to a low-score group (PCS ≤2, n = 91) and a high-score group (PCS >2, n = 57).

Close

Fig. 2 ROC analysis demonstrating the predictive value of PCS for achieving a JOA recovery rate ≥50%.

Export to PPT

Preoperative evaluations showed a mean JOA score of 9.1 ± 2.3, NDI of 38.6 ± 10.7, VAS neck pain of 5.2 ± 1.8, and VAS arm pain of 4.9 ± 1.7. The preoperative C2–C7 lordotic angle averaged 11.7° ± 7.9°, and ROM was 31.5° ± 9.4°. Four patients (2.7%) underwent secondary posterior decompression within 24 h due to marked residual dorsal compression or neurological deterioration; all four belonged to the high-score group (Fig. 3).

Close

Fig. 3 Representative imaging of patients requiring posterior reoperation within 24 h (A1 preoperative MRI with PCS grading; A2 sagittal view showing ventral–dorsal compression at C5/6; A3–A4 demonstrating absence of OPLL; B1–B2 postoperative anterior-only images revealing persistent dorsal compression with intramedullary edema; C3–C4 immediate posterior decompression within 24 h; C1–C2 three-month follow-up MRI showing focal intramedullary necrosis).

Export to PPT

To limit potential selection bias, PSM was carried out, and the matched cohorts showed no significant differences in demographic or radiological characteristics (Table 1).

Table 1 Comparison of baseline demographic and clinical characteristics between groups before and after PSM.

Variable	Unmatched High-Score Group (n = 57)	Unmatched Low-Score Group (n = 91)	P value	Matched High-Score Group (n = 53)	Matched Low-Score Group (n = 53)	P value
Age (years)	58.39 ± 10.79	56.54 ± 10.63	0.310	58.55 ± 10.65	59.23 ± 12.11	0.760
Sex	F:14; M:43	F:49; M:42	0.001	F:13; M:40	F:12; M:41	1.000
BMI	23.70 ± 1.78	24.23 ± 1.77	0.083	23.60 ± 1.71	23.41 ± 1.37	0.532
Smoking	0.35 ± 0.48	0.14 ± 0.35	0.006	0.34 ± 0.48	0.26 ± 0.45	0.402
Symptom duration (months)	13.30 ± 11.06	12.99 ± 8.52	0.857	13.42 ± 11.29	12.34 ± 7.21	0.560
K-line (+/−)	negative:14; positive:39	negative:17; positive:70	0.403	negative:14; positive:39	negative:12; positive:41	0.822
Torg–Pavlov ratio	0.81 ± 0.10	0.83 ± 0.09	0.341	0.82 ± 0.10	0.82 ± 0.11	1.000
Preoperative JOA	7.53 ± 0.73	7.75 ± 0.78	0.085	7.55 ± 0.75	7.85 ± 0.91	0.065
Preoperative NDI	33.63 ± 4.16	30.96 ± 4.07	0.000	33.62 ± 4.21	32.53 ± 3.37	0.143
Preoperative VAS—neck pain	5.44 ± 0.98	5.45 ± 1.04	0.944	5.47 ± 0.99	5.34 ± 0.83	0.459
Preoperative VAS—arm pain	5.33 ± 0.91	5.42 ± 1.10	0.614	5.38 ± 0.92	5.13 ± 1.07	0.211
C2–C7 lordotic angle (°)	10.24 ± 6.63	10.57 ± 7.03	0.773	9.90 ± 6.74	9.11 ± 7.71	0.574
ROM (°)	33.86 ± 1.69	33.64 ± 2.59	0.529	33.94 ± 1.69	33.38 ± 1.71	0.090
Follow-up duration (months)	12.53 ± 2.74	12.97 ± 3.63	0.403	12.57 ± 2.84	12.21 ± 2.40	0.485

4 Discussion

This study reviewed 148 individuals with multilevel pincer-type CSM and found that the PCS offered a dependable way to anticipate both the appropriateness of an anterior-only procedure and the likelihood of neurological improvement. When a threshold of 2 points was applied, the score produced balanced sensitivity and specificity for identifying patients who achieved a JOA recovery rate of at least 50%. Patients falling within the lower range of the scale (PCS ≤2) showed clearer gains in postoperative JOA scores, higher recovery rates, and more substantial reductions in NDI than those with higher values. Differences in VAS scores, the C2–C7 curvature, and ROM were minimal between groups. These tendencies remained stable after PSM. All four instances requiring early reoperation occurred among individuals with higher PCS values, which reinforces the clinical relevance of this score.

Debate surrounding the operative route for CSM with pronounced ventral–dorsal compression has persisted for years. Decompression from a single direction inevitably favors the side of entry, and when substantial encroachment remains elsewhere, the spinal cord often fails to recover fully. In settings marked by a high degree of circumferential narrowing—such as advanced OPLL or a pronounced pincer configuration—unilateral decompression may even precede postoperative neurological decline. Reports have referred to this pattern as white cord syndrome, with ischemia–reperfusion injury frequently considered a contributing mechanism.¹¹ Terminology aside, these observations highlight the necessity of a structured preoperative appraisal of the compressive load. The PCS introduced in this study provides a segment-by-segment depiction of dorsal pathology and offers an avenue to identify cases at risk of insufficient decompression with a single operative route. Earlier decisions often rested on the number of diseased segments, sagittal alignment, or K-line status¹²; however, such morphological descriptors do not capture posterior compression. The PCS incorporates both soft-tissue and bony components on the dorsal side and demonstrates practical reproducibility. Its performance in predicting JOA recovery, reflected by an AUC of 0.85, demonstrates its potential value in both the selection of operative approach and anticipation of postoperative neurological change.

Individuals with higher PCS values tended to show limited postoperative neurological improvement, and several mechanisms may account for this pattern. The loss of a dorsal cerebrospinal fluid buffer places the cord directly against the lamina or a markedly thickened ligamentum flavum. Even when ventral pathology is relieved through an anterior procedure, the cord may not gain enough posterior migration space, which restricts functional recovery. A high PCS frequently reflects long-standing dorsal encroachment, during which the cord may have undergone ischemic changes, demyelination, or even irreversible tissue injury—findings often paralleled by T2 hyperintensity on MRI. Although anterior fusion can re-establish vertebral height and restore cervical alignment, pronounced dorsal compression limits the extent to which these adjustments reduce the pincer configuration. The narrow operative corridor in such patients may also leave the cord under tension with compromised perfusion. Epstein et al. noted that in appropriately selected cases, posterior procedures carry a lower rate of complications than anterior operations and avoid risks associated with the esophagus, recurrent laryngeal nerve, and vertebral artery.¹³ Chen et al. described persistent under-decompression following anterior cervical discectomy and fusion in high-risk multilevel disease, particularly among patients with intramedullary signal changes or marked canal narrowing, with several ultimately requiring additional posterior decompression; combined laminoplasty and ACDF offered more pronounced neurological improvement and a higher rate of achieving a minimal clinically important difference.¹⁴ These reports suggest that in complex multilevel disease, reliance on anterior surgery alone may be inadequate. Findings from Wei et al.¹⁵ showed that among individuals with OPLL, those with PCS ≥8 benefited more from posterior laminoplasty, whereas individuals with values < 8 demonstrated poorer neurological recovery. Nonetheless, patients with PCS <8 still display heterogeneity, particularly in pincer-type CSM. When the PCS lies within the range of 1–2, ventral pathology typically predominates, and anterior decompression alone can provide a satisfactory outcome. Zhang et al., in a single-center review of 21 patients treated entirely with ACDF, observed notable gains in JOA scores after a 24-month follow-up, with a recovery rate of 65.9% and a fusion rate above 95%.¹⁶ Younger age and single-segment disease were associated with better results. These observations demonstrate that, for carefully selected patients—especially those with mild dorsal encroachment primarily composed of soft-tissue elements—ACDF remains a dependable and effective approach. Even so, indirect decompression achieved through anterior surgery may be incomplete in some situations.¹⁷ Lee et al. reported that among individuals with dorsal compression caused by hypertrophic or folded ligamentum flavum, only 58% exhibited improvement after ACDF, whereas 42% showed no relief and 3.4% experienced worsening.⁶ This indicates that indirect decompression is not consistently reliable in the cervical spine, particularly when dorsal narrowing is substantial. In this context, even patients with PCS <8 may face insufficient decompression if posterior pathology predominates. Thus, PCS values below this range should not uniformly lead to an anterior-only strategy; instead, the precise pattern and severity of compression should guide decisions, and posterior or combined routes should be considered when necessary.

For individuals with extensive multilevel disease in whom both ventral and dorsal encroachment are clearly present, neither anterior nor posterior decompression alone consistently ensures adequate relief. Qu et al. summarized that posterior laminoplasty may offer temporary symptom reduction, yet long-term follow-up frequently reveals functional decline or the need for secondary surgery. In contrast, a single-stage circumferential decompression demonstrated superior neurological outcomes and higher patient satisfaction over an 8-year follow-up period.¹⁸ These findings indicate that a combined strategy may offer a more dependable path to thorough decompression and sustained neurological benefit compared with reliance on a single operative route. In light of the present results, individuals with higher PCS values or clear bidirectional compression face a considerable risk of limited recovery when managed with an anterior-only procedure, and the possibility of residual stenosis or early reoperation becomes more prominent; in such situations, a tailored combined approach deserves careful consideration. In this cohort, patients in the lower category (≤2) achieved JOA recovery rates close to 60%, whereas those with values between 3 and 8 remained below 40%. NDI improvement followed a similar pattern. These tendencies remained consistent after PSM and in sex-stratified analyses, supporting the stability of PCS as a predictive factor. Although VAS scores improved in both categories, differences between groups were modest, which is expected since VAS largely reflects pain rather than the severity of spinal cord compromise and is more closely linked to nerve root irritation and inflammatory responses. By contrast, JOA and NDI reflect broader neurological and functional domains and therefore showed clearer group distinctions. All four early reoperations occurred among individuals with PCS values greater than 2, underscoring the importance of using PCS as a warning signal during preoperative planning. For patients who exceed this threshold, an anterior-only strategy requires explicit counseling regarding the likelihood of limited recovery and the possibility of a second-stage procedure.

This study retains several limitations. It was conducted at a single institution with a retrospective design and a modest sample size, which introduces inherent selection bias. Although PSM improved the balance of baseline characteristics, further confirmation from larger multicenter cohorts remains necessary. PCS grading relied on MRI interpretation; despite independent scoring by two clinicians and verification of agreement through κ statistics, some degree of subjectivity is unavoidable. The average follow-up period of 12.8 months restricts the ability to capture long-term outcomes, including the possibility of delayed neurological decline or fusion-related complications; extended follow-up is therefore required. In addition, PCS reflects only the dorsal component of compression and does not quantify ventral encroachment or integrate other biological indicators. Future prediction models may combine PCS with parameters such as the K-line, the Torg–Pavlov ratio, and intramedullary signal characteristics. PCS is also derived from static MRI and does not account for motion-dependent changes. As noted by Shin and colleagues,¹⁹ dynamic MRI may reveal extension-related cord compression and T2 hyperintensity more accurately and demonstrates stronger correlation with postoperative recovery, suggesting a potential role for dynamic imaging in future evaluation systems.

Advances in the understanding of pincer-type cervical pathology have encouraged the development of modified anterior techniques. Vertebral body sliding osteotomy (VBSO) and guttering osteotomy can achieve more extensive decompression in individuals with marked posterior encroachment or ossified lesions, thereby broadening the indications for anterior procedures ²⁰. Although these approaches show promise in improving neurological function and enhancing radiological decompression, they require technically demanding maneuvers and may carry elevated complication risk, and their long-term durability has yet to be clarified. Future studies should adopt multicenter, prospective designs with large sample sizes and incorporate multimodal imaging—including MRI, CT, and contrast-enhanced MR neurography—alongside automated assessment methods based on deep learning and segmentation algorithms to improve the objectivity of PCS measurement. Long-term follow-up focusing on quality of life and functional recovery is also essential to further establish the role of PCS in clinical decision-making.

5 Conclusion

The findings of this study indicate that the PCS provides a straightforward and reproducible imaging measure with practical value for assessing the suitability of an anterior-only procedure in individuals with multilevel pincer-type CSM. Patients with values of 2 or below showed markedly better neurological recovery than those with higher scores, whereas individuals in the higher range not only demonstrated limited improvement but also carried an elevated likelihood of requiring additional surgery. The score offers an objective reference for preoperative route selection and assists surgeons in striking an appropriate balance between insufficient decompression and unnecessarily extensive intervention. Prospective studies with larger cohorts, multiple centers, and extended follow-up will be essential to further clarify the clinical utility of this measure.

Ethics approval and consent

This retrospective study was conducted in accordance with the Declaration of Helsinki and was approved by the Institutional Review Board (IRB). The requirement for individual patient informed consent was waived by the IRB due to the retrospective nature of the study.

Submission statement

This manuscript is original and has not been submitted elsewhere in part or in whole.

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.