**3. Results**

#### *3.1. Demographic Data and Surgery-Related Data According to DISH Grade*

DISH was observed in 82 men and 25 women with cervical OPLL, with a co-morbidity rate of 44.8% (107/239; Table 1). Our grading system evaluation revealed that when DISH was present, grade 2 was the most common (65/107, 60.7%), followed by grade 1 (23/107, 21.5%) and grade 3 (19/107, 17.8%). There was a slight, ye<sup>t</sup> significant, correlation of DISH grade with age (r = 0.30, *p* < 0.0001; Table 1) but not with sex, BMI, or the prevalence rate of DM, FH of OPLL or TH. Only one case was found in which the bridging of OALL over four adjacent vertebral bodies was localized in the cervical spine. This case was, therefore, excluded from the analysis of patients with DISH because it did not exhibit similar bridging in the thoracic spine.


**Table 1.** Demographic data and surgery-related data for each DISH grade.

ADF, anterior decompression with fusion; APF, anterior and posterior decompression with fusion; CSF, cerebrospinal fluid; DISH, diffuse idiopathic skeletal hyperostosis; DM, diabetes mellitus; FH, family history; OPLL, ossification of the posterior longitudinal ligament; PDF, posterior decompression with fusion; PF posterior fusion.

Surgical treatment was performed in 59.4% of all cases (142/239; Table 1) in at least one of the spinal levels. The cervical spine was the most frequently treated level (129/239, 54.0%), followed by the thoracic (26/239, 10.9%) and lumbar spine (11/239, 4.6%). There was no significant difference in the rate of surgical treatment between each grade at any spinal level. Laminoplasty was the most common surgical procedure performed on the cervical spine (60/129, 46.5%) whereas posterior decompression with fusion (PDF) was more common at the thoracic spine (16/26, 61.5%). On the other hand, laminectomy and PDF were equally common at the lumbar spine (5/11, 45.5%). No remarkable differences were found in the rates of these procedures between each grade. Furthermore, all the incidences of perioperative complication were not statistically different among the grades.

#### *3.2. Association between DISH Grade and OSL*

Next, we calculated the correlation coefficient between the DISH grade and OSL for each spinal level. At the cervical level, the DISH grade was moderately correlated with the OPLL index (r = 0.45, *p* < 0.0001; Figure 3a); however, there was no correlation between DISH grade and the OLF index (r = 0.14, *p* = 0.03; Figure 3b). Moreover, the prevalence of ONL was significantly associated with DISH grade (*p* = 0.003, chi-squared test; Figure 3c). At the thoracic spine, the DISH grade was moderately correlated with the OLF and OSIL indices (OLF: r = 0.41, *p* < 0.0001, Figure 4b; OSIL: r = 0.53, *p* < 0.0001; Figure 4c), but not with OPLL (r = 0.12, *p* = 0.06; Figure 4a). There was no significant correlation of DISH grade with any OSL at the lumbar spine (OPLL: r = −0.02, *p* = 0.78, OLF: r = 0.11, *p* = 0.11, OSIL: r = 0.14, *p* = 0.03; Figure 5a–c). Finally, there were moderate to weak correlations between DISH grade and OPLL, OLF and OSIL indices in the whole spine (OPLL: r = 0.29, *p* < 0.0001, OLF: r = 0.40, *p* < 0.0001, OSIL: r = 0.50, *p* < 0.0001; Figure 6a–c).

#### *3.3. Case Presentation*

A 66-year-old man presented to one of our institutions with difficulty walking. Wholespine CT imaging showed continuous-type cervical OPLL at C3–C7, with a cervical OPLL index of 10. In addition, extensive thoracic OLF was found, with a thoracic OLF index of 9. Grade 3 DISH was distributed from C4 to L2. The level of maximum compression in the spinal canal was C3/4 with OPLL (Figure 7). Therefore, we decided to perform a two-stage surgery for cervical OPLL. First, anterior decompression with fusion (ADF) was performed from C2 to C5 with grafted bone harvested from the fibula. Two weeks after the initial surgery, an additional posterior fixation was performed from C2 to C7. Five years after the surgeries, the man's neurological symptoms have shown satisfactory improvement.

**Figure 3.** Correlation between DISH grade and OSL at the cervical spine. (**a**) Cervical OPLL index; (**b**) cervical OLF index; (**c**) prevalence of ONL. DISH, diffuse idiopathic skeletal hyperostosis; OLF, ossification of the ligamentum flavum; ONL, ossification of the nuchal ligament; OPLL, ossification of the posterior longitudinal ligament; OSL, ossification of the spinal ligaments.

**Figure 4.** Correlation between DISH grade and OSL at the thoracic level. (**a**) Thoracic OPLL index; (**b**) thoracic OLF index; (**c**) thoracic OSIL index. DISH, diffuse idiopathic skeletal hyperostosis; OLF, ossification of the ligamentum flavum; OPLL, ossification of the posterior longitudinal ligament; OSIL, ossification of the supra/interspinous ligaments; OSL, ossification of the spinal ligaments.

**Figure 5.** Correlation between DISH grade and OSL at the lumbar level. (**a**) Lumbar OPLL index; (**b**) lumbar OLF index; (**c**) lumbar OSIL index. DISH, diffuse idiopathic skeletal hyperostosis; OLF, ossification of the ligamentum flavum; OPLL, ossification of the posterior longitudinal ligament; OSIL, ossification of the supra/interspinous ligaments; OSL, ossification of the spinal ligaments.

**Figure 6.** Correlation between DISH grade and OSL in the whole spine. (**a**) Whole-spine OPLL index; (**b**) whole-spine OLF index; (**c**) whole-spine OSIL index. DISH, diffuse idiopathic skeletal hyperostosis; OLF, ossification of the ligamentum flavum; OPLL, ossification of the posterior longitudinal ligament; OSIL, ossification of the supra/interspinous ligaments; OSL, ossification of the spinal ligaments.

**Figure 7.** Illustrative case of grade 3 DISH. (**a**) Sagittal cervical CT imaging; (**b**) Sagittal thoracolumbar CT imaging. CT, computed tomography; DISH, diffuse idiopathic skeletal hyperostosis; OLF, ossification of the ligamentum flavum; OPLL, ossification of the posterior longitudinal ligament.

## **4. Discussion**

In a previous study, we reported on the distribution of DISH in patients with cervical OPLL by cluster analysis [6]. In that study, DISH was found to be gradually distributed from the thoracic to the cervical and lumbar spine, and rarely extended beyond C5 and L3 [6]. Based on these findings, we defined DISH found only in the thoracic spine as a mild case, with C5 and L3 indicating the boundaries between moderate and severe cases. The present study found a weak but significant correlation between DISH grade and age. In addition, there was only one case in which bridging of OALL over four or more vertebral bodies was found in the cervical spine but not in the thoracic spine. These findings support the rationale of our grade, that DISH mainly develops from the thoracic spine to the cervical and lumbar spine over time; therefore, our DISH grade might be a reliable tool for evaluating the severity of this pathology. However, our grade may present challenges in the clinical setting. For example, there are exceptional cases in which bridging of OALL is found outside the thoracic spine. In addition, the clinical significance of this grading system is unclear. Thus, our future research will investigate the association between DISH grade and the incidence of vertebral body fractures.

The DISH grade is correlated with the cervical OPLL index and with the thoracic OLF and OSIL indices, all of which have frequently been detected in clinical settings [5,10,22,23]. Moreover, the progression of the DISH grade correlates moderately or slightly with various OSL indices, even in the whole spine. Thus, the severity of DISH might be correlated with that of OSL in other areas in the spine in patients with cervical OPLL. Okada et al. reported that surgery was performed in about 85% of cases exhibiting a spinal fracture with DISH, of which approximately 80% underwent conventional, open posterior fixation. In addition, the presence of OPLL was associated with residual neurological paralysis at the final follow-up [12]. In contrast, Yoshii et al. analyzed data from 2353 cases with cervical OPLL, of which 1333 cases underwent ADF and 1020 cases underwent PDF. Their report revealed that at least one local complication, such as cerebrospinal fluid leakage or surgical site infection, occurred in about 6.5% and 4.7% of anterior and posterior cases, respectively [24]. In cases with symptomatic cervical OPLL and/or DISH, including those with complications, the surgical outcomes were sometimes unsatisfactory; therefore, it is necessary to carefully monitor for neurological deterioration caused by the combination of multiple OSLs.

In this study, DISH was observed in nearly 40% of subjects and the most common grade was grade 2. This is because this study targeted patients with symptomatic cervical OPLL, and the range of DISH in the spine progressed with age. In contrast, fewer patients had grade 1. Although DISH is frequently comorbid with cervical OPLL [6,25], Fujimori et al. demonstrated that healthy subjects without OPLL may occasionally have DISH [5]. Therefore, our present findings and those of previous studies indicate that patients with grade 1 DISH can be broadly divided into two categories: those with and those without cervical OPLL. Our study focused specifically on patients with OPLL which could explain why a minority of the population had grade 1 DISH. Similarly, grade 3 DISH was an uncommon finding in our study population. As our results demonstrate, OSL may progress in patients with advanced DISH, and these patients usually need to be treated surgically. However, the present study did not include patients who had undergone spinal surgery, so the number of subjects with a grade 3 DISH was relatively small.

Ankylosing spondylitis (AS) is a spinal ankylosing condition similar to DISH. The radiological hallmark of DISH is ossification flowing along the spine similarly to "melting candle wax" [26–28], whereas AS is characterized by thinner and finer syndesmophytes connecting between adjacent vertebral bodies, which is known as "bamboo spine" [26,27,29]. Although experienced spine surgeons can easily distinguish between these two ossification disorders, it is uncertain whether all diagnoses are accurate. Moreover, these two pathologies occasionally show a degree of overlap [30]. Therefore, it is possible that our subjects diagnosed as having DISH constituted a heterogeneous population that consisted mainly of cases of DISH alone but may have also included some cases with AS or both conditions.

Spinal ossification is potentially associated with various metabolic diseases. In particular, DM is frequently comorbid with OSLs [31,32]; however, no significant correlation was found between the DISH grade and the prevalence rate of DM in the present study. A previous study found that the prevalence rate of DM was neither associated with the ossification types of OPLL nor the occupying ratio of OPLL in the spinal canal [33]. Thus, the presence or absence of DM might not be related to the radiographic progression of ossification.

This study has several limitations. First, our subjects were patients with symptomatic cervical OPLL who may have been predisposed to ossification in the whole spine. Further research is necessary to clarify whether our findings apply to asymptomatic patients with DISH found incidentally. Second, our study performed evaluations using CT imaging, which is associated with the problem of radiation exposure; therefore, it would be preferable to use plain radiography rather than CT. Finally, the study had a cross-sectional design, resulting in a lower level of evidence. Thus, a longitudinal study is needed to confirm whether the severity of OSL progresses with DISH simultaneously.
