**1. Introduction**

Ossification of the posterior longitudinal ligament (OPLL) is a well-known cause of severe myelopathy and radiculopathy, especially in East Asian countries [1,2]. Patients with OPLL often experience the ossification of spinal ligaments (OSLs). Previous reports sugges<sup>t</sup> that the co-morbidity rate for diffuse idiopathic skeletal hyperostosis (DISH) and OPLL is around 25–50%, which is relatively high [3–6]. Given that DISH is usually found as a benign radiological condition that does not compress the spinal cord [7–10], this pathology has been considered clinically innocuous. However, patients with DISH are at higher risk for late-onset paralysis following ankylosing spinal fractures with minor trauma, especially in cases with spinal cord compression due to OPLL [11–13]. In addition, myelopathy frequently results from a concentration of stress factors—when spinal stenosis along with OSL is present above or below the ankylosing spine in DISH [14]. Therefore, assessing the degree of DISH is important in patients with cervical OPLL.

Despite the potentially devastating consequences of comorbid DISH and an additional OSL, such as cervical OPLL, a correlation remains to be determined between DISH severity and a predisposition to other OSL. To address this question, a tool is urgently needed for evaluating the spread of DISH. Previous studies have reported the degree of DISH according to the number of consecutive vertebral bodies involved, or the width and/or thickness of ossification on plain radiographs [15–17]; however, neither of these grading methods can accurately assess the development of ossified lesions.

In a previous study, we retrospectively examined the DISH distribution pattern in whole-spine computed tomography (CT) images for patients with cervical OPLL [6] and found that DISH developed at the thoracic level initially and extended to the cervical and/or lumbar spine over time. Therefore, we developed a novel four-point grading system that can evaluate the age-related progression of DISH (grade 0, DISH anywhere in the spine; grade 1, DISH at T3–T10; grade 2, DISH extending to the cervicothoracic junction (C6–T2) and/or thoracolumbar junction (T11–L2); grade 3, DISH extending to the cervical and/or lumbar spine beyond C5 and/or L3; Figure 1). At the Japanese Multicenter Research Organization for Ossification of the Spinal Ligament (JOSL), we established a nationwide patient registry to prospectively collect clinical and radiological data, including whole-spine CT scans of OPLL patients, with the aim of clarifying associations with the presence of each type of OSL. Accordingly, the aim of the present study was to investigate the relationship between the severity of DISH (the DISH grade [6]) and all other types of OSL based on the data collected in the patient registry.

**Figure 1.** Representative sagittal computed tomography image for DISH grades 0–3. (**a**) Grade 0 (no DISH); (**b**) Grade 1 (DISH at T3–T10); (**c**) Grade 2 (DISH at both T3–T10 and C6–T2 and/or T11–L2); (**d**) Grade 3 (DISH extending beyond C5 and/or L3). DISH, diffuse idiopathic skeletal hyperostosis.

#### **2. Materials and Methods**

#### *2.1. Patients and Methods*

This multicenter prospective observational cross-sectional study was performed by the JOSL with the assistance of the Japanese Ministry of Health, Labour, and Welfare. The inclusion criteria were as follows: over 20 years of age; diagnosis of cervical OPLL on plain radiographs; symptoms such as neck pain, numbness in the upper or lower extremities, clumsiness, or gait disturbance; presentation to 1 of 16 institutions affiliated with the JOSL between September 2015 and December 2017; and whole-spine CT images available. Patients were excluded if they had undergone surgery to treat OPLL. The study included 239 Japanese subjects (163 men and 76 women). Basic clinical data for age, sex, body mass index (BMI), presence or absence of diabetes mellitus (DM), family history (FH) of OPLL, trauma history (TH), patients with or without surgical treatment, surgical methods and perioperative complications were obtained from patient records held at participating institutions. The study was approved by the institutional review board of each participating institution and was conducted in accordance with all relevant guidelines and regulations.

#### *2.2. Radiographic Examinations*

Six senior spine surgeons (S.U., K.M., S.M., K.K., N.N. and K.T.) independently determined the incidence of OPLL, ossification of the ligamentum flavum (OLF), ossification of the supra/interspinous ligaments (OSIL), ossification of the anterior longitudinal ligament (OALL), and ossification of the nuchal ligament (ONL) in whole-spine mid-sagittal CT images (Figure 2). Before the evaluation, inter-observer agreemen<sup>t</sup> was determined by assessing the incidence of OPLL and OALL, using CT images from the same 10 patients. The average kappa (κ) coefficients of inter-observer agreemen<sup>t</sup> for OPLL and OALL were 0.83 and 0.78, respectively. The prevalence rate of ONL was calculated for DISH grades 0 to 3, as described below. We recorded the presence of OPLL, OLF and OSIL for all vertebral bodies and intervertebral disc levels of the whole spine. An ossification index was calculated according to the number of levels with OPLL (OPLL index), OLF (OLF index), or OSIL (OSIL index), as described previously [6,18–20]. OALL was considered DISH if it completely bridged at least four contiguous adjacent vertebral bodies in the thoracic spine,

according to the criteria established by Resnick and Niwayama (Figure 2) [21]. DISH was classified as follows: grade 0, no DISH at any spine level; grade 1, DISH at T3–T10; grade 2, DISH at both T3–T10 and C6–T2 and/or T11–L2; grade 3, DISH extending beyond C5 and/or L3.

**Figure 2.** Representative sagittal computed tomography image for DISH, OPLL, OLF, OSIL and 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; and OSIL, ossification of the supra/interspinous ligaments.

#### *2.3. Statistical Analysis*

All data are presented as the mean ± standard deviation. Correlations between DISH grade and age, BMI, OPLL index, OLF index, and OSIL index were analyzed using the Pearson's correlation coefficient. The chi-squared test was used to examine differences in the prevalence rate of ONL, sex distribution, the presence of DM, FH of OPLL, TH, the number of patients treated surgically, the rate of each surgical method, and each complication rate. A *p*-value of <0.01 was considered statistically significant.
