Although clinical symptoms, blood tests, and imaging modalities may suggest a diagnosis of pyogenic spondylitis, confirmation requires microbiological tests of tissue samples [1,2]. Conservative treatment, consisting of first-line antibiotic therapy, is frequently effective, although surgery may also be required. In actual clinical practice, however, the choice of optimal treatment is often difficult, due in part to the lack of clear guidelines regarding the role of surgical treatment [3].

Factors associated with the success or failure of conservative treatment include patient background, causative organisms, and changes in serum inflammatory markers [[4], [5], [6]]. However, the roles of imaging modalities in the diagnosis of pyogenic spondylitis and prediction of clinical outcomes remain unclear. Although magnetic resonance imaging (MRI) has been reported useful in diagnosing pyogenic spondylitis [[7], [8], [9]], it may not be useful in determining the effects of treatment because of discrepancies between the clinical course after conservative treatment and changes in MRI over time [10]. X-ray analysis and computed tomography (CT) can provide morphological information, which is nonspecific for pyogenic spondylitis. Moreover, CT can detect very minor bone resorption images at an early stage in patients with pyogenic spondylitis, making CT useful in determining the degree of bone destruction in the affected vertebrae [11]. However, the ability of CT findings at diagnosis to predict clinical outcomes has not been fully determined.

A novel grading system based on CT imaging in patients with pyogenic spondylitis may be useful for determining whether a conservative or surgical treatment strategy would be more beneficial. The present study therefore sought to determine factors, especially those determined by CT imaging, predictive of conversion to surgery after conservative treatment of patients with pyogenic spondylitis. Thus, the purpose of this study was to classify grades of bone destruction on CT images of patients with early stage pyogenic spondylitis, and to identify predictors of the success or failure of conservative treatment.