Our aim was to assess the role of computed tomography (CT) or magnetic resonance imaging (MRI) in the assessment of spondylodiscitis, identifying the best target structures for biopsy to increase the likelihood of positive cultures. This study was approved by our Institutional Review Board, and requirement for specific consent form was waived. In this retrospective single Institution study, we evaluated clinical and imaging records of 60 patients who underwent spine biopsy for suspected spondylodiscitis from January 2016 to May 2021. CT and MRI sensitivity and inter-reader agreement were assessed according to the phase of spondylodiscitis, defined as acute, subacute or chronic. Inter-reader agreement for the diagnosis and identification of spondylodiscitis phase was assessed using K statistics. Univariate logistic regression analysis was performed to assess any relationship between MRI/CT findings, spondylodiscitis phase and positive cultures. Sixty patients (48 males) with mean age 59, 2 ± 29 were enrolled in this study. MRI showed higher sensitivity (96% vs 65% of CT) in the diagnosis and good inter-reader agreement (k = 0.8) in the identification of the acute and subacute phase of spondylodiscitis, and moderate inter-reader agreement (k = 0.7) and lower sensitivity (80% vs 95% of CT) for the chronic phase. Univariate analysis showed as MRI-specific findings such as extensive hyperintensity of vertebral body and/or disc on Short Tau Inversion Recovery-T2w images, paravertebral collections, preserved or augmented disc height and presence of a vertebral fractures were mainly found in the acute/subacute phase and was a predictor for positive cultures (p < 0.05). MRI and CT are both able to identify the different phases of spondylodiscitis, although MRI is more sensitive in the acute phase. Findings such as extensive hyperintensity of the disc/vertebral body, fractures and paravertebral collections, represent the main targets for biopsy related to a positive culture.