Introduction: To investigate the exposure parameters for thoracic spine/(TS) radiography that allows the image acquisition at the lowest dose possible, while maintaining an adequate image quality/(IQ) to identify all relevant anatomical criteria. Methods: An experimental phantom study was conducted, and 48 different radiographs of TS (24 AP/24 lateral) were acquired. The Automatic Exposure Control/(AEC) with the central sensor was used to select the beam intensity, while Source-to-Detector-Distance/(SDD) (AP:115/125 cm; Lateral:115/150 cm), tube potential (AP:70/81/90 kVp; Lateral: 81/90/102 kVp), use of grid/no grid and focal spot (fine/broad) were manipulated. IQ was assessed by observers with ViewDEX. Effective Dose (ED) was estimated using PCXMC2.0 software. Descriptive statistics paired with intraclass correlation coefficient (ICC) were applied to analyse data. Results: The ED increased with a greater SDD for lateral-view, presenting a significant difference (p = 0.038), however IQ was not affected. For both AP and lateral, the use of grid had a significant effect on ED (p < 0.001). Despite the images acquired without grid had lower IQ scores, the observers considered the IQ adequate for clinical use. A 20% reduction in ED (0.042mSv–0.033 mSv) was observed when increasing the beam energy from 70 to 90 kVp for AP grid in. The observers ICC ranged from moderate to good (0.5–0.75) in lateral and good to excellent (0.75–0.9) for AP views. Conclusions: The optimised parameters in this context were 115 cm SDD, 90 kVp with grid for the best IQ and lowest ED. Further studies in clinical setting are necessary to enlarge the context and cover different body habitus and equipment. Implications for practice: The SDD impacts on dose for TS; Higher kVp and grid are necessary to better image quality.