Accurate assessment of pedicle morphology in Adolescent Idiopathic Scoliosis (AIS) is essential for safe pedicle screw placement. Computed tomography (CT) is the gold standard; however, routine use in adolescents raises concerns regarding cumulative radiation exposure. Plain radiographs remain the primary preoperative imaging modality, yet their diagnostic accuracy for identifying severe pedicle dysplasia remains under validated. This study aimed to validate the diagnostic performance of plain radiographs against CT for detecting severe pedicle dysplasia in AIS. We hypothesized that preoperative radiographs can reliably differentiate non severe (Grade A or B) from severe (Grade C or D) dysplastic pedicles, allowing selective use of CT imaging. We retrospectively analyzed 1,944 pedicles (T1–T12) from 81 patients with Lenke 1 and 2 AIS. All patients underwent preoperative standing radiographs and CT. Pedicles were graded and dichotomized into non severe (A/B) and severe (C/D) dysplasia. Using CT as the reference standard, sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy were calculated. Agreement was evaluated using linear weighted kappa and non-parametric correlation coefficients (Spearman rho and Kendall tau-b). Plain radiographs demonstrated a sensitivity of 73.1% and specificity of 81.4% for detecting severe dysplasia. The PPV was 62.4% and the NPV was 87.8%, with an overall accuracy of 79.0%. Level specific analysis showed highest reliability at T1 and T12, where the NPV reached 100%, while mid thoracic levels demonstrated greater variability and lower predictive values. Overall agreement was moderate, with a linear weighted kappa of 0.520 (95% CI [0.479, 0.560], p < 0.001). Correlation across pooled T1 to T12 levels was consistent, with Spearman rho and Kendall tau-b both 0.523 (p < 0.001). Plain radiographs demonstrate high specificity and strong NPV for excluding severe pedicle dysplasia in AIS. Diagnostic performance is strongest at the proximal and distal thoracic levels, particularly T1 and T12. CT may therefore be reserved for mid thoracic levels, where rotational deformity and anatomical variability may limit radiographic accuracy. These findings support a selective imaging strategy to reduce radiation exposure without compromising surgical safety.