Background: Mediastinal lesions encompass a broad spectrum of malignant and benign conditions, and histopathological confirmation is often required to guide appropriate management. Computed tomography (CT)-guided transthoracic core needle biopsy (CNB) is a less invasive alternative to surgical biopsy; however, data on its diagnostic performance and safety in Mongolia remain limited. Aim: To evaluate the diagnostic yield and safety of CT-guided transthoracic CNB for mediastinal lesions and to identify factors associated with diagnostic outcomes and procedure-related complications. Materials and Methods: This retrospective–prospective observational cohort study included 58 CT-guided transthoracic CNB procedures performed in 55 patients with mediastinal lesions. CT characteristics, histopathological diagnoses, procedural variables, and complications were analyzed. Fisher’s exact test, Pearson’s χ² test with Monte Carlo simulation, and the Mann–Whitney U test were applied, with Holm adjustment for multiple comparisons. Results: The median age was 57 years [IQR, 46–67], and 63.6% of patients were male. Malignancy was confirmed in 76.4%, while 21.8% had non-malignant disease. A specific diagnosis was established in 54 of 58 procedures, yielding a diagnostic yield of 93.1% (95% CI, 83.6–97.3%). CT lesion composition, internal characteristics, and mediastinal location were not significantly associated with diagnostic category (all p>0.05). The overall complication rate was 6.9%, with CIRSE grade ≥3 complications occurring in 1.7%. All complications occurred when the needle-path depth exceeded 7 cm, an association that remained significant after Holm adjustment (adjusted p=0.007). Conclusion CT-guided transthoracic core needle biopsy is a highly effective diagnostic procedure with a favorable safety profile for mediastinal lesions. A needle-path depth >7 cm may indicate increased procedural risk and warrants validation in larger studies.