Implant material used in spinal tumor surgery can affect dose distribution in postoperative radiotherapy planning in proportion to the atomic number of the implant. The aim of this study is to measure the dose perturbation created by Ti and PEEK implants during paravertebral radiotherapy using direct TLD-based dosimetry in a sheep cadaver model and to compare the findings with Eclipse AAA algorithm calculations. Pedicle screws were placed at two thoracic levels in a sheep cadaver to create three configurations: Reference (no implants), Ti rod, and PEEK rod. CT simulations were performed for each configuration, the PTV and critical organs were contoured, and 6 MV – 5-field IMRT plans were created using the Eclipse AAA algorithm. TLD-100 chips were placed in the vicinity of the spinal cord, screw entry points, rod surface, lamina, paraspinal muscle, skin, and radiosensitive distant organs. The measured TLD doses were compared with TPS point values. Ti implants produced a significant increase in dose in regions close to the spinal cord and implant surface (p < 0.001). Minimal dose perturbation was observed in PEEK implants, which were quite close to reference values. The AAA algorithm was unable to fully reflect the dose increase around Ti and was found to systematically underestimate the measured dose. Furthermore, dose estimation was found to be limited in regions distant from the target volume. This study uniquely confirms implant-induced dose perturbation through direct dose measurements in real biological tissue. While Ti implants cause a significant increase in dose due to their high-Z structure, PEEK implants were found to behave more like the tissue radiologically and create a more stable dose field. It is recommended that Monte Carlo-based dose calculation or direct dose verification methods be integrated into the clinical routine for implant patients.