Impact of CT image co-registration on dosimetric verification in ¹²⁵I seed implantation for lung cancer

To evaluate and compare the accuracy and differences between a pre-operative and intra-operative CT image co-registration-based dosimetric verification method and the conventional verification method in ¹²⁵I seed implantation for lung cancer. A retrospective analysis was conducted on 30 lung cancer patients who underwent ¹²⁵I seed implantation at our center between January 2021 and December 2024. Histopathological types included squamous cell carcinoma (9 cases), adenocarcinoma (13 cases), and other tumors (8 cases). All patients received a prescription dose of 140 Gy, with seed activities of 0.73 ± 0.99 mCi. All patients underwent both image co-registration-based dosimetric verification (hereinafter referred to as the co-registration group) and conventional immediate post-operative dosimetric verification (hereinafter referred to as the Conventional Group). In the co-registration group, preoperative contrast-enhanced CT images were imported into the 3D Treatment Planning System (3D-TPS), followed by the last intraoperative non-contrast CT scan. Rigid co-registration of the two CT datasets was performed within TPS, and the target volume was manually delineated on the fused images. In the Conventional Group, the last intraoperative non-contrast CT was directly imported into the TPS, and the target volume was manually delineated on these images. Following target delineation in both groups, seeds were identified and postoperative dose-volume histograms (DVHs) were generated using the TPS to obtain dosimetric parameters, including target volume (Vol), postoperative D₉₀ (minimum dose covering 90% of the target volume), V₉₀, V₁₀₀, V₁₅₀, and V₂₀₀ (representing the percentages of the target volume receiving 90%, 100%, 150%, and 200% of the prescription dose, respectively). Follow-up chest CT scans were performed at 1–6 months postoperatively. Treatment response was evaluated using the Response Evaluation Criteria in Solid Tumors (RECIST) version 1.1, with short-term efficacy and adverse reactions documented. Paired t-tests were used to analyze differences in dosimetric parameters between each verification group and the preoperative plan, as well as between the Co-registration and Conventional Groups. Bonferroni correction was applied for post hoc multiple comparisons to examine statistical significance between groups. All procedures were successfully completed, with treatment outcomes including 5 cases of complete response (CR), 14 cases of partial response (PR), 9 cases of stable disease (SD), and 2 cases of progressive disease (PD), while no adverse reactions or grade 3 or higher complications were observed. Normality testing confirmed that the target volume (Vol) and all dosimetric parameters from the 30 patients were normally distributed. Statistical comparisons revealed significant differences in Vol, D₉₀, V₁₅₀, and V₂₀₀ between the conventional post-operative verification group and the pre-operative plan, whereas the image co-registration group showed significant differences only in V₁₅₀ and V₂₀₀ compared to the pre-operative plan. A significant difference in Vol was also found between the conventional and co-registration verification groups. Bonferroni-corrected post hoc tests demonstrated a statistically significant difference in D₉₀ between the pre-operative plan and the conventional verification group, but not with the co-registration group, while significant differences in V₁₀₀ and V₂₀₀ were observed between the pre-operative plan and both verification groups, with no significant differences detected in Vol, V₉₀, or V₁₀₀. Dosimetric verification using the image co-registration method demonstrates significantly improved stability in D90 values compared to the conventional post-operative verification approach when evaluated against the pre-operative plan, suggesting its potential as a novel option for medical physicists in post-implant dose verification.