Jae Sang Ryu, Dong Wook Kim, Sungkeun Kang, Eun Jin Kim, Jin Kyong Kim, Sang Wook Kang, Jandee Lee, Jong Ju Jeong, Kee Hyun Nam, Woong Youn Chung
Recurrence after TART demonstrated compartment-specific patterns. CC recurrence was strongly associated with N1b disease and showed a symmetric subsite distribution, whereas ILN recurrence was independently associated with N1a disease. These findings suggest a dominant role of nodal biology in recurrence distribution, although a contribution from technical factors cannot be entirely excluded. Risk-adapted long-term surveillance may therefore be warranted according to baseline nodal status.
BACKGROUND: Transaxillary robotic thyroidectomy (TART) has become increasingly adopted for differentiated thyroid cancer (DTC), yet long-term locoregional recurrence patterns remain insufficiently defined. Therefore, we aimed to characterize recurrence patterns using a compartment-based anatomic mapping approach in a large TART cohort.
METHODS: We retrospectively analyzed 10,504 patients with DTC who underwent TART at a single institution from 2007 to 2024. Recurrence occurred in 97 patients (0.9%). Anatomic sites included the central compartment (CC), ipsilateral lateral neck (ILN), contralateral lateral neck (CLN), contralateral thyroid (CT), and extracompartmental (EC) regions. Temporal trends and clinicopathologic factors were evaluated. Logistic regression analyses were performed to determine factors associated with ILN and CC recurrence.
RESULTS: Recurrence was single site in 89.7% (n = 87) and multi-site in 10.3% (n = 10). ILN was the most frequent recurrence site (n = 48), followed by CT (n = 38) and CC (n = 12; 7 ipsilateral and 5 contralateral). ILN recurrence was independently associated with N1a disease (adjusted OR 9.70, 95% CI 3.22-33.08; p<0.001), whereas N1b status was not significant after adjustment. In contrast, CC recurrence was strongly associated with baseline nodal burden, particularly N1b disease (p<0.001), and demonstrated symmetric subsite distribution. EC recurrence was rare (n = 3) and occurred within 5 years postoperatively. Median follow-up was 67 months (IQR 56-91).
CONCLUSIONS: Recurrence after TART demonstrated compartment-specific patterns. CC recurrence was strongly associated with N1b disease and showed a symmetric subsite distribution, whereas ILN recurrence was independently associated with N1a disease. These findings suggest a dominant role of nodal biology in recurrence distribution, although a contribution from technical factors cannot be entirely excluded. Risk-adapted long-term surveillance may therefore be warranted according to baseline nodal status.