Abstract
Background: Papillary thyroid carcinoma in the isthmus (PTCI) remains a subject of surgical debate due to its unique anatomical location and reportedly more aggressive behavior, including higher rates of lymph node metastasis, multifocality, extrathyroidal extension, and capsular invasion. There are currently no definitive guidelines regarding the optimal extent of surgery. Objective: This study aimed to compare the three surgical approaches-total thyroidectomy, lobectomy with isthmusectomy, and isthmusectomy/extended isthmusectomy-in patients with PTCI, focusing on postoperative complications, tumor recurrence, recovery, and identifying risk factors for tumor prognosis and lymph node metastasis. Methods: We retrospectively analyzed data from 215 patients treated surgically across four medical centers from 2016 to 2022, divided into three groups based on surgical extent. We compared baseline characteristics, operative time, intraoperative blood loss, length of hospital stay, postoperative complications, and central lymph node metastasis risk factors. Propensity Score Matching (PSM) was used to create more comparable groups, so as to verify the accuracy and stability of our research results. Results: No significant differences were observed among the three groups in rates of temporary or permanent recurrent laryngeal nerve injury, permanent hypoparathyroidism, or chyle leakage (all p > 0.05). However, transient hypoparathyroidism was more common in the total thyroidectomy group (p < 0.05), which also had longer operative time, greater intraoperative blood loss, and longer postoperative hospital stay (all p < 0.05) The PSM-adjusted analyses further confirmed these findings, except that the previously observed difference in postoperative drainage volume among the three groups was no longer significant (p = 0.791). The Kaplan-Meier curves showed a similar cumulative proportion of recurrence-free survivors in the three groups with no statistically significant difference observed (p = 0.804). Univariate and multivariate logistic regression analysis identified that gender (OR = 4.405, 95%CI: 4.104-4.729, p < 0.001), multifocality (OR = 2.498, 95%CI: 1.064-5.864, p = 0.035), tumor diameter (OR = 1.096, 95%CI: 1.047-1.147, p < 0.001), capsular invasion (OR = 2.666, 95%CI: 2.547-2.791, p < 0.001), and absolute eosinophil count (OR = 1.381, 95%CI: 1.125-1.695, p = 0.002) remained significant independent predictors of central lymph node metastasis in PTCI. A multivariable logistic regression model was developed to predict CLNM, achieving an AUC of 0.777. A probability threshold of 0.50 provided the best balance between sensitivity (77.6%) and specificity (65.5%) and was selected as the clinical cut-off for stratifying high- and low-risk patients. Conclusions: Conservative procedures like lobectomy with isthmusectomy or isthmusectomy/extended isthmusectomy may represent a feasible, function-preserving option in carefully selected low-risk PTCI patients, but further validation is required. In contrast, patients with high-risk features may benefit from central lymph node dissection. The predictive model may provide supportive information for personalized surgical planning.