[A single-center study on the oncological outcomes of active surveillance of thyroid nodules measuring≤1 cm with highly suspicious ultrasound features].

Objective: To investigate the oncological outcomes of active surveillance (AS) in patients showing thyroid nodules measuring≤1 cm with highly suspicious ultrasound features. Methods: A prospective single-center cohort study. A total of 534 patients with highly suspicious thyroid nodules (2015 American Thyroid Association Nodule Sonographic Patterns and Risk of Malignancy: High Suspicion) were enrolled in this study, the patients received AS at Peking Union Medical College Hospital between January 2017 and November 2022 to assess oncological outcomes (disease progression, recurrence/metastasis rate, etc). The patients were followed up every 6 months for physical examination and neck ultrasound examination. And the value of tumor volume changes in evaluating tumor enlargement was explored too. Results: There were 413 females and 121 males in this cohort, with a mean age of (42.6±11.8) years. During a median follow-up period of 45.6 months (ranged from 3.5 to 176.0 months), disease progression occurred in 26 patients (4.9%) with highly suspicious thyroid nodules, characterized by a minimum 3-mm increase in tumor diameter in 19 patients (3.6%) and lymph node metastases in 7 patients (1.3%). Forty-seven (8.8%) patients opted for delayed surgery, with 29 patients due to a change in preference. There was no significant differences in pathologic and follow-up outcomes between patients with disease progression and preference change. Patients aged≤40 years had a higher cumulative incidence of 5-year disease progression than those aged>40 years (4.9% vs 1.9%, P=0.060). No patients experienced distant metastases or deaths. Among the 595 high-risk thyroid nodules with continuous volume assessment results and an increase in nodule diameter of less than 3 mm (including all high-risk nodules in patients with single or multiple nodules), 184 (30.9%) and 79 (13.3%) nodules exhibited volume increases of more than 50% and 100%, respectively, in multiple measurements. Among the nodules with volume changes exceeding 50% and 100%, the proportion of nodules with a baseline tumor diameter of≤0.5 cm was significantly higher than those with a diameter of>0.5 cm, at 69.0% vs 31.0% (P<0.001) and 77.2% vs 22.8% (P<0.001), respectively. Conclusions: Active surveillance in patients with highly suspicious subcentimeter thyroid nodules has good short-term oncological outcomes and can be considered a safe alternative to surgery. Due to the large variability in the measurement results of tumor volume, it is not suitable as an indicator for evaluating tumor enlargement.

Application of deep learning to identify recurrent laryngeal nerve in endoscopic thyroidectomy via breast approach / 中华内分泌外科杂志

Objective:To explore whether deep learning could apply to recognize the recurrent laryngeal nerve (RLN) in videos of endoscopic thyroidectomy (ETE) via breast approach.Methods:Videos of ETE via breast approach in Peking Union Medical College Hospital from Feb. 2020 to Aug. 2021 were collected. Videos containing RLN were selected, and the outline of RLN was marked by two thyroid surgeons. Then data were divided into a training set and a test set in a ratio of 5:1 and classified into the high and low difficulty group according to a senior thyroid surgeon’s opinion. Those pictures were input to D-LinkNet model. Precision, sensitivity and mean dice index was calculated.Results:A total of 46 videos including 153, 520 frames of pictures were included in this study. 131,039 frames of 39 videos were in the training set and 22,481 frames of 7 videos were in the test set. When the intersection over union threshold was 0.1, the sensitivity and precision was 92.9%/72.8% and 47.6%/54.9% in high and low recognition group, respectively. When the intersection over union threshold was 0.5, the sensitivity and precision turned to 85.8%/67.2% and 37.6%/43.5% in high and low difficulty group, respectively. Mean Dice index was 0.781 and 0.663 in high and low difficulty group, respectively.Conclusions:RLN recognition based on deep learning is feasible and has potential application value in ETE, which may help surgeons reduce the risk of accidental injury of RLN and improve the safety of thyroidectomy.