Study subjects
This study was part of the Multicenter Prospective Cohort Study of Active Surveillance on Papillary Thyroid Microcarcinoma which was performed in three referral hospitals in Korea (NCT02938702) (Seoul National University Hospital; SNUH, Seoul National University Bundang Hospital; SNUBH, and National cancer center; NCC). The study was approved by the institutional review boards of participating institutions. Informed written consent was obtained from all the participants.
Participants diagnosed with PTMC (≤ 1cm) without involvement of adjacent structures or lymph node/distant metastases between June 2016 and January 2021 were considered for inclusion in this study(20). Participants who met the inclusion criteria were given the option to choose between AS and immediate surgery. A total of 1,177 participants were consecutively enrolled and 699 participants underwent two or more US observations with a follow-up period > 6 months to observe the development of tumor progression, according to the predefined study protocol (21, 22).
Neck US examination procedure and image processing
At each visit, US of the thyroid and neck was performed by one expert radiologist at each institution. Gray-scale static images and video clips of the thyroid gland and each nodule were obtained. All nodules were evaluated using the transverse and longitudinal planes for three-dimensional evaluation. Color Doppler imaging was performed for each nodule with standardized parameters (23).
US image evaluation
The presence of US-diffuse thyroid disease (US-DTD) was evaluated in all participants and participants were followed up by the same radiologist who performed initial imaging to minimize interobserver variability. On gray-scale US images, DTD was determined according to parenchymal echogenicity, echotexture (with ill-defined hypoechoic area, from millimeters to geographic), size, glandular margin, and vascularity. (24, 25).
US findings of the nodule were evaluated (20). The composition of the nodule was categorized as solid or partially cystic (26). Nodule echogenicity was categorized as marked hypoechoic, mild hypoechoic (hypoechoic), isoechoic or hyperechoic. Nodule margin was categorized as smooth or ill-defined versus irregular, and orientation (shape) was categorized as parallel versus non-parallel (20, 26). Calcifications were categorized as abscent versus present. Intratumoral vascularity on Doppler was recorded as either present or absent.
Follow-up protocol and Outcome Measures
In participants who chose AS, PTMCs were monitored for tumor progression. Follow-up visits were scheduled twice a year during the first 2 years after diagnosis and then annually (21). At every evaluation session, additional imaging or biopsy was performed as necessary. Participants underwent surgery if disease progression was detected. Some participants also elected to undergo surgery in the absence of progression. When the participant refused surgery despite progression, AS was allowed. Tumor progression was defined as a size increase of ≥ 3 mm in at least one dimension, or ≥ 2 mm in at least two dimensions: suspected extrathyroidal tumor extension (direct extension into perithyroidal structures including the trachea, esophagus, nerves, and vessels on imaging), pathological diagnosis of LNM, or suspected distant metastasis (1, 3, 4, 7, 21). Any suspicious lymph nodes with a short diameter > 3mm were biopsied according to current guidelines (20).
Statistical analysis
Cox proportional hazards regression models were used to evaluate the effects of variables on tumor progression. Variables in this analysis included presence of US-DTD, tumor size, tumor location, US features of the tumor, and clinical features at baseline. Kaplan-Meier curves were plotted for significant variables to compare hazard ratios (HR) using the log-rank test. According to significant US findings, we compared participants’ and tumor-related features.
Participants were also stratified into subgroups based on US features and time-to-progression curves were plotted and compared post-hoc using the log rank test. US features and participant characteristics were analyzed for their association with certain progression criteria.