Patients
A total of 187 patients with PTC who received reoperation for locoregional cervical lymph node recurrence in Fudan University Shanghai Cancer Center (FUSCC) from 2007 to 2017. All patients underwent initial surgical therapy (thyroidectomy with or without lateral neck dissection) and RAI therapy. No evidence of distant metastasis was observed in these patients before reoperation. From the total population, two patients with no evidence of metastatic disease in the surgical specimen, and 20 patients lost to follow-up were excluded. Patients harboring positive anti-thyroglobulin antibodies (TgAb) were also not included (Figure 1).
The parameters collected in the study are listed in Table 1. This study was approved by the Institutional Review Board of FUSCC and informed consent forms were signed by all the patients.
Extent of reoperation
Reoperation for recurrent disease included central and/or lateral neck dissection (ipsilateral or bilateral). Lateral modified radical neck dissection involved levels IIa–Vb regardless of previous dissection, while central compartment complemental dissection involved ipsilateral or bilateral level VI lymph node dissection. Some reoperations also involved level VII sub-compartments, and resection of local recurrent disease with aerodigestive invasion. Nor Node plucking or berry picking techniques was involved in our reoperation.
Follow-up methods and the definition of secondary recurrence after reoperation
Follow-up modalities included physical examination, ultrasound (US), unstim-Tg, neck and chest CT scan,and 131I planer WBS with or without Single-Photon Emission Computed Tomography (SPECT). Some patients also underwent fluorodeoxyglucose positron emission tomography (FDG-PET).
Secondary clinical recurrence was defined as the locoregional reappearance of pathologically proven malignant tissue in the second or third reoperation procedure or fine needle aspirate (FNA) biopsy, or the appearance of distant metastases in the mediastinum, bone, lung, or brain as evidenced by 131I WBS or other imaging methods such as 18FDG PET-CT.
To estimate the risk of recurrence through long-term follow-up after initial treatment, Tuttle et al. established the “Response to Therapy” stratification, an approach widely employed in clinical studies (16, 17). This stratification was adopted in this study to categorize the long-term outcome for the patients presenting with secondary recurrence into 4 dimensions: (1) Excellent response (ER) for negative imaging findings and unstim-Tg <0.2 ng/mL, (2) Biochemical incomplete response (BIR) for unstim-Tg >1 ng/mL with negative imaging findings, (3) Indeterminate response (IR) for nonspecific findings on imaging studies and unstim-Tg between 0.2 and 1 ng/mL, and (4) Structural incomplete response (SIR) for structural evidence of disease regardless of Tg level.
Definitions of RAI-refractory and RAI-avid locoregional recurrence
To discern the recurrent metastatic nodules as either radioiodine avid or refractory more precisely, nine patients who received remnant ablation as the primary goal of RAI administration after thyroidectomy were further excluded as the recurrent lymph node metastases of these patients were found during late follow-up, far from the initial RAI treatment, therefore, RAI avidity was not determinable. The remaining 115 patients had cervical recurrences that were classified as either radioiodine-avid (RAI-A) or radioiodine-refractory (RAI-R) according to post-ablation and post-therapy WBS results or routine diagnostic WBS with or without SPECT at follow-up. A minority of patients had 18FDG uptake in PET-CT scanning were classified as RAI-R.
Statistics
For clinical characteristics, continuous variables are reported as the medians and ranges, and categorical variables are reported as frequencies and percentages. Cutoff values for continuous variables such as unstim-Tg level, age, the time between initial surgery and reoperation, and the frequency and dose of RAI therapy were determined by the X-tile program (Yale University School of Medicine, New Haven, CT, USA, RRID:SCR_005602). Recurrence-free survival (RFS) curves were drawn and analyzed using the Kaplan–Meier method and calculated using log-rank tests. Risk factors associated with RFS were analyzed using a Cox proportional hazards model. Analyses were performed with SPSS version 22.0 (IBM Corporation, Chicago, IL, USA, RRID:SCR_002865). A two-sided P value of <0.05 was considered statistically significant.