(P042) Patterns of Failure in Node-Positive Thyroid Cancer

M.C. Cameron, BS, K. Otto, MD, J.C. McCaffrey, MD, A.M. Trotti, MD, N. Rao, MD, B. McIver, MD, PhD, J. Russell, MD, PhD, T.A. Padhya, MD, J.J. Caudell, MD, PhD; Department of Radiation Oncology, Moffitt Cancer Center

Background: We reviewed a cohort of patients with node-positive thyroid cancer treated at a single institution surgically with or without radioactive iodine to evaluate patterns of failure.

Materials and Methods: After approval by the institutional review board (IRB), records of patients with node-positive thyroid cancer treated surgically with or without radioactive iodine between 1998 and 2012 were retrospectively reviewed. Patients with anaplastic histology, no positive nodes, or distant metastases at presentation or those who received external beam radiation in initial management were excluded. A total of 256 patients remained for analysis. Recurrences were scored if biopsy-proven or if radiographic/clinical evidence initiated a change in management. Locoregional recurrence (LRR) was assessed via Kaplan-Meier analysis. Potential predictors were compared via log-rank test. Variables with prognostic potential on univariate analysis were subjected to multivariate analysis with Cox proportional hazards regression.

Results: Median follow-up of patients alive at last contact was 40 months. Median age was 43 years (range: 14–92 yr). Females (n = 163, 63.7%) were more common than males (n = 93, 36.3%). Histology included papillary variants (n = 238, 93.4%), medullary (n = 14, 5.8%), follicular (n = 3, 1.2%), and oxyphilic (n = 1, 0.4%). Total thyroidectomy was performed in 241 (94.1%), and lobectomy was performed in 11 (4.3%); no thyroid surgery was performed in 4 (1.6%). Central neck dissections were performed in 171 (66.8%), cervical neck in 129 (50.4%), and mediastinal in 35 (13.7%). Radioactive iodine was used in 217 (84.8%) with a median dose of 148.3 mCi. American Joint Committee on Cancer (AJCC) staging was TX–2 in 128 (50%) and T3–4 in 128 (50%), with 92 cases (36%) of N1a and 164 cases (64%) of N1b. Median number of nodes examined was 16 (intraquartile range: 7–32), with a median number of positive nodes of 4 (intraquartile range: 2–10). Margins were positive in 72 (28.1%) while negative in 145 (56.6%) and unreported/unknown in 54 (21.1%). Extracapsular extension (ECE) was seen in 98 (38.3%) while not seen in 124 (48.4%) and unknown in 34 (13.3%). Lymphovascular space invasion (LVSI) was positive in 90 (35.2%) while negative in 112 (43.8%) and unknown in 54 (21.1%). Local recurrence in the thyroid bed developed in 6 (2.3%), regional in neck or upper mediastinum in 49 (19.1%), local and regional in 52 (20.3%), and distant in 6 (2.3%). Regional recurrence was more likely in the lateral neck compartment (n = 33, 67.3%) than central (n = 21, 42.8%) or mediastinal (n = 6, 12.2%); failure occurred in more than one compartment in 11 (4.3%). Actuarial risk of LRR for the entire cohort was 20.1% at 3 years, 24.3% at 5 years, and 35.2% at 10 years. LRR risk at 3 years was elevated in patients with follicular histology (100%; P = .007), T3–4 (28.7%; P = .001), N1B (24.6%; P = .006), ECE (28.3%; P = .001), LVSI (34.2%; P = .015), and increasing number of positive nodes (33.5%; P < .001 for ≥ 5 vs < 5). On multivariate analysis, only ≥ 5 lymph nodes (hazard ratio [HR] = 3.63; 95% confidence interval [CI], 1.99–6.63; P < .001) remained significant for LRR.

Conclusions: At median follow-up of 40 months, patients with node-positive thyroid cancer were most likely to recur in the neck rather than in the thyroid bed or distantly. Increasing nodal burden is associated with significantly increased risk of locoregional recurrence.