Locoregional recurrence and death from medullary thyroid carcinoma in a contemporaneous series: 5-year results

Abstract
Objective: Studies of outcome in patients with medullary thyroid carcinoma require long follow-ups to accrue enough cancer-specific events. Owing to time-dependent changes in diagnosis and therapy, long observation periods render studies susceptible to ‘time bias’ which can yield spurious results. This study was designed to investigate cancer-specific event rates after initial neck resection in contemporaneous patients recruited within less than a decade. Design: Institutional cohort study of 128 consecutive patients who underwent compartment-oriented neck surgery between 1994 and 2002 at a tertiary surgical center for hitherto untreated medullary thyroid carcinoma. Methods: The Kaplan–Meier method was used in conjunction with the log-rank test for analysis of time-dependent outcomes. Results: Follow-up was available for 120 patients (94%) including 63 rearranged during transfection (RET) carriers. There were six locoregional recurrences in the 110 patients with clear surgical margins at initial neck resection, and 12 cancer-specific deaths in the 120 patients with available follow-up. On Kaplan–Meier analysis, most clinicopathological variables were significantly associated with recurrence-free survival and cancer-specific survival. Within the median observation period of 64.5 months, patients with node-negative tumors did not develop locoregional recurrence or die from their malignancies. The low event rates precluded multivariate analyses with all clinicopathological variables. With our extensive surgical approach, median recurrence-free survival and cancer-specific survival at 5 years were 95.2 and 89.3% respectively. Discussion: Compared with literature data, our 5-year locoregional recurrence rate of 4.8% appeared very favorable, and our 5-year cancer-specific mortality rate of 10.7% was among the lowest ever reported. The growing proportion of localized medullary thyroid carcinomas among contemporaneous patients can be expected to ultimately lower the event rates, complicating future studies of outcome.