This retrospective analysis suggests that specific factors may predict disease persistence or recurrence in patients with well-differentiated thyroid carcinoma. The American Thyroid Association has recommended that follow-up treatment varies according to a patient's risk for recurrence , yet management, including the extent of surgery, necessity of post-operative ablative radioiodine, and degree of thyroid hormone suppression, has been controversial and practices are not uniform among clinicians.
An analysis of all patients with a history of WDTC in our database, regardless of the extent of surgery or use of radioiodine (n = 651), resulted in a persistence/recurrence rate of 20% (data not shown). While this is virtually identical to what was observed in a similar cohort assessed over 30 years by Mazzaferri and Jhiang (21%) , by restricting our analyses to patients who underwent a total or near-total thyroidectomy and radioiodine therapy, the rate dropped to 12%. Our Cox regression model of patient-associated factors and the global model were predictive of well-differentiated thyroid cancer persistence or recurrence, although the regression model of tumor-associated factors was not. Greater age as a continuous variable was a weakly predictive factor, but this was not consistent across all models.
Our analysis included only those who were alive at follow-up, and therefore loss of follow-up due to death [both disease-specific and all-cause mortality) was not ascertained. In addition, some subjects excluded due to the absence of thyroglobulin values or an imaging study within the HVMA database were managed by non-HVMA providers and were not available for this analysis. Finally, the unique demographics and socioeconomic status of this study population (e.g. predominantly Caucasian) may limit generalizability to other populations managed by clinicians experienced in thyroid cancer follow-up.
It is difficult to compare the relative predictive abilities of different models or staging systems, as the components in each are not uniform. Palme et al. reported that male sex, advanced initial stage, and presence of extrathyroidal spread were independent predictors of multiple recurrences of well-differentiated thyroid cancer . In a recent retrospective review of papillary microcarcinoma (≤1 cm), Mercante et al. found that capsular invasion, extrathyroidal tumor extension, and neck lymph node metastasis at presentation were the only independent risk factors for the persistence or recurrence of disease . In contrast to these and other studies [13, 14], many of the subjects in our study received their surgery at a non-HVMA site, and thus, information regarding extrathyroidal spread was not codified in a uniform manner and could not be analyzed in our cohort. The inability to capture this information may have introduced a misclassification bias of patients' disease burden, thereby altering our findings. The use of a <1 cm cutoff for tumor size in our study (as compared to ≤1 cm) may also limit direct comparison of our results to some studies. Finally, although we did not study these specific questions, it was recently reported that recurrence of papillary and follicular thyroid cancer in the first year following thyroid surgery predicts a worse outcome , and patients with micropapillary multifocal thyroid cancer have a reduced rate of recurrence following more complete thyroid surgery . The differences between the variables analyzed in this study and those of other investigators do not allow for direct comparison of the predictive models.
There are several strengths of this study, which specifically studied only patients who underwent a total or near-total thyroidectomy and radioiodine ablation to evaluate laboratory and imaging markers of WDTC persistence/recurrence. All sites of a large Boston, Massachusetts-area multispecialty group practice were included, and there was a relatively long period of follow-up of over 28 years. Many subjects contributed substantial person-time toward the primary outcome. Furthermore, this group practice, which consists of 14 sites, has utilized an electronic medical record since the late 1970s, thus permitting an easily extractable and long-term comprehensive assessment of patients followed within it with WDTC.
The findings of this study do not confirm previously published data that male gender and larger tumors have a worse prognosis for WDTC persistence/recurrence. We urge the establishment of a comprehensive national registry, created through linkage with electronic health record systems, that would uniformly capture information regarding these and other factors for the long-term monitoring of WDTC outcomes.