Role of Post-operative Thyroglobulin in Predicting Disease-recurrence in Differentiated Thyroid Cancer

Differentiated thyroid cancer (DTC) patients with intermediate- or high-risk of recurrence are commonly treated with radioactive iodine (RAI). Although the utility of serum thyroglobulin (Tg) levels after surgery and RAI ablation is the standard of care in dynamic risk stratification, its role prior to RAI ablation remains undefined. We evaluated the relationship between post-operative, pre-RAI unstimulated Tg levels and persistent or recurrent structural disease in intermediate- or high-risk patients, postulating that it may help identify patients who may not require RAI.

Patients diagnosed with DTC from three UK cancer centres were retrospectively identified from hospital electronic health records. Data collected included patient characteristics and clinical parameters such as unstimulated, post-operative and pre-RAI Tg levels and follow-up clinical and imaging results. The remaining 301 patients were analysed using univariable and multivariable logistic regression to explore the association between postoperative Tg and structural disease recurrence or persistence.

Three hundred and one patients were included in the final analysis. The cohort included 209 (69%) females and 92 (31%) males, with 21 cases of recurrent or persistent disease. Univariable analysis and multivariable logistic regression both showed that unstimulated, post-operative Tg was an independent predictor of structural disease recurrence/persistence. Receiver operator characteristic curve suggested a post-operative unstimulated Tg cutoff of 1.05 ug/L (odds ratio [OR] 1.016, 95% confidence interval [CI] 1.005 to 1.042, p = 0.011). Notably, 17 (81%) of the 21 recurrences had a post-operative Tg levels above this cut-off.

Low postoperative unstimulated Tg levels are associated with a low risk of structural recurrence in intermediate- and high-risk DTC patents. Postoperative Tg may enhance current risk stratification in patients who could safely avoid RAI, but prospective trials are needed to validate this.