While guidelines [3, 4] suggest discontinuing US surveillance in low-risk patients with remnant ablation, negative baseline US, and low serum Tg, in higher risk individuals US is still routinely recommended. This is in part due to the concern of false negative Tg in advanced disease which may be detected on US. However, other studies evaluating ATA low and intermediate risk DTC patients have reported low proportions of patients (1.7–4.8%) with US positive disease but undetectable Tg [8, 10] and it is unclear whether detecting recurrence earlier truly improves survival or quality of life [11]. Our study evaluated a group of higher risk DTC patients, as evidenced by the higher rate of recurrence, and reassuringly demonstrated reliable Tg elevation in 96% of patients with recurrence. Larger studies with a larger population of advanced DTC may help confirm this low risk of false negative Tg, especially with the use of new generation HS-Tg assay.

The optimal frequency and duration of US surveillance remains unclear. For instance, one study [5] suggested that in ATA intermediate risk patients, frequency of US be no more than every 3–5 years in the absence of suspicious clinical features. Another study [12] of lower risk patients found the mean time to recurrence was 19.2 months, with a second increased peak of recurrence at 5–6 years of follow-up; suggesting follow-up US within the first 1–2 years, then a second US at 4–6 years. A large multicentre Korean study [13] also concluded that only 1–2 US within the first 5 years of follow-up is sufficient. They noted that approximately 5% of patients may have had a delay in diagnosis compared to yearly US surveillance but did not correlate this to their clinical status or Tg levels. Our data demonstrated a median time to recurrence of 2.2 years and IQR of 1.4–5.3 years, which may support decreasing frequency of US surveillance particularly after 5 years follow-up, however prospective studies are required to confirm this practice.

False positive US findings continue to be a challenge in long term DTC surveillance. Our study found that having an expert radiologist apply structured evaluation could potentially decrease the number of reported false positive findings; in our study there was a 60% reduction in false positive findings. It is noteworthy that previous studies with even higher false positive rates (34–57%) did have all US performed at specialized academic institutions [5, 6]. Additionally, in a pediatric DTC population with all US performed in a tertiary care academic institution but without synoptic reporting, 55% of patients had at least one falsely indeterminate/suspicious US [14]. In another Canadian study, implementation of ETA guideline-based US reporting resulted in higher quality of reporting, though there was no significant difference in diagnostic accuracy [15]. Within our current healthcare system, which reflects real-world thyroid cancer care seen in most Canadian centres, it would not be realistic to implement centralized US given the limited resources as well as the cost of travel. Given the increased specificity found in our study, requesting a re-interpretation of a positive US examination may be a more feasible alternative. Re-interpretation has shown to alter patient management and be valuable for peer learning in other imaging exams [16].

Even with improvements to US reporting quality, false positive findings continue to exist and lead to earlier specialist follow-up visits as well as additional investigations, which can increase patient anxiety and stress, as well as healthcare costs. DTC patients experience higher rates of depression and anxiety than the general population, and similar health-related quality of life (HrQoL) compared with other cancer patients with worse prognoses [17]. No study has investigated the impacts of such false positive results and ensuing investigations on patients’ HrQoL. The diagnostic accuracy of Tg was excellent in our study; while we conservatively qualified five patients with false positive Tg, two of these have had ongoing rise in Tg and likely have recurrent disease not yet detected on structural imaging. Given the superior sensitivity and specificity of Tg, clinical judgment guided by Tg trend should play a role in the decision making on the appropriate interval of US.

Our study had some limitations. As the AJCC 8th edition staging was introduced in 2018, our study captured stage III and IV patients based on the 7th edition staging, and many would be reclassified with lower staging if examined today. However, the majority (76%) of patients included in our study were intermediate and high-risk patients. Focusing on ATA intermediate and high-risk patients alone in future, larger multi-centred studies would help validate our results. On retrospective review, there was decrease in false positive results by an experienced radiologist using structured evaluation; however, this radiologist was blinded to clinical and biochemical information as well as ability to compare previous imaging which is not reflective of normal practice. While inter-user variability of thyroid nodules has been demonstrated [7], the extent of inter-user variability in post-operative lymph node surveillance is not well described. Our results reflect real-world practices and may not be applicable to centers with specialized and consistent post-operative US reporting. Our findings do not translate to those with hemithyroidectomy or those with anti-Tg antibodies at the time of diagnosis; the frequency and utility of US monitoring in these settings remains unclear. Additionally, the majority of follow-up was done using conventional Tg assay measurements, necessitating use of rTSH-stimulation in early follow-up based on the existing guidelines at the time. Future studies using HS-Tg might further elucidate the true extent of false negative Tg in this population. Finally, the retrospective nature of the study spanned over a long period of time during which some practice patterns have changed.