Highlights
In a large South Korean real-world cohort of 31,861 patients with papillary thyroid carcinoma (PTC) measuring 4 cm or smaller, practice changed in the direction encouraged by the American Thyroid Association (ATA): total thyroidectomy became less common and radioactive iodine (RAI) use declined over time.
These changes were accompanied by a marked reduction in permanent hypocalcemia, falling from 2.7% to 0.2%, while transient complications remained largely stable.
Although the most recent treatment era showed the lowest crude recurrence rate, time-to-event modeling suggested a higher recurrence hazard in the latest era compared with earlier cohorts; however, 5-year restricted mean survival time and 5-year disease-free survival showed no significant difference, underscoring the importance of interpreting calendar-era analyses cautiously.
Background
Papillary thyroid carcinoma is the most common type of thyroid cancer and is often biologically indolent, especially when tumors are small and confined to the thyroid. For many years, treatment tended to be aggressive, with frequent use of total thyroidectomy and postoperative RAI even in patients at relatively low risk of recurrence.
The ATA guideline revisions in 2009 and 2015 helped drive a global shift toward de-escalation. In selected low-risk patients, lobectomy rather than total thyroidectomy may be sufficient, and routine RAI is no longer favored when the anticipated benefit is minimal. The rationale is straightforward: if disease-specific survival is already excellent, then the main opportunities for improvement lie in reducing harm, surgical morbidity, and unnecessary adjuvant therapy.
Despite these recommendations, it has been uncertain how rapidly and how completely they were adopted in real-world practice outside the guideline setting, particularly in high-volume Asian centers where thyroid cancer incidence is high and surveillance is intensive. This study addresses that gap by examining temporal changes in treatment patterns, complications, and oncologic outcomes across three guideline eras.
Study Design
This retrospective cohort study was conducted at a high-volume tertiary center in South Korea and included 31,861 patients treated for PTC tumors measuring 4 cm or smaller between 2004 and 2020.
Patients were grouped into three time periods: triad 0 (2004-2009), triad 1 (2010-2015), and triad 2 (2016-2020). To reduce confounding by case mix, the investigators performed exact matching so that clinicopathologic characteristics were balanced across groups. They then used segmented regression analysis to identify changes in management over time.
The main outcomes were postoperative complications and disease-free survival (DFS). Complications were compared using conditional logistic regression, while DFS was evaluated with stratified Cox regression. Because follow-up differed across eras, the authors also used 5-year restricted mean survival time (RMST), a useful method for comparing average survival experience over a fixed time horizon when censoring is uneven.
Key Findings
The clearest signal from this study was a sustained shift toward less extensive therapy. After the ATA guideline updates, total thyroidectomy rates declined, and RAI use fell as well. This is important because both total thyroidectomy and RAI can increase treatment burden, cost, and the risk of avoidable adverse effects in patients who may not benefit from aggressive management.
The surgical safety signal was favorable. Permanent hypocalcemia, one of the most clinically meaningful complications after thyroidectomy, decreased dramatically from 2.7% to 0.2% across eras. That is a major improvement because permanent hypocalcemia can require lifelong calcium and vitamin D supplementation and can substantially affect quality of life. Transient complications, by contrast, did not show a major change, suggesting that the main benefit of de-escalation was a reduction in lasting harm rather than in short-term postoperative events.
On the oncologic side, crude recurrence was lowest in the most recent era, at 1.3%. At first glance, this supports the safety of de-escalation. However, the stratified Cox analysis found a higher hazard ratio for recurrence in triad 2 compared with triad 0, with a hazard ratio of 1.520 and a 95% confidence interval of 1.160 to 1.980. This finding deserves careful interpretation. A hazard ratio reflects the relative rate of recurrence over time, not the absolute probability of recurrence, and it can be influenced by differences in follow-up duration, surveillance intensity, and calendar-time effects.
That is why the RMST analysis matters. When the investigators limited the comparison to 5 years, they found no significant difference in disease-free survival between eras. In practical terms, the average recurrence-free time within 5 years was similar, despite the apparent hazard difference in the full survival model. This pattern suggests that the statistically significant hazard ratio may not translate into a clinically important short-term difference in outcomes.
For clinicians, the most relevant takeaway is that de-escalation appears to have reduced overtreatment and surgical morbidity without a clear short-term loss of disease control. Still, because PTC can recur late, the absence of a 5-year difference does not prove equivalence over the long term. Longer follow-up is essential before making stronger claims about oncologic noninferiority.
Expert Commentary
This study is valuable because it examines what guidelines look like in practice, not just in theory. Many studies show that recommendations can be slow to change clinical behavior. Here, the temporal signal is clear: ATA-guided de-escalation was implemented on a large scale in a high-volume center, and it was associated with less intensive treatment and less permanent harm.
Several strengths deserve emphasis. First, the sample size is exceptionally large, which gives the study strong power to detect real shifts in practice and uncommon complications. Second, the authors used exact matching to improve comparability across eras, reducing the risk that the results were simply due to different tumor profiles in different periods. Third, the use of RMST is methodologically thoughtful, because short and unequal follow-up can make hazard-based analyses difficult to interpret in diseases with few events.
At the same time, limitations are important. This was a single-center retrospective study, so generalizability may be limited. Management decisions were likely influenced not only by guidelines but also by surgeon preference, institutional culture, changing imaging quality, and evolving pathology criteria. Surveillance practices also probably changed over time, which can affect recurrence detection. In addition, the study focused on short-term outcomes; in PTC, where recurrence can occur years later, 5-year follow-up may be insufficient to fully assess long-term oncologic safety.
The apparently conflicting findings in recurrence analysis are not unusual in oncology. A lower crude recurrence rate can coexist with a higher modeled hazard ratio when event timing, censoring, and follow-up structure differ across groups. Clinically, this means the data support de-escalation, but they do not eliminate the need for careful risk stratification. Patients with adverse features such as extrathyroidal extension, nodal disease, aggressive histology, or other higher-risk findings were not the target population for routine downstaging, and they still require individualized management.
From a practice standpoint, the study reinforces a broader principle in thyroid cancer care: the goal is not maximal therapy for everyone, but the right amount of therapy for the right patient. For low-risk PTC, a smaller operation and selective use of RAI can preserve excellent outcomes while reducing permanent complications. This is especially relevant for younger patients, patients with bilateral disease considerations, and those in whom lifelong thyroid hormone management may be influenced by the extent of surgery.
Conclusion
In this large South Korean cohort, ATA guideline updates were followed by measurable de-escalation of papillary thyroid cancer treatment, including fewer total thyroidectomies and less RAI use. The result was a substantial reduction in permanent hypocalcemia, with no significant difference in 5-year disease-free survival by RMST analysis. The study supports guideline-driven reduction of overtreatment in appropriately selected patients, while also reminding clinicians that long-term follow-up remains essential in a disease with typically favorable but sometimes delayed recurrence patterns.
Funding and ClinicalTrials.gov
Funding was not reported in the PubMed abstract. ClinicalTrials.gov registration was not applicable because this was a retrospective observational cohort study.
References
1. Ryu JS, Kim EJ, Lee IA, et al. A Temporal Analysis of Surgical Management and Outcomes Following ATA Guideline Updates in Papillary Thyroid Carcinoma: A Real-World Cohort Study of 31,861 Patients in South Korea. Thyroid. 2026-04-09:10507256261442841. PMID: 41954044.
2. Haugen BR, Alexander EK, Bible KC, et al. 2015 American Thyroid Association Management Guidelines for Adult Patients with Thyroid Nodules and Differentiated Thyroid Cancer. Thyroid. 2016;26(1):1-133. PMID: 26462967.
3. Cooper DS, Doherty GM, Haugen BR, et al. Revised American Thyroid Association Management Guidelines for Patients with Thyroid Nodules and Differentiated Thyroid Cancer. Thyroid. 2009;19(11):1167-1214. PMID: 19860577.
AI image prompt
High-resolution medical editorial illustration of a thyroid cancer surgery consultation in a modern hospital, with a thyroid gland anatomical graphic, ultrasound image, pathology slide, and a subtle downward trend arrow representing reduced overtreatment, clean blue-and-white palette, realistic clinical style, no text, professional and informative tone.