Korea now has a guideline for not operating on thyroid cancer. Who takes it up is partly a question of income

A small papillary thyroid cancer is the clearest case in medicine of a diagnosis that may not need to become a treatment. Korea, having found more of them than anywhere else, is now the country working out what to do about the ones already found.

In June 2025 the Korean Thyroid Association published the country's first guideline on active surveillance — watching a confirmed cancer rather than removing it. The evidence that this is safe is strong. The evidence about who actually ends up doing it is more uncomfortable.

What the guideline permits, and what it demands#

Surveillance is offered to adults with a pathologically confirmed Bethesda V–VI microcarcinoma and none of the features that make a small tumour dangerous: no clinical evidence of lymph node or distant metastasis, no gross extension beyond the thyroid, no invasion of the trachea or recurrent laryngeal nerve, and no aggressive histology.

It also demands things of the health system rather than only of the patient. The baseline assessment requires high-resolution neck ultrasound performed by an experienced operator, precisely to exclude those features; contrast-enhanced neck computed tomography is optional. Follow-up is neck ultrasound and thyroid function tests every six months for two years and annually thereafter, and significant tumour growth or newly detected nodal or distant metastasis triggers surgery.

The safety case#

The strongest Korean number comes from a review of 8,969 patients with papillary thyroid microcarcinoma treated at a single tertiary centre between 1994 and 2017. Over that period 107 of them died — 1.2 percent — and two died of the thyroid cancer.

Of the 107 who died, 63 — 58.9 percent — died of another malignancy. The comparison the authors draw is with papillary carcinoma larger than 1 cm, where cancer-specific death was 0.71 percent. Their conclusion is that these figures support active surveillance as an option.

A meta-analysis of 21 studies covering 9,397 patients puts numbers on the surveillance course itself: disease progressed in 14.53 percent (95 percent confidence interval 9.59 to 21.43) and 14.91 percent went on to delayed surgery, with no thyroid cancer-related mortality recorded in either the surveillance or the immediate surgery group.

The part of that meta-analysis that gets left out#

Its authors also found that patients who came to surgery after a period of surveillance had higher overall complication and recurrence rates than patients operated on immediately, along with higher rates of extrathyroidal extension and multiplicity. Their conclusion recommends surveillance and says it should be undertaken with caution, and both halves belong in any summary of it.

How much of that difference is caused by waiting is not established. Progression is not random: the tumours that declare themselves during surveillance are the ones that were going to behave worse, so the delayed surgery group is selected for difficulty before anyone chooses when to operate. What the finding does establish is that surveillance is not free for the roughly one patient in seven who ends up in theatre anyway.

Who chooses it#

The KoMPASS cohort is a Korean multicentre prospective study that enrolled 927 patients with low-risk microcarcinoma — 453 who chose surveillance and 474 who chose immediate lobectomy — and asked what predicted the choice. Mean age was 47.4 years and 72.2 percent were women.

Tumour size and family history of thyroid cancer were also associated with the decision. But income and prior awareness are of a different kind from the others, because neither is a fact about the tumour. They describe what a patient arrives with.

The quality-of-life result, read carefully#

The surveillance group reported better quality of life over follow-up than the lobectomy group. It is tempting to read that as an effect of avoiding surgery, and the study's own numbers argue against doing so.

• Quality of life scores were already higher in the surveillance group at baseline — 7.1 against 6.7 — before either group had been treated.
• By twelve months the two groups were no longer significantly different.
• The meta-analysis found quality of life and anxiety broadly similar between surveillance and immediate surgery.

A difference present before treatment and gone within a year is better explained by who chose what than by what was done to them. That reading is available in the paper; it tends not to survive the retelling.

What is not settled#

1. Median follow-up in the KoMPASS cohort was around 28 months in both groups. That is long enough to compare quality of life and far too short to compare cancer outcomes.
2. The 0.02 percent mortality figure comes from patients treated at one tertiary centre, most of whom had surgery. It describes the disease, not the surveillance strategy.
3. The guideline's baseline assessment depends on high-resolution ultrasound performed by an experienced operator. Where that operator is not available, the eligibility criteria cannot be applied as written, and nothing in the published evidence tells us how uniformly that capacity is distributed across Korea.
4. Comparisons between surveillance and immediate surgery are between patients who chose, not patients who were randomised, in every study cited here.