A thousand consecutive Korean robotic thyroidectomies report four times the nerve injury that pooled data gives the approach. The difference is who looked

Korean surgeons perform more remote-access thyroid surgery than anyone else. The operation reaches the gland through the armpit or the inside of the lip, leaving no scar on the neck, and in a country that diagnoses a great many small thyroid cancers in young women it has an obvious appeal.

This publication reported earlier today that the pooled international figure for transient nerve injury after the transaxillary approach — 1 percent — describes who was selected for the operation rather than what the operation does to the nerve. A single-institution Korean series of a thousand consecutive cases now allows that claim to be tested against Korean practice.

The Korean numbers#

Surgeons at the Catholic University of Korea reviewed 1,000 consecutive single-port transaxillary robotic thyroidectomies performed between 2021 and 2025, classifying hypoparathyroidism as transient up to twelve months and permanent beyond it, and confirming nerve injury by postoperative laryngoscopy.

Mean operative time was 79.1 minutes. Recurrence over a mean follow-up of 22.7 months was 0.8 percent with no disease-specific deaths, a result the authors themselves label short-term.

Why the gap is a measurement gap#

A transient vocal cord palsy is only counted if somebody looks for it. A unit that performs laryngoscopy on every patient after surgery finds palsies in people whose voices sound normal to them and to the surgeon. A unit that investigates only patients who complain of hoarseness records a fraction of the same events.

The cohort also proves the selection point#

Who had these operations is described precisely, and it is exactly the population that makes cross-approach comparisons unsafe.

• Mean age 40.9 years.
• 87.2 percent women.
• Mean tumour size 1.1 cm.
• Less than a total thyroidectomy in 83.0 percent; total thyroidectomy in 8.9 percent.
• Malignant on final pathology in 85.1 percent, of which papillary carcinoma accounted for 98.2 percent.

A cohort of forty-year-old women with 11 mm papillary cancers having one lobe removed is not comparable to an unselected series of thyroid operations, however it is analysed. That is not a criticism of the study, which does not claim otherwise. It is a caution about the comparison the numbers invite.

In the less-than-total group, larger tumours and gross extension beyond the thyroid were independently associated with nerve injury. In the group having total thyroidectomy, completion surgery or modified radical neck dissection, no independent risk factor for hypoparathyroidism was identified.

The other route in#

A retrospective study across four high-volume centres compared the two transoral techniques in 2,144 patients operated on between 2017 and 2024: 546 by the endoscopic vestibular approach and 1,598 robotically. Total thyroidectomy accounted for 9.6 percent and unilateral lobectomy or isthmectomy for 82.8 percent — again, mostly limited operations.

Robotic operations took substantially longer, a mean of 167.5 minutes against 121.3. The authors conclude that both are safe and oncologically sound in appropriately selected patients, that the robotic approach offers lower rates of nerve injury and hypoparathyroidism and may suit more complex cases, and that the endoscopic approach remains an effective scarless option for benign and low-risk disease. The qualifier in that conclusion — appropriately selected — is doing the same work as it does everywhere else in this literature.

What is still missing#

1. The transaxillary series is one institution and one team. High-volume single-centre results are not what a patient elsewhere should expect, in either direction.
2. A mean follow-up of 22.7 months cannot establish oncologic outcomes for a cancer whose recurrences appear over decades. The authors say so.
3. There is still no randomised comparison of remote-access against conventional thyroidectomy, so nothing here can attribute a complication rate to an approach rather than to the patients who receive it.
4. The four-centre transoral study's complication figures were not available to us beyond the authors' summary conclusion, so this article reports that conclusion and does not quantify the difference it describes.
5. Cosmetic outcome, which is the reason most of these operations are chosen, is not measured by any figure in this article.