Why the neck is still being scanned years after thyroid surgery

A patient operated on for papillary thyroid carcinoma six years ago sits down with a question that is entirely reasonable and that I am asked most weeks. The cancer was taken out. The pathology was good. Why am I still coming in for an ultrasound of my neck every year, and what exactly is being looked for?

The honest answer is specific rather than vague. The scan is looking at the lymph nodes of the neck, because that is where this cancer comes back if it comes back at all, and because a recurrence found there while it is small is a different clinical problem from one found late.

What the numbers actually look like#

A large series with long follow-up comes from Kuma Hospital in Kobe and covers 5,768 patients operated on between 1987 and 2004 with no distant metastasis at diagnosis. Follow-up ran from twelve months to 280 months and averaged 129 months, which is close to eleven years. Recurrence in the lymph nodes was found in 389 patients, recurrence in the lung in 118, and recurrence in the bone in 33. Fifty-seven patients died of the disease.

Two things follow from that table and both matter in the consulting room. The first is that the neck is where the attention belongs, because node recurrence is several times more common than everything else combined. The second is that the great majority of patients have none of these events, and a surveillance programme is not a prediction that something will happen.

Which patients the risk actually sits with#

Surveillance is more useful when it is aimed. A retrospective analysis of 189 patients operated on for papillary thyroid carcinoma compared the 33 who developed recurrent cervical nodes with the 156 who did not. Four features separated them: a tumour larger than 2 cm, extension of the tumour beyond the thyroid, higher T stage, and lymph node metastasis already present.

The Kuma series found three of those same variables on multivariate analysis — a tumour over 2 cm, nodal metastasis and extrathyroid extension — and added two more, age of 55 or over and male sex, as independent predictors of node recurrence. Of all of them, clinical node metastasis measuring 3 cm or more carried the strongest prognostic weight, and it did so for lung and bone recurrence as well as for the neck. A separate series of 259 papillary microcarcinomas worked the same question from the small end of the disease, mapping the frequency and the pattern of nodal metastasis and recurrence in order to establish what neck dissection should be doing in tumours of a centimetre or less.

Why a node recurrence is not the headline it sounds like#

The word recurrence carries the same weight in every disease when a patient hears it, and it should not. A recurrence in the cervical nodes of a papillary thyroid carcinoma is usually operable, usually found on a scan rather than because the patient noticed something, and does not generally shorten life. The authors of the Kuma analysis described node recurrence as a stressor for patients and physicians rather than as an immediate threat to life, and reserved that language for the lung and the bone.

Age is what separates those two readings. In the Kuma series, age was the strongest predictor of death from the carcinoma, stronger than any feature of the tumour itself. Which is why the same ultrasound finding in a patient of 35 and a patient of 70 does not mean the same thing, and why the follow-up schedule is not identical for the two of them.

There is a second reason the follow-up does not simply stop, and it is the most direct answer to the question this patient asked. A single-institution cohort of 269 patients with papillary thyroid carcinoma was followed for a median of 27 years. Recurrence occurred in 28 percent and cancer-specific death in 9 percent; time to recurrence averaged 8.1 years and time to death 9.0 years, both with wide variation. The figure that matters here is what happened at the far end: 11 percent of the recurrences and 17 percent of the deaths occurred more than twenty years after treatment. The authors concluded that recurrence and death from this cancer can occur more than thirty years after it is treated, and that follow-up should therefore be lifelong.

What I tell the patient at six years#

That the scan is looking at the nodes in specific compartments of the neck, not at the whole body. That the risk is concentrated in people whose original operation found certain things, and that if those things were absent the yearly scan is a formality rather than a vigil. That if something does appear in a node it is nearly always dealt with. And that the point of looking is not to find trouble but to make any trouble small.

Korea diagnoses a great deal of thyroid cancer, and carries a surveillance habit built during the years when screening ran ahead of the evidence for it. An ultrasound ordered out of habit and an ultrasound ordered for a reason look identical to the patient. The difference is whether anyone can say what it is looking for.