You don't need to be signed in to read BMJ Blogs, but you can register here to receive updates about other BMJ products and services via our site.

Primary Care Corner with Geoffrey Modest MD: Natural history of thyroid nodules (and why do we care)

18 Mar, 15 | by EBM

By: Dr. Geoffrey Modest

An Italian study looked at the natural history of apparently benign thyroid nodules, with unexpected findings (see JAMA.2015;313(9):926-935​). In this prospective, multicenter, observational study, researchers followed 992 consecutive patients with 1-4 asymptomatic nodules which were either benign by ultrasound or fine-needle aspirate (FNA), and followed them for 5 years.



–mean age of patients was 52.4 years, 82% women, half with family history of nodular thyroid disease, and none were on levothyroxine therapy. 60% of the nodules were solitary , 80% solid.

–40.2% of the nodules were benign by FNA and  59.8% had no suspicious ultrasound features [suspicious ultrasound features were at least one of: hypoechogenicity, irregular margins, taller-than-wider shape, intranodular vascular spots, and microcalcifications)

–nodule growth of >20% occurred in 153 patients (15.4%), or in 11.1% of the nodules, with a mean 5-year diameter increase of 4.9 mm (initial mean of 13.2 increasing to 18.1 mm).

–nodule growth was associated with the presence of multiple nodules, larger main nodule volume (0.2mm), male sex.   Age<45 was associated with more growth than ​age>60.

–nodules shrank spontaneously in 184 patients (18.5%), with mean shrinkage of 3.7mm

–thyroid cancer was found in only 5 of the original nodules (0.3%), and only 2 had grown during the followup period. An incidental cancer was found in 1 patient who had a thyroidectomy, but that cancer was not visualized prior to surgery.

–all of the cancers had abnormal ultrasound findings (most were solid and hypoechoic)

–93 patients developed new nodules over the course of the study (9.3%), with cancer in 2.


–a small minority of nodules increased in size, and these increases were noted within a year or so, and happened mostly in those with multinodular disease, nodule diameters >7.5mm,  and those <44 years old

–cancer was quite rare (0.3%) and FNA had a very low false negative rate (1.1%)

–in those with cancer over the 5 years, nodule growth was not a specific marker of malignancy, and all had suspicious ultrasound features initially​.

–although these areas in Italy had mild to moderate iodine deficiency (which limits the generalizability of the results to the US, which tends to be iodine sufficient), there was no difference in findings in the northern Italy area (mild deficiency) and the southern (moderate deficiency), suggesting this might not be a factor

–so, the authors recommend a redo of the current guidelines (repeat thyroid ultrasound after 6-18 months, and if nodule size is stable, then every 3-5 years (this is based not on studies, but on “expert opinion”). They suggest:

–nodules that are benign on initial FNA, or those which are subcentimeter and have no sonographic suspicious features, can be safely managed with repeat ultrasound in 1 year. And if no suspicious changes, reassessed after 5 years.  They note that this approach would apply to 85% of patients whose risk of disease progression is very low.

–closer surveillance “may be appropriate for nodules occurring in younger patients or older overweight individuals with multiple nodules, large nodules (>7.5mm), or both”

So, pretty different from the current conception. Nodule growth has been considered high risk for malignancy. Current guidelines ​suggest that people with solid nodules that are not hypoechoic have FNA when nodules are  >1-1.5cm, and those which “are growing but are benign after repeat biopsy should be considered for continued monitoring or intervention”. Although I am hesitant to apply this large study to the US, given potential differences in iodine sufficiency, it certainly makes sense to look at our data, given the very large number of thyroid nodules and how common referrals are made to endocrinologists for nodule evaluation/FNA.

Data from the American Association of Clinical Endocrinologists suggests that palpable thyroid nodules in the US population are common (3-7%), the risk of cancer (4-6.5%) is the same in palpable nodules, those picked up incidentally on other scans (eg, MRIs or CTs, where the incidence of clinically inapparent nodules/incidentalomas being 20-76%!!!), and in those with multinodular goiters. And, to my reading, there are no studies which look at the issue of the incidence of cancer in those with totally nonsuspicious thyroid nodules by ultrasound (there are data on the specificity of each of the individual abnormalities, with wide ranges — eg for hypoechogenicity it is 41.2-92.2%). The overwhelming issue to me, however, is that even though thyroid nodules are extraordinarily common (even in the US which is iodine sufficient) with malignancy predicted in the 4-6.5% range, the 2015 governmental predictions of thyroid cancer in 2015 is 62,450 cases (which seems very low given the up to 76% of incidentalomas alone), but thyroid cancer deaths is remarkably rare (1950 deaths predicted in 2015). And over the past several years the diagnosis of thyroid cancer has increased (largely from increased use of thyroid ultrasounds), but this death rate has not changed. So, it has always been unclear to me that detecting and treating thyroid cancer is in fact appropriate (ie, are people dying largely from very aggressive anaplastic cancer, or stage IV tumors that were so aggressive that early detection would not have helped, and the vast majority never would have had any problems with their cancers which happened to look malignant under the microscope).

So, it seems to me that we need data showing that screening does anything in terms of real mortality and morbidity benefit (and I do have a few patients with significant morbidity from surgery), and if so, whether there is a reasonable way to risk stratify those at high risk.  In terms of  of risk stratification, this Italian study would argue to me that the ultrasonographic characteristics are a good place to start and could dramatically decrease the followup of the vast majority of thyroid nodules, possibly the morbidity associated with FNA and possibly unnecessary surgery, and patient/provider angst.

By submitting your comment you agree to adhere to these terms and conditions

Comments are closed.

You can follow any responses to this entry through the RSS 2.0 feed.
EBM blog homepage

Evidence-Based Medicine blog

Analysis and discussion of developments in Evidence-Based Medicine Visit site

Creative Comms logo

Latest from Evidence-Based Medicine

Latest from EBM