Study suggests use of ultrasound criteria may reduce unnecessary thyroid biopsies

Certain characteristics of thyroid nodules, obtainable by ultrasound, could be used to predict the risk of thyroid cancer and reduce unnecessary biopsies, according to a recent study.

The retrospective case-control study included 8,806 patients who underwent 11,618 thyroid ultrasounds between 2000 and 2005. According to the California Cancer Registry, 105 of them were subsequently diagnosed with thyroid cancer. Researchers found three ultrasound nodule characteristics that predicted risk of thyroid cancer—microcalcifications (odds ratio [OR], 8.1; 95% CI, 3.8 to 17.3), size greater than 2 cm (OR, 3.6; 95% CI, 1.7 to 7.6) and an entirely solid composition (OR, 4.0; 95% CI, 1.7 to 9.2). Results were published online by JAMA Internal Medicine on Aug. 26.

The researchers calculated that if every nodule with any one of these characteristics was biopsied, the sensitivity would be 0.88 (95% CI, 0.80 to 0.94), but the false-positive rate would be high (0.44; 95% CI, 0.43 to 0.45) and the positive likelihood ratio would be low (2.0; 95% CI, 1.8 to 2.2), and one cancer would be found for every 56 biopsies. If two characteristics were required, sensitivity would be 0.52 (95% CI, 0.42 to 0.62), the false-positive rate would be 0.07 (95% CI, 0.07 to 0.08), the positive likelihood ratio would be 7.1 (95% CI, 6.2 to 8.2), and there would be one cancer found for every 16 biopsies.

The researchers calculated thyroid cancer risks at 2 per 1,000 for those with none of the characteristics, 18 per 1,000 for those with one, 62 per 1,000 with two characteristics and 960 per 1,000 with all three characteristics. Compared to some recommendations that currently call for biopsying all nodules over 5 mm, only biopsying those with two of the characteristics would reduce unnecessary biopsies by 90%, with only 5 per 1,000 unbiopsied patients having cancer, the study authors said. They called for a large prospective cohort study to validate these findings.

An accompanying commentary offered several critiques of the study, including that cancer prevalence was only 1.6% in the study population compared to other estimates of 8% to 15% of evaluated nodules. Current practice patterns shouldn't be altered by these findings, the editorialist said, and “nodules larger than 1.0 to 1.5 cm, especially those that are solid, are hypoechoic and contain microcalcifications should be sampled.” The study was also accompanied by a perspective article about one physician's experience with thyroid cancer diagnosis and treatment.