To the Editor In a recent issue of JAMA Internal Medicine, Smith-Bindman et al1 addressed the question of which ultrasound imaging characteristics are useful in identifying patients with thyroid nodules who are more likely to have thyroid cancer. This study adds important information for the proper clinical decision making in patients with thyroid nodules; however, some aspects need further clarification. First and foremost is the definition of noncancerous nodules, which was already addressed in the Invited Commentary.2 The definition adopted for benign nodules was based on the assumption that no cancer was identified, when in fact it is unknown how many of these patients underwent fine-needle aspiration or surgery. Indeed, the degree of uncertainty regarding this aspect is crucial and somewhat worrying because all analyses presented were based on this assumption. Second, the size of the nodule is another aspect that deserves further discussion. In this context, a receiver operating characteristic curve analysis of the size of nodules would have added value to identify possible thresholds of size associated with increased risk of harboring a thyroid cancer. This is clinically relevant because recent reports observed that approximately 19% to 30% of patients with thyroid microcarcinoma may present a more aggressive tumor behavior with regional and distant metastasis.3,4 Finally, it would be interesting to present the posttest probability of each ultrasound imaging characteristic by using the Fagan nomogram.5 By doing so, the unusual low pretest probability of thyroid cancer (1.6%) in this sample of patients with thyroid nodules would have been taken into consideration, providing better estimates of the clinical applicability of the studied ultrasonographic parameters.
Gross JL, Kramer CK, Remonti LR. Clinical Decision Making in Patients With Thyroid Nodules. JAMA Intern Med. 2014;174(6):1005-1006. doi:10.1001/jamainternmed.2014.42