Response to “American Thyroid Association Statement on Surgical Application of Molecular Profiling for Thyroid Nodules: Current Impact on Perioperative Decision Making”

Dear Editor:

We read with interest the statement of the Surgical Affairs Committee of the American Thyroid Association (ATA) (1). We are pleased that the ATA recognizes the positive and growing impact of molecular diagnostic tests on the treatment of patients with cytologically indeterminate thyroid nodules.

We are, however, concerned that the statement lacks evaluation of study design, places undue emphasis on some test metrics over other equally important metrics, and lacks certain conflict disclosures that could create bias in the reader's interpretation.

Clinical care guidance is typically based on a hierarchy of high-quality analytical and clinical data. Data from double-blind, multicenter studies with expert pathology review are most likely to reflect the true performance of any molecular diagnostic test and therefore should be given greater weight over those from unblinded studies conducted among a small number of (or single) centers and local pathology review. The ATA statement lacks such an evaluation, implying that the described studies are comparable in terms of depth of evidence, thereby failing to provide readers with important information.

Another concerning element of the statement is the emphasis on knowing malignancy prevalence within individual institutions in order to estimate negative predictive value and positive predictive value. The authors do not emphasize that sensitivity and specificity, equally important metrics, are not influenced by malignancy prevalence and do not change from institution to institution. Without this key distinction, readers would miss the significance of many of the results summarized in the statement, which include findings from several studies with alarmingly low sensitivity to detect malignancy across the indeterminate cytology subtypes. The Surgical Affairs Committee recognizes the need for personalized care in parts of their statement, but this is overshadowed by the more insistent message that clinicians should use the institutional pretest risk of malignancy for thyroid nodule management. Unfortunately, this approach ignores individual patient clinical characteristics and directs clinicians to treat based on institutional averages—a step backward in an era of personalized medicine.

Lastly, it is surprising that the statement authors did not go further to report dualities of interest. The first and last authors on this statement are employees of the University of Pittsburgh, which commercially provides thyroid molecular testing independently and through a business relationship with CBL Pathology and therefore should have been included in the Author Disclosure Statement. We are not opposed to individuals with commercial relationships being included as authors; rather, we object to the bias introduced when these relationships are not disclosed and when authors representing other commercial tests are excluded from the process.

In a 2013 statement on molecular testing, the ATA Clinical Affairs Committee concluded that consensus from the ATA Guidelines Task Force was needed before any recommendation could be provided. It is unclear why the Surgical Affairs Committee has been able to circumvent this process. Aside from creating concerns about the validity of organizational guidelines going forward, there is a more concerning implication for patients when practice-changing guidance is published without the due diligence, systematic review process, and careful management of potentially conflicting interests that have traditionally characterized ATA guideline development.