Comparison of Local Histopathology and a Central Pathology Panel in Diagnostic Thyroid Nodule Surgery from a Multicenter, Blinded Study


Interview with Michael Shanik, M.D., Endocrine Associates of Long Island

Dr. Michael Shanik is an endocrinologist and partner at Endocrine Associates of Long Island, a private group practice in Smithtown, New York. He spoke to us about a poster he is presenting at ENDO 2017 regarding the challenges of accurate histopathology diagnosis of thyroid nodules.

Q: What was the purpose of this study?

Dr. Shanik: We wanted to see to what degree local and a central, expert panel of histopathologists agreed on the diagnosis when reviewing thyroid nodule samples.

Q: How did you go about doing this?

Dr. Shanik: We looked at data for 492 nodules where the patient underwent surgery and a diagnosis was made by local histopathology. In each case, a panel of three expert pathologists, blinded to the local diagnosis and clinical findings, reviewed the slides and also made a diagnosis. We derived the data from the 47-center ENHANCE trial, which utilizes a biorepository of paired cytology, genomic and histopathology samples from thyroid nodule patients. Most of the nodules we reviewed had indeterminate cytopathology results.

Q: What did you find?

Dr. Shanik: We found a significant discrepancy between local and central histopathology. Specifically, among the 194 nodules that were classified by the central panel as malignant, local pathologists disagreed with the central panel’s diagnosis nearly 30 percent of the time. There was even discordance among the central panel, with only 78 percent of cases having “high confidence” calls in which there was strong panel-member agreement. Among benign nodules, there was greater local-central concordance. In this category, 94 percent of the local diagnoses agreed with the central panel.

Q: What is the significance of these findings in terms of patient care?

Dr. Shanik: This study reinforces the complexity of diagnosing thyroid nodules that are indeterminate by cytology as definitively benign or malignant – even among a panel of three expert academic thyroid pathologists. I think that molecular testing can provide a more objective assessment to help physicians better classify these nodules at the pre-surgical stage so that they know what to do next for their patients. This can help patients avoid surgery that they don’t need and, conversely, can help guide them toward surgery when it is warranted. Ultimately, everyone involved, including the patient, can make earlier and better-informed decisions.

Our findings also tell us that there is a limit to the performance that any molecular test for indeterminate thyroid nodules can achieve. If pathologists cannot agree 100 percent on the “clinical truth,” then no test can achieve 100 percent sensitivity and specificity. The performance of any test can only be as good as the reference standard.

Q: Is there anything else you’d like to add?

Dr. Shanik: From the patient’s perspective, genomic testing that can provide better answers and enable a more accurate diagnosis – earlier in the process, before surgery is needed – is helpful, especially because it can reduce the anxiety that often accompanies having an indeterminate thyroid nodule.

Download PDF