| Abstract: |
Background: Tall-cell variant (TCV) is widely believed to be a more aggressive subtype of papillary thyroid carcinoma (PTC). Despite the significance of TCV with respect to risk stratification and therapeutic decision making, its diagnosis is subject to inter-observer variability. This study aimed to determine the level of agreement among expert pathologists in the identification and reporting of TCV. Methods: Seventeen surgical resections for thyroid cancer containing the diagnostic term "tall cell" in their pathology reports and 22 cases diagnosed as classical PTC were selected. Cases were digitalized, and 14 expert pathologists reviewed the scanned slides blinded to the original interpretation. Each pathologist designated each case as TCV or not and answered multiple questions about diagnostic histopathologic features of TCV. Results: The overall strength of agreement for identifying TCV was fair (Fleiss kappa 0.34), and the proportion of observed agreement was 0.70. Of 22 cases originally diagnosed as PTC classical variant, 15 (68%) were reclassified as TCV by at least one expert pathologist. It was noted that four different definitions for TCV were used by the participants based on various combinations of cell height to width (H:W) ratio and the percentage of tumor cells showing that specific ratio. All pathologists agreed that the diagnosis of TCV does not rely solely on a specific H:W ratio. Conclusions: Pathologic reporting of TCV varies among pathologists. This disagreement is a result of the lack of unanimous diagnostic criteria and variation in individual pathologists' interpretations. These discrepancies lead to over- and under-diagnosis of TCV, which has significant implications in patient management. It is imperative to understand this variability in diagnosis TCV as it relates to risk stratification and interpretation of clinical studies related to this histologic subtype of PTC. Further studies are needed to reach consensus on the diagnostic criteria of TCV. © Copyright 2017, Mary Ann Liebert, Inc. 2017. |
