Re: Extracapsular tumor extension in cervical lymph nodes: Reconciling the literature and seer data

To the Editor: We would like to take the opportunity to comment on the recently published article by Brannon et al, ‘‘Extracapsular Tumor Extension in Cervical Lymph Nodes: Reconciling the Literature and SEER Data.’’ In the literature, the risk of extracapsular tumor spread (ECS) increases with a larger nodal size; however, ECS can be also be detected in smaller nodes. As the authors discussed, the published literature represents findings from studies consisting only of small numbers of patients and retrospectively obtained data. Therefore, this study analyzed the relationship between nodal size and the frequency of ECS in a larger, more representative, prospectively collected database. For the collection of data, the authors consulted the Surveillance, Epidemiology, and End Results (SEER) registry. A total of 1684 patients eventually could be studied. As the article showed, the association between ECS and nodal size in the current literature was clearly different than the findings after review of the SEER database in this study. Looking for a possible explanation Brannon et al suggested a possible influence of the individual evaluation of the lymph nodes by the pathologist or the lack of a structured protocol for reporting ECS and registration in the SEER. In the literature, the reported incidence of microscopic ECS varies between 21% and 85%, and the incidence of macroscopic ECS varies between 17% and 54%. As the authors discussed, it is unclear whether these different percentages reflect differences between patient populations studied or interobserver variation between pathologists on the interpretation of ECS. Because the presence of ECS is important in determining the treatment protocol, it is important to report on this topic in a universal way. Therefore, we recently performed a study to examine the level of agreement among pathologists on the presence or absence of ECS. The interobserver and intraobserver agreement among 10 pathologists on the diagnosis of ECS in 41 tumor-positive lymph nodes was evaluated. All pathologists evaluated twice the 41 lymph nodes for the presence or absence of ECS. The pathologists were instructed to conduct their evaluation on the basis of their own criteria. The kappa value of the interobserver agreement among pathologists varied between 0.14 and 0.75, and the overall kappa value was 0.42 and 0.49 in the 2 scoring sessions. The intraobserver kappa value varied between 0.49 and 0.95. With respect to these findings, we concluded that the intraobserver and interobserver agreement among pathologists in the assessment of the presence of ECS was low. Taking into account these results, we believe the differences found by Brannon et al probably were caused by interobserver variation between different pathologists determining ECS in the SEER database. In conclusion, because of the widely accepted prognostic significance and therapeutic consequences of ECS, we agree with the authors that there is a need for internationally accepted reproducible criteria for the histopathologic assessment of ECS in metastatic lymph nodes in the neck.