Characterization of a Novel Entity of G3 (High-grade Well-differentiated) Colorectal Neuroendocrine Tumors (NET) in the SEER Database.

Abstract

To the Editor: I have read with interest the recent article published in AJCO entitled “Characterization of a Novel Entity of G3 (High-grade Well-differentiated) Colorectal Neuroendocrine Tumors (NET) in the SEER Database,”1 and I would like to thank the authors for their efforts. That being said, the core concept of the article is based, in my opinion, on a misunderstanding on how Surveillance, Epidemiology, and End Results (SEER) database information is being organized and reported. In the methods section, the authors described their method in defining the study cohort as follows “We grouped patients on the basis of tumor grade and differentiation into the following cohorts: G1-2 NET (defined as low-intermediate grade [grades 1-2] well-differentiated histology [ICD 0-3= carcinoid]), G3 NET (defined as HG [grades 3-4] well-differentiated histology [ICD 0-3= carcinoid]), and NEC (defined as HG [grade 3] poorly differentiated NETs [ICD 0-3= small cell neuroendocrine]).” I am not sure from where the authors concluded that the 4-tiered grading system reported in the SEER database is equivalent to World Health Organization (WHO) grading published in 2010, 2017, or 2019 (which is based on Ki67 and/or mitotic index). I am hereby summarizing my main objections to the conclusions of this study: (1) Ki67 and mitotic index are not reported within the SEER database. These are the pillars of the grading system according to WHO 2010 classification. How can we trust any subsequent conclusions from this study? (2) The study duration is from 2000 to 2015 (ie, almost two-thirds of the study population were recruited before 2010); so, it is impossible that they could have been graded by the WHO grading system of 2010 or its later modifications. (3) As defined in the SEER documentation itself, grading is based on differentiation only (not Ki67 or mitotic index) (https://seer.cancer.gov/tools/ grade/). So, the whole premise of the study (which discusses discordance between differentiation and KI67-based grading) is wrong. (4) WHO 2010 classification and its subsequent revisions recommended not using the term carcinoid. So, having cases labeled as carcinoid indicate probably that they either have been diagnosed before the WHO 2010 classification or they have been diagnosed by a pathologist not knowledgeable about WHO classification. In both cases, this information cannot be used to derive any credible conclusions.