Re: David J. McConkey, Woonyoung Choi, Colin P.N. Dinney. New Insights into Subtypes of Invasive Bladder Cancer: Considerations of the Clinician. Eur Urol 2014;66:609–10

Abstract

Following the Cancer Genome Atlas marker paper on invasive urothelial bladder cancer (UC) in early 2014 [1], two groups affiliated with the consortium published separate papers on molecular subtypes of muscle-invasive UC (MIUC) [2,3]. All three papers cite our work on defining a molecular taxonomy for UC [4] and use themuscle-invasive (MI) tumors (n = 93 of 308) from our 2012 study as a validation resource. In the original work by the group at MD Anderson Cancer Center (MDA), the interrelations between the three-tier MDA classification system and our five-tier taxonomy are discussed in some detail and are shown to have broad commonalities [2]. However, in a recent Platinum Journal opinion article highlighting the potential usefulness of molecular subtyping in the context of UC management, the taxonomy based on both non– muscle-invasive (NMI)UC(NMIUC)andMIUCwasomitted in favor of the twoMIUC-derived classifiers [5]. In a subsequent review, the same authors expand on the connections [6] but seemtoharbor thebelief that intrinsic subtypes ofMI tumors are distinctly different from those of NMIUC. This view is difficult to reconcile with the concept of progression to MI disease and does not make biologic sense, because the underlyingbiologyof a tumor—that is, the intrinsic subtype— is not contingent on the pT stage, a feature previously highlighted [4]. By focusing on theMI subset of UC, theMDA-based group defined a classification scheme that may seem superficially distinct from our taxonomy but also removed a large proportion of the full biologic and clinical spectrum of UC. It is in this light that we object to some aspects of the work by the group at MDA. As recently acknowledged [6], the group from MDA approached us in early 2012 and asked us to apply our (Lund) classification system to amixed set of NMI and MI tumors run on the same Illumina platform used to derive our taxonomy; applying the Lund classification system, the MDA and Lund cohorts showed excellent agreement in terms of subtype distribution and underlying biology. This agreement, of course, is a consequence of the