A Revolution in Classifying Cancer?

Andre Goy, MD

Editor-in-Chief of

Oncology & Biotech News

Chairman and Director Lymphoma Division Chief John Theurer Cancer Center at HackensackUMC Chief Science Officer and Director of Research and Innovation Regional Cancer Care Associates Professor of Medicine, Georgetown University

In a recent online edition of Cell, a very critical paper appeared reporting on a study conducted as part of The Cancer Genome Atlas (TCGA).

This is the largest study of its kind to date, looking at molecular and genetic characteristics of more than 3500 tumor samples of 12 different cancer types using a multi-platform approach. This is important since previous molecular studies have mostly focused on 1 technology at a time and on pathologically defined tumor types identified as ‘‘within-a-tissue/ organ’’ disease subtypes.

This new approach looked at comprehensive genomics using 6 different platforms— mostly genomics such as DNA and RNA sequencing, plus a protein expression profiling.

An extensive bioinformatics analysis was then performed across all platforms to identify subgroups with similarities (clusters), which led to defining distinct molecular subtypes of cancer (different pathways). A total of 11 cancer types were identified, which provided independent and clinically relevant prognostic information above and beyond tumor stage and primary tissue-of- origin. Based on this study, 1 in 10 cancer patients would be classified differently by this new molecular taxonomy versus our current tissue-of-origin tumor classification system.

This is obviously potentially relevant for the management of these patients (particularly in the era of novel targeted therapies), since the proportion of potentially misclassified tumors is comparable to the rate of EGFR mutations in unselected non—small cell lung cancers (treated with radically different therapy).

Also in the study, the divergence of cancer within one subtype or the overlap among different cancers (from different organs) was provocative. For example, at least three subtypes of bladder cancer were identified with different outcomes; one of these subtypes was virtually indistinguishable from lung adenocarcinomas, and another was most similar to squamous cell cancers arising from head-and-neck tumors.

The research highlighted and confirmed known differences between subtypes of breast cancer, but also revealed the new and surprising finding that basal-like breast cancers actually constitute their own cancer class. Basal-like breast cancers, which are commonly referred to as triple-negative breast cancers, are particularly aggressive and more prevalent among African-American and younger women. At the molecular level, these basal-like breast cancers might have more in common with ovarian cancers and cancers of squamous cell origin than with other subtypes of breast cancer.

Though obviously this still needs validation, the potential of these data is nothing less than revolutionary—as the number of cases studied increases (reflecting more of the cancer diversity), the authors of the Cell paper suggest that up to 30% to 50% of cases might eventually be reclassified! The new classification approach may (should) shift the field to focus more on the development of drugs targeting larger groups of cancers with genomic similarities, as opposed to a single tumor type. We should be thankful to this remarkable work, which will no doubt challenge our views…I believe for the better.

Hoadley KA, Yau C, Wolf DM, et al. Multiplatform analysis of 12 cancer types reveals molecular classification within and across tissues of origin [published online August 7, 2014]. Cell. doi:10.1016/j.cell.2014.06.049.