Consensus Classifier Algorithm Developed to Aid in Tumor Subtype Identification
Posted: Tuesday, October 9, 2018
According to a study published in Clinical Cancer Research, a novel ovarian tumor subtype classifier has been identified by Levi Waldron, PhD, of the City University of New York (CUNY) School of Public Health, and colleagues from the University of Toronto and Dana-Farber Cancer Center, Boston. It is their hope that the classifier will help to clarify the relevant subtypes in terms of risk factors, treatments, and outcomes.
“We hope this study will help put to rest the ongoing questions of how to define subtypes of ovarian cancer and will provide a model for how to define and compare molecular subtypes for other cancer,” stated Dr. Waldron in a CUNY press release.
The consensus classifier represents the subtype classifications of tumors based on the consensus of multiple methods, explained the researchers regarding the study design. “Using our compendium of expression data, we examine the possibility that a subset of tumors is unclassifiable based on currently proposed subtypes.”
In a meta-analysis of publicly available expression data, the researchers used a compendium of 15 microarray data sets that included 1,774 high-grade serous ovarian tumors. The authors found that overall survival seemed to depend on the molecular subtype, “with the worse prognosis subtype having only half the survival rate of the best prognosis subtype,” added Dr. Waldron.
The study proposes the consensus classifier algorithm identifies only the tumor subtypes that have been assigned at a high confidence interval. Ovarian tumors that fall outside the criteria for the consensus identification require ongoing research to determine whether their ambiguity is due to a combination of subtypes or other reasons not yet known.