Abstract
PURPOSE: Many cancers can be treated with targeted therapy. Almost inevitably, tumors develop resistance to targeted therapy, either from pre-existence or by evolving new genotypes and traits. Intratumor heterogeneity serves as a reservoir for resistance, which often occurs as a result of the selection of minor cellular subclones. On the level of gene expression, clonal heterogeneity can only be revealed using high-dimensional single-cell methods. We propose using a general diversity index (GDI) to quantify heterogeneity on multiple scales and relate it to disease evolution. MATERIALS AND METHODS: We focused on individual patient samples that were probed with single-cell RNA (scRNA) sequencing to describe heterogeneity. We developed a pipeline to analyze single-cell data via sample normalization, clustering, and mathematical interpretation using a generalized diversity measure, as well as to exemplify the utility of this platform using single-cell data. RESULTS: We focused on three sources of patient scRNA sequencing data: two healthy bone marrow (BM) donors, two patients with acute myeloid leukemia-each sampled before and after BM transplantation, four samples of presorted lineages-and six patients with lung carcinoma with multiregion sampling. While healthy/normal samples scored low in diversity overall, GDI further quantified the ways in which these samples differed. Whereas a widely used Shannon diversity index sometimes reveals fewer differences, GDI exhibits differences in the number of potential key drivers or clonal richness. Comparison of pre- and post-BM transplantation acute myeloid leukemia samples did not reveal differences in heterogeneity, although biological differences can exist. CONCLUSION: GDI can quantify cellular heterogeneity changes across a wide spectrum, even when standard measures, such as the Shannon index, do not. Our approach can be widely applied to quantify heterogeneity across samples and conditions.