Bioinformatics roadmap for therapy selection in cancer genomics

Abstract

Tumour heterogeneity is one of the main characteristics of cancer and can be categorised into inter- or intratumour heterogeneity. This heterogeneity has been revealed as one of the key causes of treatment failure and relapse. Precision oncology is an emerging field that seeks to design tailored treatments for each cancer patient according to epidemiological, clinical and omics data. This discipline relies on bioinformatics tools designed to compute scores to prioritise available drugs, with the aim of helping clinicians in treatment selection. In this review, we describe the current approaches for therapy selection depending on which type of tumour heterogeneity is being targeted and the available next-generation sequencing data. We cover intertumour heterogeneity studies and individual treatment selection using genomics variants, expression data or multi-omics strategies. We also describe intratumour dissection through clonal inference and single-cell transcriptomics, in each case providing bioinformatics tools for tailored treatment selection. Finally, we discuss how these therapy selection workflows could be integrated into the clinical practice.

Keywords: bioinformatics; drug prioritisation; next-generation sequencing; precision oncology; treatment selection; tumour heterogeneity.

Fig. 1

Roadmap for drug prioritisation from different omics profiles. The roadmap is represented as an underground map in which each next‐generation sequencing (NGS) technique is a different line and each step in the workflow is a station. Common steps between workflows are displayed as interchange stations. The names of the available tools are preceded by coloured symbols that indicate in which technique they can be applied. BQSR, base quality score recalibration; DGE, differential gene expression; QC, quality control.

Fig. 2

Concept of clonetherapy. The conventional approach in cancer treatment is to target the major subclone, since it is the most represented in a bulk sample. However, if the drug does not hit clonal alterations, other subclones might survive the treatment and expand. Clonetherapy aims to hit both clonal and subclonal alterations identified in a deconvoluted data set in order to target all subclones, thus avoiding potential relapse.

Fig. 3

Integration of drug prioritisation methods for clinical decision‐making during the cancer patient journey. Integrated bioinformatics analysis of clinical and multi‐omics data from individual cancer patients would generate a report that includes tumour genomics profiling during the patient journey. Based on such profiles, drug prioritisation methods would provide predictions to propose tailored treatments during the different stages of disease progression. The genomics report would be completed with retrospective treatment response information obtained by comparison with other patients with similar clinical and genomic profiles. Data retrieved at each step of the patient journey would be stored in federated databases for aiding future clinical decisions. DFS, disease‐free survival.