Clinical Utility of Molecular Genetic Cancer Profiling

Abstract

The ability to sequence entire genomes and transcriptomes is revolutionising cancer biology and, more slowly, cancer medicine. The evidence that large‐scale sequencing has clinical utility, that it improves outcomes, is slowly accumulating. A major controversy has centred on whether or not all tumours refractory to standard therapy should undergo therapy selection by molecular genetic cancer profiling (MGCP). Despite several completed or ongoing clinical trials, this issue is not settled; however, other forces now ensure that MGCP will become a routine diagnostic tool. Biopsies are becoming smaller, providing greater safety for patients, but yielding less DNA/RNA while there are now sufficiently many molecularly targeted agents approved in major cancers that testing by single‐gene assays is no longer practical or economical. Key Concepts A clinical test does not have inherent clinical utility. The clinical utility of molecule genetic cancer profiling varies with the intended clinical use: diagnosis, prognosis, treatment selection, and cancer risk assessment. Assessment of clinical utility does NOT include consideration of cost. There is incomplete consensus on which genes should be in a molecular genetic cancer profile but in practice panels cover at most several hundred genes (ie much smaller than an exome). Most molecular genetic cancer profiling has used tumor genomic DNA but assays are increasingly including analysis of RNA for translocation detection and gene expression. The traditional method for determining clinical utility, the randomized clinical trial, is evolving to facilitate screening of many genes and many molecularly targeted agents rapidly. Large‐scale clinical trials are underway to assess the clinical utility of applying molecular genetic panels to all cancers in a histology‐agnostic fashion. Clinical trials for molecularly targeted agents predicted to be beneficial by molecular genetic cancer profiling are always conducted on patients with advanced disease who have already failed several established ‘lines of therapy’. Likelihood of response to a new class of therapy, immune checkpoint inhibitors, might be predicted by microsatellite instability and, possibly, by the Tumor Mutation Burden. Response to another new class of therapy, CAR‐T cells, might be mutation agnostic.