The pros and cons of incorporating transcriptomics in the age of precision oncology

The treatment of cancer continues to evolve toward personalized therapies based on individual patient and tumor characteristics. Our successes and failures in adopting a precision-oncology approach have demonstrated the utmost importance in identifying the proper predictive biomarkers of response. Until recently, most biomarkers were identified using immunohistochemistry for protein expression or single-gene analysis to identify targetable alterations. With the rapid propagation of next-generation sequencing to evaluate tumor tissue and "liquid biopsies," identification of genomic biomarkers is now standard, particularly in non-small cell lung cancer, for which there is now an extensive catalog of biomarker-directed therapies with more anticipated to come. Despite these great strides, it has also become apparent that using genomic biomarkers alone will be insufficient, as it has been consistently shown that at least one-half of patients who undergo tumor genomic profiling have no actionable alteration. This is perhaps to be expected given the remarkable breadth of nongenetic factors that contribute to tumor initiation and progression. Some have proposed that the next logical step is to use transcriptome profiling to define new biomarkers of response to targeted agents. Recently, results from the WINTHER trial were published, specifically investigating the use of transcriptomics to improve match rates over genomic next-generation sequencing alone. In this review, we discuss the complexities of precision-oncology efforts and appraise the available evidence supporting the incorporation of transcriptomic data into the precision-oncology framework in the historical context of the development of biomarkers for directing cancer therapy.