In precision oncology we are seeking to deliver treatments that are appropriate for an individual patient based on specific molecular or clinical features of the disease or of the patient. The most precise and scientifically clear-cut examples are the use of a specific drug treatment that targets a clearly defined molecular characteristic of a tumour. This treatment works in patients whose tumours carry that characteristic, and does not work in patients whose tumours do not. One of the earliest examples of precision oncology was the introduction of trastuzumab for the treatment of human epidermal growth factor receptor 2 (HER2)-positive breast cancers.1 Our increasing knowledge of the molecular pathology in cancers through application of genomic and proteomic technologies has resulted in the recognition of tumour subtypes, with particular molecular signatures, within most cancer sites. Some molecular signatures are found in cancers derived from multiple different organs.
Biomarker-guided therapeutics in precision oncology can depend on nucleic acid or protein markers, imaging or clinical characteristics. An added layer of complexity arises from the use of multiple biomarkers (e.g. complex genetic signatures or proteomic profiles) or the integration of biomarkers (e.g. molecular markers with imaging) to predict efficacy of treatment and patient outcome. Against this complex background, the direct link between a biomarker and the mechanism of action of a specific treatment may be obscured. Some aspects of precision oncology have an empirical basis without a clear mechanistic link between marker, treatment and target.
*These authors contributed equally.
The challenges of clinical trials in precision oncology
The increasing availability of biomarkers and treatments for use in precision oncology has generated substantial challenges in the design of clinical trials to ensure they produce robust answers, obtained efficiently with achievable patient numbers and at affordable cost. Key challenges include:
The availability of reliable, reproducible, standardized and valid biomarker assays that can be measured promptly and cheaply to inform the choice of therapy.
Identification of the mechanisms of treatments used in precision oncology, and adequate insight into 'off-target' effects.
The availability of high-quality tissue samples representative of the whole tumour.
The availability of study designs that can robustly identify the clinical utility of biomarker-treatment combinations in appropriate patient populations.
Identifying the correct time and place to deploy new technologies such as next generation sequencing (NGS) or circulating tumour DNA (ctDNA).
Patient selection strategies and trial designs that can provide robust answers about the value of biomarkers present in small proportions of patients with a given tumour type.
Pitfalls of study design in precision oncology include the failure to distinguish between a biomarker that selects a therapy that has improved efficacy in the biomarker-positive subpopulation, and a biomarker that only indicates a group of patients with a better prognosis. This is especially true when all patients in the biomarker study have been exposed to the therapy in question. When a ...