Bringing Blood-Based Molecular Testing to the Clinic

In this issue of Clinical Cancer Research, Chicard and colleagues (1) reported results of genomic copy number profiling of cell-free DNA (cfDNA) in neuroblastoma. The discovery of oncogenic molecular aberrations in cancer has provided key insights into the mechanisms of tumorigenesis, tumor biology, and prognosis and has led to the development of therapies directed at specific druggable targets (2). Some of the most significant examples of genomically targeted therapeutics include BRAF inhibitors in melanoma harboring BRAF mutations, ABL kinase inhibitors in BCR-ABL–positive chronic myelogenous leukemia, and EGFR tyrosine kinase inhibitors in non–small cell lung cancer (NSCLC) bearing EGFR mutations (2–4). Traditionally, oncogenic aberrations have been tested using DNA isolated from archival tumor tissue samples; however, inadequate samples are often a limiting factor that can preclude molecular analysis. In addition, the molecular profile of the primary tumor tissue or of an isolated metastasis does not necessarily reflect the genetic makeup of other metastatic disease due to tumor heterogeneity (5). Indeed, different oncogenic alterations can even occur in different areas of a primary tumor (6). In addition, in at least 15% to 30% of cases, there is variability between the mutation status of primary tumor and of distant metastases, suggesting that cancer genotype can evolve over time (6, 7). Therefore, making therapeutic decisions using molecular data from historical archival tumor samples can be problematic. Having a technique that can elucidate the molecular profile of cancer in real time is of special importance for furthering precision cancer therapy.

cfDNA is released into the circulation from cells undergoing apoptosis or necroptosis in primary or metastatic cancer lesions or in the tumor microenvironment and can be identified in the blood samples collected from patients with cancer (Fig. 1; ref. 8). Unlike tissue biopsies, obtaining samples of plasma-derived cfDNA is a minimally invasive approach with minimal risk and cost to patients. Plasma cfDNA can be used to assess molecular profile at different time points and provide valuable information about genetic changes that occur during the disease trajectory, as cancer progression is not a static process. In addition to identification of molecular targets for cancer therapy, molecular testing of cfDNA can provide additional information about prognosis, evaluate response to therapy, reveal disease progression or recurrence, and detect early emergence of molecular abnormalities that drive resistance to systemic therapy (8). Agreement rate between the molecular profile of cfDNA and archival tumor tissue is deemed to be acceptable and ranges from 70% to 100%, depending on the timing of collection, technology (PCR vs. next-generation sequencing), number of genes tested, and biological factors, such as heterogeneity (8–12). Recently, both the European Medicines Agency and the FDA approved a PCR-based cfDNA test to detect EGFR mutations in NSCLC as an alternative to molecular testing of tumor tissue.

Chicard and colleagues (1) investigated the utility of genomic copy number profiling of cfDNA in samples from pediatric patients with neuroblastoma of all stages. The authors are to be commended for conducting a study on a dataset of samples from 70 patients, which can help to bring cfDNA-based molecular profiling to pediatric patients with neuroblastoma and, if translated to clinical care, could improve treatment outcomes. This can be of special importance as copy number alterations have prognostic impact in neuroblastoma and can dictate different therapeutic approaches. The authors focused on detecting numerical chromosome alterations, segmental chromosome alterations, and MYCN amplifications. Because of the nature of these abnormalities, the authors decided to use a relatively simple and cost-effective method, the molecular inversion probe–based OncoScan array, which was designed to detect small fragments of DNA. Results of molecular testing of cfDNA for the above-mentioned copy number alterations were compared with the results of standard comparative genomic hybridization array in paired tumor samples. Overall, agreement between tests was found in 47 (76%) of 62 cases with copy number alterations in the tumor tissue. However, cfDNA testing did not reveal copy number alterations in 10 samples with known alterations in the tumor tissue, and five samples of cfDNA revealed copy number alterations not present (n = 4) or different (n = 1) than those detected in the tumor tissue. The authors suggest that these differences are due to tumor heterogeneity, which is a plausible explanation. However, it is also conceivable that these differences reflect technical factors related to the limits of detection and/or the optimal ratio between sensitivity and specificity, all of which are common issues in cfDNA studies. In general, accurate assessment of copy number alterations in cfDNA has been technically challenging due to our lack of capability to differentiate circulating tumor DNA from nonmalignant cfDNA. Studies with capture-based next-generation sequencing of plasma cfDNA from patients with diverse advanced cancers demonstrated higher than expected rates of copy number alterations in oncogenes, such as MET (10). Although it is certainly possible that tumor cells with MET amplification can be more likely to shed cfDNA into the circulation, observed differences may also be attributable to the algorithm used for cfDNA assessment. Therefore, studies comparing cfDNA and standard tumor tissue testing for copy number alterations are of utmost importance.

In summary, testing for copy number alterations in plasma cfDNA from patients with neuroblastoma offers an attractive tool for molecular diagnostics and other applications, such as determination of prognosis. Presented retrospective data suggested acceptable sensitivity. However, the clinical utility remains to be proven in prospective studies. Also, cfDNA occurs in short fragments, which can further complicate molecular analysis (8). Finally, cfDNA originates from dying cells, and it is unclear to what extent these cells represent the genotype of the prevailing population of malignant cells. Therefore, alternative blood-based sources of tumor nucleic acids need to be investigated (8). Circulating tumor cells contain full-length DNA, which can be used for molecular detection of copy number alterations; however, it remains to be shown whether or not circulating tumors cells have genotypic features representing the population of tissue tumor cells. Alternatively, exosomes are small vesicles, which are actively secreted into the circulation from tumor and normal cells. Exosomes contain proteins, an abundance of RNA and full-length DNA. Therefore, using exosomal RNA or DNA can provide material for molecular testing, which might be less challenging than cfDNA.

F. Janku is a consultant/advisory board member for Guardant Health and reports receiving commercial research grants from Biocartis and Trovagene. R. Kurzrock is an employee of and has ownership interest (including patents) in Curematch, Inc. and Novena, Inc., is a consultant/advisory board member for Actuate Therapeutics, Sequenom, and Xbiotech, and reports receiving commercial research grants from Foundation Medicine, Genentech, Guardant, Merck Serono, Pfizer, and Sequenom. No other potential conflicts of interest were disclosed.

Conception and design: R. Kurzrock

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): F. Janku, R. Kurzrock

Writing, review, and/or revision of the manuscript: F. Janku, R. Kurzrock

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): F. Janku