The most fundamental event in the development of cancer is the alteration of genetic information in the cancer cells that leads to autonomous growth. Changes may include especially mutations, copy number changes, and gene rearrangements. Gene expression may also be altered in similar fashion by epigenetic changes. Specific DNA alterations may or may not be relevant to the biology of the tumor, and this will be reflected in the utility of these alterations as DNA biomarkers. Driver alterations can be used to select treatment. The overall mutation rate may also be a relevant marker, as reflected in the use of mutational burden as a marker of response to checkpoint immunotherapies [7].

In a typical example of identifying an important DNA biomarker in a PDX model, Kortmann et al. established a PDX derived from a patient with ovarian serous carcinoma and a germline BRCA2 mutation. The PARP inhibitor olaparib alone and in combination with carboplatin markedly inhibited growth in this model but not in a second serous carcinoma PDX with normal BRCA status [8]. A subsequent randomized clinical trial showed that BRCA status can be used to enrich ovarian cancer patients responsive to olaparib [9].

The new frontier in DNA biomarkers is the measurement of DNA alterations in the blood of patients with cancer. Tumor DNA spills into the bloodstream with the natural turnover of tumor cells, and this circulating cell-free tumor (ct)DNA can be identified within the background of plasma DNA derived from normal cells and hematopoietic cells [10–13]. In its simplest form, a recent study showed that the amount of detectable ctDNA correlates with the patient tumor burden [14, 15]. ctDNA can, however, be exploited for much more with next-generation sequencing, which allows precise determination of tumor-specific DNA alterations. Since the ctDNA should reflect the entire tumor burden, it may overcome some of the limitations of tumor heterogeneity that arise with tissue sampling. Because it only requires a blood draw, ctDNA is an assay that can be repeated longitudinally to monitor disease progression and response to therapy. Treatment-induced alterations in the genomic landscape can be identified, and appropriate, rational changes in therapy can be instituted. This is discussed below in the context of precision oncology.

Synonymous with the decoding of genes has been the translational function of messenger RNA (mRNA), converting the genetic information into functional proteins [16]. However in the last two decades, the discovery of many different regulatory nonprotein-coding RNA (ncRNA) has revolutionized the understanding of fundamental biological mechanisms and consequences of dysregulation resulting in disease [17]. ncRNA has also become the focus of diagnostic and therapeutic biomarker development. The FDA has approved several RNA sequencing tests (Table 9.1) [18].

Proof of principle studies to validate in vitro biomarker discoveries are being performed using PDX models. These models closely reflect the patient tumor microenvironment but enable biomarker investigation without burdening the patients directly. Crea et al. demonstrated this principle in PDX models of prostate cancer [19]. In RNA sequencing of paired metastatic and nonmetastatic PDX, they identified the long non-coding RNA PCAT18 as the most highly upregulated transcript. Cancer-specific upregulation of PCAT18 was confirmed in an independent prostate cancer patient cohort. PCAT18 was also detectable in plasma samples, and levels increased with more advanced disease.

The same group in another study compared differential RNA expression between a metastatic and nonmetastatic prostate cancer PDX and discovered a circulating microRNA signature that differentiated localized from metastatic prostate cancer [20]. A subsequent analysis of patient specimens showed some overlap with the PDX-derived signature [21].

A more recent study has taken a more comprehensive approach to the identification of RNA biomarkers in PDX [22]. A group linked to AstraZeneca performed RNA sequencing on 79 PDX models from various solid tumors. Since the human stroma of PDX is replaced by mouse stroma within a few passages, species-specific RNA sequencing allowed this group to perform comprehensive analysis of interactions between the human tumor and the murine microenvironment. They were able to establish independent tumor and stromal biomarkers. The clinical relevance of this approach remains to be determined.

As DNA and RNA biomarkers have strongly influenced our understanding of cancer dynamics in terms of disease identification, progression, treatment modulation, and development of resistance to treatment, so has the development of proteomics. The proteome represents the entire set of proteins modified or produced by an organism [23]. To some degree, the proteome represents the final product of the innumerable events that happen at the DNA and RNA level. Analysis of the human genome has identified over 20,000 protein-coding genes (an integrated encyclopedia of DNA elements in the human genome [24]). Posttranslational modifications add to the complexity of the proteome. A wide variety of protein cancer biomarkers has been described, but only a small number has been approved for clinical use by the FDA (Table 9.2).