Nontargeted circulating tumor DNA (ctDNA) whole-genome sequencing is a novel strategy for genomic characterization of high-grade serous ovarian cancer. Changes in ctDNA levels are a sensitive indicator of disease burden with an average lead time of 6 months to clinical progression. This presents a unique opportunity to identify pathways driving progression as molecular vulnerabilities for clinical drug development.
See related article by Paracchini et al., p. 2549
In this issue of Clinical Cancer Research, Paracchini and colleagues (1) explore shallow whole-genome sequencing (sWGS) of circulating tumor DNA (ctDNA) in patients with high-grade serous ovarian cancer (HGSO). Across cancer types, tumor genomic analyses have already provided critical insights into molecular pathways driving cancer progression, identifying high-priority avenues for drug development. The most striking example of this in the management of HGSO, has been defining the role of PARP inhibitors in patients with tumors harboring defects in homologous recombination (HR). The success of PARP inhibitors has been a dramatic advance for women with pathogenic BRCA mutations, but there remains an acute need to identify tailored strategies for women with HR-proficient tumors. There is optimism that translational initiatives involving longitudinal tumor genomics and characterization of clonal tumor evolution, will contribute to these efforts. Contemporary next-generation sequencing strategies, with their broader coverage and lower costs, have improved our ability to apply genomic analyses to patients in the clinic. However, repeat tumor biopsies can be challenging and may inadequately represent the heterogeneity of a particular tumor genome. Circulating tumor DNA (ctDNA) assays are a promising option to circumvent some of these limitations (2).
Paracchini and colleagues (1) describe sWGS in plasma samples acquired in an aptly named, “convenience cohort” of 46 patients with HGSO. This was a representative, albeit high-risk, population of primarily stage IIIC/IV patients, with 22% harboring a pathogenic germline BRCA1/2 variant. Roughly half underwent upfront surgery followed by adjuvant chemotherapy, while the remainder received neoadjuvant chemotherapy and interval surgery. Over half (24 patients, 52%) had platinum-sensitive disease, 10 (22%) were platinum-refractory/resistant with the remainder defined as partially platinum sensitive, providing a cohort of patients covering the spectrum of relevant HGSO biology. Because plasma and tumor biospecimens were collected at nonuniform timepoints, the authors cleverly cohorted samples into three groups to address a few important questions.
Before considering Paracchini's data, it is important to note that the majority of completed HGSO ctDNA studies to date have involved targeted strategies, primarily focusing on mutational analysis, most often centered on TP53 variants. These studies have demonstrated a prognostic relevance of ctDNA levels, and an improved sensitivity for early detection of disease progression over standard metrics like CA125 or cross-sectional imaging. Unfortunately, the wide range of TP53 mutations across HGSO necessitates this being a personalized assay, individualized for each patient and based on their unique tumor genomics (3). In contrast, Paracchini and colleagues applied an innovative, strategy of sWGS to report on somatic copy-number alterations (CNA). This is a sensible approach because HGSO is defined by chromosomal instability and a high frequency of CNA. Importantly, it is nontargeted, independent of tumor analysis, and can therefore be applied broadly without baseline tumor analysis and assay individualization.
In an initial validation, 57% CNAs were detected in both plasma and tumor, a concordance level comparable with what has been reported with other analyses (4). Furthermore, an informative analysis of the 35 most highly recurrent regions of genomic gains/losses, demonstrated that 33 were hallmark features of HGSO, with the majority (22/33) detectable in both plasma and tumor, confirming that sWGS of ctDNA can provide a good general overview of the tumor genome of HGSO. Tumor fraction (TF) appeared to be prognostic, again aligning with previous reports, and lending confidence that ctDNA sWGS provides a clinically relevant assessment of tumor burden.
The most interesting findings from this work revolve around the temporal dynamics of ctDNA in patients on treatment, and in posttreatment surveillance. It is disconcerting, although not entirely surprising, to recognize that most patients had persistently detectable ctDNA after completing first-line treatment. This was in spite of CA125 normalization, signifying the presence of occult disease, below the detection threshold of standard assays. Among the patient anecdotes presented, the case of patient 21650 (Fig. 5 of Paracchini and colleagues; ref. 1) is particularly instructive, as the dramatic decline and normalization of CA125 in the early phase of treatment, was accompanied by a persistent rise of ctDNA throughout adjuvant chemotherapy. Low CA125 levels likely reflected the success of surgical debulking, with the increasing ctDNA levels indicating the emergence of platinum-refractory disease. In the whole patient cohort, rising ctDNA levels preceded clinical/radiographic progression in the majority of patients by approximately 6 months. These findings highlight the fact that CA125 monitoring alone may provide incomplete information around a patient's cancer burden, and like TP53 monitoring, genomic ctDNA analysis could function as a more sensitive indicator of disease burden in HGSO (3, 5).
Identifying disease progression prior to detectable radiographic progression, allows for the earlier discontinuation of ineffective therapies, sparing patients and health care systems from cumulative physical and financial toxicity. It also allows a window within which, screening for “next-line” clinical trials can be considered. However, unless the earlier detection of progression can be paired with a rationale prioritization of next line of therapy based on evolving treatment-resistant clones, it is unlikely to impact on patient survival, and more likely to confound outcome metrics by lead-time bias. Unfortunately, detection of TP53 ctDNA, although a sensitive indicator of disease burden, is unable to provide biologically relevant information of the dominant pathways driving progression to be considered for therapeutic targeting. This is where the work by Paracchini and colleagues (1) provides some intriguing insights worthy of further exploration.
Posttreatment plasma samples demonstrated lower levels of genomic complexity (copy number instability or CNI), with only 12 of the 35 highly recurrent CNA detected in baseline samples also identified in patient's plasma after treatment. Among these, two specific cytobands were highlighted because of their high frequency of detection: 11q13.3 and 19p13.11. One can hypothesize that a more comprehensive analysis of these regions might identify specific pathways which may be targeted as unique molecular vulnerabilities. For example, the 11q13.3 cytoband is already recognized as encoding genes for several proteins relevant in tumor growth and progression including cell-cycle mediators, folate receptors and angiogenic factors, with drugs targeting several of these already being in the evaluation phase for patients with HGSO (6). The approximately 6-month lead time between ctDNA detection of disease progression and the development of clinically evident disease, provides a unique window to complete genomic analyses and obtain ancillary data in real-time to help guide treatment decisions. One could envision a future where patients might then be stratified into one of several potential, matched treatment options based on their ctDNA make-up as illustrated in Fig 1. Although 11q13.3 has already been reasonably characterized in cancer, the emergence of the 19p13.11 cytoband in treatment-resistant disease represents a novel finding in HGSO, underscoring a mechanism whereby sWGS of ctDNA might provide discovery opportunities for novel therapeutic approaches.
In summary, Paracchini and colleagues (1) describe the novel and practical application of untargeted sWGS of plasma ctDNA in patients with advanced HGSO. Although these preliminary observations still require independent validation, the potential of this approach to facilitate discovery of novel treatment targets and for clinical trial patient selection, are an exciting step forward in advancing precision medicine in HGSO.
No disclosures were reported.