The concept of oligometastatic cancer was originally presented in the 1990’s by Hellman and Weichselbaum (1). This theory posits that a population of patients with truly limited sites of metastatic disease exists in which aggressive local therapy may improve outcomes and, in some cases, even effect a cure. The evidence for the potential curability of some patients with oligometastatic disease derived originally from series of liver metastasectomy in colorectal cancer (2) as well as lung metastasectomy with several types of primary cancer (3).

Stereotactic body radiation therapy (SBRT), or stereotactic ablative radiotherapy (SABR), represents a popular option for treatment of oligometastatic disease given its non-invasive nature compared to metastasectomy, and its delivery of ablative doses of radiation. SBRT is characterized by hypofractionated (limited number of treatments) external delivery of radical irradiation dose with highly conformal treatment planning targeted at a well-defined target volume, employing modern techniques of stereotactic or image-guided patient setup (4). Multiple prospective studies have now demonstrated that SBRT may benefit the population with oligometastatic disease (5, 6). Presently open studies (including the STOP and HALT studies) intend on assessing the effect of the addition of SBRT to standard systemic therapy for oligoprogressive cancer (7). Future directions include the assessment of the role of SBRT in the population of oligometastatic patients receiving immunotherapy given the potential for release of tumor antigens and stimulation of an immune response to tumor (8).

A controversy that has emerged for the management of oligometastatic disease is how to truly define it clinically. While a common definition utilized historically is five or fewer metastases (9), various definitions have been employed both for clinical trial entrance criteria in recent studies and for defining this entity to assess for appropriateness of SBRT (10). Multiple recent studies have employed a more liberal definition with allowance for >5 total metastases including brain metastases, consistent with the general observation that protocols have generally focused on extent of extracranial disease in their definitions of oligometastatic disease (11, 12). Newer clinical trials of metastatic NSCLC have employed a cutoff of up to 6 extracranial metastatic sites for classification of oligometastatic disease; this was irrespective of extent of intracranial metastatic disease, so long as patients (a) did not have leptomeningeal carcinomatosis (aggressive metastatic involvement of meningeal regions) and (b) underwent brain-specific treatment via whole-brain radiation therapy (WBRT) or stereotactic radiosurgery (10, 13, 14).

Despite efforts to properly define oligometastatic disease, there is likely a dichotomy between patients with truly a limited metastatic disease burden, and those who will progress to diffuse widespread disease in a short period of time. This dichotomy is likely driven by biology, and several attempts have been made to identify biologically the patients with truly limited metastatic disease. Thus far, these attempts have focused predominantly on measurement of serum levels of circulating markers, and this has been complicated by the sensitivity of detection (15).

A recently published series demonstrated that genomic profiling acquired from serum samples could predict the responses to systemic therapy in advanced NSCLC (16). This profiling was able to successfully segregate patients into genomic clusters that were prognostic for treatment response. Given the predictive power of this genomic profiling technique and the non-invasiveness of a liquid biopsy platform, we decided to use this technique in an attempt to identify a genomic profile for oligometastatic extracranial disease. We sought to do so using a population of patients with primary lung cancer, focusing on NSCLC We focused on patients with metastatic disease secondary to NSCLC, the most common type of lung cancer and a leading cause of brain metastases worldwide (17).

In summary, we present a gene mutational scoring system derived from liquid biopsy that purports to risk stratify patients based on the likelihood of oligometastatic extracranial disease. In this study, the oligometastatic signature predicted not only for the presence of oligometastatic extracranial disease but also for oligoprogression, defined as progression in 5 or fewer extracranial metastatic sites. While the interval validation presented herein is derived from patients with known brain metastases who underwent SRS for definitive treatment of their intracranial metastatic disease, the aim of this study is to develop a genomic risk score to guide treatment decisions pertaining to their extracranial disease.

The potential advantage of identifying an accurate biomarker for oligometastatic extracranial disease is to distinguish the multiple phenotypes within the oligometastatic population, and properly triage their management based on the predicted benefit of treatment. At the present time, clinical trials continue to treat oligometastatic patients as a homogeneous population. However, nearly 50% of patients who present with oligometastatic NSCLC will experience distant progression within 2 years with many of these representing polymetastatic progression (31). Early data suggests that a proportion of patients with oligometastatic disease do benefit from more aggressive treatment (32, 33). If a biomarker could accurately identify the population who would experience early polymetastatic failure, then those patients could be spared aggressive local therapy. Conversely, patients with a true oligometastatic phenotype could be treated with consolidative ablative therapies such as SBRT.

Unfortunately, identification of a clinically useful biomarker for oligometastatic disease in NSCLC has to date proven elusive. The ability of present imaging to determine oligometastatic patients who will have early polymetastatic progression is suboptimal (34). There have been recent efforts to identify cytogenetic biomarkers for true oligometastatic disease. Attempts have included correlation of disease burden with levels of circulating tumor DNA (ctDNA) derived from liquid biopsies to the burden of metastatic disease (35). At this time, however, ctDNA sensitivity for oligometastatic disease is problematic due to the extremely low levels found in the peripheral blood. Another strategy for biomarker identification has been to attempt detection of circulating miRNAs, as these molecules are thought to be regulators of metastatic development and have even been proposed as the molecular basis of prevention of polymetastatic progression (36). The limitations of using circulating miRNAs as a biomarker for oligometastatic disease include issues with variability of miRNA levels from one patient to another given an equivalent burden of disease (37).

The present effort to identify an oligometastatic biomarker uses a genomic signature derived from non-invasive liquid biopsy. Unlike previously described attempts at using liquid biopsy to determine oligometastatic status, the present study uses multiple biomarkers to derive a risk score. The genomic platform used in the study has been validated for use in NSCLC and is widely available clinically (38). The CCNE1 and ARID1A genes were found to be associated with polymetastatic progression. Both ARID1A (39) and CCNE1 (40) play important roles in regulating the cell cycle. Approximately 10% of NSCLC patients have ARID1A gene mutations and these are associated with a poor prognosis (41). CCNE1 is a cyclin that modulates activity of cyclin-dependent kinases for progression through multiple phases of the cell cycle. CCNE1 mutation has been demonstrated in NSCLC to be increasingly overexpressed in higher stage disease, which leads to a worsened disease specific survival (42).

The genes that more associated with oligometastatic phenotype were ATM, JAK2, MAP2K2 and NTRK1, and these play important roles in cellular signaling pathways (43–46). ATM is involved in DNA damage response, JAK2 in cytokine signaling, MAP2K2 in the MAPK pathway, and NTRK1 in neurotrophic signaling. The role of these genes in cellular signaling suggests that dysregulation may contribute to the establishment of an oligometastatic state, wherein cancer cells exhibit a more controlled and localized spread. This distinction highlights the potential dichotomy between the molecular mechanisms driving polymetastatic (cell cycle driven) and oligometastatic disease (cell signaling-driven) in NSCLC.

For any genomic signature for oligometastatic disease, it is important to demonstrate whether the signature is associated with oligometastatic (as opposed to polymetastatic) progression. In the present series, those with a positive risk score for oligometastatic disease were twice as likely as those with neutral or negative risk score for oligometastatic disease to have an oligometastatic pattern of progression. In a randomized phase II study of patients with NSCLC with oligometastatic disease, the median progression free survival was significantly improved in the arm receiving SBRT to oligometastatic disease (33). While the trial by Gomez was not powered to assess the pattern of progression, patients who were not treated with SBRT were more likely to experience failure in known sites of disease (oligometastatic progression). Similar results were found with subsequent phase II prospective studies of oligometastatic NSCLC patients (5, 6). That the present study showed that the oligometastatic signature led to greater oligometastatic failure leads to a potential clinical utility of this genomic biomarker to triage patients to SBRT.

There are several limitations to the present study. First of all, the limited sample size leads to potential that a larger dataset may arrive with a more robust model. The dataset was derived from patients with brain metastases, and it is unclear whether a validation dataset of patients without brain metastases would provide a similar genomic signature. While oligometastatic disease is a prognostic factor for patients with brain metastases in general (11), it is possible that the subset of brain metastasis patients with NSCLC and oligometastatic disease may have their own pattern of behavior and progression. In spite of these limitations, the study did show that patients with the oligometastatic genomic signature were also more likely to fail in an oligometastatic pattern, whereas those oligometastatic patients without the oligometastatic signature had a greater degree of polymetastatic failure - serving as a form of internal validation. In this study, we suggest a trend towards improved survival (at 6 months) with the positive oligometastatic risk score, consistent with our finding that oligometastatic patients on presentation exhibited superior survival outcomes compared to their non-oligometastatic counterparts across time. At the same time, patients with the oligometastatic genomic signature tended to exhibit faster rates of oligoprogression. In a possible interpretation of these seemingly disparate findings, our data may indirectly support the notion of oligoprogression as a distinct entity that may be more amenable to local therapies associated with improved survival (25). As such, future directions include validation with a larger and more diverse dataset, potentially in patients without brain metastases. Ideally this should be done at an independent institution from which the primary data were derived. Should this genomic signature prove valid, it would provide a clinical tool that could be used for management of oligometastatic patients.