Recent findings from studies confirm the promises of ctDNA in this setting, but unanswered questions still remain about timing and methodology
In colorectal cancer (CRC), liquid biopsy has emerged as a powerful and attractive tool for detecting minimal residual disease (MRD), monitoring early relapse, profiling molecular characteristics, and predicting therapeutic response. The utility of circulating tumour DNA (ctDNA) has expanded from its initial use in the advanced/metastatic context for molecular profiling and detecting acquired resistance mechanisms, to the identification of MRD and early detection. While its use can now be extended to any type of tumour, CRC has the most compelling clinical evidence for the application of liquid biopsy. (J Clin Oncol. 2022;40:2846-2857; Ann Oncol. 2022;33:750-768).
Up to one third of patients with localised CRC experience a relapse despite receiving curative-intent treatment based on surgery and adjuvant chemotherapy (ACT). The TNM staging system, which currently forms the basis for our therapeutic decisions, does not appear to be sufficient for stratifying patients according to the risk of relapse and selecting those who will benefit from additional treatment. For this reason, there is a particular interest in the utility of ctDNA to detect MRD and avoid overtreatment of some CRC patients while ensuring that others are not undertreated. Numerous retrospective studies have demonstrated that the detection of ctDNA, not only immediately after surgery but also after completing ACT, is associated with worse disease-free survival (DFS) and may indicate resistance to standard treatment (JAMA Oncol. 2019;5:1710-1717; Ann Oncol. 2019;30:1804-1812; Clin Cancer Res. 2022;28:507-517). Additionally, the detection of ctDNA after surgery precedes radiological relapse by a median of approximately 8.2 months.
Another important finding that emerged from these studies is that a high percentage of patients with detectable ctDNA after surgery do not show negative ctDNA after completing treatment. These findings have been confirmed in prospective studies (N Engl J Med. 2022;386:2261-2272; Nat Med. 2023;29:127-134), suggesting that hat the detection of MRD may define a subset of patients with an aggressive phenotype and that standard chemotherapy may not be sufficient to eradicate MRD.
The DYNAMIC trial was the first prospective, randomised study to demonstrate that ctDNA can guide therapeutic decisions in patients with stage II CRC. Among the patients randomised to the ctDNA-guided management arm, those with detectable ctDNA received ACT, while those without detectable ctDNA were observed. The study achieved its primary endpoint, showing that a ctDNA-guided approach reduced the use of ACT without compromising DFS.
At ESMO 2022, the same research group reported that among patients without ctDNA detection, recurrences were predominantly local, in contrast to patients with detectable ctDNA where recurrences occurred in distant organs (Abstract 318MO). Additionally, the GALAXY study, an observational study within the CIRCULATE-JAPAN trial, showed that the absence of ctDNA clearance after completing ACT was also associated with worse DFS, and the rate of ctDNA that did not become undetectable after treatment was 32% (Nat Med. 2023;29(1):127-134).
More clinical trials based on the utility of ctDNA to guide therapeutic decisions after surgery are currently underway, and solid conclusions are eagerly awaited.
More recently, new data from the same study have highlighted the importance of ctDNA status at the MRD time point post-surgery as a prognostic factor for patient outcomes. This factor proves to be the most significant risk factor, regardless of BRAF V600E status (JCO. 2023.41.16_suppl.3521)
Also, the INTERCEPT programme has provided confirmation to our current knowledge that the sensitivity to detect MRD increases during follow-up and only a small percentage of patients during surveillance can be considered MRD+ (47%), as 53% experience radiological relapse at the same time. These findings indicate that we need to improve the sensitivity to detect MRD after surgery or upon completion of ACT, or even reconsider the follow-up approach for patients with CRC (JCO.2023.41.16_suppl.3522)
Despite the promises of liquid biopsy, we should not forget it still has some limitations. At best, the assay has a sensitivity of 87.5% (considering mutation tracking) and a specificity close to 100%. Currently, there are two approaches to addressing MRD detection, employing a tumour-informed approach or a tumour-agnostic approach. There are no studies comparing both assays, but some doubts have emerged about the incorporation of methylation/epigenomic markers that could limit the sensitivity of the plasma-based assay alone (JCO Precis Oncol. 2022;6:e2100298; ESMO Open. 2022;7:100547).
So, many important questions remain to be addressed, specifically about which are the best assay to detect MRD - a tumour-informed approach or a tumour-agnostic approach -, the best timing for treating MRD - immediately after surgery or at any point during follow-up - and whether targeted therapy should be used instead of standard chemotherapy.
There are many studies ongoing in this setting and in the coming years we will likely witness a significant shift in the therapeutic paradigm of CRC.