Minimal residual disease (MRD) refers to a small number of cancer cells remaining in a patient’s body after oncological treatment with curative intent. MRD does not cause clinical symptoms and is not detectable by traditional methods such as imaging (e.g., CT, MRI, PET, ultrasound, X-ray, etc.) or abnormal blood serum protein levels. MRD assays must be sensitive enough to detect as little as 1 cancerous cell in a background of 1 million noncancerous cells. This extreme sensitivity makes MRD assay development complex, but overcoming this barrier imbues MRD assays with the potential for much earlier detection of cancer recurrence than traditional methods, and thus earlier intervention with adjuvant chemotherapy or second and third-line treatments.

Detection of MRD indicates the failure of treatment to eliminate all cancerous cells which implies greater probability of future disease recurrence. Monitoring MRD signals during the remission period allows early detection of disease progression (1). Assaying for MRD, therefore, offers a means to monitor, predict, and detect cancer recurrence, as well as stratify patients for adjuvant therapy and predict their response to treatment ( Figures 1A-C ) (2, 3). However, with the lack of a solid tumor to sample, traditional biopsy methods cannot be applied to MRD assays. Instead, MRD detection must make use of circulating tumor-derived entities such as circulating tumor cells (CTCs), extracellular vesicles, and circulating tumor DNA (ctDNA). These circulating molecules and cells shed by tumors are a rich source of information on cancer biology, featuring much of the genetic and epigenetic abnormalities that drive tumorigenesis and treatment response ( Figure 1D ) (4, 5). Such assays can be used to match patients to effective treatments by detecting clinically actionable variations (e.g., drug resistance) within the detected cancer cells and their molecular products, thus avoiding invasive procedures and the morbidities of ineffectual therapies. These novel diagnostic assays have utility in predicting both MRD and cancer recurrence. A positive MRD test after surgery can inform the decision for: (i) further surgery with curative intent; (ii) adjuvant chemotherapy; and/or (iii) increased surveillance, including imaging such as an FGD-PET scan. Whereas positive recurrence indicated by a quantitatively increasing signal over time can inform the decision for: (i) increased surveillance; and (ii) a switch to secondary/tertiary therapy based on the clonal evolution of the tumor burden, as determined by the assayed biomarkers being tracked (i.e., somatic mutations).

The marketplace for MRD liquid biopsy assays is new, but several MRD diagnostic tests have now received early regulatory approvals ( Table 1 ). In 2021, the Natera Signatera test for MRD was granted Breakthrough Device Designation (BDD) status by the FDA for an intended use in Stage I-IV colorectal cancer. This approach uses whole-exome sequencing of primary tumors along with matched normal blood samples to prepare bespoke personalized multiplex PCR assays. In total, 16 patient-specific somatic variants are selected to track using next-generation sequencing (NGS) (39, 40). The Inivata RaDaR test for residual disease and recurrence has received a Conformité Européenne (CE) mark for use in Europe and was granted FDA BDD status in 2021 (41, 42). Again, exome-sequencing informs the design of a secondary somatic mutation PCR panel and tracks up to 48 mutations in blood using NGS. These two-staged approaches are stated by their respective companies to take between 2 and 4 weeks for assay design and up to 1 week turnaround on subsequent blood draws. Finally, the FoundationOne Tracker test was awarded BDD status in 2022 and combines Foundation Medicine’s patient-specific NGS tumor profiling with the multiplex PCR design of Natera (43). This test is used to detect and track MRD, assess a patient’s response to therapy, and monitor for relapse following curative intent therapy.

While there are only a few somatic mutation-based MRD panels in the marketplace, there are a host currently in development and undergoing clinic trials. For example, Strata Oncology are also exploring a two-staged tumor-informed panel design in their Sentinel trial for cancer recurrence (44). Burning Rock Biotech has developed brPROPHET, a personalized capture panel-based approach for MRD detection and are currently working with BeiGene to progress this assay to clinical studies (45). Foresight Diagnostics (46) claims that Phased Variant Enrichment & Detection Sequencing (PhasED-seq), where two or more mutations occur on the same strand of DNA, offers more sensitivity for MRD. The company intends to develop a CLIA validated assay and market the assay as a lab-development test (LDT). Exact Sciences is running the 750 patient CORRECT-MRD II study to validate their MRD and recurrence test for stage II and III colorectal cancer. This is based on the Target Digital Sequencing (TARDIS) approach that Exact Sciences licensed from The Translational Genomics Research Institute (47).

Notably, genetic mutations are not broadly shared across cancers and the two-stage solutions to this shortcoming are limited by tumor heterogeneity and the volume of a tissue biopsy. That is, because only a subset of a tumor is sequenced there is still a high probability that mutations will be missed. However, a multi-omics approach can augment this shortcoming. The greater frequency and universality of epigenetic changes makes them more sensitive and universal markers for evidence of cancer after curative intent treatment, and somatic mutation testing highlights actionable variants that can inform subsequent treatment. This promise of MRD assays that incorporate epigenetics with somatic mutation testing is being realized. Guardant Health has combined somatic mutations, DNA methylation and fragmentomics into their LUNAR panel, which they have progressed into clinical trials for both primary detection and MRD of early-stage colorectal cancer (48).

Multiple companies are also progressing purely epigenetic-based MRD panel approaches. MethylGene is undertaking a clinical trial on multiple myeloma patients that will utilize DNA methylation sequencing of ctDNA for MRD detection (49). While GRAIL has concentrated on bringing the DNA methylation-based Galleri test to market for early detection of cancer, in 2021 the company also formed a deal with Amgen, AstraZeneca, and Bristol Myers Squibb to evaluate its technology for the detection of MRD and early recurrences (50). This expansion of use case by GRAIL points to the versatility of ultra-sensitive liquid biopsy assays. Fast turnaround and inexpensive serial epigenetic tests can identity MRD and recurrence early. A positive result provides the evidence to undertake imaging and to order large NGS panels to study clonal evolution of the tumor and indicate modes of treatment.

While most product development for MRD is concentrated around sequencing large panels of markers, simple DNA methylation-based PCR tests can also be highly efficacious. Colvera, developed by Clinical Genomics, is available as an LDT in the USA and has Medicare coverage for MRD and recurrence monitoring of colorectal cancer. This methylated ctDNA test detected 66.0% of recurrence, significantly higher than the 31.9% sensitivity of carcinoembryonic antigen (CAE), the current standard of care (51), and also has potential for identifying MRD due to treatment failure (52). Another PCR test is Bladder EpiCheck offered by Nucleix. This multiplex DNA methylation-based PCR assay is for monitoring of tumor recurrence in conjunction with cystoscopy in patients previously diagnosed with bladder cancer (53, 54). Also in development, Singlera Genomics is recruiting colorectal cancer patients to run clinical trials on their ColonAiQ six-plex methylation PCR test for MRD and cancer recurrence (55).

Interestingly, a simple well-designed PCR test can have adequate diagnostic power when compared to a large NGS panel. This is illustrated when comparing the rates of primary detection of colorectal cancer across the GRAIL-sponsored circulating cell-free genome atlas (CCGA) (27) study (Trial identifiers: NCT02889978 and NCT03085888) and Colvera trial data (56) (Trial identifier: ACTRN12611000318987). Both are large prospective observational studies and both report similar rates of detection across the four cancer stages, despite the GRAIL technology sequencing the entire ccfDNA repertoire and the Clinical Genomics test being only a two-gene PCR assay. This suggests the liquid biopsy detection of cancer, including MRD detection, may be primarily limited by the presence of any ctDNA in the peripheral blood, not the features nor breadth of the detection technology.

Liquid biopsy is a powerful, multifaceted, and minimally invasive method for MRD detection and hence for monitoring therapeutic response, disease recurrence, or patient resistance to therapy. Although a relatively new field, liquid biopsy assays are showing promising results when compared to current standards of care. Two-staged somatic mutation approaches, where tumor sequencing informs MRD assay targets, allow clinicians to track actionable mutations and monitor for signs of treatment resistance. However, the time required to establish theses assays does not fit well with the patient journey, as clinicians need to know sooner than 5-6 weeks whether a patient is not responding to treatment. Therefore, epigenetic-based MRD assays are preferable at treatment onset, as epigenetic changes are typically more widespread and likely to be shared among a greater number of cancers than somatic mutations. With the recent advent of enzymatic methylation conversion, targeted sequencing approaches combining DNA methylation, fragmentomics and machine learning will likely come to predominate. However, DNA methylation-based PCR tests currently offer the most cost-effective solution within this niche. In fact, a compelling paradigm may be the combination of inexpensive serial testing with a DNA methylation-based assay, then following a positive MRD/recurrence result with a broad NGS-based somatic mutation panel to identify actionable mutations.

AJ, JR, CM, KF, and WL wrote and revised this review. All authors contributed to the article and approved the submitted version.