What is MRD

MRD, namely, molecular residual lesions (also known as micro residual lesions or measurable residual lesions), has its concept derived from hematological tumors, and refers to tumor-derived molecular abnormalities that cannot be detected by imaging or traditional laboratory methods after treatment, but are found through liquid biopsy (ctDNA with abundance ≥0.02% can be stably detected in peripheral blood, including driver genes or other Class I/II gene variants), representing the persistent existence of tumors and the possibility of clinical progress ^[1]. In a broad sense, the detection content of MRD can include specific markers derived from tumor cells in body fluids (usually blood), including but not limited to ctDNA, ctRNA, CTC, exosome and other substances, and the detection scope includes genomics and proteomics ^[2][3]. NGS-based ctDNA mutation detection is currently the most commonly used method for MRD detection of solid tumors. In colorectal cancer, non-small cell lung cancer and breast cancer, there is sufficient evidence that MRD has a prognostic stratification effect, and its clinical value is relatively clear; In other solid tumors such as pancreatic cancer, liver cancer, esophageal cancer, and gastric cancer, there are also some evidences suggesting that MRD has a prognostic stratification effect ^[4].

So when does the MRD really start? How? This article will be divided into two parts by MRD analysis strategy, Landmark detection time point, clinical significance, detection difficulties, etc.

I. Analysis strategy of MRD

The MRD detection of solid tumors is divided into two analysis strategies, i.e., prior analysis of tumor tissue (tumor-informed) and unknown analysis of tumor tissue (Tumore-agnostic/naive) according to whether or not the mutation information of tumor tissue is referred.

Prior analysis of tumor tissues (Tumor-informed assays) requires the detection of tumor tissues with WES (or large panel)+ personalized panel. The Tumor-informed strategy first requires high-throughput sequencing (mainly WES) of tumor tissues of patients to identify tumor-specific mutations of each patient, and select a certain number of high-abundance trunk mutations to customize the panel (usually only including 16–50 tumor-specific mutations), and finally detect these mutations in plasma ctDNA of patients, which is equivalent to the detection of WES (or large panel)+ personalized panel. The analysis strategy greatly reduces the risk of false positives caused by technical and biological backgrounds (such as clonal hematopoiesis) due to the small number of detection mutation targets, and therefore extremely high-depth sequencing can be performed, and the detection sensitivity is improved. At present, MRD products in the market are mostly based on this strategy.

Tumor-informed also has a fixed Panel assay strategy: after the baseline tissue assay, a fixed panel is used for subsequent monitoring. Panel is usually designed based on common driver genes and targeted drug action sites, which limits its applicability and is rarely seen on the market.

Tumor tissue unknown analysis (tumor-agnostic/naive) does not need to obtain tumor tissue and directly carries out immobilized panel detection. The strategy mainly uses a universal large panel with three to four hundred genes or more, which is equivalent to the detection of immobilized panel. The use of a fixed mutant panel eliminates the need to obtain a patient's tumor tissue in advance for sequencing, which can greatly simplify the process, reduce costs, and shorten the patient's MRD status assessment cycle. In the decision-making process of patient adjuvant therapy, the delay of adjuvant therapy may reduce the curative effect of treatment. Therefore, it is also important to quickly assess the MRD status of patients so as to make clinical decisions as early as possible ^[5].

Which of these two strategies is better?

Comparison of three strategies ^[6][7]

In the MEDAL study, the world's first prospective clinical study comparing three MRD detection strategies for head-to-head detection of early stage lung cancer, 98.2% of all selected variants were non-standard, and 99.9% of them were unique to one patient. If only standard mutations were monitored, nearly half of landmark's MRD-positive samples would not be recognized.

At the same time, it was found that Prophet's ability to predict DFS and OS at baseline was higher than that of Tumor-agnostic and Tumor-informed immobilized Panel tests. The consensus on the detection of molecular residual lesions (MRD) in solid tumors also explicitly mentions that the Tumore-Informed analysis based on tumor tissue mutation has better sensitivity and specificity compared with the Tumore-Agnostic/Naive analysis, and the Tumore-Informed analysis strategy is preferred.

II. Landmark Detection Time Point

01. New adjuvant therapy

CTONG1804 was a prospective, multicenter, Phase II study evaluating the clinical efficacy of neoadjuvant nivolumab alone (N) and a combination of nivolumab- chemotherapy (N/C) based on PD-L1 expression.

In plasma samples collected before treatment in 38 patients, ctDNA was detected in 89.5%(34/38) of pretreatment samples, including 21.1%(8/38) of stage II disorders and 68.4%(26/38) of stage III disorders. CtDNA detection declined during neoadjuvant therapy, from 89.5% before (T0) to 34.2% after neoadjuvant therapy (T2), and then continued to decline to 27.6% after surgery (T3). Specifically, neoadjuvant immune monotherapy N and N/C reduced the positive rate of ctDNA by 42.9%(T3) and 18.2%(T3), respectively.

In all patients with local and distant relapses, detectable ctDNA was observed at at least at one point. The 18-month EFS rates in patients with CT DNA/MRD–(T2 and T3) and ctDNA/MRD+(T2 or T3) were 93.8% and 47.3%, respectively. These results suggest that ctDNA was negative at both time points after neoadjuvant therapy and surgery, predicting pCR and survival benefits, which need to be further confirmed in prospective clinical trials.

CT DNA/MRD–(T2 and T3) vs. ctDNA/MRD+(T2 or T3) EFS^[8]

The evaluation value of MRD on the efficacy of neoadjuvant therapy has also been reported in the literature for bladder cancer ^[9], colorectal cancer ^[10], and breast cancer ^[11]. Existing evidence suggests that ctDNA clearance after neoadjuvant therapy is a potentially effective marker for predicting distant recurrence. If the patients' ctDNA is positive after receiving neoadjuvant therapy, the prognosis may be poor, and it is not suitable for organ function preservation, and timely adjustment of treatment regimen should be considered. CtDNA-based MRD status assessment can potentially assist in clinical screening of advantageous populations suitable for non-surgical treatment, and provide new ideas for meticulous and holistic management of patients in progressive stage. If patients with gastric cancer receive adjuvant treatment, MRD test is recommended before treatment, which is helpful to judge the prognosis and formulate further treatment follow-up strategy. Dynamic monitoring of MRD may be considered in patients with locally advanced gastric cancer receiving neoadjuvant therapy during the treatment ^[12].

However, preoperative ctDNA-MRD results are not always correlated with postoperative results. A study has found that among the 47 patients with recurrent non-small cell lung cancer, 14 patients were negative for pre-operative MRD but positive for post-operative monitoring ^[13]. This indicates that preoperative ctDNA-MRD results do not affect their conclusions in postoperative monitoring, and further clarify the importance of postoperative MRD monitoring.

Our study found that the preoperative ctDNA-MRD status of our patient was correlated with many factors, including tumor size, tumor stage, pathological type, and genotype. Specifically, the larger the tumor volume and the later its staging are, the higher the preoperative ctDNA-MRD positive rate will be ^[14][15]. No consensus has been reached regarding the prognostic role of preoperative ctDNA-MRD findings. Preoperative ctDNA-MRD is heterogeneous in different cancer species or different pathological types, and more studies are needed to clarify its prognostic predictive value in different cancer species.

02. After radical treatment, before adjuvant treatment

Currently, most initial MRD testing occurs at this time.

NCT02965391 studied the acquisition of 10 mL of plasma samples immediately before surgery (time A) and at prespecified time points after tumor resection [time B(5 minutes), time C(30 minutes) and time D(2 hours)] and after surgery [time P1(1 day), time P2(3 days) and time P3(1 month)]. Rapid attenuation of ctDNA after radical tumor resection. There were no significant differences in mean RFS and mean OS in both MRD-positive and MRD-negative patients at 1 postoperative day; There were significant differences in RFS and OS between MRD-positive and MRD-negative patients 3 days and 1 month after surgery. These results suggest that MRD findings at 3 days and 1 month after surgery are more relevant to our patient's tumor recurrence and survival.

Mean RFS and mean OS in patients with positive and negative MRD 1 day after surgery ^[16]

Mean RFS and mean OS in patients with positive and negative MRD 3 days after surgery ^[16]

At present, there is no unified standard for sampling timing and detection timing of baseline analysis of MRD. Experts consensus on non-small cell lung cancer molecular residual lesions recommends that the time window for the first MRD detection should be within 1 week to 1 month after radical treatment of NSCLC. Detection and clinical application of molecular residual lesions in gastric cancer The China Expert Consensus (Version 2023) recommends that MRD testing may be considered for advanced gastric cancer after radical surgery or during the first postoperative follow-up. US experts recommend that the initial blood samples be collected two weeks after the operation for MRD based on ctDNA. If the results are negative, the monitoring time can be shortened (such as one month later) to reconfirm the negative results, but this should be done before the start of adjuvant therapy ^[17].

In summary, combined with the consensus, the author recommends that the first detection point of MRD in patients with solid tumors should be within 1 week to one month after the radical surgery, and before the start of adjuvant therapy. After radical chemoradiotherapy, before consolidation treatment or at half the progress of radiotherapy ^[18]; It is appropriate for advanced patients who have complete remission after systemic treatment. Preoperative detection can be used to evaluate the effect of neoadjuvant therapy. The prognosis should be mainly evaluated by the postoperative MRD status.

03.Longitudinal Test Duration

Longitudinal refers to continuous monitoring performed at a series of time points after radical treatment, such as three months and six months after surgery, to comprehensively assess the disease progression potential of the patient and the dynamic changes in MRD status. Patients who were all negative for MRD during this period were defined as negative for Longitudinal; At any monitoring time point, a patient who was positive for MRD was defined as positive for Longitudinal. Summary analysis from multiple studies has shown that postoperative Landmark single-point MRD positivity has always been closely associated with a poor prognosis in patients with a specificity greater than 90%, but a relatively low sensitivity with a median sensitivity of 56%. These results indicated that Longitudinal had better sensitivity and specificity for patient stratification, and the sensitivity reported in the literature reached 100%, with the median sensitivity of 89%^[19]. These findings suggest that long-term dynamic monitoring is more accurate in assessing the risk of recurrence.

In the CIRCULATE-Japan study, the researchers divided the colorectal cancer patients into four groups of continuous positive, positive to negative, negative to positive, and continuous negative according to their MRD results at 4 and 12 weeks after operation. The results showed that the DFS in the "positive to negative" group was significantly higher than that in the "positive to positive" group (81.4%vs33.8%). The importance of long-term monitoring is re-emphasized.

DFS comparison of four groups ^[20]

So how long will Longitudinal last? In the research of adaptive reduction therapy for patients with advanced lung cancer by Professor Wu Yilong's team, the median follow-up period of the patients was 19.2 months, and the median progression-free survival (PFS) of the whole population was 18.4 months ^[21]. Studies with NSCLC have shown that 12 to 18 months after surgery is the peak time range for MRD testing. The risk of recurrence gradually decreases if the patient remains negative for MRD for more than 18 months.

In summary, we recommend that patients undergo MRD every 3–6 months for at least 2 years. The specific detection time can be combined with the patient's return follow-up time. Some people also think that it is recommended to review stage 1B and later every 3 months until 3 years after operation. It is recommended that patients prior to Stage 1B be re-examined 3-6 months after surgery with simultaneous follow-up for MRD testing. However, the source is not available for reference only.

III. SpaceGen Detection Strategy

The solid tumor molecular residual lesion (MRD) detection product independently developed by SpaceGen adopts the customized Panel detection strategy. First, the full exon sequencing is performed on the primary tumor tissue to determine the patient-specific gene mutation result, and then the customized Panel performs subsequent ctDNA detection on a certain number of mutation sites in the tissue. And has obvious advantage compared with similar product.

• We have independently developed the tumor version of WES (whole exon sequencing) detection, and added probe coverage for tumor hot spot mutation regions, intron regions related to common gene fusion, non-coding region ClinVar pathogenicity locus, SNP skeleton and MSI region on the basis of WES; Baseline sample test results covering targeted therapy, immunotherapy, and genetic risk assessment;

• The customized panel adopted the self-developed sequencing library construction technology of one-sided primer amplification combined with UMI.

• Automatic monitor site screening is carry out, and that monitoring sensitivity is improved from site screening by weight scoring indexes such as the frequency of mutation sites, the function of genes, mutation types, mutation grades and the like;

• Ultra-high-depth sequencing (100000X) combined with UMI analysis enabled 0.05% mutation detection sensitivity, meeting the detection performance requirements of MRD.

• Baseline analysis+multiple dynamic monitoring, and full-cycle MRD dynamic monitoring of tumor diagnosis and treatment;

• Flexible full cycle management, multi-directional selection;

• It is suitable for prognosis prediction, recurrence risk assessment and curative effect monitoring of various solid tumors.

References

[1] Expert consensus on molecular residual lesions of non-small cell lung cancer

[2] Application and progress of detection of minimal residual lesions in solid tumors

[3] Research progress and application of minimal residual lesions in non-small cell lung cancer

[4] Consensus on the detection of molecular residual lesions (MRD) of solid tumors

[5] MRD Testing Industry Report (2023)

[6] Cancer Discov 2021 Dec 1; 11(12):2968-2986.

[7] Cancer Cell. 2023 Oct 9; 41(10):1749-1762.e6.

[8] Signal Transduct Target Ther 2023 Dec 6; 8(1):442.

[9] AEur Urol. 2022 Aug; 82(2):212-222.

[10] PLoS Med 2021 Aug 31; 18(8):e1003741.

[11] Cancer Cell. 2023 Jun 12; 41(6):1091-1102.e4.

[12] China expert consensus on the detection and clinical application of molecular residual lesions in gastric cancer (2023 edition)

[15] Nature. 2023; 616(7957):553-562.

[16] Clin Cancer Res. 2019; 25(23):7058-7067.

[17] Nature 2023 Jul; 619(7969):259-268.

[18] Cancer Cell 2023 Oct 9; 41(10):1763-1773.e4.

[20] Nat Med 2023 Jan; 29(1):127-134.

[21] JAMA Oncol 2024 Jul 1; 10(7):932-940.

Statement:

This article is only for sharing, if it involves copyright issues, please contact us as soon as possible, we will correct the first time, thank you!