Introduction

Colorectal cancer (CRC) is the third most prevalent cancer worldwide and the second leading cause of cancer-related deaths (Keum and Giovannucci 2019). Approximately 150,000 new cases of CRC and 52,580 CRC-related deaths were reported in the United States in 2022 (Siegel et al. 2022). Up to 17.7–22.0% of patients newly diagnosed with CRC acquire metastatic CRC (mCRC), and 19.6–20.6% of patients newly diagnosed with stage I–III CRC eventually develop mCRC (Elferink et al. 2015; Väyrynen et al. 2020; Nors et al. 2023; Siegel et al. 2023; Tsai et al. 2023).

In addition to conventional chemotherapy, several drugs that target the molecular drivers of CRC—such as epidermal growth factor receptor (EGFR) and vascular endothelial growth factor pathways—have been used extensively to improve the survival of patients with mCRC (Cremolini et al. 2015; Venook et al. 2017; Bennouna et al. 2019; Heinrich et al. 2023) Current mCRC treatment guidelines recommend screening for mutations in genes such as KRAS, NRAS, and BRAF before selecting the chemotherapy regimen. RAS status is crucial for mCRC treatment decisions (Van Cutsem et al. 2016; Benson et al. 2021; Yoshino et al. 2023). If the RAS gene is mutated, certain targeted therapies, such as anti-EGFR agents, may be less effective. Therefore, determining RAS status is useful for tailoring treatment plans and increasing the likelihood of favorable outcomes in patients with mCRC.

Recent research suggests that temporal changes may occur in RAS status (Osumi et al. 2021; Nicolazzo et al. 2023). In CRC, the condition “NeoRAS wild-type (NeoRAS-WT)” refers to a situation in which a patient initially had a RAS mutation, but after treatment, the status reverted to wild-type (non-mutated). This change can impact treatment decisions because patients with NeoRAS-WT tumors may respond differently to certain therapies compared to those with persistent RAS mutations (Cremolini et al. 2019; Osumi et al. 2021; Nakajima et al. 2021). Meanwhile, “NeoRAS-WT” status remains a controversial concept. In recent years, RAS status is being tested using blood samples (Patelli et al. 2021, 2023). The circulating tumor DNA (ctDNA) RAS assay, developed using liquid biopsy technology, can reveal inter- and intra-tumor heterogeneity, thereby improving diagnosis by facilitating early detection and screening, monitoring, and follow-up; moreover, personalized therapeutic strategies can improve treatment outcomes (Reinert et al. 2019; Patelli et al. 2023). In a previous study, we reported that the frequency of mutation loss in a single ctDNA RAS assay was 43.5% in patients with RAS mutations. However, due to dynamic changes in metastatic organs and chemotherapy regimens over the treatment course, it is unclear how RAS responds to multiple ctDNA tests over time and what the optimal timing is for ctDNA assay. Determining these parameters is crucial for effective treatment of CRC with RAS mutations. Therefore, in this study, we aimed to determine the temporal dynamics of RAS status during treatment and its clinical significance using ctDNA RAS assay in patients with mCRC harboring mutant RAS.

Methods

Study population

This single-center retrospective cohort study was conducted at the Kanagawa Cancer Center, Yokohama, Japan. The study was conducted in accordance with the principles of the Declaration of Helsinki and was approved by the Institutional Review Board of Kanagawa Cancer Center (Approval Number: 2023-85). The requirement for written informed consent was waived due to the retrospective nature of the study, and participants were given the opportunity to opt-out. The study was registered in the Japanese Clinical Trials Registry, with registration number UMIN000053017. Consecutive patients who had RAS mutant-type (RAS-MT) mCRC between January 1, 2013, and April 30, 2023, were included in this study and their RAS status was monitored. Patient data, including demographic, clinical, and pathological information, were extracted from their medical records. Patients who met the following inclusion criteria were enrolled: (1) histologically diagnosed with adenocarcinoma, and (2) RAS-MT (KRAS/NRAS exon 2, 3, or 4) confirmed in primary tumor tissue before beginning chemotherapy. Patients with colon cancer were excluded from this study.

Testing and monitoring of RAS genotypes

To determine RAS status before starting treatment, tumor tissue taken surgically or by biopsy from the primary lesion was tested using the MEBGEN RASKET™-B kit (Medical & Biological Laboratories Co., Tokyo, Japan). Genomic DNA was extracted from formalin-fixed paraffin-embedded tissue sections, and the DNA concentrations ranged from 10 to 20 ng/µL. Polymerase chain reaction was performed with reverse sequence-specific oligonucleotides for all the samples. We examined the presence of the following mutations for exon 2 (G12S, G12C, G12R, G12D, G12V, G12A, G13S, G13C, G13R, G13D, G13V, and G13A), exon 3 (A59T, A59G, Q61K, Q61E, Q61L, Q61P, Q61R, and Q61H), and exon 4 (K117N, A146T, A146P, and A146V) of KRAS and NRAS.

After the initiation of standard-of-care regimens, the OncoBEAM™ RAS CRC assay (Sysmex Inc., Kobe, Japan) was used to monitor RAS mutations over time. The OncoBEAM™ RAS assay is a molecular diagnostic test designed to detect mutations in RAS genes, specifically KRAS and NRAS. This assay uses liquid biopsy technology, whereby ctDNA shed by cancer cells into the plasma is analyzed. During the course of treatment, ctDNA was assayed with OncoBEAM™ at the discretion of the physician. Before the ctDNA RAS assay, the patients were treated with standard-of-care regimens. The use of anti-EGFR agents for patients with aRAS-ML by ctDNA RAS assay was decided on a case-by-case basis. Clinicopathological background, including age, sex, tumor location, and treatment regimen, was compared between patients who acquired RAS mutation loss (aRAS-ML) during treatment and those who retained a RAS mutation. The rate of conversion to aRAS-ML over time was analyzed, and the prognostic impact of mutation loss and optimal testing time was assessed.

Definition of aRAS-ML

In this study, aRAS-ML was defined as the absence of RAS mutations after treatment initiation in patients who were diagnosed with RAS-MT (KRAS exons 2, 3, 4 or NRAS) upon pre-treatment tissue examination; mutation loss was determined based on the findings of the transient OncoBEAM™ RAS CRC assay. Patients with a mutation detected in the first ctDNA RAS assay were tested again during the treatment course, and were classified as aRAS-ML if no RAS mutation was detected in subsequent tests.

Statistical analysis

Continuous variables were analyzed using the Mann–Whitney U test and described using the median and interquartile range. Differences between categorical variables were analyzed using Fisher’s exact probability test. Survival outcomes were assessed using the Kaplan–Meier method, and group data were compared using the log-rank test. A multivariate Cox proportional hazard model was used to analyze the impact of acquired mutation loss on survival. Factors significant for overall survival in the univariate analysis were entered into the model using the forced entry method. Statistical analyses were carried out using EZR (Saitama Medical Center, Jichi Medical University, Saitama, Japan), a graphical user interface for R (The R Foundation for Statistical Computing, Vienna, Austria), with a significance level of p < 0.05.

Results

Patient characteristics

A total of 82 patients with RAS-MT mCRC (median age: 66 years; range: 26–86 years) were eligible for this study. Of these, 33 patients (40.2%) acquired mutation loss on RAS during the treatment course. The clinicopathological characteristics of the patients are listed in Table 1. No differences were observed between the aRAS-ML and RAS-MT groups in terms of patient background, including age, sex, and tumor location.

Table 1 Demographic and clinicopathological characteristics of patients with metastatic colorectal cancer
Full size table

At the beginning of the standard-of-care regimen, the most common target organs were the lungs, liver, peritoneum, and lymph nodes in the aRAS-ML group, and the liver, lungs, and peritoneum in the RAS-MT group. The most common RAS mutation found in pre-treatment tissue was KRAS codon 12, in both the aRAS-ML and RAS-MT groups (supplement).

Results of plasma ctDNA RAS monitoring

The median time between tissue RAS testing and the first ctDNA RAS assay was 24 months; 21 months (range: 16–43 months) in the aRAS-ML group and 25 months (range: 15–34 months) in the RAS-MT group, with no significant differences (p = 0.73). aRAS-ML was detected in 32 (39.0%) patients during the first ctDNA RAS assay. A second ctDNA RAS assay was performed in 22 patients who were shown to have a RAS mutation in the first assay, with one patient (4.5%) having aRAS-ML. In this case, the interval between the first and second ctDNA assays was 11 months. A third ctDNA assay was performed in six patients, and RAS mutation was detected in all cases. In addition, ctDNA assays were performed in all patients detected to harbor aRAS-ML at the time of the first- or second-line standard regimens, and no patients were detected to harbor aRAS-ML after the third-line regimen or later (Fig. 1). A trend towards increasing single-organ metastases in the target organ for treatment was observed in the aRAS-ML group at the first ctDNA assay, when compared with that in the RAS-MT group (84.8% vs. 59.2%, p = 0.02) (Table 2). There was no trend in RAS mutation loss according to the type of chemotherapy regimen administered. Among the patients with aRAS-ML, seven patients (21.2%) received anti-EGFR therapy after mutation loss and the median progression-free survival was 8 months (range: 4–12 months). Of these seven patients, six (85.7%) had mutations detected in the ctDNA RAS assay after treatment. There was no significant difference in overall survival between those who received anti-EGFR therapy and those who did not in patients with aRAS-ML (p = 0.8).

Fig. 1
figure 1

Temporal diversity of RAS mutations

Full size image
Table 2 Summary of treatment course in patients with RAS mutations and patients without mutations
Full size table

Relationship between aRAS-ML and prognosis

The median follow-up duration was 35 months. The 3-year overall survival rate was significantly better in the aRAS-ML group (85.9%) compared with that in the RAS-MT group (57.2%; p = 0.0037) (Fig. 2). Table 3 shows the results of univariate analysis of overall survival for each clinicopathologic factor. Multivariate analysis showed that persistent ctDNA RAS mutation during the treatment was an independent prognostic factor for overall survival (hazard ratio: 2.7, 95% confidence interval: 1.1–6.3, p = 0.02) (Table 4).

Fig. 2
figure 2

Overall survival with acquired RAS-mutation loss (aRAS-ML) versus ctDNA RAS-MT

Full size image
Table 3 Analysis of overall survival by clinicopathological factors in metastatic colorectal cancer with RAS mutation
Full size table
Table 4 Multivariate analysis of predictors of overall survival
Full size table

Discussion

The present study revealed the temporal dynamics of RAS in mCRC. To the best of our knowledge, this is the first study to analyze and report the long-term changes in RAS and aRAS-ML detection rate. We showed that the detection rate of aRAS-ML decreased during the course of treatment. Moreover, our results indicate that patients with mutation loss may have better prognosis than patients with persistent RAS mutations.

The outcome of ctDNA RAS testing is a valuable indicator for challenging anti-EGFR antibody treatment in patients with RAS-WT tissues (Cremolini et al. 2019; Patelli et al. 2023). The CHRONOS trial, a phase 2 study that reported the results of anti-EGFR antibody rechallenge following ctDNA evaluation, reported partial responses in 30% and disease control in 63% cases with RAS-WT status (Sartore-Bianchi et al. 2022). The VELO trial reported by Napolitano et al. analyzed plasma ctDNA in patients with CRC having RAS/BRAF-WT before treatment; in their study, the median progression-free survival was 4.5 months in the panitumumab plus trifluridine-tipiracil group (95% confidence interval: 2.2–6.8 months) vs. 2.6 months in the trifluridine–tipiracil only group (95% confidence interval: 1.0–4.3 months) (Napolitano et al. 2023). Thus, although some studies have reported EGFR antibody rechallenge using ctDNA assays in recent years, the clinical significance of aRAS-ML in mCRC is not well established. Mutation loss after induction chemotherapy in patients with RAS-MT was between 21.5 and 90% (Klein-Scory et al. 2020; Sunakawa et al. 2022; Wang et al. 2022; Osumi et al. 2023).

Sunakawa et al. reported the rate of early disappearance of RAS mutations in patients receiving FOLFOXIRI plus bevacizumab as first-line treatment for CRC, as monitored by performing liquid biopsy (Sunakawa et al. 2022). After 8 weeks of treatment, they observed a RAS-MT disappearance rate of 78% based on ctDNA testing results. In the present study, there was a lower frequency of RAS mutation loss in more late-line treatment and fewer cases of mutation loss in patients with multi-organ metastases. It can be assumed that more cases with multi-organ metastases occur in late-line treatment due to disease progression, and that the frequency of mutation loss is lower due to tumour heterogeneity caused by multi-organ metastases, rather than the type of regimen. In view of this, it is possible that a more aggressive regimen at an earlier stage of treatment, with a reduction in the number of target lesions, may reduce tumour heterogeneity and result in aRAS-ML, thus improving the long-term prognosis. The higher rate of mutation loss observed in the report by Sunakawa et al. with the potent regimen FOLFOXIRI plus bevacizumab may provide support for this hypothesis.

A clinical problem with standard systemic regimens for mCRC is the oncological heterogeneity of target metastases (Kim et al. 2015; Lu et al. 2016; Adua et al. 2017; Morris and Strickler 2021; Nicolazzo et al. 2021). In addition to differences in the metastatic organs such as the liver, lung, lymph nodes, and peritoneum, the heterogeneity of multiple metastases within an organ can lead to inconsistent chemotherapy efficacy and indications. Unlike conventional organ biopsies, liquid biopsy testing for RAS mutations may represent the overall disease status. Therefore, ctDNA testing by liquid biopsy in mCRC may reflect the disease status of mCRC in real-time, considering the heterogeneity of multiple metastases. Our findings showed that the presence or absence of aRAS-ML status (as determined by temporal ctDNA testing) may be a prognostic marker. Hence, monitoring the status of RAS mutations during the course of treatment may help to identify groups of patients with a better prognosis.

In this study, loss of RAS mutation was an independent prognostic factor. Additionally, the aRAS-ML group tended to have more cases with a single target organ at the time of ctDNA collection. This suggests that the disappearance of the RAS mutation in the blood may be a marker of the therapeutic efficacy in patients with tissue RAS-MT. In mCRC, tumor heterogeneity within each metastatic lesion or within a metastatic tumor may lead to differential response to standard chemotherapy (Adua et al. 2017; Wang et al. 2022). This is supported by the higher incidence of single-organ metastases in the aRAS-ML group in our study. Oncological heterogeneity may be higher in cases of multiple-organ metastases than in cases of single-organ metastases. Therefore, if at the time of ctDNA RAS testing, a patient shows single-organ metastasis, RAS-MT clearance may be increased, resulting in aRAS-ML. In other words, patients with single-organ metastases are more likely to have aRAS-ML and should be actively monitored, and the possibility of EGFR antibody challenge should be considered. Meanwhile, there is currently no evidence to suggest that anti-EGFR agents has any benefit for patients with aRAS-ML. The C-PROWESS study, a phase II trial evaluating the safety and efficacy of panitumumab in combination with irinotecan for the management of NeoRAS-WT colorectal cancer, is currently underway in Japan and the results of this trial are awaited(Osumi et al. 2022). Challenges related to assessing difficult-to-image metastases, such as micro-metastases, and the lack of evidence for the therapeutic efficacy of EGFR antibodies in aRAS-ML remain to be overcome in the future studies.

The NeoRAS-WT concept refers to the loss of detectable RAS mutation in the presence of continued evidence of mCRC. The present study analyzed the temporal dynamics of the ctDNA RAS and suggests that it may serve as a marker for treatment response. When patients respond to treatment, all ctDNA level decrease and, in the best responses, become undetectable. For ctDNA to be in a biological NeoRAS-WT state, it must remain detectable at a level where the RAS mutation can be detected. Therefore, this marker may be associated with improved survival as it identifies patients who respond well to treatment. Consistent with this, all patients who had a third assessment in the later stages of treatment had detectable RAS mutation, and no patients had mutation loss at the third assessment, suggesting a rebound in detection with an increase in tumor volume as treatment becomes less effective. In other words, the ctDNA RAS assay may not show a loss of chemotherapy-induced RAS mutation, but rather a decrease in the detection of ctDNA. It is important to note that the concept of NeoRAS-WT may differ from the clinical significance of loss of ctDNA RAS mutation loss during treatment. Meanwhile, temporal RAS status is also important in the decision to use EGFR therapies and may therefore be a significant marker.

A notable issue regarding the characterization of NeoRAS-WT revolves around the inherent challenge in detecting loss of RAS mutation or the absence of detectable ctDNA. As a result, ascertaining whether ctDNA could not be detected owing to a diminutive tumor volume and inability to determine metastatic involvement in specific organs remains a significant hurdle. Osumi et al. pointed out that conversion to NeoRAS-WT may be more common in patients without liver metastases (Osumi et al. 2023). NeoRAS-WT is defined as the absence of RAS mutation coupled with trunk mutations of genes such as TP53, APC, or methylated genes (Moati et al. 2020; Nicolazzo et al. 2020, 2021). Nicolazzo et al. reported that the most common actionable mutations detected in patients with NeoRAS-WT were PIK3CA (35.7%), RET (11.9%), IDH1 (9.5%), KIT (7%), EGFR (7%), MET (4.7%), ERBB2 (4.7%), and FGFR3 (4.7%) (Nicolazzo et al. 2023). The prevailing notion is that concurrent genetic alterations serve as corroborative evidence for the presence of tumor DNA in the plasma. Hence, further investigations in aRAS-ML cases utilizing next-generation sequencing, methylated genes, or other serum tumor markers are imperative to clarify this; it is one of the limitations of our present study. Furthermore, a method remains to be established for reliable identification of tumor-derived ctDNA in blood.

The present study has some other limitations. This is a retrospective study with a small sample size; therefore, the treatment course and standard-of-care regimens were not standardized. Additionally, the study did not examine the impact of chemotherapy response (i.e. partial response, stable disease or progressive disease) on ctDNA RAS status. Further investigation is required to determine whether differences in chemotherapy response are associated with differences in RAS mutation loss rates. Larger clinical trials are required to determine the optimal time for testing ctDNA RAS, the choice of regimen, the choice of treatment line (i.e., first-line, second-line, third-line, or later), and how to detect the presence of tumor-derived ctDNA. The importance of anti-EGFR agents in RAS-WT mCRC is further strengthened by the results of the PARADIGM trial(Watanabe et al. 2023). In daily clinical practice, mCRCs that are RAS-MT at the start of treatment and convert to aRAS-ML during the course of treatment may challenge anti-EGFR agents that were previously considered to be of no benefit in RAS-MT colorectal cancer, potentially significantly influencing treatment strategies. As each case of mCRC varies significantly in terms of tumor volume, metastatic organs, and other parameters, several challenges persist for “standardizing”. In this regard, tracking ctDNA RAS mutations by liquid biopsy may be useful to determine overall tumor status.

Conclusions

The rate of ctDNA mutation loss in patients with RAS-MT mCRC decreases over time. Therefore, ctDNA RAS assay should be performed early in treatment to challenge the use of EGFR regimens.