Circulating tumor DNA as a predictive biomarker for treatment response and survival in metastatic colorectal cancer

Abstract

Purpose

To explore the potential of circulating tumor DNA (ctDNA) as a prognostic biomarker to predict treatment response and survival outcomes in patients with metastatic colorectal cancer (mCRC).

Methods

A retrospective analysis was conducted on 134 patients with mCRC who were treated between January 2020 and December 2021. The patients were classified into ctDNA-negative and ctDNA-positive groups based on plasma ctDNA detection. Demographic, clinical, and laboratory parameters, treatment response, survival outcomes, and adverse events were recorded and analyzed.

Results

No significant differences were observed in baseline characteristics between the two groups. Compared to the ctDNA-positive patients, ctDNA-negative patients exhibited superior outcomes, including a higher objective response rate (65.22% vs. 46.15%), disease control rate (81.16% vs. 63.08%), progression-free survival (8.24 ± 1.02 vs. 7.86 ± 0.91 months), overall survival (24.58 ± 3.58 vs. 23.27 ± 3.46 months), and 1-year survival rate (73.91% vs. 55.38%). The ctDNA-positive group had a significantly higher incidence of adverse events. Correlation analyses revealed significant associations between ctDNA status, tumor markers, treatment response, and survival outcomes.

Conclusions

ctDNA is a promising noninvasive biomarker for predicting treatment response, survival, and adverse events in mCRC, potentially guiding personalized therapeutic strategies.

Supplementary information

The online version contains supplementary material available at 10.1007/s00384-024-04785-7.

Introduction

Metastatic colorectal cancer (mCRC) remains a significant global health challenge owing to its increasing incidence, disease prognosis, and treatment outcomes [1–3]. Despite advances in therapeutic strategies, the heterogeneous nature of mCRC outcomes highlights the need for robust biomarkers to guide personalized treatment decisions. Accurate prognostic and predictive biomarkers are essential for optimizing patient care, improving clinical outcomes, and enhancing resource allocation in the management of mCRC [4–6]. Traditional biomarkers in mCRC, including tumor characteristics and histopathological features, have provided valuable insights into disease progression and therapeutic responses [7, 8]. However, the dynamic nature of mCRC and inherent tumor heterogeneity underscores the limitations of relying solely on conventional markers for prognostication and treatment decision-making. These challenges highlight the need for innovative, noninvasive biomarkers capable of providing real-time information on tumor burden, genetic evolution, and treatment response. Such biomarkers could facilitate the delivery of personalized and precise medicine in the context of mCRC [9, 10].

Circulating tumor DNA (ctDNA), derived from tumor cells and detectable in peripheral blood, has emerged as a promising noninvasive biomarker. ctDNA reflects the genetic and epigenetic landscape of a tumor, providing a comprehensive molecular snapshot of tumor heterogeneity and evolution [11–13]. The clinical utility of ctDNA lies in its ability to detect genetic mutations, copy number variations, and methylation patterns, offering valuable insights into tumor biology and therapeutic resistance mechanisms. In the context of mCRC, ctDNA has the potential to enhance our understanding of disease dynamics, treatment response, and the development of resistance mechanisms, which might guide treatment decisions, monitor disease progression, and predict therapeutic responses. Despite its promise, the integration of ctDNA into routine clinical practice requires comprehensive evaluation, particularly in the context of treatment response, survival outcomes, and occurrence of treatment-related adverse events. Previous studies have suggested the potential of ctDNA as a prognostic and predictive biomarker in various cancer types, including colorectal cancer [14, 15]. However, there remains a need for robust evidence to validate its clinical utility and to explore its integration into routine clinical practice.

This study aimed to evaluate the potential of ctDNA as a prognostic biomarker for mCRC, focusing on its association with treatment response, survival outcomes, and adverse events. By providing real-world evidence, we hope to contribute to the growing body of research that supports the role of ctDNA in precision oncology and its application in mCRC management.

Materials and methods

Study design

This retrospective case–control study involved 134 patients diagnosed with mCRC and treated at our institution between January 2020 and December 2021. The patients were categorized into groups based on their ctDNA status, distinguishing between ctDNA-negative (n = 69) and ctDNA-positive (n = 65). Due to the retrospective nature of the study, a one-to-one matching procedure was not feasible. On average, each case had 1 control. This study adhered to the STROBE reporting recommendation.

Inclusion and exclusion criteria

Inclusion criteria: patients who were histologically confirmed to have colorectal cancer with distant metastasis, aged between 18 and 75 years, had an ECOG performance status score of 2 or less, had an expected survival period of more than 3 months, had complete medical records, had normal mental and cognitive function, and were able to undergo four years of follow-up investigation.

Exclusion criteria: patients with concomitant autoimmune diseases or hematologic disorders; patients with a history of infection within one month prior to the start of treatment; patients with severe cardiovascular diseases, and those with severe hepatic dysfunction [3].

ctDNA detection method

Five milliliters of venous blood was drawn into an EDTA tube and then centrifuged at 1500 × g for 15 min to acquire plasma aliquots, which were subsequently preserved at − 80 °C. The plasma samples were sequenced using the Geneseeq PrimeTM 425 gene panel. Because of the low concentration of ctDNA released from the local tumor, we initially performed pre-amplification of 10 ng of ctDNA using six cycles of PCR with Q5 Hot Start High-Fidelity DNA Polymerase (NEB) and primer/probe mix (Taqman SNP). If the variant allele frequency (VAF) of ctDNA exceeded the predetermined limit of detection (LOD), the ctDNA test was considered positive. Conversely, the absence of detectable ctDNA mutations was defined as negative.

Data collection

General patient information included age, sex, Eastern Cooperative Oncology Group (ECOG) performance status, smoking history, alcohol consumption, BMI, primary tumor site, number of metastatic sites, hypertension, and diabetes status. ECOG performance status assesses the patient’s overall health and tolerance to treatment based on their physical abilities. A score of 0 indicates normal activity with no evidence of disease, whereas a score of 1 indicates the ability to perform light work activities. A score of 2 indicated self-care ability but an inability to work. A score of 3 implies partial self-care during the day, with limited ability to perform activities. A score of 4 represented a completely bedridden status and an inability to self-care, and a score of 5 denoted death. Baseline laboratory values, including CEA, CA 19–9 levels, hemoglobin, albumin (ALB), and LDH levels, were recorded. Treatment response parameters such as objective response rate (ORR), progression-free survival (PFS), disease control rate (DCR), overall survival (OS), and 1-year survival rate were also documented. Additionally, adverse effects following therapy, such as neutropenia, fatigue, diarrhea, and peripheral neuropathy, were observed.

Blood testing

Fasting venous blood samples were collected at 7:00 AM, and 5 mL was drawn from the median cubital vein. The samples were then centrifuged at a radius of 6.0 cm and speed of 3000 rpm for 10 min. The resulting upper layer of serum was carefully separated and stored in a freezer until further analysis. CEA and CA19-9 levels were determined using the Roche Cobas E411 electrochemiluminescence immunoassay analyzer and the corresponding reagents, following the manufacturer’s instructions. Serum ALB levels were assessed by enzyme-linked immunosorbent assay using a kit provided by R&D Systems (Shanghai Xitang Branch, batch number 200907261). Hemoglobin levels were measured using a Beckman Coulter DxH800 hematology analyzer. LDH levels were determined using a Toshiba TBA-120FR automated biochemical analyzer and reagents from Hunan Yonghe Yangguang Biotechnology Co., Ltd., based on the rate method.

Treatment response

The effectiveness of clinical treatments for both patient groups was evaluated based on the “Response Evaluation Criteria in Solid Tumors (RECIST) guidelines version 1.1.” The assessment was based on the product of the two largest perpendicular diameters of lesions: an increase of at least 25.00% indicated disease progression, an increase of less than 25.00% or a reduction of less than 50.00% in the product of the two diameters indicated stable disease, a reduction of at least 50.00% in the product of the two diameters with no new lesions and lasting for more than 30 days indicated partial response (PR), and the disappearance of the tumor for more than 30 days indicated complete response (CR). The ORR was calculated as the sum of the CR and PR rates. The DCR was determined by adding the number of patients with CR, PR, and stable disease cases, and then dividing this sum by the total number of cases. This value was then multiplied by 100%. PFS was characterized as the duration from the commencement of treatment to the point of disease progression or death. OS was specified as the duration between enrollment and the occurrence of death.

Statistical methods

Data analysis was conducted using SPSS statistical software (version 29.0; SPSS Inc., Chicago, IL, USA). Categorical variables are expressed in [n (%)] format. The chi-squared test was used with the basic formula for sample sizes ≥ 40 and theoretical frequency T ≥ 5. When the sample size was ≥ 40 but the theoretical frequency ranged from 1 to < 5, the corrected formula for the chi-squared test was applied. For sample sizes < 40 or theoretical frequency T < 1, Fisher’s exact probability method was employed for statistical analysis. The normal distribution of continuous variables was assessed using the Shapiro–Wilk test. Continuous variables with a normal distribution are presented as mean ± SD and analyzed using the corrected variance t-test, while non-normally distributed variables are expressed as median (25th percentile, 75th percentile) and analyzed using the Wilcoxon rank-sum test. Statistical significance was defined as a two-tailed P-value < 0.05. Independent sample t-tests were used for continuous variables, and chi-squared tests were used for categorical variables. Additionally, Pearson’s correlation analysis was employed for continuous variables, and Spearman’s correlation analysis was used for categorical variables.

Results

Demographic and clinical characteristics

A total of 134 patients with mCRC were included in the analysis, with 69 classified as ctDNA-negative and 65 as ctDNA-positive. There were no statistically significant differences in baseline demographic or clinical characteristics between the two groups, including age, sex distribution, ECOG performance status, smoking history, alcohol consumption, body mass index, primary tumor site, number of metastatic sites, and comorbidities, such as hypertension and diabetes (Table 1, P > 0.05). These findings suggest that the baseline characteristics of the two groups are comparable, which is important for interpreting the subsequent results related to treatment response and survival outcomes.

Table 1.

Demographic and clinical characteristics of patients withmCRC

Parameter	ctDNA negative group (n = 69)	ctDNA positive group (n = 65)	t/χ2	P
Age (years)	58.24 ± 6.78	59.86 ± 7.32	1.324	0.188
Gender (male/female)	38 (55.07%) / 31 (44.93%)	38 (58.46%) / 27 (41.54%)	0.049	0.825
ECOG performance status	1.22 ± 0.33	1.25 ± 0.22	0.753	0.453
Smoking history (pack-years)	21.34 ± 4.16	22.64 ± 5.23	1.594	0.113
Alcohol consumption (drinks/week)	4.78 ± 1.53	4.92 ± 1.45	0.539	0.591
Body mass index (kg/m2)	25.61 ± 3.17	26.25 ± 3.52	1.104	0.272
Primary tumor location (colon/rectum)	47 (68.12%) / 22 (31.88%)	42 (64.62%) / 23 (35.38%)	0.060	0.806
Number of metastatic sites	2.18 ± 0.96	2.35 ± 1.08	0.963	0.338
Hypertension (Y/N)	29 (42.03%) / 40 (57.97%)	31 (47.69%) / 34 (52.31%)	0.235	0.628
Diabetes (Y/N)	15 (21.74%) / 54 (78.26%)	16 (24.62%) / 49 (75.38%)	0.036	0.850
Previous chemotherapy (Y/N)	58 (84.06%) / 11 (15.94%)	52 (80.00%) / 13 (20.00%)	0.150	0.699
Previous radiotherapy (Y/N)	25 (36.23%) / 44 (63.77%)	26 (40.00%) / 39 (60.00%)	0.073	0.786
Surgery for primary tumor (Y/N)	63 (91.3%) / 6 (8.70%)	61 (93.85%) / 4 (6.15%)	0.053	0.818

Baseline laboratory values

Significant differences were observed in certain baseline laboratory parameters between ctDNA-negative and ctDNA-positive groups (Table 2). Patients with detectable ctDNA exhibited higher levels of CEA (64.32 ± 15.37 ng/mL vs. 72.58 ± 18.23 ng/mL, t = 2.827, P = 0.005) and CA 19–9 (123.45 ± 26.48 U/mL vs. 135.87 ± 30.56 U/mL, t = 2.508, P = 0.013). No significant differences were noted in hemoglobin, albumin, or LDH levels (P > 0.05). These findings provide important insights into the association between ctDNA status and specific baseline laboratory parameters, which may have implications for the treatment response and survival outcomes in patients with mCRC.

Table 2.

Baseline laboratory values of patients withmCRC

Parameter	ctDNA negative group (n = 69)	ctDNA positive group (n = 65)	t/χ2	P
CEA level (ng/mL)	64.32 ± 15.37	72.58 ± 18.23	2.827	0.005
CA 19–9 level (U/mL)	123.45 ± 26.48	135.87 ± 30.56	2.508	0.013
Hemoglobin (g/dL)	12.54 ± 1.39	12.27 ± 1.57	1.067	0.288
ALB level (g/dL)	3.84 ± 0.41	3.74 ± 0.53	1.254	0.212
LDH level (U/L)	325.73 ± 40.65	332.51 ± 45.86	0.903	0.368

ALB albumin

Treatment response

Patients in the ctDNA-negative group demonstrated superior treatment responses compared with their ctDNA-positive counterparts (Table 3): the objective response rate (ORR) and disease control group (65.22% vs. 46.15%, P = 0.041), as was the ctDNA-negative group (DCR; 81.16% vs. 63.08%, P = 0.032). Furthermore, progression-free survival (PFS) was longer in the ctDNA-negative group (8.24 ± 1.02 months vs. 7.86 ± 0.91 months, P = 0.022), as was overall survival (OS; 24.58 ± 3.58 months vs. 23.27 ± 3.46 months, P = 0.033) (Fig. 1). The 1-year survival rate was also higher in the ctDNA-negative group (73.91% vs. 55.38%, P = 0.039) (Fig. 2). These findings suggest a potential association among ctDNA status, treatment response, and survival outcomes in patients with mCRC.

Table 3.

Treatment response in patients withmCRC

Parameter	ctDNA negative group (n = 69)	ctDNA positive group (n = 65)	t/χ2	P
ORR (%)	45 (65.22%)	30 (46.15%)	4.193	0.041
DCR (%)	56 (81.16%)	41 (63.08%)	4.608	0.032
PFS (months)	8.24 ± 1.02	7.86 ± 0.91	2.310	0.022
OS (months)	24.58 ± 3.58	23.27 ± 3.46	2.153	0.033
1-year survival rate (%)	51 (73.91%)	36 (55.38%)	4.265	0.039

ORR objective response rate, DCR disease control rate, PFS progression-free survival, OS overall survival

Fig. 1.

Overall survival

Fig. 2.

Progression-free survival

Adverse events

The ctDNA-positive group experienced a significantly higher incidence of treatment-related adverse events than the ctDNA-negative group (Table 4). These included neutropenia (13.85% vs. 2.90%, P = 0.046), fatigue (16.92% vs. 4.35%, P = 0.036), diarrhea (15.38% vs. 2.90%, P = 0.026), and peripheral neuropathy (12.31% vs. 1.45%, P = 0.030). These findings suggest a potential association between ctDNA status and incidence of specific adverse events in patients with mCRC, which may have implications for treatment tolerability and patient management.

Table 4.

Adverse events

Parameter	ctDNA negative group (n = 69)	ctDNA positive group (n = 65)	t/χ2	P
Neutropenia (%)	2 (2.90%)	9 (13.85%)	3.970	0.046
Fatigue (%)	3 (4.35%)	11 (16.92%)	4.393	0.036
Diarrhea (%)	2 (2.90%)	10 (15.38%)	4.960	0.026
Peripheral Neuropathy (%)	1 (1.45%)	8 (12.31%)	4.685	0.030

Correlation analysis

Correlation analyses revealed significant associations between ctDNA status and the key clinical parameters (Table 5). Positive correlations were observed between ctDNA levels and tumor markers, such as CEA (r = 0.24, P = 0.005) and CA 19–9 (r = 0.214, P = 0.013). In contrast, negative correlations were identified between ctDNA status and treatment response parameters, including ORR (r = − 0.192, P = 0.026), DCR (r = − 0.202, P = 0.019), PFS (r = − 0.196, P = 0.023), OS (r = − 0.184, P = 0.033), and 1-year survival rate (r = − 0.194, P = 0.025). Associations were also observed between ctDNA status and the incidence of adverse events including neutropenia (r = 0.199, P = 0.021), fatigue (r = 0.205, P = 0.017), diarrhea (r = 0.219, P = 0.011), and peripheral neuropathy (r = 0.217, P = 0.012). These findings underscore the potential utility of ctDNA as a predictive biomarker for treatment response, survival outcomes, and occurrence of specific adverse events in patients with mCRC.

Table 5.

Correlation analysis ofctDNAwith treatment response and survival in patients withmCRC

Parameter	r	R2	P
CEA level	0.24	0.058	0.005
CA 19–9 level	0.214	0.046	0.013
ORR	− 0.192	0.037	0.026
DCR	− 0.202	0.041	0.019
PFS	− 0.196	0.039	0.023
OS	− 0.184	0.034	0.033
1-year survival rate	− 0.194	0.038	0.025
Neutropenia	0.199	0.04	0.021
Fatigue	0.205	0.042	0.017
Diarrhea	0.219	0.048	0.011
Peripheral neuropathy	0.217	0.047	0.012

ORR objective response rate, DCR disease control rate, PFS progression-free survival, OS overall survival

Discussion

Colorectal cancer remains a critical global health burden, with metastatic disease posing considerable challenges in terms of prognosis and treatment outcomes [16, 17]. The identification of reliable prognostic and predictive biomarkers was crucial for optimizing patient care and therapeutic decision-making in the management of mCRC [18–20]. ctDNA has emerged as a promising noninvasive biomarker with the potential to revolutionize personalized cancer medicine by offering real-time information on tumor burden, genetic heterogeneity, and treatment response [21–23].

Our findings revealed that ctDNA-negative patients exhibited superior treatment responses than ctDNA-positive patients. The ctDNA-negative patients achieved significantly higher objective response rates, disease control rates, ORR, and DCR than the ctDNA-positive patients. These findings are consistent with previous research implicating ctDNA as a potential predictive biomarker for treatment response in various cancer types, including CRC [24–26]. The ability to stratify patients based on ctDNA status may facilitate tailored treatment approaches, potentially leading to improved clinical outcomes and resource utilization in the management of mCRC.

In addition to the treatment response, our study demonstrated a significant association between ctDNA status and survival outcomes in patients with mCRC. The observed median PFS of 8.24 months in the ctDNA-negative group compared with 7.86 months in the ctDNA-positive group underscores the potential prognostic value of ctDNA in predicting disease progression and subsequent survival outcomes. Similarly, the ctDNA-negative group demonstrated a median OS of 24.58 months, whereas the ctDNA-positive group exhibited a median OS of 23.27 months, further highlighting the potential role of ctDNA as a prognostic biomarker for survival outcomes in patients with mCRC. These findings align with published studies reporting that prognostic significance of ctDNA in various cancer types, which emphasize the potential of ctDNA as a dynamic, real-time biomarker for monitoring disease progression and treatment response [27–29].

Importantly, our study identified specific adverse events associated with the ctDNA status in patients with mCRC. The incidence of adverse events, including neutropenia, fatigue, diarrhea, and peripheral neuropathy, was significantly higher in the ctDNA-positive group than that in the ctDNA-negative group. These findings suggest a potential association between ctDNA status and treatment tolerability, highlighting the clinical relevance of ctDNA for guiding patient management and supportive care strategies. The ability to anticipate and mitigate treatment-related adverse events based on ctDNA status may contribute to improved patient outcomes and quality of life during the course of mCRC treatment [30, 31].

Our correlation analysis further revealed significant associations between ctDNA and specific clinical parameters, including CEA and CA 19–9 level, and the incidence of adverse events. Positive correlations were observed between ctDNA status, CEA levels, and CA 19–9 level, supporting the potential of ctDNA as a complementary biomarker for traditional tumor markers in mCRC. Additionally, negative correlations were observed between ctDNA status and treatment response parameters (ORR, DCR, PFS, OS, and 1-year survival rate), further validating the prognostic significance of ctDNA in mCRC. These findings emphasize the potential of ctDNA as a comprehensive biomarker with implications for treatment response, survival outcomes, and adverse event profiles, warranting further exploration of personalized mCRC management.

Although our study provides valuable insights into the role of ctDNA as a predictive biomarker in mCRC, certain limitations should be acknowledged. The retrospective nature of the study and relatively modest sample size may have introduced inherent biases and limited the generalizability of the findings. Prospective studies are needed to validate these results and better control for confounding variables. The study was conducted at a single institution, and the relatively modest sample size of 134 patients may limit the statistical power and ability to detect subtle but clinically significant differences. Larger multi-center studies are required to confirm our findings and ensure that the results are applicable to a broader population. Additionally, this study focused on a specific ctDNA detection method and gene panel, warranting further validation in diverse patient populations and treatment settings. Future prospective studies encompassing multi-center collaborations and longitudinal ctDNA monitoring are warranted to provide robust evidence for the clinical utility of ctDNA as a predictive biomarker in the context of personalized mCRC management. Moreover, the integration of ctDNA analyses into routine clinical practice mandates considerations related to cost-effectiveness, assay standardization, and data interpretation, necessitating multidisciplinary efforts to optimize its implementation.

Conclusion

In summary, our research highlights the potential of ctDNA as a prognostic biomarker for both treatment response and survival in individuals with mCRC. The associations among ctDNA status, treatment response, survival outcomes, and adverse events highlight the multifaceted clinical utility of ctDNA in guiding personalized treatment strategies and patient management. As the paradigm of precision oncology continues to evolve, the integration of ctDNA analysis into routine clinical practice holds considerable promise for optimizing therapeutic decision-making and improving patient outcomes in the era of mCRC management.

Supplementary information

Below is the link to the electronic supplementary material.

Author contribution

All authors contributed to the study conception and design. Material preparation and data collection and analysis were performed by M. K. and Y. D. The first draft of the manuscript was written by M. K., and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.

Data availability

The original data involved in the present study can be provided under reasonable request.

Declarations

Ethics approval

This study was approved by the Ethics Committee of Maoming People’s Hospital in accordance with regulatory and ethical guidelines pertaining to retrospective research studies.

Consent to participate

Informed consent was waived for this retrospective study due to the exclusive use of de-identified patient data, which posed no potential harm or impact on patient care.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.