Chemotherapy plus bevacizumab with or without anti-programmed death 1 immunotherapy as the second-line therapy in colorectal cancer

Abstract

BACKGROUND

Patients with microsatellite stable (MSS) metastatic colorectal cancer (mCRC) typically exhibit an immunosuppressive tumor microenvironment and demonstrate a low response rate to immunotherapy. Reports suggest that chemotherapy and anti-angiogenic therapy may have the potential to enhance the response to immunotherapy in these patients. This study aims to evaluate the effectiveness and safety of chemotherapy combined with bevacizumab with or without anti-programmed death 1 (PD-1) immunotherapy as the second-line regimen for MSS mCRC.

AIM

To evaluate the effectiveness and safety of chemotherapy combined with bevacizumab with or without anti-PD-1 immunotherapy as the second-line regimen for MSS mCRC.

METHODS

A retrospective analysis was conducted on patients with MSS mCRC diagnosed at Peking University First Hospital and Jilin Cancer Hospital from January 2020 to December 2024. The patients were divided into two groups: The experimental group receiving second-line chemotherapy combined with bevacizumab and anti-PD-1 immunotherapy, and the control group receiving chemotherapy combined with bevacizumab. Propensity score matching was applied to balance potential prognostic factors, including age, gender, Eastern Cooperative Oncology Group score, number of metastases, and primary tumor site. The progression-free survival, overall survival, disease control rate, objective response rate, and treatment-related adverse reactions were compared between the two groups. Kaplan-Meier analysis and log-rank test were used to compare survival outcomes. Inverse probability of treatment weighting was used for sensitivity analysis.

RESULTS

Propensity score matching resulted in 103 matched eligible patients. The median follow-up period was 13.9 months in the matched cohort. The objective response rate was 11.5% and 9% for the experimental and control groups, respectively (P = 0.710), while the disease control rate was 76.9% and 53.2%, respectively (P = 0.058). The median progression-free survival in the experimental group was 8.27 months [95% confidence interval (CI): 6.7-14.7 months], significantly higher than that in the control group, which was 4.63 months (95%CI: 3.9-5.67 months) (hazard ratio = 0.4143, 95%CI: 0.2462-0.6972, P = 0.00066). There was a trend towards the higher median overall survival in the experimental group compared to the control group (hazard ratio = 0.4504, 95%CI: 0.1897-1.07, P = 0.064). The incidences of adverse events were similar between the two groups.

CONCLUSION

Compared with the standard second-line chemotherapy combined with bevacizumab regimen, second-line therapy that combines chemotherapy with bevacizumab and anti-PD-1 immunotherapy has demonstrated promising efficacy in the treatment of MSS mCRC, while exhibiting a similar safety profile.

Key Words: Microsatellite stable; RAS mutation; Metastatic colorectal cancer; Immune checkpoint inhibitors; Programmed death 1

Core Tip: This manuscript addresses to evaluate the effectiveness and safety of chemotherapy combined with bevacizumab with or without anti-programmed death 1 immunotherapy as the second-line regimen for microsatellite stable metastatic colorectal cancer. As there is currently no clinical data on the second-line treatment of advanced colorectal cancer with the combination of immunotherapy, anti-angiogenic drugs, and anti-programmed death 1 immunotherapy, we conducted a multicenter retrospective cohort clinical study to explore the safety and efficacy of this triplet therapy in second-line treatment of advanced colorectal cancer patients.

INTRODUCTION

The latest statistical data show that both the incidence and mortality rates of colorectal cancer are on the rise. In 2022, colorectal cancer ranked third in terms of new cases and second in terms of deaths among all malignant tumors worldwide[1]. For metastatic colorectal cancer (mCRC), the standard first-line recommended regimen currently involves chemotherapy combined with targeted therapy, which can improve tumor control rates in advanced patients to some extent. The first- and second-line chemotherapy options mainly involve sequencing folinic acid, 5-fluorouracil, oxaliplatin (FOLFOX) with folinic acid, 5-fluorouracil, irinotecan (FOLFIRI) or vice versa, with no impact on treatment efficacy[2]. In terms of targeted drug selection, for patients with wild-type RAS/BRAF genes, bevacizumab is the standard choice after cetuximab resistance; however, for those with primary resistance to bevacizumab in the first line, current clinical studies suggest that switching to cetuximab does not provide additional benefit compared to continuing bevacizumab[3,4]. Therefore, for second-line treatment of mCRC, regardless of gene status and primary tumor location, bevacizumab combined with second-line chemotherapy is the standard treatment option recommended by current guidelines[5]. Nevertheless, after first-line treatment failure, the overall response rates (ORR) of second-line chemotherapy combined with bevacizumab (doublet regimen) was only 5%-36%, with the median progression-free survival (PFS) of just 4-7 months[6-12]. Therefore, improving the efficacy of second-line treatment for mCRC is a major challenge in current clinical practice.

Immune checkpoint inhibitors have made significant breakthroughs in the treatment of multiple tumors, particularly in microsatellite instability-high mCRC, where anti-programmed death 1 (PD-1) immunotherapy has demonstrated outstanding efficacy[13,14]. However, this patient population accounts for less than 5% of all mCRC cases. For the majority of patients with microsatellite stable (MSS) mCRC, immunotherapy is largely ineffective, limiting their options for chemotherapy and targeted therapy. Therefore, new combination therapy regimens are needed to improve the response to immunotherapy for this subtype. Previous studies have shown that anti-angiogenic therapy can reverse the immunosuppressive tumor microenvironment by normalizing blood vessels and inducing T-cell infiltration and activation[15]. The phase II CheckMate 9X8 study compared the efficacy of nivolumab combined with modified FOLFOX6 (mFOLFOX6) (plus bevacizumab vs standard treatment (mFOLFOX6 plus bevacizumab) in first-line treatment of mCRC patients[16]. Subgroup analysis revealed that patients with consensus molecular subtype (CMS)3[17] could benefit from the addition of nivolumab to mFOLFOX6 plus bevacizumab. The BBCAPX study demonstrated that sintilimab combined with capecitabine and oxaliplatin (CAPEOX) and bevacizumab in first-line treatment improved disease response in patients with RAS-mutant MSS mCRC, with controllable adverse reactions and good safety[18]. The METIMMOX study was a phase II clinical trial comparing the efficacy of chemotherapy combined with nivolumab vs chemotherapy alone as first-line treatment for MSS mCRC[19]. Results showed a median PFS of 6.6 months in the combination immunotherapy group vs 5.6 months in the chemotherapy-alone group. This study suggested that short-course oxaliplatin-based chemotherapy in MSS mCRC patients may alter tumor immunogenicity, potentially inducing responsiveness to immune checkpoint inhibitors.

Based on these findings, chemotherapy combined with bevacizumab and anti-PD-1 immunotherapy (triplet regimen) has a solid theoretical foundation and has shown preliminary efficacy and safety in MSS mCRC patients. Therefore, the optimal combination of these three therapeutic agents - chemotherapy, anti-angiogenic drugs, and anti-PD-1 immunotherapy - should be an important research direction for changing the current treatment landscape of mCRC. As there is currently no clinical data on the second-line treatment of advanced colorectal cancer with the combination of chemotherapy, anti-angiogenic drugs, and anti-PD-1 immunotherapy, we conducted a multicenter retrospective cohort clinical study to explore the safety and efficacy of this triplet therapy in second-line treatment of advanced colorectal cancer patients.

MATERIALS AND METHODS

Study design and participants

This study employed a multicenter retrospective cohort research design. Patients with advanced colorectal cancer who were treated at Peking University First Hospital and Jilin Cancer Hospital between January 1, 2020 and December 30, 2024 were enrolled. The experimental group consisted of patients who received second-line treatment with chemotherapy combined with bevacizumab and anti-PD-1 immunotherapy, while the control group comprised patients who received conventional treatment (chemotherapy combined with bevacizumab).

The inclusion criteria were: (1) Histologically or cytologically confirmed unresectable mCRC (stage IV according to the American Joint Committee on Cancer Staging Manual 8^th edition) with measurable lesions based on the Response Evaluation Criteria In Solid Tumors (RECIST) 1.1 criteria; (2) Progression after prior first-line standard two-drug chemotherapy regimen with or without targeted therapy; (3) Proficient mismatch repair (pMMR) or MSS, BRAF wild-type; (4) Patients receiving chemotherapy combined with bevacizumab and anti-PD-1 immunotherapy or chemotherapy combined with bevacizumab; (5) Having not undergone radiotherapy or having completed radiotherapy more than 4 weeks ago; and (6) Eastern Cooperative Oncology Group (ECOG) score ≤ 2.

The exclusion criteria were: (1) Patients with deficient DNA mismatch repair or microsatellite instability-high, or BRAF mutations; (2) Presence of symptomatic brain metastases; (3) Uncontrolled active infection; (4) Dysphagia, intractable vomiting, or known drug absorption disorders; and (5) Patients with symptomatic or high-risk obstruction, bleeding, or perforation, or those who have undergone intestinal stent placement to relieve intestinal obstruction.

In most cases, the chemotherapy regimen comprised an oxaliplatin-based doublet (FOLFOX, CAPEOX, or raltitrexed and oxaliplatin) or a topoisomerase inhibitor-based (FOLFIRI or CAPEOX: Capecitabine and irinotecan (IRI), or raltitrexed and IRI). anti-PD-1 immunotherapy included penpulimab, pembrolizumab, sintilimab, tislelizumab, and toripalimab. The anti-angiogenic agent was bevacizumab.

This study was conducted in compliance with the postulates of Declaration of Helsinki and approved by the Ethics Committee of Peking University First Hospital and Jilin Cancer Hospital. The requirement for patient approval or informed consent was waived by the Human Ethics Committee of Peking University First Hospital and Jilin Cancer Hospital, owing to the retrospective nature of the study and because the analysis used anonymous clinical data. The flowchart of patient selection is shown in Figure 1.

Figure 1 Flow chart of patient inclusion. MSS: Microsatellite stable; bev: Bevacizumab; pMMR: Proficient mismatch repair; MSI-H: Microsatellite instability-high.

Procedures

Follow-up data were collected through hospital records, telephone interviews, outpatient visits, and rehospitalizations. The data included age, sex, height, weight, ECOG status, primary tumor location, number of metastatic sites, tumor differentiation grade, percentage reduction in tumor volume, PFS, overall survival (OS), follow-up duration, and survival status. Additionally, peripheral blood indicators within 7 days prior to the initial triplet regimen were collected, encompassing absolute leukocyte count, absolute neutrophil count (ANC), absolute lymphocyte count (ALC), platelet (PLT) count, absolute monocyte count, absolute eosinophil count, albumin (ALB), lactate dehydrogenase (LDH), carcinoembryonic antigen and carbohydrate antigen 199. Neutrophil-to-lymphocyte ratio (NLR) was calculated as ANC/ALC. Lymphocyte-to-monocyte ratio was calculated as the ratio of ALC/absolute monocyte count. Platelet-to-lymphocyte ratio was calculated as the ratio of PLT/ALC, body mass index (BMI) was defined as weight (kg) divided by height squared (m^²) [weight (kg)/height (m)²], advanced lung cancer inflammation index was composed of BMI, ALB, and NLR, with the specific formula being BMI (kg/m²) × ALB level (g/dL)/NLR. Systemic immune-inflammation index, as an evaluation index of systemic inflammatory response, the calculation formula is PLT × ANC/ALC.

The last follow-up date was in January 2025. PFS was the primary outcome, defined as the time from enrollment to the first documented disease progression according to RECIST version 1.1, or death from any cause, whichever occurred first. Secondary outcomes included OS, ORR, disease control rate (DCR), and safety evaluation. OS was calculated from the date of enrollment to the date of death from any cause, with censored cases defined by the last available follow-up. ORR was defined as the proportion of patients with a best objective response of complete response or partial response according to RECIST criteria (version 1.1). DCR was defined as the proportion of patients with complete response, partial response, or stable disease according to RECIST criteria (version 1.1). Treatment-related adverse events were evaluated according to the Common Terminology Criteria for Adverse Events version 4.0.

Statistical analysis

The PSM method was used to eliminate potential confounding factors that could influence the therapeutic effect between the experimental and control groups. PSM analysis was conducted using the nearest-neighbor method with a caliper of 0.018 and a 1:4 matching ratio to balance characteristics such as age, sex, ECOG status, metastasis, and tumor location using the MatchIt package[20]. Categorical variables were compared using the χ² test or Fisher’s exact test, while continuous variables were assessed using the Mann-Whitney test. Kaplan-Meier estimates were obtained to compare the actuarial survival and the two treatment efficacy endpoints between the two groups. Independent prognostic factors, hazard ratios (HRs), and 95% confidence intervals (CIs) were evaluated using the Cox proportional hazards (PH) model. The cutoff values for continuous variable hematological indicators in predicting patient survival were set at the median. Based on this cutoff value, patients were categorized into high-expression and low-expression groups. The predictive value of hematological indicators for the triplet regimen in MSS mCRC was assessed using receiver operating characteristic curve analysis. A two-sided P value of < 0.050 was considered statistically significant. All statistical analyses in our study were performed using R software (version 4.4.2).

Sensitivity analysis

To address potential biases, three sensitivity analyses were conducted. First, we performed the Schoenfeld residual test on the original data to assess whether the covariates satisfied the PHs assumption, using the cox. zph function from the survival package[21]. If the PH assumption was met, we proceeded with univariate and multivariate regression analyses. Second, PSM analysis with varying matching ratios and inverse probability of treatment weighting are employed to adjust for baseline characteristics and evaluate treatment outcomes of PFS and OS. Third, to evaluate the robustness of the results, we both excluded four patients who received single-agent chemotherapy combined with bevacizumab and immunotherapy, and focused our analysis on patients treated with IRI-based regimens.

RESULTS

Patients

After one-to-four lines of PSM, the experimental group receiving triplet regimen included 26 patients, while the control group comprised 77 patients. Among the 103 patients (73 males and 30 females), 45 patients were over 60 years old. There were 29 cases (28.2%) of right-sided colon cancer, 85 patients underwent primary tumor surgery, and 69 patients (67%) had metastases in two or more organs. Sixty-eight patients had RAS mutations, and all cases were pMMR (Table 1). The main baseline characteristics of eligible patients were well-balanced between the two groups (Supplementary Figure 1A).

Table 1 Baseline characteristics and treatment details of patients in the matched cohort, n (%).

Characteristics	Levels	Control group (n = 77)	Experiment group (n = 26)	P value
Age	> 60	33 (42.9)	12 (46.2)	0.949
	≤ 60	44 (57.1)	14 (53.8)
Gender	Male	55 (71.4)	18 (69.2)	1.000
	Female	22 (28.6)	8 (30.8)
ECOG	0-1	69 (89.6)	22 (84.6)	0.739
	2	8 (10.4)	4 (15.4)
Primary tumor location	Right colon	20 (26)	9 (34.6)	0.552
	Left colon and rectum	57 (74)	17 (65.4)
Primary tumor surgery	No	15 (19.5)	3 (11.5)	0.533
	Yes	62 (80.5)	23 (88.5)
Number of metastatic organs	1	26 (33.8)	8 (30.8)	0.968
	≥ 2	51 (66.2)	18 (69.2)
Liver metastasis	No	19 (24.7)	8 (30.8)	0.724
	Yes	58 (75.3)	18 (69.2)
Lung metastasis	No	41 (53.2)	11 (42.3)	0.461
	Yes	36 (46.8)	15 (57.7)
RAS mutation type	Unknown	15 (19.5)	5 (19.2)	0.872
	Wild-type	12 (15.6)	3 (11.5)
	Mutation	50 (64.9)	18 (69.2)

The control group received chemotherapy combined with bevacizumab; the experimental group received chemotherapy combined with bevacizumab and anti- programmed death 1 immunotherapy. ECOG: Eastern Cooperative Oncology Group.

Effectiveness

As of December 2024, the median follow-up period was 13.9 months in the matched cohort. The median PFS was 5.33 months, (95%CI: 4.6-6.33 months), and the median OS was 23 months. In the experimental group, the median PFS was 8.27 months (95%CI: 6.7-14.7 months), and the median OS was 8.6 months. In the control group, the median PFS was 4.63 months (95%CI: 3.9-5.67 months). The PFS in the experimental group was superior to that in the control group (HR = 0.414, 95%CI: 0.2462-0.6972, P = 0.00066) (Figure 2A). There was a trend towards a higher median OS in the experimental group compared to the control group (HR = 0.4504, 95%CI: 0.1897-1.07, P = 0.064) (Figure 2B). The ORR was 11.5% in the experimental group and 9% in the control group (P = 0.710), while the DCR was 76.9% and 53.2%, respectively (P = 0.058) (Table 2).

Figure 2 After propensity score-matching analysis of progression-free survival and overall survival. A: After propensity score-matching analysis of PFS (ratio = 4); B: After propensity score-matching analysis of overall survival (ratio = 4).

Table 2 Comparison of short-term efficacy between two groups of microsatellite stable metastatic colorectal cancer.

Group	Number	PR	SD	PD	DCR (%)
Control group	77	7	34	36	57.14
Experimental group	26	3	17	6	76.92
χ2 value	-	-	2.706	3.585	3.5847
P value	-	0.711	0.1	0.058	0.058

¹Fisher’s exact test.

The control group received chemotherapy combined with bevacizumab; the experimental group received chemotherapy combined with bevacizumab and anti- programmed death 1 immunotherapy. PR: Partial response; SD: Stable disease; PD: Progressive disease; DCR: Disease control rate.

Sensitivity analysis

For the original cohort, the PFS in the experimental group was superior to that in the control group (HR = 0.414, 95%CI: 0.2462-0.6972, P = 0.0054) (Figure 3A) and the OS did not show a statistically significant difference between the two groups (HR = 0.547, 95%CI: 0.2328-1.287, P = 0.167) (Figure 3B). The independent variables conformed to the PHs assumption model (P > 0.05). Cox univariate and multivariate analyses revealed that immunotherapy combined with chemotherapy (P = 0.004) and patients without liver metastases (P = 0.002) were associated with better PFS (Table 3). Robustness of survival analysis results: insensitivity to changes in PSM ratio, treatment regimen exclusions, and cohort restrictions in PFS and OS (Figures 4 and 5). The inverse probability of treatment weighting also reached the same conclusion for PFS (P = 0.0014) (Figure 4D) and OS (P = 0.1505) (Figure 5D). Meanwhile, after applying weights, the variables between the two groups maintained a substantial balance at the baseline level (Supplementary Figure 1). Adjusting the PSM ratio in PFS (Figure 5A-C) and OS (Figure 6A-C) did not alter the final results; neither did excluding patients treated with single-agent triplet regimens (Figures 4E and 5E). Similarly, restricting the cohort to IRI-based chemotherapy patients had no effect on the conclusions (Figures 4F and 5F).

Figure 3 Kaplan-Meier curve of original cohort. A: In the original cohort (progression-free survival); B: In the original cohort (overall survival).

Figure 4 Sensitivity analysis of progression-free survival. A: After propensity score-matching analysis in progression-free survival (PFS) (ratio = 1); B: After propensity score-matching analysis in PFS (ratio = 2); C: After propensity score-matching analysis in PFS (ratio = 3); D: After inverse probability of treatment weighting analysis; E: Kaplan-Meier curve after excluding 4 patients who received single-agent chemotherapy combined with bevacizumab and anti- programmed death 1 immunotherapy in PFS; F: Kaplan-Meier curve restricting the cohort to irinotecan-based chemotherapy patients in PFS.

Figure 5 Sensitivity analysis of overall survival. A: After propensity score-matching analysis in overall survival (OS) (ratio = 1); B: After propensity score-matching analysis in OS (ratio = 2); C: After propensity score-matching analysis in OS (ratio = 3); D: After inverse probability of treatment weighting analysis (OS); E: Kaplan-Meier curve after excluding 4 patients who received single-agent chemotherapy combined with bevacizumab and anti-programmed death 1 immunotherapy in OS; F: Kaplan-Meier curve restricting the cohort to irinotecan-based chemotherapy patients in OS.

Figure 6 Forest plots depict the hazard ratios and 95% confidence intervals for progression-free survival by subgroup. CI: Confidence interval; ECOG: Eastern Cooperative Oncology Group.

Table 3 Univariate and multivariate Cox analysis of the effect of prognostic factors in the original cohort, n (%).

Dependent: Survival (PFS/30, status)		All patients (n = 131)	HR (univariable)	HR (multivariable)
Gender	Male	76 (58.0)	-	-
	Female	55 (42.0)	0.87 (0.60-1.26, P = 0.458)	-
Age	> 60	66 (50.4)	-	-
	≤ 60	65 (49.6)	0.73 (0.50-1.07, P = 0.110)	-
ECOG	0-1	116 (88.5)	-	-
	2	15 (11.5)	1.34 (0.73-2.47, P = 0.340)	-
Number of metastatic organs	1	55 (42.0)	-	-
	≥ 2	76 (58.0)	1.96 (1.32-2.90, P < 0.001)	1.90 (1.27-2.84, P = 0.002)
Primary tumor location	Right colon	32 (24.4)	-	-
	Left colon and rectum	99 (75.6)	0.92 (0.61-1.40, P = 0.697)	-
Liver metastasis	No	46 (35.1)	-	-
	Yes	85 (64.9)	1.57 (1.04-2.36, P = 0.032)	1.28 (0.84-1.96, P = 0.254)
Lung metastasis	No	70 (53.4)	-	-
	Yes	61 (46.6)	1.23 (0.85-1.78, P = 0.274)	-
RAS mutation type	Unknown	25 (19.1)	-	-
	Wild type	20 (15.3)	0.97 (0.51-1.85, P = 0.930)	-
	Mutation	86 (65.6)	1.37 (0.84-2.24, P = 0.202)	-
Group	Control group	105 (80.2)	-	-
	Experimental group	26 (19.8)	0.51 (0.31-0.82, P = 0.006)	0.49 (0.30-0.80, P = 0.004)

n = 131, events = 115, likelihood ratio test = 22.56 on 3 df (P < 0.001). ECOG: Eastern Cooperative Oncology Group; PFS: Progression-free survival; HR: Hazard ratio.

Safety

No substantial differences in adverse events were observed between the two groups. The incidence of grade 1-2 fatigue was slightly higher in the experimental group compared to the control group (30.8% vs 11.7%, P = 0.0332), but the incidence of grade 3-4 adverse events was similar between the two groups (Table 4).

Table 4 Treatment-emergent adverse events in the 103 patients of the matched dataset, n (%).

Toxicities	Experimental group (n = 26)			Control group (n = 77)			P value for grade 1-2	P value for grade 3-4
	Grade 0	Grade 1-2	Grade 3-4	Grade 0	Grade 1-2	Grade 3-4
Anemia	8 (30.8)	15 (57.7)	3 (11.5)	47 (61.0)	29 (37.7)	1 (1.3)	0.1197a	0.2451b
Neutropenia	18 (69.2)	7 (26.9)	1 (3.8)	58 (75.3)	14 (18.2)	5 (6.5)	0.4996a	1b
Leukocytopenia	14 (53.8)	11 (42.3)	1 (3.8)	55 (71.4)	18 (23.4)	4 (5.2)	0.1088a	0.3667b
Thrombocytopenia	17 (65.4)	9 (34.6)	0 (0.0)	62 (80.5)	15 (19.5)	0 (0.0)	0.1902a	1b
Proteinuria	21 (80.8)	5 (19.2)	0 (0.0)	70 (90.9)	7 (9.1)	0 (0.0)	0.2984a	1b
Aspartate transaminase increased	18 (69.2)	7 (26.9)	1 (3.8)	56 (72.7)	20 (26.0)	1 (1.3)	1a	1b
Alanine transaminase increased	20 (76.9)	5 (19.2)	1 (3.8)	65 (84.4)	12 (15.6)	0 (0.0)	0.8985a	1b
Alkaline phosphatase increased	21 (80.8)	5 (19.2)	0 (0.0)	69 (89.6)	7 (9.1)	1 (1.3)	0.1732b	1b
Blood bilirubin increased	20 (76.9)	5 (19.2)	1 (3.8)	61 (79.2)	16 (20.8)	0 (0.0)	1a	1b
Triglycerides increased	20 (76.9)	4 (15.4)	2 (7.7)	67 (87.0)	10 (13.0)	0 (0.0)	0.7475b	0.4967 b
Nausea	18 (69.2)	8 (30.8)	0 (0.0)	47 (61.0)	30 (39.0)	0 (0.0)	0.6076a	1b
Vomiting	19 (73.1)	7 (26.9)	0 (0.0)	61 (79.2)	16 (20.8)	0 (0.0)	0.7054a	1b
Fatigue	18 (69.2)	8 (30.8)	0 (0.0)	68 (88.3)	9 (11.7)	0 (0.0)	0.0332b	1b
Fever	24 (92.3)	2 (7.7)	0 (0.0)	73 (94.8)	4 (5.2)	0 (0.0)	0.6407b	1b
Diarrhea	24 (92.3)	1 (3.8)	1 (3.8)	71 (92.2)	6 (7.8)	0 (0.0)	0.6759b	1b
Peripheral neurotoxicity	25 (96.2)	1 (3.8)	0 (0.0)	77 (100.0)	0 (0.0)	0 (0.0)	0.2524b	1b
Hand-foot syndrome	23 (88.5)	3 (11.5)	0 (0.0)	73 (94.8)	4 (5.2)	0 (0.0)	0.3648b	1b
Hypertension	24 (92.3)	1 (3.8)	1 (3.8)	72 (93.5)	4 (5.2)	1 (1.3)	1b	1b

^aP values were calculated by χ² test.

^bP values were calculated by Fisher’s exact test.

The control group received chemotherapy combined with bevacizumab; the experimental group received chemotherapy combined with bevacizumab and anti-programmed death 1 immunotherapy.

Subgroup analysis

Subgroup analysis showed that patients younger than 60 years old (HR = 0.25, 95%CI: 0.12-0.56, P = 0.001), those without liver metastases (HR = 0.14, 95%CI: 0.04-0.48, P = 0.002), and those with lung metastases (HR = 0.27, 95%CI: 0.13-0.58, P = 0.001) are more likely to benefit from triplet regimen (Figure 6).

Exploratory biomarker analysis

We analyzed the baseline hematological indicators associated with PFS in patients receiving triplet regimen (Table 5). Univariate analysis revealed that patients with advanced age, right colorectal cancer (CRC), normal carbohydrate antigen 199 levels, high advanced lung cancer inflammation index, and low LDH could benefit from triplet regimen. Multivariate analysis further identified liver metastasis (HR = 8.15, 95%CI: 1.39-47.83, P = 0.020) and LDH (HR = 4.11, 95%CI: 1.02-16.55, P = 0.046) as independent prognostic risk factors (Table 6).

Table 5 Baseline hematological prognostic indicators markers for patients receiving chemotherapy combined with bevacizumab and anti-programmed death 1immunotherapy, mean ± SD.

Characteristics	Stats	Normal range
Height (cm)	168.3 ± 8.0	140-190
Weight (kg)	67.7 ± 12.0	40-100
ALLC (109/L)	6.0 ± 1.9	3.5-9.5
RDW (%)	15.2 ± 3.8	11.6-14.8
PLT (109/L)	178.2 ± 63.7	125-350
ANC (109/L)	4.1 ± 1.4	1.8-6.3
ALC (109/L)	1.4 ± 0.5	1.1-3.2
AMC (109/L)	0.4 ± 0.1	0.1-0.6
AEC (109/L)	0.2 ± 0.2	0.02-0.52
ALB (g/L)	41.4 ± 4.0	40-55
LDH (IU/L)	240.0 ± 154.9	109-245
FIB (g/L)	3.6 ± 0.7	2-4
Dimer (ng/ml)	257.3 ± 647.3	0-500
CEA (ng/mL), n (%)	6 (24.0%)	0-5
	19 (76.0%)	-
CA199 (IU/mL), n (%)	13 (52.0%)	0-37
	12 (48.0%)	-

ALLC: Absolute leukocyte count; RDW: Red blood cell distribution width; PLT: Platelet; ANC: Absolute neutrophil count; ALC: Absolute lymphocyte count; AMC: Absolute monocyte count; AEC: Absolute eosinophil count; ALB: Albumin; LDH: Lactate dehydrogenase; FIB: Fibrinogen; CEA: Carcinoembryonic antigen; CA199: Carbohydrate antigen 199.

Table 6 Univariate and multivariate Cox analysis of prognostic factors in microsatellite stable metastatic colorectal cancer patients receiving chemotherapy combined with bevacizumab and anti-programmed death 1 immunotherapy.

Dependent: Survival (PFS/30, status)		All	HR (univariable)	HR (multivariable)
Age	> 60	11 (44.0)	-	-
	≤ 60	14 (56.0)	0.20 (0.07-0.62, P = 0.005)	0.27 (0.05-1.48, P = 0.133)
Gender	Male	18 (72.0)	-	-
	Female	7 (28.0)	1.01 (0.40-2.54, P = 0.979)	-
ECOG	1	21 (84.0)	-	-
	2	4 (16.0)	1.01 (0.23-4.55, P = 0.985)	-
Location	Right colon	9 (36.0)	-	-
	Left colon and rectum	16 (64.0)	0.73 (0.31-1.74, P = 0.477)	-
Liver metastasis	No	7 (28.0)	-	-
	Yes	18 (72.0)	5.23 (1.48-18.50, P = 0.010)	8.15 (1.39-47.83, P = 0.020)
Lung metastasis	No	10 (40.0)	-	-
	Yes	15 (60.0)	0.62 (0.25-1.51, P = 0.291)	-
RAS mutation type	Unknown	5 (20.0)	-	-
	Wild type	3 (12.0)	0.75 (0.13-4.20, P = 0.742)	-
	Mutation	17 (68.0)	2.30 (0.74-7.14, P = 0.151)	-
CEA	Normal	6 (24.0)	-	-
	Abnormal	19 (76.0)	0.59 (0.21-1.67, P = 0.316)	-
CA199	Normal	13 (52.0)	-	-
	Abnormal	12 (48.0)	3.16 (1.12-8.88, P = 0.029)	1.06 (0.23-4.90, P = 0.942)
NLR	≤ 2.9	13 (52.0)	-	-
	> 2.9	12 (48.0)	1.45 (0.60-3.55, P = 0.410)	-
LMR	≤ 3.63	13 (52.0)	-	-
	> 3.63	12 (48.0)	0.57 (0.23-1.39, P = 0.216)	-
PLR	≤ 121	13 (52.0)	-	-
	> 121	12 (48.0)	1.43 (0.58-3.48, P = 0.436)	-
BMI	≤ 24.9	12 (48.0)	-	-
	> 24.9	13 (52.0)	1.02 (0.43-2.42, P = 0.963)	-
ALI	≤ 260.9	13 (52.0)	-	-
	> 260.9	12 (48.0)	0.30 (0.10-0.85, P = 0.023)	0.50 (0.15-1.64, P = 0.252)
SII	≤ 426.8	13 (52.0)	-	-
	> 426.8	12 (48.0)	1.74 (0.72-4.25, P = 0.220)	-
ALLC	≤ 6.0	13 (52.0)	-	-
	> 6.0	12 (48.0)	1.31 (0.53-3.20, P = 0.559)	-
ANC	≤ 4.1	13 (52.0)	-	-
	> 4.1	12 (48.0)	1.31 (0.53-3.20, P = 0.559)	-
ALC	≤ 1.4	13 (52.0)	-	-
	> 1.4	12 (48.0)	0.61 (0.25-1.50, P = 0.285)	-
AMC	≤ 0.4	13 (52.0)	-	-
	> 0.4	12 (48.0)	0.71 (0.28-1.82, P = 0.473)	-
AEC	≤ 0.2	13 (52.0)	-	-
	> 0.2	12 (48.0)	1.59 (0.65-3.89, P = 0.309)	-
RDW	≤ 15.2	13 (52.0)	-	-
	> 15.2	12 (48.0)	1.04 (0.42-2.57, P = 0.938)	-
LDH	≤ 240.0	13 (52.0)	-	-
	> 240.0	12 (48.0)	3.09 (1.16-8.22, P = 0.024)	5.72 (1.58-20.72, P = 0.008)

n = 25, events = 21, likelihood ratio test = 25.81 on 5 df (P < 0.001). ECOG: Eastern Cooperative Oncology Group; PFS: Progression-free survival; HR: Hazard ratio; CEA: Carcinoembryonic antigen; CA199: Carbohydrate antigen 199; NLR: Neutrophil-to-lymphocyte ratio; LMR: Lymphocyte-to-monocyte ratio; PLR: Platelet-to-lymphocyte ratio; BMI: Body mass index; ALI: Advanced lung cancer inflammation index; SII: Systemic immune-inflammation index; ALLC: Absolute leukocyte count; ANC: Absolute neutrophil count; ALC: Absolute lymphocyte count; AMC: Absolute monocyte count; AEC: Absolute eosinophil count; RDW: Red blood cell distribution width; LDH: Lactate dehydrogenase.

The area under curve values of LDH in predicting the efficacy of second-line triplet regimen for MSS mCRC patients at 6 months, 9 months, and 12 months were 0.80, 0.79, and 0.72 respectively (Figure 7A). Based on the median value of LDH, MSS mCRC patients were divided into high groups and low groups. Survival analysis demonstrated that patients in the high-value group had significantly prolonged PFS compared to those in the low-value group (P = 0.019) (Figure 7B).

Figure 7 Lactate dehydrogenase as a predictive biomarker of progression-free survival. A: Receiver operating characteristic curves of lactate dehydrogenase (LDH) at 9 months, 12 months, and 15 months before patients receiving chemotherapy combined with bevacizumab and anti-programmed death 1 immunotherapy; B: Survival curves of LDH before patients receiving chemotherapy combined with bevacizumab and anti-programmed death 1 immunotherapy (high group: LDH ≥ 240 mmol/L and low group LDH < 240 mmol/L). ACU: Area under curve.

DISCUSSION

Multiple Phase III clinical studies have demonstrated that, following the failure of first-line chemotherapy combined with bevacizumab in advanced colorectal cancer, switching the chemotherapy regimen while continuing anti-angiogenic therapy provides additional survival benefits compared to chemotherapy alone[7,8,22]. However, the role of immunotherapy in second-line sequential treatment for mCRC remains unestablished. Previous studies have shown that IRI-based chemotherapy for second-line treatment of advanced CRC yields an ORR of 4%-18.8%, with the median PFS of 2.5-5.8 months and the median OS of 9.9-19.5 months[8,22-25]. When IRI-based chemotherapy is combined with anti-angiogenic targeted therapy for second-line treatment of advanced CRC, the ORR ranges from 14%-19.8%, with median PFS and OS ranging from 3.5-6.9 months and 11.9-13.5 months, respectively[8,22,24]. This study revealed an ORR of 11.5% and a DCR of 76.9% with triplet regimen. The triplet regimen for second-line treatment of advanced CRC showed superior PFS compared to the doublet regimen (P = 0.00066). Although our analysis demonstrated a trend toward OS benefit (P = 0.064), this finding requires cautious interpretation. Whether PFS improvement and depth of response will ultimately translate into definitive OS gains remains uncertain and necessitates validation through larger sample sizes and prospective cohort studies.

The results of chemotherapy combined with anti-PD-1 immunotherapy for first-line treatment of patients with MSS mCRC are unsatisfactory, with limited overall efficacy improvement. Further consideration is needed regarding the combination and value of this treatment modality. The CHECKMATE-9X8[16] study explored the efficacy of nivolumab plus mFOLFOX6/bevacizumab [nivolumab + standard-of-care (SOC)] vs mFOLFOX6/bevacizumab (SOC) alone for first-line treatment of mCRC. The ORR was 60% in the nivolumab group vs 46% in the SOC group, with a median PFS of 11.9 months in both groups. The median OS was 29.2 months in the nivolumab group, and not reached in the SOC group. Exploratory subgroup analysis of this study indicated that a higher proportion of patients with CMS1 and CMS3 tumor types remained progression-free at 12 months with nivolumab treatment. The AtezoTRIBE randomized Phase II trial, comparing FOLFOXIRI and bevacizumab with or without atezolizumab, showed that adding atezolizumab prolonged PFS in the overall population[26]. However, in the MSS patient subgroup, the addition of atezolizumab did not significantly improve PFS. In multivariate analysis, high tumor mutational burden and high immune score were independently associated with prolonged PFS with atezolizumab treatment. The KEYNOTE-651 study evaluated the long-term safety and efficacy of pembrolizumab combined with oxaliplatin, leucovorin, and fluorouracil (mFOLFOX7 regimen) for first-line treatment or pembrolizumab combined with IRI, leucovorin, and fluorouracil (FOLFIRI regimen) for second-line treatment of MSS/pMMR mCRC. The results suggested that pembrolizumab combined with chemotherapy is safe and effective for both first-line and second-line treatment of MSS/pMMR mCRC[27]. The NIVACOR trial assessed the efficacy and safety of nivolumab combined with FOLFOXIRI and bevacizumab in RAS/BRAF-mutant mCRC patients for first-line treatment. In the MSS patient subgroup analysis, the ORR was 78.9%, the DCR was 96.2%, and the median PFS was 9.8 months (95%CI: 8.18-15.24)[28]. However, there is a lack of data on second-line immune combination therapy for MSS mCRC. We used the triplet regimen and found that it had certain advantages in terms of the median PFS for second-line treatment of MSS mCRC.

Basic studies have shown that liver metastases from CRC are in an immunosuppressive environment[29]. Subgroup analysis of our study found that patients with liver metastases had worse prognoses than those without liver metastases, while patients with lung metastases had better prognoses with triplet regimen. This is similar to the findings of the REGONIVO study[30], which reported better prognoses for patients with lung metastases compared to those with liver metastases. In contrast, subgroup analysis of the bevacizumab and CAPEOX study found that patients with liver metastases had better prognoses in the triplet regimen[18]. Further basic research is needed to explore the efficacy of anti-PD1 immunotherapy in mCRC with different metastatic sites.

The results of this study indicate that baseline LDH level is an independent prognostic factor for PFS following the triplet regimen in MSS mCRC. From a metabolic perspective, tumor cells tend to prioritize glycolysis for energy production, where glucose is metabolized to pyruvate and then converted to lactate via LDH instead of undergoing aerobic oxidation through the mitochondrial tricarboxylic acid cycle, a phenomenon known as the “Warburg effect”[31]. Research has shown that in metastatic cervical cancer patients receiving combination immunotherapy, higher LDH levels are associated with lower survival rates[32]. Additionally, patients with normal LDH levels before receiving camrelizumab treatment for esophageal squamous cell carcinoma exhibit longer OS[33].

We acknowledge that our multicenter retrospective cohort study has several limitations. First, as a retrospective analysis with a relatively small sample size in the experimental group, data was collected from only two participating centers, and the findings require validation in larger cohorts and more research institutions. Second, the heterogeneity in treatment regimens - including diverse chemotherapy protocols and five different anti-PD-1 agents - may have confounded efficacy comparisons. The lack of standardization in both chemotherapy and immunotherapy regimens introduces variability that complicates direct comparisons. Third, the immaturity of PFS and OS data in this retrospective analysis, coupled with suboptimal patient management practices, may have introduced biases. Fourth, the study did not evaluate programmed death-ligand 1 expression levels in patients, precluding analysis of the association between programmed death-ligand 1 expression and the efficacy of anti-PD-1 immunotherapy.

CONCLUSION

In the second-line treatment of MSS mCRC, chemotherapy combined with bevacizumab and anti-PD-1 immunotherapy is superior to the traditional regimen of chemotherapy with bevacizumab. This indicates that the triplet regimen is a promising therapeutic strategy, which is expected to provide more clinical benefits. Subgroup analysis shows that patients younger than 60 years old, those without liver metastases, and those with lung metastases may benefit more significantly.

ACKNOWLEDGEMENTS

We are grateful to the patients and their families for supporting the study.