Cancer-derived inflammation is associated with bevacizumab resistance and inferior clinical outcomes in patients with metastatic colorectal cancer

Abstract

Background:

Cancer-derived inflammation has been shown to attenuate the efficacy of chemotherapy (CT) in colorectal cancer. However, its role in clinical response to bevacizumab (Bev) remains unclear in patients with metastatic colorectal cancer (mCRC).

Objectives:

To investigate the clinical significance of the fibrinogen-to-prealbumin ratio (FPR) as a biomarker in mCRC patients receiving first-line Bev plus CT (Bev/CT).

Design:

A retrospective, single-center, observational study.

Methods:

Patients treated with first-line Bev/CT were enrolled across discovery (n = 249), internal validation (n = 115), and external validation (n = 109) cohorts, along with a patient group receiving CT alone (n = 175). Propensity score matching was performed to balance baseline characteristics between the Bev/CT- and CT-treated groups. The primary endpoints were disease control rate (DCR) and progression-free survival (PFS), and 1-year overall survival (OS) was designated as a secondary endpoint.

Results:

Elevated pretreatment FPR was significantly associated with increased tumor burden. FPR-H patients exhibited an unfavorable DCR compared to the FPR-L patients in the overall Bev/CT-treated population (adjusted odds ratio (OR) = 2.18, 95% confidence interval (CI) = 1.32–3.61, p < 0.01). FPR-H was also independently associated with shorter PFS in the discovery (p_log-rank < 0.01, adjusted hazard ratio (HR) = 1.51, 95% CI = 1.11–2.05), internal validation (p_log-rank < 0.01, adjusted HR = 2.47, 95% CI = 1.43–4.27), and external validation cohorts (p_log-rank < 0.001, adjusted HR = 2.28, 95% CI = 1.43–3.64), regardless of KRAS status, particularly in the microsatellite stable subgroup. Furthermore, FPR-H patients had significantly worse 1-year OS than the FPR-L subgroup in the overall population (p_log-rank < 0.001; adjusted HR = 2.86, 95% CI = 1.78–4.61). Dynamic monitoring of FPR demonstrated utility in tracking disease progression during Bev/CT treatment, with earlier detection capability compared to imaging-based assessment. In propensity score-matched analysis, patients receiving Bev/CT showed superior outcomes relative to those receiving CT alone (p_log-rank < 0.001), especially among the FPR-L subgroup (p_log-rank < 0.01). In addition, Bev/oxaliplatin-treated FPR-L cases achieved better outcomes than those receiving the Bev/irinotecan regimen, particularly in the subgroup undergoing palliative resection.

Conclusion:

Cancer-derived inflammation appears to play a crucial role in mediating both intrinsic and acquired resistance to Bev. The FPR emerges as a robust, independent biomarker for predicting response to Bev, monitoring disease progression, and informing personalized therapeutic decision-making in patients with mCRC.

Graphical abstract

Keywords

bevacizumab colorectal cancer fibrinogen-to-prealbumin ratio inflammation therapeutic resistance

Introduction

Colorectal cancer (CRC) poses a significant public health burden in China, ranking as the second most frequently diagnosed malignancy and the fourth leading cause of cancer-related mortality.¹ Due to postoperative minimal residual disease and therapeutic resistance, approximately 50% of CRC patients develop distant metastasis,² with around 20% presenting with metastatic CRC (mCRC) at initial diagnosis.^3,4 Recently, substantial improvement in survival rates among mCRC patients has been attributed to the advent of targeted therapies and immunotherapy.^5,6 Bevacizumab (Bev) has become a cornerstone of first-line treatment when combined with chemotherapy (Bev/CT). This regimen has demonstrated durable clinical responses and improved disease control rate (DCR).^7,8 However, its effectiveness is often limited by intrinsic or acquired resistance.^9,10

Bev, a monoclonal antibody targeting vascular endothelial growth factor A (VEGFA), could suppress angiogenesis by blocking VEGF/VEGFR signaling.¹¹ Nevertheless, conventional biomarkers such as VEGF, CD31, and KRAS/BRAF mutations have proven limited utility in predicting clinical response to Bev therapy.^12–15 Although emerging biomarkers such as circulating tumor DNA and microbial DNA show promise, their widespread clinical application is hindered by high costs and technical challenges.^16,17 Therefore, there remains an urgent need for cost-effective, accessible, and reliable biomarkers to identify mCRC patients who are most likely to benefit from the first-line Bev/CT regimen.

Cancer-derived inflammation is a well-established hallmark of CRC, promoting tumor progression, metastasis, and drug resistance.¹⁸ Chronic inflammation processes reshape the tumor microenvironment and contribute to therapeutic resistance,¹⁹ thereby diminishing responsiveness to both CT and targeted therapy in CRC patients.²⁰ Our previous studies identified the fibrinogen-to-prealbumin ratio (FPR) outperforms other commonly inflammatory ratios such as neutrophil-to-lymphocyte ratio (NLR), lymphocyte-to-monocyte (LMR), and systemic inflammation response index, and systemic immune-inflammation index (SII) as an optimal and sensitive biomarker in reflecting chronic inflammation burden.^21,22 Circulating FPR has been shown to stratify subgroups benefiting from CT in CRC,^23–26 monitor recurrence and disease progression among NSCLC,²⁷ and serve as an independent prognostic factor across various malignancies, including CRC.²⁸ Despite this evidence, the role of inflammation in modulating Bev/CT efficacy in mCRC remains poorly characterized. Moreover, no study to date has systematically evaluated the roles of FPR in assessing Bev therapeutic efficacy, predicting clinical prognosis, monitoring disease dynamics, or guiding therapeutic selection in mCRC.

In this study, we enrolled a total of 648 eligible mCRC patients. Among them, 473 patients received first-line Bev/CT and were allocated to a discovery cohort (n = 249), an internal validation cohort (n = 115), and an external validation cohort (n = 109), and the remaining 175 patients received CT alone. The objectives of this study were threefold: (1) to investigate the association between baseline circulating FPR and clinical outcomes; (2) to evaluate the utility of dynamic FPR changes in monitoring disease progression and informing treatment decisions; and (3) to explore the relationship with cancer-derived inflammation and Bev resistance in mCRC.

Materials and methods

Population

The Ethics Committee of the Second Affiliated Hospital of Nanchang University approved this retrospective study (O-MRE[2025] No. 65). We screened and selected eligible participants based on the following inclusion criteria: (1) patients who were first diagnosed with mCRC through imaging, laboratory, and gastrointestinal endoscopy between January 2018 and December 2024 at our hospital; (2) patients aged over 18 years at the time of diagnosis; and (3) newly diagnosed patients who were willing to receive further treatment at our hospital. Patients meeting any of the following exclusion criteria were excluded from the study: (1) diagnosis of more than two primary malignancies; (2) those with bacterial or viral infection within 1 month prior to diagnosis; (3) use of anti-inflammatory drugs or receipt of CT before diagnosis; and (4) loss to follow-up for three consecutive months without reaching any study endpoint.

Clinical treatments and response evaluations

All enrolled patients were categorized into groups based on whether they received CT or Bev/CT as first-line treatment, following shared decisions-making between patients and clinicians. The treatment regimens included the following: (1) oxaliplatin-based CT (FOLFOX/XELOX, CT-OX), (2) irinotecan-based CT (FOLFIRI/ XELIRI, CT-IRI), (3) Bev combined with oxaliplatin-based CT (Bev/OX), and (4) Bev combined with irinotecan-based CT (Bev/IRI). Clinical response was assessed according to RECIST v1.1 criteria²⁹ within the 3-month period after the completion of at least two treatment cycles. The DCR was defined as the proportion of patients achieving complete response (CR), partial response (PR), or stable disease (SD).

Follow-up

The primary endpoints were treatment response and progression-free survival (PFS), with 1-year overall survival (OS) designated as the secondary endpoint. Follow-up assessments were conducted every 3 months during the first 2 years and every 6 months thereafter, and the follow-up deadline was June 2025. PFS was defined as the time from initial diagnosis to disease progression, death, or the last follow-up. OS was defined as the time from initial diagnosis to death or the last follow-up, whichever occurred first. Common imaging modalities such as computed tomography (CT), magnetic resonance imaging (MRI), and single-photon emission CT were selected to assess disease status according to clinical requirements for optimal visualization of specific metastatic sites. All imaging assessments were centrally reviewed. Progressive disease (PD) was primarily determined by radiological imaging confirmation of new or progressive metastatic lesions. Serum tumor biomarker monitoring was performed for each patient, and a twofold increase in serum carcinoembryonic antigen (CEA) or carbohydrate antigen 19-9 (CA19-9) levels from the nadir was used as a supplementary criterion for PD.³⁰ The study adheres to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guideline for cohort (Appendix S1).³¹

Laboratory detection

Laboratory parameter data were extracted from hospital records at the following clinically defined time points: pre-/post-treatment, at disease control (On-DCR), and at disease progression (On-PD). Fibrinogen (Fib) was measured using the Clauss assay (Sysmex CA-7000; Sysmex, Tokyo, Japan), and prealbumin (pAlb) was quantified by immunoturbidimetry (Olympus AU5400; Beckman Coulter, Tokyo, Japan). Serum CEA and CA19-9 levels were simultaneously analyzed by chemiluminescence immunoassay (Siemens ADVIA Centaur XP; Siemens, Erlangen, Germany). The FPR was calculated as Fib × 1000/pAlb. All inter- and intra-assay coefficients of variation were below 10%.

KRAS mutation and immunohistochemical detection of mismatch repair proteins

Formalin-fixed paraffin-embedded (FFPE) tumor samples were used to analyze KRAS mutations and mismatch repair protein expression (MLH1, MSH2, MSH6, and PMS2). Genomic DNA was extracted using the QIAamp DNA FFPE Kit (Qiagen; Hilden, Germany), and samples with OD260/280 ratios between 1.8 and 2.0 were included. KRAS mutations (G12C, G12S, G12R, G12V, G12D, G13A, and G13D) were detected by amplification-refractory mutation system PCR using the AmoyDx kit (Amoy Diagnostics, Xiamen, China). Immunohistochemistry was performed by a trained technician and evaluated independently by two pathologists. Specimens showing positive expression of all four proteins were classified as microsatellite stable (MSS), while loss of expression in any of the proteins was defined as microsatellite instability (MSI).³²

Statistical analysis

The optimal FPR cutoff for PFS was determined using X-tile software.³³ Group comparisons were performed using the Mann–Whitney, Kruskal–Wallis, Chi-square, or Fisher’s tests. Treatment efficacy was evaluated through logistic regression analysis, with results presented as odds ratio (OR) and 95% confidence intervals (CIs). PFS and OS were estimated using Kaplan–Meier curves (log-rank test) and Cox proportional hazards regression (hazard ratio (HR), 95% CI). Associations between continuous variables were assessed using Pearson correlation and restricted cubic spline (RCS) analysis. Propensity score matching was applied to balance the clinicopathological characteristics between the Bev/CT- and CT-treated groups. All statistical analyses were carried out using SPSS 27.0 software (IBM Corp, Armonk, NY, USA), GraphPad Prism 9.0 (GraphPad Software, San Diego, CA, USA), and MSTATA online (https://www.mstata.com/). Statistical significance was defined as p < 0.05.

Results

Clinical characteristics of the eligible patients

The detailed flowchart of eligible cases is presented in Figure 1. Based on the inclusion and exclusion criteria, a total of 648 patients were enrolled, including 175 patients who received first-line CT (CT group) and 473 patients who underwent first-line Bev/CT (Bev/CT group). Among patients in the Bev/CT group diagnosed before June 2023, participants were randomly assigned to the discovery (249 cases) and internal validation cohort (115 cases) in an approximately 2:1 ratio; those diagnosed on or after this date were allocated to the external validation cohort (109 cases). The median follow-up duration was 9 months (range, 1–36 months). Baseline characteristics are summarized in Tables S1 and S2. Among patients receiving Bev/CT, baseline characteristics differed across cohorts, with fewer palliative surgeries performed in the discovery cohort (p < 0.01). The majority of patients (n = 311) received Bev/OX, with the remainder receiving Bev/IRI (Table S1). Before propensity score matching, the Bev/CT group had significantly higher prevalence of hypertension (45.67% vs 13.71%, p < 0.001), diabetes (17.55% vs 6.29%, p < 0.001), right-sided mCRC (33.62% vs 24.00%, p = 0.02), and palliative surgery (69.56% vs 48.57%, p < 0.001). However, no significant differences were observed in the distribution of liver-limited or multisite metastases between the two groups. Following propensity score matching, all clinicopathologic characteristics were balanced (Table S2).

Figure 1.

The flowchart of screening and identification of eligible patients in the present study.

FPR and clinical characteristics

Based on PFS in the Bev/CT cohort, the optimal cutoff value for the FPR was determined as 24.1. Accordingly, we categorized the cohort into high-grade inflammation (FPR-H) and low-grade inflammation groups (FPR-L). Postoperative FPR levels were markedly lower in patients who underwent palliative surgery compared to those who did not receive surgical intervention (p < 0.001; Figure 2(a)). Similarly, patients with single-site mCRC exhibited lower FPR than those with multi-site metastases (p = 0.07; Figure 2(b)). Furthermore, a significant positive correlation was observed between FPR and primary tumor size in palliative surgical patients (Pearson’s r = 0.40; Figure 2(c)).

Figure 2.

Associations of FPR with clinical characteristics and clinical outcomes in Bev/CT-treated patients with mCRC. (a) FPR comparison between pre- and post-palliative surgical patients. (b) FPR comparison among the patients with single- and multiple-metastatic sites. (c) Association of FPR with tumor size. (d, e) Association of FPR with DCR, PFS in the discovery cohort. (f, g) Association of FPR with DCR, PFS in internal validation cohort. (h, i) Association of FPR with DCR, PFS in external validation cohort. (j, k) Restricted cubic spline between FPR with DCR, PFS in the overall population. (l, m) Association of FPR with DCR, PFS in the overall population with MSS mCRC.

FPR and clinical outcomes

After two cycles of first-line Bev/CT, only one patient achieved a CR. The rates of SD and PD were 76.53% and 19.24%, respectively, yielding an ORR of 4.23% (Table S1). Patients with FPR-H exhibited a low DCR compared to those classified as FPR-L across cohorts (Discovery: 72.94% vs 82.32%, p = 0.08; Internal validation: 57.69% vs 85.39%, p < 0.01; External validation: 78.79% vs 89.47%, p = 0.14; Figure 2(d), (f), and (h)). Multivariable logistic regression analyses indicated that FPR-H (adjusted OR = 2.18, 95% CI = 1.32–3.61, p < 0.01) and multi-site metastases (adjusted OR = 2.07, 95% CI = 1.13–3.79, p = 0.02) were significantly associated with a notably higher PD in the overall Bev/CT population (Table 1). Furthermore, multivariate RCS analysis revealed a nonlinear association between FPR and clinical response in the overall population (p-nonlinear = 0.02, Figure 2(j)).

Table 1.

Association between the baseline characteristics and clinical response in the overall Bev/CT-treated population.

Characteristics	DCR (N = 382)	PD (N = 91)	Univariate		Multivariate
Characteristics	DCR (N = 382)	PD (N = 91)	OR (95% CI)	p-Value	OR (95% CI)	p-Value
Sex (male)	234 (61.26%)	65 (71.43%)	0.63 (0.38–1.04)	0.07	0.58 (0.34–0.10)	0.05
Age (>65 years)	109 (28.53%)	25 (27.47%)	1.05 (0.63–1.76)	0.84	1.08 (0.61–1.90)	0.80
Smoking (yes)	41 (10.73%)	11 (12.09%)	1.14 (0.56–2.32)	0.71	0.94 (0.38–2.34)	0.89
Drinking (yes)	28 (7.33%)	9 (9.89%)	1.39 (0.63–3.05)	0.42	1.43 (0.50–4.06)	0.50
Hypertension (yes)	174 (45.55%)	42 (46.15%)	1.02 (0.65–1.62)	0.92	1.01 (0.59–1.73)	0.96
Diabetes (yes)	65 (17.02%)	18 (19.78%)	1.20 (0.67–2.15)	0.53	1.17 (0.63–2.16)	0.62
Tumor site (right)	130 (34.03%)	29 (31.87%)	1.10 (0.68–1.80)	0.70	1.09 (0.65–1.84)	0.74
Palliative surgery (yes)	273 (71.47%)	56 (61.54%)	0.64 (0.40–1.03)	0.07	0.79 (0.48–1.33)	0.38
Liver-limited metastasis (yes)	144 (37.70%)	25 (27.47%)	0.63 (0.38–1.04)	0.07	0.95 (0.49–1.83)	0.87
Multisite metastasis (⩾2)	125 (32.72%)	46 (50.55%)	2.10 (1.32–3.34)	<0.01	2.07 (1.13–3.79)	0.02
Treatment regimen (Bev/OX)	251 (65.71%)	60 (65.93%)	0.99 (0.61–1.60)	0.97	1.10 (0.66–1.84)	0.72
FPR (>24.1)	103 (26.96%)	41 (45.05%)	2.22 (1.39–3.56)	<0.01*	2.18 (1.32–3.61)	<0.01

Tumors site at the caecum, ascending colon, and transverse colon were defined as right-sided colorectal cancer, and those located within the splenic flexure and beyond were defined as left-sided colorectal cancer. Multivariate analysis included the following parameters: sex, age, smoking, drinking, hypertension, diabetes, tumor site, palliative surgery, liver-limited metastasis, multisite metastasis, and treatment regimens.

p < 0.001, for logistic regression analysis.

Bev/CT, bevacizumab combined chemotherapy; Bev/OX, Bev combined oxaliplatin-based CT; CI, confidence interval; DCR, disease control rate; FPR, fibrinogen-to-prealbumin ratio; OR, odds ratio; PD, progressive disease.

In the discovery cohort, the FPR-H patient had significantly worse PFS than the FPR-L cases (p_log-rank < 0.01; Figure 2(e)). After multivariable adjustment, elevated FPR remained independently associated with poorer PFS (adjusted HR = 1.51, 95% CI = 1.11–2.05, p < 0.01). This finding was consistently validated in both the internal (p_log-rank < 0.01, adjusted HR = 2.47, 95% CI = 1.43–4.27; Figure 2(g)) and external (p_log-rank < 0.001, adjusted HR = 2.28, 95% CI = 1.43–3.64; Figure 2(i)) validation cohorts (Table 2). Similarly, FPR-H patients had significantly inferior 1-year OS compared to FPR-L patients in the overall population (p_log-rank < 0.001; adjusted HR = 2.86, 95% CI = 1.78–4.61; Figure S2, Table S3). Moreover, multivariate RCS analysis indicated an inverted L-shaped dose–response relationship between FPR and PFS among first-line Bev/CT-treated mCRC patients (Figure 2(k)).

Table 2.

Association between the baseline characteristics and progression-free survival in the three cohorts.

Characteristics	Discovery cohort			Internal validation cohort			External validation cohort
	p-Value	Univariate	Multivariate	p-Value	Univariate	Multivariate	p-Value	Univariate	Multivariate
	p-Value	HR (95% CI)	HR (95% CI)	p-Value	HR (95% CI)	HR (95% CI)	p-Value	HR (95% CI)	HR (95% CI)
Sex (male)	0.30	0.86 (0.65–1.14)	0.81 (0.59–1.10)	0.24	0.78 (0.52–1.18)	0.78 (0.49–1.24)	0.12	0.70 (0.45–1.09)	0.80 (0.49–1.29)
Age (>65 years)	0.70	0.94 (0.69–1.28)	0.10 (0.70–1.41)	0.69	1.09 (0.72–1.66)	1.43 (0.87–2.35)	0.51	0.86 (0.55–1.35)	0.57 (0.31–1.04)
Smoking (yes)	0.52	1.16 (0.73–1.84)	1.30 (0.68–2.46)	0.97	0.99 (0.53–1.86)	0.37 (0.08–1.65)	0.27	0.70 (0.37–1.31)	0.73 (0.33–1.58)
Drinking (yes)	0.60	1.15 (0.68–1.94)	0.93 (0.45–1.93)	0.05	1.99 (0.99–4.00)	4.04 (0.75–21.83)	0.67	1.17 (0.56–2.42)	1.93 (0.81–4.61)
Hypertension (yes)	0.39	1.13 (0.86–1.48)	1.03 (0.75–1.41)	0.33	1.22 (0.82–1.81)	1.45 (0.88–2.38)	0.78	0.94 (0.62–1.42)	1.11 (0.66–1.84)
Diabetes (yes)	0.12	1.32 (0.93–1.87)	1.33 (0.93–1.92)	0.49	1.20 (0.71–2.04)	1.58 (0.88–2.86)	0.92	1.03 (0.62–1.70)	1.07 (0.57–2.01)
Tumor site (right)	0.75	1.05 (0.79–1.39)	1.09 (0.81–1.47)	0.81	1.05 (0.69–1.62)	0.92 (0.58–1.45)	0.69	0.91 (0.58–1.43)	0.92 (0.52–1.62)
Palliative surgery (yes)	<0.01	0.63 (0.48–0.83)	0.72 (0.53–0.98)	0.44	0.83 (0.52–1.33)	1.05 (0.63–1.73)	0.01	0.55 (0.34–0.89)	0.57 (0.33–0.97)
Liver-limited metastasis (yes)	0.04	0.75 (0.56–0.99)	0.93 (0.65–1.35)	0.32	0.81 (0.54–1.22)	1.16 (0.67–2.01)	0.52	1.15 (0.75–1.78)	1.62 (0.92–2.85)
Multisite metastasis (⩾2)	<0.01	1.75 (1.32–2.31)	1.66 (1.15–2.41)	0.02	1.66 (1.10–2.49)	1.71 (0.99–2.96)	0.17	1.34 (0.88–2.05)	1.57 (0.91–2.72)
Treatment regimes (Bev/OX)	0.16	1.23 (0.92–1.65)	1.39 (1.02–1.91)	0.37	1.20 (0.80–1.80)	1.34 (0.86–2.10)	0.33	1.23 (0.81–1.87)	1.05 (0.67–1.64)
FPR (>24.1)	<0.01	1.50 (1.13–1.99)	1.51 (1.11–2.05)	0.01	1.86 (1.17–2.95)	2.47 (1.43–4.27)	<0.01*	2.15 (1.40–3.32)	2.28 (1.43–3.64)

Tumor sites at the caecum, ascending colon, and transverse colon were defined as right-sided colorectal cancer, and those located within the splenic flexure and beyond were defined as left-sided colorectal cancer. Multivariate analysis included the following parameters: sex, age, smoking, drinking, hypertension, diabetes, tumor site, palliative surgery, liver-limited metastasis, multisite metastasis, and treatment regimens.

p < 0.001, for log-rank test within Kaplan–Meier curve.

Bev/CT, bevacizumab combined chemotherapy; Bev/OX, Bev combined oxaliplatin-based CT; CI, confidence interval; FPR, fibrinogen-to-prealbumin ratio; HR, hazard ratio.

Exploratory analysis: FPR, KRAS mutation, MSI, and clinical outcomes

Given limited tumor tissue availability, KRAS and MSI status were analyzed in 101 and 188 Bev/CT-treated patients, respectively. Among these, 22 patients were identified as KRAS wild-type, 79 patients harbored a KRAS mutated genotype, and 182 patients exhibited MSS. Owing to the small sample size, no significant difference in FPR was observed between KRAS wild-type and KRAS-mutated patients (19.76 vs 14.28, p = 0.20; Figure S1(A)). Notably, FPR-H patients demonstrated consistently lower DCR compared to FPR-L counterparts across KRAS subtypes (wild-type: 71.43% vs 86.67%, p = 0.56; mutated: 72.73% vs 85.96%, p = 0.20; Figure S1(B) and (C)). Kaplan–Meier analysis showed that FPR-H was marginally associated with unsatisfactory PFS in both KRAS wild-type (mPFS: 4 vs 10 months, p_log-rank = 0.04; Figure S1(D)) and mutated subgroups (mPFS: 5 vs 10 months, p_log-rank < 0.01; Figure S1(E)). MSS patients with FPR-H showed a significantly reduced DCR (65.71% vs 85.71%, p = 0.01; Figure 2(l)) and shorter PFS (mPFS: 5 vs 12 months, p_log-rank < 0.001; Figure 2(m)) compared to FPR-L cases.

Dynamic monitoring of FPR and progression of Bev/CT-treated mCRC

In this study, we observed that pre-treatment FPR levels in patients with PD were significantly higher than those of patients with DCR (median: 22.20 vs 16.47, p < 0.01; Figure 3(a)). Longitudinal analysis showed a marked reduction in FPR from pre-treatment to On-DCR time point (median: 17.05 vs 13.99, p < 0.001); conversely, FPR showed a significant increase at On-PD compared to baseline levels (median: 21.10, p = 0.02; Figure 3(b)).

Figure 3.

Dynamic changes in FPR for monitoring disease progression in mCRC patients treated with first-line Bev/CT. (a) Pre-treatment FPR comparison in DCR and PD patients. (b) Longitudinal FPR changes across pre-treatment, on-DCR, on-PD time-points. (c) Median time to progression comparison monitored by FPR and imaging detection. (d, e) Temporal FPR dynamics in discovery cohort. (f, g) Temporal FPR dynamics in internal validation cohort. (h, i) Temporal FPR dynamics in external validation cohort. (j–l) Kaplan–Meier analysis of FPR dynamic change in the discovery (j), internal validation (k), and external validation (l) cohorts.

In the discovery cohort, most patients (60.5%) maintained continuous FPR-L, of whom 84% achieved DCR. Among those transitioning from pre-treatment FPR-H to post-treatment FPR-L, 87% exhibited SD or PR (Figure 3(d)). Conversely, 45% of PD patients had baseline FPR-H, and 52% maintained persistently high FPR (Figure 3(d)). Longitudinal analysis showed that proportions of FPR-H were 35%, 9%, and 37% at pretreatment, On-DCR, and On-PD time-points, respectively (Figure 3(e)). Furthermore, post-treatment FPR-L was associated with significantly longer PFS compared to FPR-H (mPFS: 9 vs 4 months, p_log-rank < 0.01; Figure 3(j)). In the internal validation cohort, 24 patients showed PD as a clinical response to Bev/CT therapy, with 46% presenting baseline FPR-H, and 8 patients displayed persistently elevated FPR at the On-PD time point (Figure 3(f)). Of patients transitioning from FPR-H to FPR-L, 81% achieved DCR (Figure 3(f)). Sixty patients included in longitudinal analysis showed 80% of them with pre-treatment FPR-L, 92% and 18% of them harbored FPR-L and FPR-H at On-DCR and On-PD time-point, respectively (Figure 3(g)). Persistent FPR-H was associated with poor PFS (mPFS: 3 vs 12 months, p_log-rank < 0.001), while continuous FPR-L correlated with optimal survival (Figure 3(k)).

These findings were also observed in the external validation cohort, where 74 patients showed DCR as a clinical response to Bev/CT therapy, with 78% presenting baseline FPR-L, and 56 patients maintaining persistent FPR-L throughout treatment (Figure 3(h)). Ten patients transitioned from FPR-H to FPR-L, with 90% achieving SD (Figure 3(h)). FPR-L prevalence shifted characteristically with disease status changes (pre-treatment: 71%; On-DCR: 88%; On-PD: 61%; Figure 3(i)). Persistent FPR-H predicted markedly worse PFS versus post-treatment FPR-L (mPFS: 5 vs 9 months, p_log-rank = 0.03), while continuous FPR-L was associated with the best outcome (12 months; Figure 3(l)). Moreover, the median positive time of FPR-H appearance was significantly earlier than CT/MRI detection (2 vs 3 months, p < 0.001; Figure 3(c)).

FPR and therapeutic decision

All PFS comparisons between first-line CT and Bev/CT were evaluated using data after 1:1 propensity score matching. In the CT cohort, median PFS was significantly longer in FPR-L than in FPR-H patients (9 vs 5 months, p_log-rank = 0.02; Figure 4(a)). The Bev/CT cohort showed significantly improved overall PFS versus CT (Figure S3(A)), particularly in the FPR-L subgroup (p_log-rank < 0.01; Figure 4(a)). However, no significant PFS benefit was observed with Bev/CT in FPR-H patients compared to CT alone (p_log-rank = 0.46; Figure 4(a)). In FPR-L patients, first-line Bev/OX-treated cases harbored superior survival compared to those with treatment of CT-OX (p_log-rank = 0.01; Figure 4(b)). No significant difference in PFS was observed between IRI-treated regimens with or without Bev in FPR-L cases, or between CT-OX and CT-IRI in FPR-H patients (Figure 4(b) and Figure S3(B)).

Figure 4.

Comprehensive analysis of Bev-based therapy outcomes stratified by FPR in overall populations. (a) PFS comparison between Bev/CT- and CT-treated mCRC patients stratified by FPR following propensity score matching. (b) PFS comparison between FPR-L mCRC patients treated with OX or IRI with or without Bev following propensity score matching, (c, d) Clinical response to Bev/CT regimen and PFS in right- and left-sided mCRC patients stratified by FPR. (e, f) Clinical response to Bev/CT regimen and PFS in Bev/OX- and Bev/IRI-treated mCRC patients stratified by FPR. (g, h) Clinical response to Bev/CT regimen and PFS in non-palliation surgical- and surgical-treated mCRC patients stratified by FPR. (i–l) PFS comparison between FPR-L and FPR-H mCRC patients treated with OX or IRI with or without Bev following propensity score matching (i, j: FPR-L patients with or without palliative surgery; k, l: FPR-H patients with or without palliative surgery).

We subsequently assessed the treatment benefits of Bev/CT within the original, non-matched Bev/CT population and found that among left-sided mCRC patients, those with FPR-H showed significantly lower DCR than those with FPR-L (67.78% vs 85.27; p < 0.001; Figure 4(c)). However, these differences did not reach statistical significance in right-sided mCRC patients. PFS was consistently better in FPR-L patients regardless of tumor location, while no significant PFS difference was observed between left- and right-sided tumors within the same FPR subgroup (Figure 4(d)). No significant difference in DCR was found between Bev/OX and Bev/IRI treatment in either FPR-L (85.64% vs 83.46%) or FPR-H (71.56% vs 71.43%) subgroups. However, both regimens achieved higher DCR in FPR-L patients compared to FPR-H (Figure 4(e)). Bev/OX-treated FPR-L patients exhibited superior PFS compared to all other subgroups, while Bev/OX- and Bev/IRI-treated FPR-H patients showed similar outcomes (Figure 4(f)). Patients undergoing palliative resection with FPR-L status exhibited significantly higher DCR compared to those with FPR-H (87.15% vs 70.51%; p < 0.01) and non-surgical patients (87.15% vs 77.50%; p = 0.048; Figure 4(g)). Surgical FPR-L patients also demonstrated the best PFS (mPFS: 11 months), followed by non-surgical FPR-L patients (8 months; p_log-rank < 0.01), while non-surgical FPR-H patients had the poorest outcomes (7 months; p_log-rank < 0.001; Figure 4(h)).

In the propensity score-matched analysis, surgical FPR-L patients receiving Bev/OX showed significantly longer PFS compared to CT-OX or Bev/IRI (13 vs 8.5 vs 10 months, respectively; p_log-rank < 0.01; Figure 4(i)). However, this survival advantage was not observed in non-surgical FPR-L patients (Figure 4(j)) or in FPR-H patients regardless of surgical status (Figure 4(k) and (l)). No survival difference was detected between Bev/IRI and CT-IRI in any subgroup (Figure 4(i)–(l)).

Discussion

Cancer-derived inflammation has been shown to induce resistance to CT and tyrosine kinase inhibitors.^27,34 However, its role in Bev-based therapy for mCRC patients remains unclear. In this study, we demonstrated that cancer-derived high-grade inflammation is associated with both intrinsic and acquired resistance to Bev. The FPR reliably reflected tumor burden and emerged as a practical biomarker for evaluating treatment response, predicting prognosis, monitoring disease progression, and guiding therapeutic decision-making in Bev/CT treatment.

FPR, which integrates information on inflammation, nutrition, and coagulation,³⁵ was significantly correlated with primary tumor size, metastatic burden, and surgical history in this study, which are consistent with our previous findings in early-stage CRC.²⁵ Although palliative surgery is associated with 1-year OS, its impact on long-term OS remains undefined. Importantly, it showed no consistent association with PFS or treatment response across any cohort. This indicates that palliative surgery does not confound the primary findings of our study regarding the predictive value of the FPR biomarker for initial treatment efficacy. Notably, elevated FPR was consistently associated with approximately a twofold increased risk of disease progression across all cohorts and subgroups, including exploratory analyses stratified by KRAS status and MSS, supporting its robustness as a prognostic biomarker. Furthermore, high FPR remains an independent risk factor for 1-year mortality in the overall population.

Several systemic inflammatory biomarkers, such as NLR, platelet-to-lymphocyte ratio, LMR, and modified Glasgow prognostic score, have emerged as informative markers for predicting treatment efficacy and prognosis in mCRC.^36–40 Encouragingly, we also found that FPR could effectively monitor disease progression during Bev/CT treatment. Moreover, the median positive time of FPR elevation was significantly earlier than that of traditional imaging-based detection methods, suggesting its potential utility for early intervention. Furthermore, among surgical mCRC patients within the FPR-L subgroup, those receiving Bev/OX treatment exhibited superior DCR and PFS compared to those treated with Bev/IRI. By contrast, propensity score matching analysis revealed that patients with FPR-H derived limited benefit from Bev regardless of the CT backbone, suggesting that CT-OX or CT-IRI may represent a more appropriate therapeutic strategy for this subgroup.

Although Bev/CT is the standard first-line therapy for mCRC, approximately 10%–20% of patients exhibit intrinsic resistance.^41–43 We confirmed that elevated FPR levels are associated with innate resistance to Bev, effectively negating the survival benefit of adding Bev to CT. Furthermore, increasing FPR levels were correlated with reduced treatment response, and nearly one-third of patients experiencing disease progression showed elevated FPR at the time of progression, implicating cancer-derived inflammation in both intrinsic and acquired resistance mechanisms.

To the best of our knowledge, this is the first study to establish the FPR as a comprehensive biomarker of inflammation-associated Bev resistance in mCRC. However, several limitations should be addressed. First, as a single-center study with a limited sample size, the statistical power was insufficient to perform robust subgroup analyses, particularly those stratified by MSI and KRAS status. Second, given the current immaturity of the 3-year and 5-year OS data, we consequently selected 1-year OS as a key secondary endpoint. The results consistently demonstrated that elevated FPR is an independent risk factor for disease progression. Active follow-up of patients is ongoing, and the long-term survival outcomes will be reported in a future publication. Third, although an FPR cutoff value of 24.1 was identified and internally validated across several Bev-based and CT-only cohorts, its generalizability requires confirmation in external populations. Finally, while our previous work suggested the superiority of FPR over other inflammatory indices (NLR, LMR, and SII), the current study did not include direct comparative analyses among these biomarkers. Therefore, to fully address these limitations and validate our conclusions, a large-scale, multicenter, prospective randomized controlled trial is warranted.

Conclusion

Cancer-derived inflammation attenuates the efficacy of Bev in mCRC. FPR serves as a noninvasive and clinically applicable biomarker that can evaluate Bev response, predict prognosis, monitor disease progression, and guide therapeutic decisions in mCRC patients receiving first-line Bev/CT.

Supplemental Material

sj-doc-2-tam-10.1177_17588359251411086 – Supplemental material for Cancer-derived inflammation is associated with bevacizumab resistance and inferior clinical outcomes in patients with metastatic colorectal cancer

Supplemental material, sj-doc-2-tam-10.1177_17588359251411086 for Cancer-derived inflammation is associated with bevacizumab resistance and inferior clinical outcomes in patients with metastatic colorectal cancer by Ya-Nan Li, Yang Wu, Feng-Wen Deng, Ying Lu, Guo-Bin Song, Lin Xiang, Tian Peng, Xue-Xin Cheng and Hou-Qun Ying in Therapeutic Advances in Medical Oncology

Supplemental Material

sj-docx-1-tam-10.1177_17588359251411086 – Supplemental material for Cancer-derived inflammation is associated with bevacizumab resistance and inferior clinical outcomes in patients with metastatic colorectal cancer

Supplemental material, sj-docx-1-tam-10.1177_17588359251411086 for Cancer-derived inflammation is associated with bevacizumab resistance and inferior clinical outcomes in patients with metastatic colorectal cancer by Ya-Nan Li, Yang Wu, Feng-Wen Deng, Ying Lu, Guo-Bin Song, Lin Xiang, Tian Peng, Xue-Xin Cheng and Hou-Qun Ying in Therapeutic Advances in Medical Oncology

Footnotes

Acknowledgements

We extend our sincere gratitude to Dr Ming-Feng Fu for his support during this study. We also deeply appreciate all the patients who participated in this work.

Declarations

ORCID iDs

Ya-Nan Li

Yang Wu

Feng-Wen Deng

Ying Lu

Guo-Bin Song

Lin Xiang

Tian Peng

Xue-Xin Cheng

Hou-Qun Ying

Supplemental material

Supplemental material for this article is available online.

References

Han

Zheng

Zeng

, et al. Cancer incidence and mortality in China, 2022. J Natl Cancer Cent 2024; 4(1): 47–53.

Hernandez-Dominguez

Yilmaz

Steele

SR.

Stage IV colorectal cancer management and treatment. J Clin Med 2023; 12(5): 2072–2098.

Ciardiello

Martini

, et al. Clinical management of metastatic colorectal cancer in the era of precision medicine. CA Cancer J Clin 2022; 72(4): 372–401.

Fernandez-Montes

Alonso

Aranda

, et al. SEOM-GEMCAD-TTD clinical guidelines for the systemic treatment of metastatic colorectal cancer (2022). Clin Transl Oncol 2023; 25(9): 2718–2731.

Ohishi

Kaneko

Yoshida

, et al. Current targeted therapy for metastatic colorectal cancer. Int J Mol Sci 2023; 24(2): 1702–1721.

Weng

Zhu

, et al. Exploring immunotherapy in colorectal cancer. J Hematol Oncol 2022; 15(1): 95–123.

Wang

Fei

Vanderlaan

, et al. Biological activity of bevacizumab, a humanized anti-VEGF antibody in vitro. Angiogenesis 2004; 7(4): 335–345.

Marques

Duarte

Sterrantino

, et al. Triplet (FOLFOXIRI) versus doublet (FOLFOX or FOLFIRI) backbone chemotherapy as first-line treatment of metastatic colorectal cancer: a systematic review and meta-analysis. Crit Rev Oncol Hematol 2017; 118: 54–62.

Bergers

Hanahan

Modes of resistance to anti-angiogenic therapy. Nat Rev Cancer 2008; 8(8): 592–603.

10.

Itatani

Kawada

Yamamoto

, et al. Resistance to anti-angiogenic therapy in cancer-alterations to anti-VEGF pathway. Int J Mol Sci 2018; 19(4): 1232–1250.

11.

Hurwitz

Integrating the anti-VEGF-A humanized monoclonal antibody bevacizumab with chemotherapy in advanced colorectal cancer. Clin Colorectal Cancer 2004; 4(Suppl. 2): S62–S68.

12.

Luo

RH.

Predictive and prognostic biomarkers with therapeutic targets in advanced colorectal cancer. World J Gastroenterol 2014; 20(14): 3858–3874.

13.

Petrelli

Coinu

Cabiddu

, et al. Prognostic factors for survival with bevacizumab-based therapy in colorectal cancer patients: a systematic review and pooled analysis of 11,585 patients. Med Oncol 2015; 32(2): 456–466.

14.

Kopetz

Hoff

Morris

, et al. Phase II trial of infusional fluorouracil, irinotecan, and bevacizumab for metastatic colorectal cancer: efficacy and circulating angiogenic biomarkers associated with therapeutic resistance. J Clin Oncol 2010; 28(3): 453–459.

15.

Seufferlein

Lausser

Stein

, et al. Prediction of resistance to bevacizumab plus FOLFOX in metastatic colorectal cancer—results of the prospective multicenter PERMAD trial. PLoS One 2024; 19(6): e0304324.

16.

Yang

Wang

Jin

, et al. Circulating tumor DNA detectable in early- and late-stage colorectal cancer patients. Biosci Rep 2018; 38(4): BSR20180322.

17.

Messaritakis

Koulouris

Boukla

, et al. Exploring gut microbiome composition and circulating microbial DNA fragments in patients with stage II/III colorectal cancer: a comprehensive analysis. Cancers 2024; 16(10): 1923–1941.

18.

Formica

Luccchetti

Cunningham

, et al. Systemic inflammation, as measured by the neutrophil/lymphocyte ratio, may have differential prognostic impact before and during treatment with fluorouracil, irinotecan and bevacizumab in metastatic colorectal cancer patients. Med Oncol 2014; 31(9): 166–174.

19.

Zhang

Davis

Shah

, et al. IL-33 promotes growth and liver metastasis of colorectal cancer in mice by remodeling the tumor microenvironment and inducing angiogenesis. Mol Carcinog 2017; 56(1): 272–287.

20.

Rossi

Basso

Strippoli

, et al. Are markers of systemic inflammation good prognostic indicators in colorectal cancer? Clin Colorectal Cancer 2017; 16(4): 264–274.

21.

Ying

Liao

Sun

, et al. The role of cancer-elicited inflammatory biomarkers in predicting early recurrence within stage II–III colorectal cancer patients after curable resection. J Inflamm Res 2021; 14: 115–129.

22.

Sun

Tan

Gao

, et al. Circulating fibrinogen to pre-albumin ratio is a promising biomarker for diagnosis of colorectal cancer. J Clin Lab Anal 2019; 33(1): e22635.

23.

Chen

Zhang

Sun

, et al. Elevated FPR confers to radiochemoresistance and predicts clinical efficacy and outcome of metastatic colorectal cancer patients. Aging 2019; 11(6): 1716–1732.

24.

Zhang

Wang

, et al. Potential prognostic factors for predicting the chemotherapeutic outcomes and prognosis of patients with metastatic colorectal cancer. J Clin Lab Anal 2019; 33(8): e22958.

25.

Ying

Sun

Liao

, et al. The value of circulating fibrinogen-to-pre-albumin ratio in predicting survival and benefit from chemotherapy in colorectal cancer. Therap Adv Med Oncol 2021; 13: 17588359211022886.

26.

Hailun

Huang

Yuan

, et al. Prognostic significance of preoperative fibrinogen-to-prealbumin ratio in patients with stage I–III colorectal cancer undergoing surgical resection: a retrospective cohort study. BioMed Res Int 2021; 2021: 3905353.

27.

Ying

Liao

Luo

, et al. Cancer-elicited inflammation attenuates response and outcome in tyrosine kinase inhibitor naive patients with advanced NSCLC. Pharmacol Res 2021; 170: 105734–105744.

28.

You

Sun

, et al. Albumin to fibrinogen ratio and fibrinogen to pre-albumin ratio are economical, simple and promising prognostic factors for solid malignancy. J Thorac Dis 2019; 11(Suppl. 15): S2036–S2038.

29.

Therasse

Arbuck

Eisenhauer

, et al. New guidelines to evaluate the response to treatment in solid tumors. European Organization for Research and Treatment of Cancer, National Cancer Institute of the United States, National Cancer Institute of Canada. J Natl Cancer Inst 2000; 92(3): 205–216.

30.

Manojlovic

Savic

Nikolic

, et al. Dynamic monitoring of carcinoembryonic antigen, CA19-9 and inflammation-based indices in patients with advanced colorectal cancer undergoing chemotherapy. World J Clin Cases 2022; 10(3): 899–918.

31.

Von Elm

Altman

Egger

, et al. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. J Clin Epidemiol 2008; 61(4): 344–349.

32.

Niu

Wang

Feng

, et al. Correlation between microsatellite instability and RAS gene mutation and stage III colorectal cancer. Oncol Lett 2019; 17(1): 332–338.

33.

Camp

Dolled-Filhart

Rimm

DL.

X-tile: a new bio-informatics tool for biomarker assessment and outcome-based cut-point optimization. Clin Cancer Res 2004; 10(21): 7252–7259.

34.

Brockmueller

Samuel

Mazurakova

, et al. Curcumin, calebin a and chemosensitization: how are they linked to colorectal cancer? Life Sci 2023; 318: 121504.

35.

Banaste

Rousset

Mercier

, et al. Preoperative nutritional risk assessment in patients undergoing cytoreductive surgery plus hyperthermic intraperitoneal chemotherapy for colorectal carcinomatosis. Int J Hyperthermia 2018; 34(5): 589–594.

36.

Kuramochi

Yamada

Yoshida

, et al. The pre-treatment lymphocyte-to-monocyte ratio predicts efficacy in metastatic colorectal cancer treated with TAS-102 and bevacizumab. Anticancer Res 2021; 41(6): 3131–3137.

37.

Boukovala

Modest

Ricard

, et al. Evaluation of the inflammation-based modified Glasgow prognostic score (mGPS) as a prognostic and predictive biomarker in patients with metastatic colorectal cancer receiving first-line chemotherapy: a post hoc analysis of the randomized phase III XELAVIRI trial (AIO KRK0110). ESMO Open 2024; 9(5): 103374.

38.

Roxburgh

McMillan

DC.

Role of systemic inflammatory response in predicting survival in patients with primary operable cancer. Future Oncol 2010; 6(1): 149–163.

39.

Okugawa

Toiyama

Yamamoto

, et al. Lymphocyte-C-reactive protein ratio as promising new marker for predicting surgical and oncological outcomes in colorectal cancer. Ann Surg 2020; 272(2): 342–351.

40.

Muneoka

Shirai

Sasaki

, et al. Change of blood neutrophil-to-lymphocyte ratio predicts survival in patients with metastatic colorectal cancer. Gan To Kagaku Ryoho 2017; 44(4): 313–317.

41.

Bai

Zhang

, et al. Clinical outcomes of Chinese patients with metastatic colorectal cancer receiving first-line bevacizumab-containing treatment. Med Oncol 2015; 32(2): 469–479.

42.

Hurwitz

Fehrenbacher

Novotny

, et al. Bevacizumab plus irinotecan, fluorouracil, and leucovorin for metastatic colorectal cancer. N Engl J Med 2004; 350(23): 2335–2342.

43.

Arnold

Eggers

Uhlig

, et al. Treatment of metastatic colorectal carcinoma with bevacizumab in first-line and beyond first progression: the KORALLE non-interventional cohort study. Oncol Res Treat 2022; 45(10): 576–587.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.64 MB

0.03 MB