Skip to main content
Download PDF
Download ePub

Safety, efficacy, and cost-effectiveness evaluation of systemic treatments for refractory colorectal cancer: a systematic review and modeling study

Health Economics Review volume 15, Article number: 33 (2025) Cite this article

Abstract

Objectives

To conduct pooled estimates and comparative evaluations of safety and efficacy, alongside cost-effectiveness and value-based pricing analyses, for systemic treatments recommended by the National Comprehensive Cancer Network in refractory colorectal cancer.

Methods

A comprehensive search for related randomized controlled trials was conducted on PubMed, EMBASE, the Cochrane Library, and ClinicalTrials.gov. Safety was evaluated by aggregating treatment-related adverse events (TRAEs) and performing Bayesian network meta-analysis (NMA) for indirect comparisons. Pooled survival estimates of overall survival (OS) and progression-free survival (PFS) were conducted to assess treatment efficacy. For NMA of OS and PFS, time-variant fractional polynomial models were employed as the primary analysis, with Cox proportional hazards models used for result validation. Economic evaluations were performed using partitioned survival models from the US public sector perspective. Clinical parameters were sourced from meta-analyses; cost parameters included drug treatment, follow-up and administration, end-of-life care, and adverse event management expenses, which were obtained from the Federal Supply Schedule, public databases or published literature. Utility values were sourced from the CORRECT trial. Price simulations were also conducted. Robustness of results was confirmed by sensitivity and scenario analyses

Results

We included nine studies comprising 3,978 patients and incorporating six treatments recommended by NCCN, including best supportive care (BSC), regorafenib, regorafenib dose optimization (REDo), trifluridine/tipiracil (TAS-102), TAS-102 with bevacizumab (TAS-BEV), and fruquintinib. Targeted treatments increased serious TRAEs and grade 3 + TRAEs compared to BSC. However, no significant safety differences were found among the targeted therapies. Regarding efficacy, REDo led in median OS, while fruquintinib led in median PFS. NMA indicated that TAS-BEV had the greatest PFS and OS survival benefit, followed by fruquintinib and REDo. Cost-effectiveness analysis favored BSC as the least expensive and the most cost-effective profile. TAS-BEV had the greatest effectiveness, with TAS-102 being the most cost-effective among targeted therapies. For cost-effectiveness against BSC, the price reductions of TAS-102, fruquintinib, REDoS, regorafenib, and TAS-BEV were 39%, 24%, 14%, 8%, and 7%, respectively.

Conclusions

Targeted therapies have comparable safety; TAS-BEV is highly effective, TAS-102 is the top cost-effective targeted therapy. Treatment choice should balance individual patient needs with safety, efficacy, and cost.

Key points

Our pioneering meta-analysis introduces reconstructed IPD to derive robust survival metrics for refractory CRC, setting a new standard by accounting for event censoring. Consistent OS and PFS estimates from high-quality studies provide precise treatment insights and aid in designing future trial sample sizes.

As the first to apply NMA with reconstructed IPD for refractory CRC, we overcome the limitations of proportional hazards assumptions. Our dynamic model delivers timely efficacy assessments of various treatments, bolstering the clinical relevance and trustworthiness of our results.

Our analysis is the first to compare the effectiveness and cost-efficiency of all NCCN-recommended regimens for refractory CRC from a US public payer perspective. With price simulations for value-based pricing, we offer crucial evidence for optimizing clinical drug use and resource allocation, facilitating informed healthcare decisions.

Introduction

Globally, cancer is a primary cause of mortality and represents a significant obstacle to enhancing life expectancy across all nations [1]. In the United States, among all cancers, colorectal cancer (CRC) is the second leading cause of cancer-related mortality, and is the most common cause of death for men under the age of 50 [2]. The impact of cancer on society is substantial and transcends economic boundaries, affecting both more and less economically developed nations [1].

Advanced CRC treatment is customized based on individual disease features and treatment history [3]. First-line CRC treatment commonly consists of chemotherapy like FOLFOX (folinic acid, fluorouracil, and oxaliplatin) or FOLFIRI (folinic acid, fluorouracil, and irinotecan), paired with bevacizumab generally, or cetuximab/panitumumab for non-mutated RAS patients [4,5,6]. Pembrolizumab is the selected immunotherapy for patients with high microsatellite instability (MSI-H) or mismatch repair deficiency (dMMR) [7, 8]. Once disease progresses, second-line treatments are tailored, possibly changing chemotherapy, maintaining bevacizumab, or starting anti-EGFR antibodies in RAS wild-type tumors if not yet used [9, 10]. For refractory CRC, the recommended in the National Comprehensive Cancer Network (NCCN, Version 1.2025)-approved systemic treatments encompass regorafenib, regorafenib dose optimization (REDo), trifluridine/tipiracil (TAS-102), TAS-102 with bevacizumab (TAS-BEV), and fruquintinib. Additionally, biomarker-driven therapies are recommended for specific molecular subtypes, including PD-1 inhibitors (pembrolizumab, nivolumab ± ipilimumab) for MSI-H/dMMR tumors, encorafenib + cetuximab for BRAF V600E-mutant tumors, and HER2-directed therapies (trastuzumab + tucatinib or trastuzumab + pertuzumab) for HER2-positive tumors [11, 12]. Per the CORRECT trial, regorafenib improved median overall survival (mOS) to 2.8 months over BSC’s 1.8 months [13]. The REDoS study indicated REDo increased OS over regorafenib and was safer [14]. The RECOURSE trial demonstrated that TAS-102 significantly improved OS compared to BSC, with an HR of 0.68 and a 95% confidence interval (CI) of 0.58 to 0.81 [15]. Recent SUNLIGHT study results revealed that TAS-BEV notably enhanced survival over TAS-102 alone, with an mOS of 10.8 months compared to 7.5 months [16]. Recent data from the FRESCO and FRESCO-2 trials showed that fruquintinib improved OS compared with BSC, with HRs of 0.65 and 0.66 respectively [17, 18]. As a result, the U.S. Food and Drug Administration (FDA) has approved fruquintinib for treating patients with refractory CRC.

Employing individual patient data (IPD) meta-analysis, regarded as the best practice for analyzing time-to-event data, allows for nuanced treatment outcome evaluation while addressing data censoring [19]. In the context of refractory CRC, trials such as CONCUR and CORRECT have investigated the efficacy of a consistent treatment. Leveraging IPD meta-analysis in this case can yield more precise survival estimates. Furthermore, while numerous clinical studies have explored different treatment options, BSC frequently served as the control arm in these trials [13,14,15,16,17,18, 20,21,22]. Without direct comparisons of treatments, guideline committees are reluctant to prefer one drug over others. Hence, conducting a network meta-analysis (NMA) is essential to determine the relative effectiveness of different treatment options [23]. Given the substantial differences in drug prices ($11,826 for TAS-102 versus $25,200 for fruquintinib per 28-day cycle) and the modest survival gains over BSC (e.g., TAS-102 extends the mOS by only 1.8 months), cost considerations are critical [15]. Economic evaluations of different regimens should be evidence-based to establish their value.

Accordingly, we aimed to merge data to analyse the pooled clinical efficacy and safety of different treatments and to use NMAs to determine their relative effectiveness and safety. Finally, variations in cost-effectiveness among regimens were determined, and value-based pricing strategies were derived.

Methods

Search strategy

Literature synthesis was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses for Individual Patient Data (PRISMA-IPD) guidelines [24]. The search strategy is provided in Additional File Part 1. Researchers conducted a systematic search for relevant trials in PubMed, EMBASE, the Cochrane Library, and ClinicalTrials.gov up to December 2023.

Study selection and data extraction

Two researchers, Zhao and Jiang, initially screened the titles and abstracts of articles. Discrepancies were resolved through discussion with a panel including an oncology expert. The inclusion criteria were guided by the PICOS criteria:

  1. (1)

    Population: Adults with histologically or cytologically confirmed colon or rectal adenocarcinoma, refractory or intolerant to at least two prior chemotherapy lines.

  2. (2)

    Interventions and comparisons: Treatment options recommended by the NCCN for refractory CRC patients unresponsive to chemotherapy include regorafenib, ReDO, TAS-102, TAS-BEV, fruquintinib, and BSC.

  3. (3)

    Outcomes: The primary outcomes of this study were PFS, OS, and adverse events (adverse events [AEs] of grade 3 or higher, and serious AEs). The inclusion criteria mandated published Kaplan‒Meier curves for OS or PFS to facilitate IPD reconstruction from survival curves.

  4. (4)

    Study design: Priority was given to phase II-III RCTs that provided the necessary data. We included only the most recent and comprehensive trials to prevent data duplication.

  5. (5)

    The extracted data included PFS, OS, and AE. Furthermore, details such as the publication year, first author, trial registration, participant demographics (age, sample size, countries), and intervention specifics (treatments, endpoint outcomes) were also gathered.

Quality assessment

The quality of the studies included was assessed using the Cochrane Collaboration’s risk of bias (ROB) tool [25]. The quality of the eligible studies was classified as high, low, or unclear. Egger’s regression test was used to assess publication bias, with p values less than 0.05 indicating publication bias [26].

Statistical analyses

Guyot’s method was employed to acquire IPD [27]. We digitalized published Kaplan-Meier curves via the GetData Graph Digitizer software version 2.24.

Pooled estimates of safety and survival, and network meta-analysis

After reconstructing IPD, a pooled analysis of OS and PFS was performed via the MetaSurv approach. This involved aggregating arcsine-transformed survival probabilities with a random-effects model to produce a summary survival curve without assuming a specific distribution [28]. Given that AEs reported in various studies are categorized as treatment-related or of all-cause, with a notable absence of all-cause AEs reported in the majority of studies [13, 15, 20,21,22]. To enable a more thorough comparison, we compared treatment-related adverse events (TRAEs). Moreover, we employed a random-effects model from the Meta package to amalgamate the incidence rates of TRAEs.

For NMAs of PFS and OS, both fixed-effects and random-effects models were used, with the random-effects model accounting for inter-study variability. The presence of low heterogeneity prompted the reporting of results from the fixed-effects models. Time-to-event data analysis revealed that hazard ratios (HRs) often varied over time, and deviations from the proportional hazards (PH) assumption were noted, as exemplified by the OS curve in the FRESCO-2 [18]. Hence, relying exclusively on HRs derived from Cox-PH models as effect size measures in NMA is unsuitable. To address this issue, we estimated time-varying HRs and projected long-term survival rates using Bayesian fractional polynomial (FP) models [29], first-order FP models were fitted using power parameters − 2, -1, -0.5, 0.5, 1, 2, and 3. Model fit was evaluated using the Akaike information criterion (AIC) and visual inspection. In the sensitivity analyses, Asian trials were excluded. This exclusion was conducted to account for potential differences in patient characteristics, treatment responses, healthcare systems, and drug pricing structures between Asian and Western populations, which could influence both clinical and economic outcomes [30, 31]. FP models were executed using the ‘survival’ package. Survival outcomes and life-years gained (LYG) were the selected outcome measures. LYG was calculated using Wiksten’s method, employing a two-step approach for NMA [29]. OS and PFS curves for the reference treatment were developed and extended using parametric models such as the exponential, gamma, Gompertz, Weibull, generalized gamma, log-normal, log-logistic, FP, restricted cubic spline (RCS), and Royston-Parmar spline (RP) models. Model selection hinged on the AIC and visual assessment. Next, time-varying HRs between various treatments and the reference were derived using the FP model, enabling the generation of survival curves to estimate life-years for each regimen. Additionally, Cox-PH models were applied via the ‘Netmeta’ package. This method yielded cautious results, confirming the robustness of the primary analyses. The Cox-PH models used HRs with 95% CIs. For safety assessment, Bayesian NMAs were performed using the ‘BUGSnet’ package, and ORs with 95% CIs represented the effect sizes.

We performed subgroup analyses of OS and PFS stratified by RAS mutation status (wild-type vs. mutant) and the number of prior systemic treatments (≤ 3 vs. >3). Given the unavailability of subgroup survival curves, we obtained only HR estimates using Cox-PH models.

Heterogeneity was gauged using Cochran’s Q test and the I2 statistic. An I2 greater than 40% or a Q test p-value less than 0.1 denoted significant heterogeneity [32]. Model inconsistency was evaluated using the edge-splitting method, which incorporates both direct and indirect evidence [33, 34]. Markov chain convergence was checked using trace plots and the Gelman-Rubin statistic [35]. Data analysis was performed using R 4.1.0.

Cost-effectiveness analysis

Model structure

A partitioned survival model with three health states was utilized to calculate total costs and QALYs for each treatment option (BSC, RSD, ReDO, TAS-102, TAS-BEV, and fruquintinib). Patients began in the PFS state and transitioned to progressive disease (PD) or death. The model used a 28-day cycle over 10 years, covering 99% of patient transitions to death. Outcomes were measured from the US public sector perspective (Veterans Affairs, VA), discounting costs and QALYs annually at 3% [36]. A half-cycle correction was implemented.

Clinical and economic parameters

All parameters are detailed in Table S1. Data on the efficacy and safety for various regimens were derived from the initial part of this study [37]. AEs of grade 3 + with incidence rates of 1% or higher were incorporated into the model. The utility values for the PFS and PD states were sourced from the CORRECT trial, which provided EuroQol 5-Dimension Health Questionnaire (EQ-5D) scores of 0.73 for PFS and 0.59 for PD [15]. Treatment duration for different regimens was equated to time in PFS, indicated by time until treatment discontinuation [38]. The economic parameters included drug treatment expenses, AE management, follow-up and administrative costs, and end-of-life care costs. Drug costs were based on Federal Supply Schedule (FSS) pricing, reflecting VA-negotiated prices with drug firms. Compared with list drug prices, FSS drug costs more accurately reflect the true treatment burden [36]. Direct drug costs were calculated for an 80 kg patient with a body surface area of 1.78 m2. The BSC costs were derived from existing literature [39]. For follow-up and administrative cost calculations, it was assumed that patients attended biweekly physician visits and underwent complete blood counts with differential and comprehensive metabolic panel tests [40]. Furthermore, following the NICE guidelines, patients were assumed to receive a CT scan quarterly [41]. Prices were derived from Medicare’s Current Procedural Terminology (CPT) codes [42]. Post-progression costs were assessed for refractory CRC patients, upon advancing to the PD state, they received BSC, adhering to NICE guideline recommendations [41]. In scenario analysis, we accounted for the possibility that patients in the PD state might continue receiving active systemic therapy rather than exclusively receiving BSC. AE treatment costs were primarily derived from the Healthcare Cost and Utilization Project by the Agency for Healthcare Research and Quality [43], and were applied in the initial cycle of the model. The end-of-life care cost originated from real-world data of cancer patients from the perspective of the United States [44]. Costs were adjusted to 2023 U.S. dollars using the Personal Consumption Expenditures Price Index for healthcare services [45]. The willingness-to-pay threshold for each QALY was set at $150,000 ($100,000-200,000) [40]. For more details, refer to Table S1.

Sensitivity analyses, scenario analyses and price simulation

Deterministic sensitivity analysis (DSA) was used to assess the effects of key parameters on the incremental net monetary benefit (INMB), and the results are displayed in tornado diagrams. Parameters varied within set ranges from the literature or databases, or within ± 20% of base-case values where specific limits were unavailable. A probabilistic sensitivity analysis (PSA) using Monte Carlo simulation with 10,000 iterations was conducted on the base-case. Costs were modeled with a gamma distribution, and probabilities, proportions, and utilities were modeled with a beta distribution. The cost-effectiveness acceptability curves (CEAC) explored cost-effectiveness probabilities over a $0 to $500,000 willingness-to-pay threshold. Table S1 details the sources of uncertainty.

Our study included scenario analyses to address uncertainties in model structure, parameter and utility values, survival extrapolation, and treatment patterns. The following scenarios were considered: (1) excluding Asian RCTs from the NMA; (2) assuming PH, modeling used Cox-PH model outcomes; (3) in the PD state, active treatment primarily consisting of chemotherapy was permitted (predominantly chosen subsequent active treatments were informed by RCTs); [13,14,15,16,17,18, 20,21,22] (4) utility values were derived from the UK advanced CRC patients; [46] (5) according to the NICE guidelines, BSC was assumed to incur no cost [41].

From a cost-effectiveness standpoint, reasonable pricing strategies were evaluated using BSC and TAS-102 as benchmarks, with assumed intervention prices ranging from no reduction to a full discount. The willingness-to-pay threshold was set at $150,000/QALY in the price simulation.

Results

Characteristics of the included studies

Figure S1 presents the flowcharts. Of the 3979 records found in the specified databases, 3614 were excluded initially due to the selection criteria. Subsequently, 365 studies were considered for full-text review. After applying the eligibility criteria, 9 studies were included in our network, 3 of which were exclusively from the Western Pacific region [17, 20, 22]. This study included 3978 patients and six NCCN-recommended treatments: regorafenib, REDo, TAS-102, TAS-BEV, and fruquintinib [47]. Additional characteristics are summarized in Table 1. The evidence network plot is shown in Fig. 1.

Fig. 1
figure 1

Evidence network plot. BSC, Best Supportive Care; N, number of patients; RCT, Randomized Controlled Trial; REDo, regorafenib dose optimization; TAS-102, trifluridine/tipiracil; TAS-BEV, TAS-102 with bevacizumab

Full size image
Table 1 Characteristics of the included studies
Full size table

Risk of Bias

The assessment of ROB is presented in Additional File Part 2. All RCTs had a low risk of bias, although some open-label trials raised concerns about blinding and allocation concealment [14, 16, 21]. The Egger test results suggested no publication bias within our network, and funnel plots can be found in Additional File Part 3.

Safety results

The pooled safety estimates, illustrated in Fig. 2A, revealed that the incidence of serious TRAEs across targeted therapies, ranked from lowest to highest were fruquintinib (0.06, 95% CI [0.04–0.10]), REDo (0.11, 0.04–0.23), regorafenib (0.19, 0.05–0.51), TAS-102 (0.31, 0.16–0.5), and TAS-BEV (0.45, 0.3–0.6). For grade 3 + TRAEs, fruquintinib had the lowest incidence (0.41, 95% CI [0.31–0.51]), followed by TAS-102 (0.46, 0.4–0.52), and regorafenib (0.54, 0.5–0.58).

Fig. 2
figure 2

Summarized results of safety analysis. A Pooled Incidence of Adverse Events; B. Network Meta-Analysis Results of Adverse Events. BSC, Best Supportive Care; REDo, regorafenib dose optimization; TAS-102, trifluridine/tipiracil; TAS-BEV, TAS-102 with bevacizumab; TRAE, treatment-related adverse event

Full size image

The NMA results (Fig. 2B) indicated that, for serious TRAEs, targeted treatments do not significantly increase the risk compared to BSC. The safety profiles, from best to worst, were as follows: TAS-102 (OR 1.02, 95% CI[0.63–1.7]), REDo (1.03, 0.31–3.38), fruquintinib (1.12, 0.83–1.57), TAS-BEV (1.18, 0.45–3.13), and regorafenib (1.24, 0.92–1.67). In terms of Grade 3 + TRAEs relative to BSC, the safety ranking was: fruquintinib (OR 6.63, 95% CI [1.55–31.57]), followed by regorafenib (7.34, 1.62–32.9), and TAS-BEV (7.51, 0.88–65.32). The league table revealed no significant differences in the safety profiles of targeted therapies for either serious or grade 3 + AEs (Table S2).

Efficacy and effectiveness

The pooled survival estimates (Fig. 3) indicated that the mOS for the regimens varied, ordered from the lowest to highest: BSC (5.92 months, 95% CI [4.76–7.09]), TAS-102 (7.23, 6.32-8), regorafenib (7.37, 5.18–9.11), fruquintinib (8.19, 6.36–10.51), TAS-BEV (8.27, 5.08–11.73), and REDo (9.8, 7.5–11.9). In terms of PFS, from lowest to highest were BSC (1.58 months, 95% CI [1.51–1.67]), regorafenib (2.24, 1.87–3.31), TAS-102 (2.52, 1.91–2.91), REDo (2.8, 2–5), TAS-BEV (3.58, 2.08–5.23), and fruquintinib (3.62, 3.34–3.82). Heterogeneity was low in this analysis, all under 40%.

Fig. 3
figure 3

Pooled estimates of overall survival and progression-free survival. BSC, Best Supportive Care; REDo, regorafenib dose optimization; TAS-102, trifluridine/tipiracil; TAS-BEV, TAS-102 with bevacizumab

Full size image
Fig. 4
figure 4

Summary of network meta-analysis results for PFS and OS analysis. (A) OS; (B) PFS. BSC, Best Supportive Care; OS, overall survival; PFS, progression-free survival; REDo, regorafenib dose optimization; TAS-102, trifluridine/tipiracil; TAS-BEV, TAS-102 with bevacizumab

Full size image
Fig. 5
figure 5

Network meta-analysis results based on the cox proportional hazards model for OS and PFS. BSC, Best Supportive Care; N, number of patients; HR, Hazards ratios; OS, Overall Survival; PFS, Progression-free survival; RCT, Randomized Controlled Trial; REDo, regorafenib dose optimization; TAS-102, trifluridine/tipiracil; TAS-BEV, TAS-102 with bevacizumab

Full size image

For NMA, OS and PFS curves for BSC were extrapolated using log-normal and log-logistic models, respectively. For the second step, we applied a first-order FP model (P=-2) to derive time-varying HRs for OS and PFS across treatments versus BSC. The model fit results are shown in Tables S3-S4. Figure 4 shows 60-month OS LYG, ranked from most to least: TAS-BEV (17.04 months), fruquintinib (15.27), TAS-102 (12.5), REDo (11.45), regorafenib (8.87), and BSC (8.18); For 24-month PFS LYG, from highest to lowest: TAS-BEV (10.85 months), fruquintinib (7.03), regorafenib (6.29), TAS-102 (5.69), REDo (4.78), and BSC (2.05). Excluding Asian-only trials, 60-month OS life-year gains ranked as follows: TAS-BEV (18.3 months), fruquintinib (13.84), REDo (12.03), TAS-102 (11.27), regorafenib (9.32), and BSC (8.21). For PFS at 24 months, LYG ranked as follows from highest to lowest: TAS-BEV (9.9 months), fruquintinib (6.02), regorafenib (5.75), REDo (4.87), TAS-102 (4.85), and BSC (2.05). Assuming that the PH assumption held, the results based on the Cox-PH model indicated that TAS-BEV had a notable lead in both OS and PFS. For OS, the rankings from highest to lowest against TAS-BEV were: REDo (HR 1.18, 95% CI [0.7–1.99]), fruquintinib (1.53, 1.14–2.04), regorafenib (1.64, 1.22–2.22), TAS-102 (1.66, 1.35–2.05), and BSC (2.33, 1.81–2.99). For PFS, the rankings from highest to lowest against TAS-BEV were: fruquintinib (HR 1.45, 95% CI [1.11–1.9]), REDo (1.85, 1.16–2.98), regorafenib (2.21, 1.69–2.9), TAS-102 (2.26, 1.88–2.72), and BSC (4.89, 3.89–6.15). Similarly, low heterogeneity was observed in both PFS and OS, as detailed in Fig. 5. See Table S5 for subgroup analysis results.

Cost-effectiveness

Clinical experts assessed the model’s structure, assumptions, data sources, and outcomes for face validity. PFS and OS survival rates matched trial data, confirming model robustness. The right-censoring percentages for all OS and PFS curves were under 20% (except for TAS-BEV in SUNLIGHT at 30%) [16], suggesting minimal uncertainty in extrapolation [13,14,15,16,17,18].

Baseline analysis revealed that treatment costs increased from BSC at $41,242 to TAS-102 at $140,656, REDo at $177,069, regorafenib at $186,803, TAS-BEV at $320,452, and fruquintinib topping at $334,799. The effectiveness ranged from 0.69 QALYs for BSC to 0.85 for regorafenib, 0.99 for REDo, 1.1 for TAS-102, and 1.39 for fruquintinib to 1.59 QALYs for TAS-BEV. BSC has historically been the most cost-effective option, with ICERs for other regimens relative to BSC escalating from TAS-102 at $349,254/QALY to TAS-BEV at $451,853/QALY, fruquintinib at $588,347/QALY, REDo at $682,534/QALY, and regorafenib at $1,317,383/QALY. TAS-102 emerged as the leading cost-effective active treatment, more details are provided in Table 2. The DSA results shown in Fig. 6 revealed that treatment costs, including drug pricing and patient body size factors, were the main drivers behind the highest INMB across regimens. Nevertheless, the foundational conclusions of the baseline analysis held firm: the five treatments, when compared to BSC, failed to be cost-effective when parameters varied within their specified ranges. Moreover, the cost-effectiveness advantage of TAS-102 over the other four targeted therapies remained unchanged amid variations in parameter values. The PSA results indicated that when the willingness-to-pay threshold was within the range of $100,000-200,000, BSC consistently remained the most cost-effective option (Fig. 7A); comparative results among targeted therapies showed (Fig. 7B) that TAS-102 was the most cost-effective option within the specified threshold range. Similarly, after excluding Asian trials, the conclusions remained unchanged (Fig. 7C-D). The scenario analysis results (Table 2) also indicated that the baseline analysis findings were stable. Among the five scenarios, BSC was the lowest-cost option, while TAS-BEV was the most effective choice. Additionally, BSC was the most cost-effective option, and TAS-102 was the most cost-effective targeted therapy.

Table 2 Summary of cost-effectiveness analysis results
Full size table
Fig. 6
figure 6

One-way sensitivity analysis tornado diagram. BSC, Best Supportive Care; DSA, Deterministic Sensitivity Analysis; REDo, regorafenib dose optimization; TAS-102, trifluridine/tipiracil; TAS-BEV, TAS-102 with bevacizumab

Full size image
Fig. 7
figure 7

Cost-effectiveness acceptability curves from probabilistic sensitivity analysis. A. Primary Analysis, All Treatments; B. Primary Analysis, Excluding BSC; C. Scenario Analysis, Excluding Asian Trials, All Treatments; D. Scenario Analysis, Excluding Asian Trials, Excluding BSC. BSC, Best Supportive Care; REDo, regorafenib dose optimization; TAS-102, trifluridine/tipiracil; TAS-BEV, TAS-102 with bevacizumab

Full size image

At a willingness-to-pay threshold of $150,000/QALY, the pricing simulation suggested that, compared to BSC, TAS-102, fruquintinib, REDoS, regorafenib, and TAS-BEV would need to decrease their prices to 39%, 24%, 14%, 8%, and 7% of their original prices, respectively, to potentially be cost-effective. In comparison to TAS-102, a price reduction of 40% for REDoS, 56% for fruquintinib, and 59% for regorafenib would be required to potentially achieve economic viability (Fig. 8).

Fig. 8
figure 8

Evidence-based pricing simulation outcomes. A. VS. BSC; B. VS. TAS-102. BSC, Best Supportive Care; REDo, regorafenib dose optimization; TAS-102, trifluridine/tipiracil; TAS-BEV, TAS-102 with bevacizumab

Full size image

Discussion

Main findings

In this study, we systematically evaluated 9 studies involving 3978 patients to explore the safety, efficacy, and cost-effectiveness of different treatment regimens for refractory CRC. This study examined six NCCN-endorsed treatments: BSC, regorafenib, REDo, TAS-102, TAS-BEV, and fruquintinib [47]. Regarding safety, our meta-analysis showed the pooled incidences of serious TRAEs and grade 3 + TRAEs across treatment regimens. Targeted treatments, compared with BSC, variably raised serious TRAEs and grade 3 + TRAE incidence rates. For serious TRAEs, the incidence rates ranked from highest to lowest were as follows: TAS-BEV (45%), TAS-102 (31%), regorafenib (19%), REDo (11%), and fruquintinib (6%). However, the NMA analysis indicated that, among targeted therapies, there were no significant differences in either serious TRAEs or grade 3 + TRAEs. The ranking discrepancy between direct incidence rates and NMA-derived Odd Ratios resulted from heterogeneity across trials, particularly the large variation in serious TRAEs incidence rates in the BSC arms. In the FRESCO trial, the BSC group had an SAE incidence of only 1.5%, while in TERRA and CONCUR, it was 23% and 4.4%, respectively.

Regarding efficacy, treatment regimens significantly affect the OS and PFS. BSC had the shortest median survival times for OS and PFS. REDo reported the longest mOS at 9.8 months, followed by TAS-BEV at 8.27 months. For PFS, fruquintinib led to a median survival of 3.63 months, closely followed by TAS-BEV at 3.58 months. The NMA results suggested that for OS, TAS-BEV offered the greatest survival benefit, with fruquintinib and REDo trailing behind. In terms of PFS, TAS-BEV again surpassed other regimens, with fruquintinib being its closest competitor. Assuming PH, the results once again validated the significant advantages of TAS-BEV over other regimens in terms of OS and PFS. Fruquintinib and REDo outperformed regorafenib and TAS-102 but without statistical significance. Moreover, we observed low heterogeneity across treatment regimens, enhancing the reliability of our analysis findings.

From a public payer’s view, baseline analysis revealed BSC to be the lowest-cost, most cost-effective choice, with TAS-BEV being most effective. Among the five targeted therapies, TAS-102 was the most cost-effective option. DSA results indicated a stable baseline analysis that was impervious to parameter shifts, model choices, or base assumptions. Similarly, PSA demonstrated that within $100,000/QALY to $200,000/QALY thresholds, BSC was consistently the most cost-effective, with TAS-102 as the optimal active treatment. Price-simulation results indicated that, for cost-effectiveness compared to BSC, TAS-102, fruquintinib, REDoS, regorafenib, and TAS-BEV should reduce prices to 39%, 24%, 14%, 8%, and 7% of their original values, respectively. Moreover, REDoS, fruquintinib, and regorafenib required 40%, 56%, and 59% price cuts, respectively, for economic viability against TAS-102.

This study revealed substantial efficacy of TAS-BEV over alternative regimens, with corroborative evidence from the C-TASKFORCE and SUNLIGHT clinical trials [16, 21]. According to SUNLIGHT, the superiority of TAS-BEV over TAS-102 was confirmed across most subgroups, including factors like age, sex, primary disease location, metastatic sites count, and RAS mutation status [16]. The efficacy of the TAS-BEV comes from TAS-102’s dual action—combining trifluridine, which attacks cancer cells, with tipiracil, which prevents trifluridine’s breakdown—and bevacizumab’s ongoing VEGF blockage, a clinically proven approach to extend survival in refractory CRC [9, 48]. Fruquintinib is an oral, selective, and potent inhibitor of VEGF receptors 1, 2, and 3, which are crucial for angiogenesis and thus for tumor growth and metastasis [49]. As the latest drug approved by the FDA, its success is tied to the publication of the global multicenter FRESCO-2 study [18]. Fruquintinib was initially launched in China for refractory colorectal cancer treatment, following the findings of the FRESCO study [17]. Unlike the SUNLIGHT study, where 24% of patients were VEGF-naïve, those in the FRESCO-2 study were treated with a more advanced line of therapy [18]. Therefore, this study may have somewhat underestimated the efficacy of fruquintinib. However, fruquintinib has retained its status as a leading choice for efficacy and safety among targeted single-agent treatments. Furthermore, patient diversity in the FRESCO and FRESCO-2 studies suggested a broader application for fruquintinib [17, 18]. However, its higher price in the United States has somewhat restricted its widespread clinical use [50]. Compared to prior regorafenib, REDo had lower rates of side effects such as fatigue, hand-foot skin reaction, and hypertension, and it offered marginal OS and PFS benefits and enhanced patients’ quality of life [14]. Additionally, REDo has somewhat decreased treatment expenses, making it an effective alternative to regorafenib when considering these factors.

Balancing safety, efficacy, and cost-effectiveness is key in refractory CRC treatment. While targeted therapies improve survival, they also raise adverse event risks and costs. Thus, treatment selection should consider disease traits, patient health, prior treatments, side effect tolerance, and patient preferences [51]. For instance, despite TAS-BEV’s effectiveness in life extension, it also carries a risk of heightened adverse events [16], like severe neutropenia. Hence, TAS-BEV might not suit patients valuing quality of life over extended survival, or those with severe comorbidities or frail health [52]. On the other hand, TAS-BEV may appeal to patients aiming for the longest survival, particularly if they are healthy and can manage side effects. For those concerned with costs or in resource-limited healthcare settings, BSC might be favored for its affordability. Moreover, TAS-102, a cost-effective targeted regimen, could balance patient expenses with treatment effectiveness. When choosing TAS-102, it is crucial for doctors and patients to jointly consider the effectiveness-side effect trade-off and its cost-effectiveness relative to alternatives. TAS-BEV showed notable survival benefits, indicating that it is an optimal option for patients who can afford it. Doctors needed to ensure that the higher costs were in line with patients’ financial means and treatment goals. Effective communication was key so that patients comprehended the benefits, risks, and financial implications of their options. Ongoing monitoring of treatment efficacy and side effects is also crucial for doctors to adapt plans to enhance patient quality of life and overall well-being.

A previous study by Cho et al. assessed the cost-effectiveness of regorafenib and TAS-102 compared to BSC from a US payer perspective, finding that neither was cost-effective at the standard willingness-to-pay threshold of $150,000/QALY [53]. Sang et al.‘s study analyzed ReDO against regorafenib, TAS-102, and TAS-BEV from a US payer standpoint and concluded that ReDO was the most cost-effective option at the $150,000/QALY threshold [40]. The discrepancies between these results and those of the current article may stem from limitations in Sang et al.‘s study. First, they sourced efficacy data directly from clinical trials without indirect comparison adjustments, resulting in an ostensibly greater effectiveness for ReDO over TAS-BEV, which was arguably implausible [36, 38]. Second, our analysis included additional clinical data, such as the SUNLIGHT study, which impacted the results [16]. Finally, we employed a more suitable parametric extrapolation model, unlike prior studies that relied only on exponential models [54].

While our analysis is based on a WTP threshold of $150,000/QALY, it is important to note that WTP varies significantly across countries, influencing cost-effectiveness assessments. For example, the UK applies a threshold of $25,000-$38,000/QALY [55], while Germany does not have a fixed WTP threshold but often considers values around $55,000/QALY in price negotiations [56]. Japan generally applies a threshold of $35,000-$50,000/QALY, whereas Australia typically considers $33,000/QALY in health technology assessments [57]. China often adopts a threshold of $12,000-$36,000/QALY, based on 1–3 times GDP per capita [58]. In contrast, the US does not have a standardized WTP threshold, but estimates range from $100,000 to $150,000/QALY, with higher values sometimes considered for severe or rare diseases [59]. Given these differences, the cost-effectiveness of TAS-102 and similar therapies may vary across healthcare systems, depending on local WTP thresholds, healthcare budgets, and reimbursement policies. Additionally, drug pricing strategies differ across countries, with some nations implementing centralized price negotiations, while others, like the US, rely on market-driven pricing. These factors highlight the importance of country-specific economic evaluations to ensure that cost-effectiveness assessments align with local healthcare priorities and budget constraints.

Strengths and limitations

This study not only provides critical insights into the safety and efficacy of various targeted treatment regimens but also presents a nuanced cost-effectiveness analysis pivotal for clinical decision-making. The findings are vital for crafting treatment approaches and distributing resources, especially in settings constrained by healthcare availability. The value of this article is multifaceted: This article is the first of its kind to implement a meta-analysis of reconstructed IPD for survival data in refractory CRC patients, setting a gold standard by effectively accounting for event censoring. The pooled estimates for OS and PFS showed minimal heterogeneity, and all incorporated studies were of high quality. These robust, homogeneous survival data can accurately reflect the performance of treatment regimens and provide a basis for calculating sample sizes in future clinical trials. Second, this study is the first to conduct an NMA for refractory CRC using reconstructed IPD, a noteworthy advancement since traditional NMA based on HR requires the PH assumption which was proven inapplicable here. By employing a time-varying model, we leveraged IPD to evaluate regimen efficacy at various time points, enriching the use of clinical evidence and bolstering result reliability [60]. Third, for the first time in this study, an effectiveness and cost-effectiveness comparison of all regimens recommended by the NCCN guidelines was conducted from the perspective of a US public payer. This study provides valuable evidence for rational clinical drug use, rational allocation of healthcare resources, and decision-making. In addition, price simulations were conducted to inform value-based drug pricing.

Like most modeling study, our research also has several limitations. First, when evaluating regimen safety, due to limited evidence, we only used grade 3 + TRAEs and ignored other dimensions of safety indicators, such as grade 1–2 AEs and all-cause AEs. Second, the evidence for safety analysis exhibited significant heterogeneity, necessitating a cautious interpretation of some results. Furthermore, to enhance evidence quality, this study exclusively included phase II-Ш RCTs, excluding single-arm studies and disregarding real-world evidence. Therefore, the relevant findings require further validation in real-world clinical practice. Third, indirect comparisons might encounter variations in baseline patient characteristics across different RCTs. Lacking access to individual patient data, our IPD meta-analysis could not account for factors beyond survival status. Consequently, further head-to-head RCTs comparing targeted therapies are essential to confirm our study’s conclusions. Fourth, we only used a VA perspective and did not consider other payers, further research is needed to supplement this in the future. Finally, this study did not study the availability and affordability, implying that further research is required.

Conclusion

Compared to BSC, targeted therapies markedly increase AEs without significant safety differences among themselves. In terms of efficacy, TAS-BEV leads substantially, with furqutinib and REDo as secondary choices. For cost-effectiveness, BSC is the frontrunner, with TAS-102 as the preferred targeted therapy. Regimens differ in safety, efficacy, and cost-effectiveness, necessitating a balanced decision by healthcare providers based on individual patient circumstances and therapeutic aims.

Data availability

Data sharing statement Full data set is available, on request from the corresponding author, e-mail, tokammy@cpu.edu.cn.

Abbreviations

AEs:

Adverse events

AIC:

Akaike information criterion

BSC:

Best Supportive Care

C:

Control group

CRC:

Colorectal cancer

I:

Intervention group

HR:

Hazard ratio

N:

Number of patients

NA:

Not available

NCCN:

National Comprehensive Cancer Network

NMA:

Network meta-analysis

OS:

Overall survival

PD:

Progressive disease

PFS:

Progression-free survival

QALY:

Quality-Adjusted life year

RCT:

Randomized controlled trial

REDo:

Regorafenib dose optimization

ROB:

Risk of bias

TAS-102:

Trifluridine/tipiracil

TAS-BEV:

TAS-102 with bevacizumab

TRAE:

Treatment-related adverse event

References

  1. Sung H, et al. Global cancer statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin. 2021;71(3):209–49.

    Article  PubMed  Google Scholar 

  2. Siegel RL, et al. Colorectal cancer statistics, 2023. CA Cancer J Clin. 2023;73(3):233–54.

    Article  PubMed  Google Scholar 

  3. Dekker E, et al. Colorectal cancer. Lancet. 2019;394(10207):1467–80.

    Article  PubMed  Google Scholar 

  4. Venook AP, et al. Effect of First-Line chemotherapy combined with cetuximab or bevacizumab on overall survival in patients with KRAS Wild-Type advanced or metastatic colorectal cancer: A randomized clinical trial. JAMA. 2017;317(23):2392–401.

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  5. Watanabe J, et al. Panitumumab vs bevacizumab added to standard First-line chemotherapy and overall survival among patients with RAS Wild-type, Left-Sided metastatic colorectal cancer: A randomized clinical trial. JAMA. 2023;329(15):1271–82.

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  6. Douillard JY, et al. Panitumumab-FOLFOX4 treatment and RAS mutations in colorectal cancer. N Engl J Med. 2013;369(11):1023–34.

    Article  PubMed  CAS  Google Scholar 

  7. Ganesh K, et al. Immunotherapy in colorectal cancer: rationale, challenges and potential. Nat Rev Gastroenterol Hepatol. 2019;16(6):361–75.

    Article  PubMed  PubMed Central  Google Scholar 

  8. André T, et al. Pembrolizumab in Microsatellite-Instability-High advanced colorectal cancer. N Engl J Med. 2020;383(23):2207–18.

    Article  PubMed  Google Scholar 

  9. Bennouna J, et al. Continuation of bevacizumab after first progression in metastatic colorectal cancer (ML18147): a randomised phase 3 trial. Lancet Oncol. 2013;14(1):29–37.

    Article  PubMed  CAS  Google Scholar 

  10. Bennouna J, et al. Continuation of bevacizumab vs cetuximab plus chemotherapy after first progression in KRAS Wild-Type metastatic colorectal cancer: the UNICANCER PRODIGE18 randomized clinical trial. JAMA Oncol. 2019;5(1):83–90.

    Article  PubMed  Google Scholar 

  11. National Comprehensive Cancer Network. NCCN Clinical Practice Guidelines in Oncology: Colon Cancer. Version 1.2025. Accessed February 26. 2025. Available at: https://www.nccn.org/guidelines/guidelines-detail?category=1amp;id=1428.

  12. National Comprehensive Cancer Network. NCCN Clinical Practice Guidelines in Oncology: Rectal Cancer. Version 1.2025. Accessed February 26. 2025. Available at: https://www.nccn.org/guidelines/guidelines-detail?category=1amp;id=1428.

  13. Grothey A, et al. Regorafenib monotherapy for previously treated metastatic colorectal cancer (CORRECT): an international, multicentre, randomised, placebo-controlled, phase 3 trial. Lancet. 2013;381(9863):303–12.

    Article  PubMed  CAS  Google Scholar 

  14. Bekaii-Saab TS, et al. Regorafenib dose-optimisation in patients with refractory metastatic colorectal cancer (ReDOS): a randomised, multicentre, open-label, phase 2 study. Lancet Oncol. 2019;20(8):1070–82.

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  15. Mayer RJ, et al. Randomized trial of TAS-102 for refractory metastatic colorectal cancer. N Engl J Med. 2015;372(20):1909–19.

    Article  PubMed  Google Scholar 

  16. Prager GW, et al. Trifluridine-Tipiracil and bevacizumab in refractory metastatic colorectal cancer. N Engl J Med. 2023;388(18):1657–67.

    Article  PubMed  CAS  Google Scholar 

  17. Li J, et al. Effect of fruquintinib vs placebo on overall survival in patients with previously treated metastatic colorectal cancer: the FRESCO randomized clinical trial. JAMA. 2018;319(24):2486–96.

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  18. Dasari A, et al. Fruquintinib versus placebo in patients with refractory metastatic colorectal cancer (FRESCO-2): an international, multicentre, randomised, double-blind, phase 3 study. Lancet. 2023;402(10395):41–53.

    Article  PubMed  CAS  Google Scholar 

  19. Stewart LA, Tierney JF. To IPD or not to IPD? Advantages and disadvantages of systematic reviews using individual patient data. Eval Health Prof. 2002;25(1):76–97.

    Article  PubMed  Google Scholar 

  20. Xu J, et al. Results of a randomized, Double-Blind, Placebo-Controlled, phase III trial of Trifluridine/Tipiracil (TAS-102) monotherapy in Asian patients with previously treated metastatic colorectal cancer: the TERRA study. J Clin Oncol. 2018;36(4):350–8.

    Article  PubMed  CAS  Google Scholar 

  21. Pfeiffer P, et al. TAS-102 with or without bevacizumab in patients with chemorefractory metastatic colorectal cancer: an investigator-initiated, open-label, randomised, phase 2 trial. Lancet Oncol. 2020;21(3):412–20.

    Article  PubMed  CAS  Google Scholar 

  22. Li J, et al. Regorafenib plus best supportive care versus placebo plus best supportive care in Asian patients with previously treated metastatic colorectal cancer (CONCUR): a randomised, double-blind, placebo-controlled, phase 3 trial. Lancet Oncol. 2015;16(6):619–29.

    Article  PubMed  CAS  Google Scholar 

  23. Tian J, et al. Progress and challenges of network meta-analysis. J Evid Based Med. 2021;14(3):218–31.

    Article  PubMed  Google Scholar 

  24. Stewart LA, et al. Preferred reporting items for systematic review and Meta-Analyses of individual participant data: the PRISMA-IPD statement. JAMA. 2015;313(16):1657–65.

    Article  PubMed  Google Scholar 

  25. Higgins JP, et al. The Cochrane collaboration’s tool for assessing risk of bias in randomised trials. BMJ. 2011;343:d5928.

    Article  PubMed  PubMed Central  Google Scholar 

  26. Muka T, et al. A 24-step guide on how to design, conduct, and successfully publish a systematic review and meta-analysis in medical research. Eur J Epidemiol. 2020;35(1):49–60.

    Article  PubMed  Google Scholar 

  27. Guyot P, et al. Enhanced secondary analysis of survival data: reconstructing the data from published Kaplan-Meier survival curves. BMC Med Res Methodol. 2012;12:9.

    Article  PubMed  PubMed Central  Google Scholar 

  28. Jackson D, White IR, Thompson SG. Extending dersimonian and Laird’s methodology to perform multivariate random effects meta-analyses. Stat Med. 2010;29(12):1282–97.

    Article  PubMed  Google Scholar 

  29. Wiksten A, et al. Nonproportional hazards in network Meta-Analysis: efficient strategies for model Building and analysis. Value Health. 2020;23(7):918–27.

    Article  PubMed  Google Scholar 

  30. Yousef M, Yousef A. Molecular, socioeconomic, and clinical factors affecting Racial and ethnic disparities in colorectal cancer survival. JAMA Oncol. 2024;10(11):1519–29. https://doiorg.publicaciones.saludcastillayleon.es/10.1001/jamaoncol.2024.3666.

    Article  PubMed  Google Scholar 

  31. Malick A, Ferris JS. Racial, ethnic, and sex differences in Incidence-Based mortality of aggregate upper Gastrointestinal cancers. Clin Transl Gastroenterol. 2024;15(8):e00745. https://doiorg.publicaciones.saludcastillayleon.es/10.14309/ctg.0000000000000745.

    Article  PubMed  PubMed Central  Google Scholar 

  32. Zhao B, et al. A bayesian network meta-analysis regarding the comparative efficacy of therapeutics for ALK-positive, brain metastatic non-small cell lung cancer. Pharmacol Res. 2021;174:105931.

    Article  PubMed  CAS  Google Scholar 

  33. van Valkenhoef G, et al. Automated generation of node-splitting models for assessment of inconsistency in network meta-analysis. Res Synth Methods. 2016;7(1):80–93.

    Article  PubMed  Google Scholar 

  34. Dias S, et al. Checking consistency in mixed treatment comparison meta-analysis. Stat Med. 2010;29(7–8):932–44.

    Article  PubMed  CAS  Google Scholar 

  35. Brooks SP, Gelman A. General methods for monitoring convergence of iterative simulations[J]. Journal of Computational and Graphical Statistics. 1998;7(4):434–55.

  36. Yoo M, et al. Cost-effectiveness analysis of 7 treatments in metastatic hormone-sensitive prostate cancer: a public-payer perspective. J Natl Cancer Inst. 2023;115(11):1374–82.

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  37. Kearns B, et al. Generalized linear models for flexible parametric modeling of the hazard function. Med Decis Mak. 2019;39(7):867–78.

    Article  Google Scholar 

  38. Kuznik A, et al. Cost-Effectiveness of Cemiplimab versus standard of care in the united States for First-Line treatment of advanced Non-small cell lung cancer with programmed Death-Ligand 1 expression ≥ 50. Value Health. 2022;25(2):203–14.

    Article  PubMed  Google Scholar 

  39. Chu JN, et al. Cost-effectiveness of immune checkpoint inhibitors for microsatellite instability-high/mismatch repair-deficient metastatic colorectal cancer. Cancer. 2019;125(2):278–89.

    Article  PubMed  CAS  Google Scholar 

  40. Cho SK, et al. Value-Based analysis of therapies in refractory metastatic colorectal cancer in US. Clin Colorectal Cancer. 2022;21(4):277–84.

    Article  PubMed  Google Scholar 

  41. National Institute for Health and Care Excellence. Regorafenib for previously treated metastatic colorectal cancer [ID4002]. https://www.nice.org.uk/guidance(Accessed Feb 2, 2024).

  42. CMS,https:/. /www.cms.gov/Medicare/Medicare-Fee-for-Service-Payment,Accessed Feb 1, 2024.

  43. AHRQ. Healthcare Cost and Utilization Project. Available at: https://hcupnet.ahrq.gov/.

  44. Yi D, et al. Drivers of care costs and quality in the last 3 months of life among older people receiving palliative care: A multinational mortality follow-back survey across England, Ireland and the united States. Palliat Med. 2020;34(4):513–23.

    Article  PubMed  Google Scholar 

  45. https://www.bls.gov/data/inflation_calculator.htm

  46. Hoyle M, et al. Cost-effectiveness of cetuximab, cetuximab plus Irinotecan, and panitumumab for third and further lines of treatment for KRAS wild-type patients with metastatic colorectal cancer. Value Health. 2013;16(2):288–96.

    Article  PubMed  Google Scholar 

  47. https://www.nccn.org/guidelines/nccn-guidelines

  48. Tabernero J et al. Effect of Trifluridine/tipiracil in patients treated in RECOURSE by prognostic factors at baseline: an exploratory analysis. ESMO Open. 2020;5(4).

  49. Sun Q, et al. Discovery of fruquintinib, a potent and highly selective small molecule inhibitor of VEGFR 1, 2, 3 tyrosine kinases for cancer therapy. Cancer Biol Ther. 2014;15(12):1635–45.

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  50. Information for Connecticut Prescribers of Prescription Drugs:FRUZAQLATM. (fruquintinib) Oral Capsules extension: http://bfdogplmndidlpjfhoijckpakkdjkkil/pdf/viewer.html?file=https%3A%2F%2Fassets-dam.takeda.com%2Fimage%2Fupload%2Flegacy-dotcom%2Fsiteassets%2Fen-us%2Fhome%2Fcorporate-responsibility%2Fculture-of-compliance%2Fstate%2Fct%2FCT_fruzaqla-price-form.pdf

  51. Biller LH, Schrag D. Diagnosis and treatment of metastatic colorectal cancer: A review. JAMA. 2021;325(7):669–85.

    Article  PubMed  CAS  Google Scholar 

  52. Zielińska A, et al. Management of pain in colorectal cancer patients. Crit Rev Oncol Hematol. 2021;157:103122.

    Article  PubMed  Google Scholar 

  53. Cho SK, Hay JW, Barzi A. Cost-effectiveness analysis of regorafenib and TAS-102 in refractory metastatic colorectal cancer in the united States. Clin Colorectal Cancer. 2018;17(4):e751–61.

    Article  PubMed  Google Scholar 

  54. Lakshmi A, et al. NICE 2022 guidelines on the management of melanoma: update and implications. J Plast Reconstr Aesthet Surg. 2023;85:401–13.

    Article  PubMed  Google Scholar 

  55. Claxton K, Martin S, Soares M, et al. Methods for the Estimation of the NICE cost-effectiveness threshold. Health Technol Assess. 2015;19(14):1–504.

    Article  PubMed  PubMed Central  Google Scholar 

  56. Santos AS, Guerra-Junior AA, Godman B, Morton A, Ruas CM. Cost-effectiveness thresholds: methods for setting and examples from around the world. Expert Rev Pharmacoecon Outcomes Res. 2018;18(3):277–88.

    Article  PubMed  Google Scholar 

  57. Clement FM, et al. Using effectiveness and cost-effectiveness to make drug coverage decisions: a comparison of Britain, Australia, and Canada. JAMA. 2009;302(13):1437–43.

    Article  PubMed  CAS  Google Scholar 

  58. Liu GG, Hu SL, Wu JH, et al. China guidelines for Pharmacoeconomic evaluations 2020. China Market; 2020.

  59. Neumann PJ, Cohen JT, Weinstein MC. Updating cost-effectiveness—The curious resilience of the $50,000-per-QALY threshold. N Engl J Med. 2014;371(9):796–7. https://doiorg.publicaciones.saludcastillayleon.es/10.1056/NEJMp1405158.

    Article  PubMed  CAS  Google Scholar 

  60. Freeman SC, Sutton AJ, Cooper NJ. Uptake of methodological advances for synthesis of continuous and time-to-event outcomes would maximize use of the evidence base. J Clin Epidemiol. 2020;124:94–105.

    Article  PubMed  PubMed Central  Google Scholar 

Download references

Funding

This work was supported by the General Program of National Natural Science Foundation of China (grant no. 72174207).

Author information

Author notes

  1. Mingye Zhao and Yunlin Jiang contributed equally to this work.

Authors and Affiliations

  1. Department of Pharmacoeconomics, School of International Pharmaceutical Business, China Pharmaceutical University, Nanjing, Jiangsu, China

    Mingye Zhao & Wenxi Tang

  2. Center for Pharmacoeconomics and Outcomes Research, China Pharmaceutical University, Nanjing, Jiangsu, China

    Mingye Zhao & Wenxi Tang

  3. Yueyang Hospital of Integrated Traditional Chinese and Western Medicine, Shanghai University of Traditional Chinese Medicine, Shanghai, China

    Yunlin Jiang

  4. School of Public Health, Faculty of Medicine, The Chinese University of Hong Kong, Ma Liu Shui, China

    Taihang Shao

  5. Center for Pharmacoeconomics and Outcomes Research, Department of Public Affairs Management, School of International Pharmaceutical Business, China Pharmaceutical University, Nanjing, Jiangsu, China

    Wenxi Tang

Authors
  1. Mingye Zhao
    View author publications

    You can also search for this author inPubMed Google Scholar

  2. Yunlin Jiang
    View author publications

    You can also search for this author inPubMed Google Scholar

  3. Taihang Shao
    View author publications

    You can also search for this author inPubMed Google Scholar

  4. Wenxi Tang
    View author publications

    You can also search for this author inPubMed Google Scholar

Contributions

Prof. Tang, Mr Zhao, Dr Jiang and Mr Shao had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Mr Zhao and Dr Jiang contributed equally to this work. Concept and design: Tang; Acquisition of data: Zhao, Jiang; Analysis and interpretation of data: Zhao, Shao; Drafting of manuscript: Zhao, Jiang; Critical revision of the manuscript for important intellectual content: Tang, Zhao, Jiang, Shao; Statistical analysis: Zhao, Jiang; Obtaining funding: Tang; Administrative and technical support: Tang; Supervision: Tang.

Corresponding author

Correspondence to Wenxi Tang.

Ethics declarations

Ethical approval

This study does not involve human participants or animal subjects.

Patient consent for publication

Not required.

Provenance and peer review

Not commissioned; externally peer reviewed.

Patient and public involvement

No patient involved.

Permission to reproduce material from other sources

NA.

Competing interests

The authors declare no competing interests.

Additional information

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Supplementary Material 1

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Zhao, M., Jiang, Y., Shao, T. et al. Safety, efficacy, and cost-effectiveness evaluation of systemic treatments for refractory colorectal cancer: a systematic review and modeling study. Health Econ Rev 15, 33 (2025). https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s13561-025-00622-x

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s13561-025-00622-x

Keywords

Health Economics Review

ISSN: 2191-1991

Contact us