Efficacy, safety and cost-effectiveness of chemo-immunotherapy combinations for recurrent or metastatic nasopharyngeal carcinoma: an updated systematic review and cost-effectiveness analysis

Introduction

Nasopharyngeal carcinoma (NPC) is a malignant epithelial tumor that originates in the nasopharynx and is highly prevalent in East and Southeast Asia (1). At the time of primary diagnosis, approximately 6% to 10% of NPC cases exhibit distant metastasis (2,3). Despite undergoing systemic treatment, patients with locally advanced NPC continue to face significant challenges, including local recurrence in 10% of cases and distant metastasis in 10–20% of cases (4). The prognosis for patients diagnosed with recurrent or metastatic nasopharyngeal carcinoma (R/M NPC) remains dismal, as evidenced by a median overall survival (mOS) of only around 10–15 months (1), highlighting the urgent need to improve outcomes for this specific patient population.

Since 2016, the GP regime has emerged as the standard first-line treatment for R/M NPC due to positive outcomes in the GEM20110714 phase III trial (5,6). However, its efficacy is limited, indicated by a median progression-free survival (mPFS) of just 7.0 months in data released in 2016. The application of immune checkpoint inhibitors (ICIs) brings hope for extending the survival of cancer patients (7-9). Clinical trials such as Keynote 040, Keynote 048, Eagle, Condor, Checkmate 141, have demonstrated the clinical benefits of ICIs for patients with R/M head and neck carcinoma (10-14). Similarly, toripalimab, camrelizumab and tislelizumab plus gemcitabine and cisplatin (TorGP, CamGP and Tis ) have received first-line approval for R/M NPC based on the findings of clinical trials including CAPTAIN-1^st (15), JUPITER-02 (16,17), and RATIONALE 309 (18). All three were randomized, double-blind, placebo-controlled Phase III trials conducted across multiple centers in NPC endemic regions.

In recent years, Chinese researchers have increasingly focused on globalization. Through rigorous scientific research, several programmed cell death protein 1 (PD-1) inhibitors have successfully penetrated international markets and obtained regulatory approvals for specific indications. For instance, based on the findings from the RATIONALE 302 study, tislelizumab has been granted approval in the European Union for treating adult patients with unresectable, locally advanced or metastatic esophageal squamous cell carcinoma (ESCC) who have previously undergone platinum-based chemotherapy. This serves as an affirmation to the quality of the original Chinese medication. While immunotherapy may prolong survival and improve quality of life, it raises treatment-related costs due to the need for prolonged maintenance therapy and immune-related toxicity management (19,20). Evaluating the value of any therapeutic intervention necessitates an economic evaluation to ensure justifiable expenses.

Han et al. conducted a study evaluating both the efficacy and cost-effectiveness of first-line chemo-immunotherapy regimens for R/M NPC (21). Although existing network meta-analysis (NMA) and cost-effectiveness analysis (CEA) in R/M NPC have demonstrated that TorGP regimen represents the most cost-effective therapeutic option, critical evidence gaps remain that limit their current clinical applicability (21). Specifically: (I) these studies lacked comprehensive safety comparisons across the three key clinical trials; (II) their economic evaluations were conducted prior to the significant price reduction of PD-1 inhibitors in 2023, fundamentally altering cost-benefit calculations. Our study bridges these evidence gaps through: (I) our study establishes a hierarchical ranking of regimens based on efficacy, safety, and cost-effectiveness, providing clinicians with a practical framework to weigh therapeutic benefits against toxicity and cost; (II) conducting updated pharmacoeconomic modeling that incorporates both the post-price-reduction drug costs and three different analytical models for scenario analyses.

Consequently, we conduct an NMA and CEA to evaluate the efficacy, safety, and cost-effectiveness of first-line chemo-immunotherapy regimes for patients with R/M NPC from a public health standpoint in China. This comprehensive evaluation provides valuable information to clinicians, enabling them to improve outcomes for this challenging patient population. We present this article in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) reporting checklist (22) (available at https://tcr.amegroups.com/article/view/10.21037/tcr-2025-415/rc).

Methods

The study protocol has been formally registered with PROSPERO under the registration number CRD42023422137. This study systematically evaluates efficacy, safety, and cost-effectiveness parameters across three pivotal randomized controlled trials (RCTs) in R/M NPC (JUPITER-02, RATIONALE 309, and CAPTAIN-1st).

Efficacy and safety

Data sources and search methodology

A comprehensive systematic search was conducted then across several databases, including PubMed, Embase, the Cochrane Library, and ClinicalTrials.gov, covering all records from their inception through June 20, 2025. Additionally, online proceedings from annual conferences were searched to ensure up-to-date outcomes. The approach for the literature search is described in Table S1. The criteria for inclusion and exclusion are outlined in Table S2.

Data retrieval and evaluation of quality

Two investigators conducted independent literature screening and data extraction, with arbitration by a senior researcher to resolve discrepancies.

After deduplication in EndNote X9.3.3, two-stage screening was implemented: initial exclusion based on title or abstract review, followed by full-text assessment of potentially eligible studies. Data extraction covered: (I) basic information including study phase, methodological design, bibliographic source with publication year, and registered National Clinical Trials identifier; (II) demographic parameters encompassing randomization, sample size, Eastern Cooperative Oncology Group (ECOG) performance score, histologic type, ethnic background, and the distribution of patients by age and gender (Table S3); (III) therapeutic protocols detailing pharmacological agents and administration regimens; (IV) outcome indicators including hazard ratios (HRs) with 95% confidence intervals (CIs) for OS, PFS and objective response rate (ORR) (Table S4), incidences of treatment-emergent adverse events (TEAEs) (Table S5) and immune-related TEAEs (irTEAEs) (Table S6). Methodological rigor of incorporated RCTs was assessed utilizing the Cochrane Risk of Bias Tool (RoB 2.0) (23).

Statistical analysis

Statistical computations were executed with RStudio version 4.2.3, employing the gemtc package (version 1.0-0). The outcome indicators included survival data (PFS and OS) and binary variable data [ORR and adverse events (AEs)], in which the primary outcome was PFS, and the others were secondary outcomes. HRs with 95% CIs were used as the effect sizes for OS and PFS, while odds ratios (ORs) with 95% CIs were used for ORR and grade ≥3 TEAEs (24,25).

Rankings were derived using the surface under the cumulative ranking curve (SUCRA), a probabilistic transformation of mean rank values. A greater SUCRA value signified that the treatment held a better rank (26,27).

Assumptions

In order for NMA to be valid, it must simultaneously satisfy three assumptions: homogeneity, transitivity, and consistency (28).

• Consistency: when studies meet the assumption of consistency, the results from both direct and indirect comparisons can be merged. If the consistency hypothesis is not met, it is necessary to investigate the reasons for the inconsistency, such as potential differences in clinical and methodological characteristics. Subgroup analysis should be conducted if necessary.
• Homogeneity: the degree of heterogeneity across studies was assessed using the inconsistency index (I²) text (29). Consistent with Cochrane guidelines, substantial heterogeneity was defined as I²>50%. In such cases, we applied random-effects modeling and conducted sensitivity analyses to address variability. Otherwise, it indicated an acceptable level of heterogeneity and a fixed effects model would be applied.
• Transitivity: there is no universally accepted statistical test that mandates the similarity of key characteristics for accurate analysis.

CEA

The CEA was guided by Consolidated Health Economic Evaluation Reporting Standards (CHEERS) 2022 Explanation and Elaboration, a standardized guideline for reporting economic evaluations (30). The health care system was chosen as the perspective of the evaluation. Both costs and outcomes were discounted at an annual rate of 5% (31-33). The analysis evaluated effectiveness by measuring quality-adjusted life-years (QALYs). The economic evaluation employed lifetime incremental cost-effectiveness ratios (ICERs), using a willingness-to-pay (WTP) threshold of $40,354.20 per QALY, which corresponds to three times China’s per capita gross domestic product (GDP) in 2024 (34,35).

Model construction

The partitioned survival model (PSM) was developed to evaluate cost-effectiveness utilizing Microsoft Excel (version 2019) (36), whose cycle length was set at 3 weeks, with a 10-year horizon according to the treatment cycle. Three states, PFS, progressive disease (PD) and death, were employed to model the developmental progression of R/M NPC. The PFS phase was further segmented into three intervals—0 to 4.5 months, 4.5 to 28.5 months, and beyond 28.5 months—based on the treatment regimen and frequency of follow-up assessments.

Health state proportions were calculated as follows: for PFS patients, the number was derived from the PFS curves; for PD patients, the proportion was computed as the difference between OS and PFS curves; for death, the proportion of patients who had died was calculated as one minus the proportion of patients who were still alive based on the OS curves. To adjust for OS and PFS, age-specific mortality rates of the general population were obtained from the seventh national census data provided by the National Bureau of Statistics.

We proposed that patients with R/M NPC were male, 50 years old, 65 kg in weight, and had a body surface-area of 1.72 m² (15,16,18). Patients were randomly assigned to four groups: TorGP, CamGP, TisGP, and GP. To facilitate the modelling process, drug doses, administration methods, and treatment durations for the four groups were obtained from three RCTs (15,16,18) (Table S7).

Model transition and survival estimates

Although the final data on JUPITER-02 were reported at the American Association for Cancer Research (AACR) Annual Meeting (17), to minimize the effect of median follow-up time on the HR and 95% CI of mPFS and mOS, we selected similar median follow-up times for the extracted data (Table S1). For survival fitting, we used the JUPITER-02 trial as the anchor due to its maximum sample size and a relatively high degree of maturity.

The proportional hazard (PH) assumption is verified using Schoenfeld’s test, with a P value >0.05 indicating that the assumption was met. Therefore, our study is consistent with the PH tests (Figure S1). Data points were extracted using GetData Graph Digitizer software (version 2.25) to rebuild individual patient data.

It is important to note that the likelihood of a cure in advanced cancer is generally low. Therefore, the mixture cure model may tend to overestimate treatment effectiveness in patients with advanced cancer. The landmark model needs to divide the population into several subgroups according to the characteristics of the patients, which is not explored because there is no relevant data available at present (37).

A range of standard parametric survival models, the Log-normal, Weibull, Exponential, Log-logistic and Gompertz distributions, were fit to the Kaplan-Meier (KM) curves (38,39). In the scenario analysis, we incorporated the restricted cubic splines (RCS) and mixture cure model (MCM).

However, heterogeneity between studies may increase if different types of distribution parameters are used in each study, so care needs to be taken to ensure applicability and comparability of the chosen distribution parameter type. In this study, the most appropriate practice was followed, where the lowest Akaike’s information criterion (AIC) and Bayesian information criterion (BIC) were selected as measures of the best fit quality.

Standard parametric survival model

For the OS curves of TorGP, we obtained the smallest AIC and BIC values in the exponential and Log-logistic, respectively, and we subsequently selected the best-fitting KM curves by visual examination (Figure S2). Consequently, for the TorGP group our selections were Log-logistic for OS and Weibull for PFS, respectively. Similarly, for the GP group our selections were Log-normal for OS and Log-logistic for PFS (Table S8).

Gianluca Baio made a point that Maximum Likelihood Estimation (MLE) is simpler than Bayesian methods in curve extrapolation (40). We use MLE for KM curve extrapolation and subsequently adjusted the curves for CamGP and TisGP according to the HR values, and the results were listed in Figure S3.

Restricted cubic splines models

Similarly, we selected the model with the smallest AIC value, which is an indicator of goodness of fit (Table S9).

Costs and utilities

This analysis exclusively accounted for direct medical costs from the standpoint of the Chinese healthcare system. Medication expenses were sourced from Menet (https://www.menet.com.cn). Management expenses, encompassing costs associated with intravenous infusions and hospitalization, as well as follow-up expenses, including laboratory and imaging examinations, were extracted from the database of our institution in 2025.

In clinical practice, interventions and treatments are usually focused on TEAEs of grade ≥3. Therefore, in the included trials, TEAEs of grade ≥3 with an incidence rate greater than 5% were considered, taking into account adverse reactions associated with different drugs (15,16,18). The management of TEAEs was based on literatures and expert opinions.

The incidence of grade ≥3 TEAEs in the GP group was calculated as the mean across three RCTs (Table S5). We assumed that all AEs occurred one time at the initiation of the simulation. Hospice care expenditures were derived from a cost-utility analysis of palliative care (41). All expenses were converted into US dollars (1 USD =7.15 RMB, July, 2025). Each PD-1 inhibitor was given every three weeks and was available for up to 2 years, or until the disease progressed or severe AEs occurred. Details are shown on Table S7.

The follow-up treatment regime for R/M NPC patients when the disease progresses has not been clearly defined in the three RCTs. In response to the 2024 Chinese Society of Clinical Oncology (CSCO) guideline updates that equally recommend capecitabine, docetaxel, and gemcitabine as Category 2A options for second-line therapy, we have conducted comprehensive scenario analyses to ensure the robustness of our findings (42-45).

A phase 3 trial in metastatic NPC, NCT02460419, demonstrated the efficacy of capecitabine maintenance therapy with an improved mPFS of 27.7 months compared to the best supportive care group (35.9 vs. 8.2 months, HR =0.44, 95% CI: 0.26–0.74) (46). Therefore, we modeled the use of capecitabine mono-chemotherapy as the follow-up treatment when disease progression occurred. To mitigate the oral toxicity of capecitabine, oral care was supplemented twice daily on days 1–14. Additionally, this study incorporated docetaxel and gemcitabine as second-line therapy options in the scenario analysis. The recommended regimens for second-line treatment of R/M NPC include docetaxel administered at 30 mg/m² via intravenous infusion on days 1, 8, and 15 of a 28-day cycle; concurrently, gemcitabine is recommended at 1,000 mg/m² as a 30-minute intravenous infusion on the same schedule.

Detailed usage information can be found in Table S7.

Given regional heterogeneity in China’s medical insurance reimbursement schemes, we did not consider preferential policies.

Utility measures patients’ quality of life on a scale ranging from 1, representing perfect health, to 0, representing death. A utility value of 0.65 was assigned to patients in the PFS state, while a utility value of 0.52 was assigned to patients in the PD state, based on data from an CEA for R/M NPC. In addition, to incorporate the impact of AEs on utility values, a decrement in utility was calculated using disutility values for AEs. The parameters for the disutility of each AEs were obtained from published literature (47). Summary of model inputs and ranges are shown on Table 1.

Sensitivity analysis

We conducted both one-way and probabilistic sensitivity analyses (PSA) to evaluate the robustness of the model (48,49). One-way sensitivity analyses were performed to evaluate the impact of each input parameter individually, with the results displayed using tornado diagrams. Additionally, a PSA was carried out to assess all input parameters simultaneously, using 1,000 Monte Carlo simulations based on their statistical distributions. The outcomes were presented as scatter plots and cost-effectiveness acceptability curves (CEAC).

Results

Characteristics of the included studies

Our initial literature search yielded 278 records from various databases. After screening abstracts and eliminating duplicate and irrelevant studies, 12 articles met our inclusion criteria. Three RCTs (JUPITER-02, CAPTAIN-1st, and RATIONALE 309) were selected after a full-text review. Updated data from JUPITER-02 were presented at an AACR meeting (15-18). The flowchart is shown in Figure S4. The evaluation of risk of bias by RoB2 is shown in Figure S5.

Comparisons of efficacy

The NMA assessed four treatment regimes, TorGP, CamGP, TisGP, and GP, for efficacy in terms of primary outcome (PFS) and secondary outcomes (OS and ORR).

Regarding PFS, chemo-immunotherapy consistently demonstrated better PFS than standard chemotherapy. In this case, since the I² value is 33%, which is less than 50%, the fixed effects model was used for the analysis. Among the chemo-immunotherapy regimes, TisGP provided the best PFS benefit compared to GP (HR =0.5, 95% CI: 0.37–0.68), followed by CamGP (HR =0.51, 95% CI: 0.37–0.69) and TorGP (HR =0.52, 95% CI: 0.36–0.74), despite there being no statistically significant difference, as shown by the league table (Figure 1).

Figure 1 Efficacy and safety profiles of the Bayesian network meta-analysis in patients with R/M NPC. Hazard ratios and 95% CIs for overall survival and progression-free survival, and a hazard ratio <1.00 provides better survival benefits. ORs and 95% CIs for objective response rate and grade ≥3 adverse events, and an OR <1.00 indicates a better efficacy or safety (from top to bottom). CIs, confidence intervals; ORs, odds ratios; R/M NPV, recurrent or metastatic nasopharyngeal carcinoma; TEAEs, treatment-emergent adverse events

For OS, chemo-immunotherapy regimes showed greater OS benefit than chemotherapy alone, with TorGP offering marked benefits compared to GP (HR =0.6, 95% CI: 0.36–0.99). SUCRA ranking indicated that TisGP, TorGP, CamGP, and GP ranked from first to fourth, respectively (Figure 1).

For ORR, the fixed effects model was used due to I²=17%. Although there were no statistically significant differences, TisGP was ranked first, followed by CamGP, TorGP and GP according to the SUCRA ranking (Figure 1). TisGP was observed to be the best compared to GP (OR =0.54, 95% CI: 0.32–0.9), followed by CamGP (OR =0.6, 95% CI: 0.3–1.17) and TorGP (OR =0.61, 95% CI: 0.37–1.02).

Safety and toxicity

Firstly, the pooled incidence of grade 1–5/3–5 AEs in TorGP, CamGP, and TisGP were summarized, which were 100%/89%, 100%/94%, and 100%/80.9%, respectively. TEAE leading to death and permanent discontinuation of all treatments in the three regimes were 2.7%/7.5%, 4%/9%, and 3.8%/1.5% as shown in Table S10. Since the probability of grade 1–5 TEAEs was almost 100% in both experimental and control groups in the three RCTS, only grade ≥3 TEAEs were included in our NMA. As I² was 17%, the fixed effects model was used. The safety results showed that TisGP, TorGP, GP, and CamGP were ranked from first to fourth (Figure 1). Surprisingly, the results showed that immune-chemotherapy combinations, except for CamGP, did not increase the risk of AEs compared to standard chemotherapy, as shown in Figure 1.

The TEAEs that were frequently reported for the chemo-immunotherapy included bone marrow suppression, nausea and vomiting. The frequently reported irTEAEs were hypothyroidism, rash, pruritis, alanine aminotransferase (ALT) increased, aspartate aminotransferase (AST) increased and anemia, as shown in Figure 2.

Figure 2 Incidence of all grade (A), grade ≥3 (B) and immune related (C) TEAEs. ALTI, alanine aminotransferase increased; ASTI, aspartate aminotransferase increased; ATAP, anti-thyroid antibody positive; BCI, blood creatinine increased; BTSHI, blood thyroid stimulating hormone increased; Cam, camrelizumab; NCD, neutrophil count decreased; NSD, nervous system disorders; PCD, platelet count decreased; RCEP, reactive capillary endothelial proliferation; TEAEs, treatment-emergent adverse events; Tis, tislelizumab; Tor, toripalimab; WBCCD, white blood cell count decreased.

Differences in the probabilities of these specific TEAEs exist among chemo-immunotherapy regimes. CamGP exhibited the highest rates of bone marrow suppression and hypothyroidism, and it was the only regime associated with reactive capillary endothelial proliferation. TorGP exhibited the highest probability of causing vomiting, abnormal liver function and pyrexia. In contrast, TisGP was associated with the least risk of almost all recorded TEAEs (Figure 2A). The incidence of all grade TEAEs such as hypokalemia, increased blood creatinine, hypochloremia, and lymphopenia had the narrowest incidence spectrums as depicted in Figure 2A. As shown in Figure 2C, when it came to irTEAEs, CamGP was associated with more occurrences, whereas TisGP not only occurred in fewer types but also occurred less frequently.

Rankings

SUCRA values for each group were presented in a radar chart (Figure S6). From the picture, it can be seen that the performance of TisGP ranked first in terms of PFS, OS, ORR, and safety.

Heterogeneity, inconsistency, and transitivity assessment

Since the I² results were found to be less than 50%, there was no significant heterogeneity detected. As our experimental model was an open-loop design, inconsistency detection was not necessary (27). Similarity in the clinical characteristics was observed across all the included studies, indicating that the transitivity of the trials was acceptable.

The trajectory and density plots were used to evaluate the convergence results of the model. From the trajectory plot in Figure S7, it can be seen that when the number of iterations reaches more than 5,000, the Markov Chain Monte Carlo (MCMC) chain exhibits stable fluctuations with good overlap. According to the density plot, when the number of iterations reaches 20,000, the bandwidth tends to 0 and achieves stability, indicating that the model converges well. The Brooks-Gelman-Rubin diagnostic also revealed the stability and replicability of the inferential iterations for each MCMC chain, as shown in Figure S8.

Base case results of CEA

K-M and parametric survival distributions are shown at Figure S9.

CamGP, TorGP and TisGP were respectively associated with an ICER of $24,331, 18,776 and 20,762/QALY compared to GP. TorGP and TisGP provided an ICER of $10,567 and 5,880/QALY when compared with CamGP. On comparing with TorGP, TisGP provided an incremental cost of $1,700 and additional 0.03 QALYs, resulting in an ICER of $67,564/QALY, with a WTP threshold of $40,354.20/QALY. Hence, TorGP demonstrated the best cost-effectiveness in the base case analysis. We reach the same conclusion in RCS and MCM models that TorGP is the most cost-effectiveness. The consistent findings from our RCS and MCM approaches robustly identified TorGP as the optimal cost-effective therapeutic strategy (Table 2).

Sensitivity analysis of CEA

The one-way sensitivity analysis was illustrated using a tornado diagram (Figure S10). The discount rate, utility of PD, and utility of PFS were found to have a significant influence on the result of cost-effectiveness. Other factors considered in the analysis, such as the cost of administration, cost of end-of-life care, and incidence of AEs, had relatively little impact on the outcomes.

The PSA results are depicted in cost-effectiveness scatterplot (Figure 3A) and CEAC (Figure 3B). The scatter plots representing the ICERs of TorGP, CamGP, and TisGP were all above 1-time GDP per capita and below 3-time GDP per capita. This indicated that the probability for these regimes to be deemed cost-effective when compared with GP regime was 0% at a WTP threshold of 1-time GDP per capita but be cost-effective compared to GP at a WTP threshold of 3-time GDP per capita. As demonstrated in the CEAC, the probability that TorGP was cost-effective increased as the WTP for additional QALY rose. The probability of TorGP to be cost-effective under a WTP of 3-time GDP per capita was 100%. Consistent with the advantage shown in the base-case analysis, TorGP appears to have the greatest potential for cost-effectiveness in the PSA.

Figure 3 Probabilistic sensitivity analysis. Cam, camrelizumab; GDP, gross domestic product; GP, gemcitabine and cisplatin; QALYs, quality-adjusted life-years; Tis, tislelizumab; Tor, toripalimab.

Scenario analysis for second-line therapy

Evidence evaluation frameworks consistently classify capecitabine, docetaxel, and gemcitabine as Category 2A recommended regimens for second-line management of R/M NPC. Recognizing that clinical selection among these regimens often depends more on individual patient characteristics and institutional prescribing patterns than on substantial efficacy differences, our study incorporated docetaxel and gemcitabine agents into comprehensive scenario analyses. The results demonstrated that the choice of second-line therapy had limited impact on clinical outcomes, with consistent conclusions indicating TorGP as the most cost-effective treatment strategy (Table S11).

Discussion

Our research suggests that TisGP shows a clinically favorable trend in efficacy and safety, TorGP demonstrates superior cost-effectiveness with an incremental cost-effectiveness ratio well below the willingness-to-pay threshold.

When searching on PubMed using the keywords “cost-effectiveness analysis and recurrent metastatic nasopharyngeal carcinoma”, we identified six relevant articles (21,47,50-53). One was excluded because it involved protocols that were excluded from the guidelines (53). Among the remaining five articles, four articles did not include all chemo-immunotherapy regimes (47,50-52).

In contrast to Han et al.’s study (21), firstly, within our study, we conducted a comprehensive evaluation of the drug toxicity of chemo-immunotherapy. Safety rankings of the four regimens and the probabilities of different adverse effects were summarized generically. Secondly, regarding efficacy, in the context where final data from relevant studies have not yet been released, this study’s methodological design attempts to address the issue of varying follow-up durations among included studies, aiming to provide clinicians with more accurate treatment recommendations. Thirdly, in light of recent price negotiations for ICIs, the present analysis utilizes updated 2025 drug pricing to reflect current conditions. Finally, during the model construction process of CEA, we opted for a dynamic partitioned survival model instead of a Markov model. The dynamic partitioned survival model offers advantages by not being bound by transition probabilities between different health states. Hence, it is considered the preferred model for economic evaluations of advanced or metastatic cancer drugs (54). Additionally, we incorporated RCS and MCM as complementary scenario analyses to evaluate the robustness of our findings.

In terms of efficacy, our study provides evidence supporting the use of chemo-immunotherapy as a viable first-line treatment option for patients with R/M NPC. Notably, in our NMA, TisGP demonstrated potential advantages over other regimes. Reasons may be as follows: firstly, tislelizumab, which targets PD-1 and has been engineered to minimize binding to Fcγ receptors on macrophages and reduce the occurrence of antibody-dependent cellular phagocytosis, resulting in a more potent and durable anti-tumor response to anti-PD-1 therapy (55). Additionally, the RATIONALE 309 study allowed the control group to receive sequential crossover immunotherapy after disease progression. In that study, 49% of patients in the control group received single-agent treatment with tislelizumab as subsequent therapy (18). However, this sequential immunotherapy was not permitted in the JUPITER-02 and CAPTAIN-1 trials, but approximately 13% of patients in the control arms of those trials subsequently received immunotherapy (15,16). Despite the differences in trial design, TisGP demonstrated potential advantages in terms of OS compared to the other two regimes. This evidence strengthens the case for considering TisGP as a potentially more effective treatment option for R/M NPC.

In terms of toxicity, a previous study indicated that combining anti-DNA repair agents, such as gemcitabine, with ICIs may be a prospective approach due to its potential to mitigate toxicity associated with the latter (56). In our analysis, chemo-immunotherapy regimes, except for CamGP, did not show a higher risk of AEs than standard chemotherapy, and TisGP may be the best safety profile among all regimes evaluated. Several hypotheses have been proposed to explain this finding, including the role of the tumor microenvironment, expression of immune-related genes, and individual response to ICI drugs. However, the underlying mechanisms driving these hypotheses remain unclear and require further investigation (57-59). Analyzing the toxicity spectrum of immunotherapeutic drugs is crucial in predicting the potential risk for AEs in specific patient populations. In our study, we observed the top three irTEAEs by incidence in each group, with TorGP showing hypothyroidism, rash, and pruritus (17.8%, 8.9%, and 6.2%, respectively). CamGP showed reactive capillary endothelial proliferation, hypothyroidism, and rash as the order of highest incidence (58%, 43%, and 25%, respectively). In the TisGP group, hypothyroidism, rash, and pruritus were observed as the top three irTEAEs (13.7%, 3.8%, and 1.5%, respectively). As we know, radiation exposure after radiotherapy for NPC may induce delayed hypothyroidism (60,61). Therefore, in the case of R/M NPC, almost all patients have undergone radiotherapy of the nasopharynx and neck. It is crucial to pay closer attention to thyroid function when using immune drugs to minimize the risk of exacerbating hypothyroidism. TisGP emerges as a safer treatment option for this particular patient population. As a result, when considering immunotherapy for patients, clinicians should take into account their underlying conditions. It is crucial to analyze the toxicity profile of immunotherapeutic drugs in order to guide clinicians in selecting suitable treatment regimens and closely monitoring patient responses during therapy to minimize the occurrence of AEs. This comprehensive approach ensures that the benefits of immunotherapy are maximized while minimizing potential risks.

There are four main types of economic evaluations: cost-minimization (CMA), CEA, cost-utility (CUA), and cost-benefit analysis (CBA). CMA measure the difference in costs between alternative interventions. When comparing the interventions, the assumption is made that the alternatives are equally effective, and therefore the difference is in costs only (62). ANDREW highlighted that the use of CMA is rarely appropriate as a method of analysis (63). Additionally, in two books, authors emphasized that CMA should not be considered a suitable form of full economic evaluation. One said, “it is sometimes argued that if the two or more alternatives under consideration achieve the given outcome to the same extent, a CMA can be performed. However, it is not appropriate to view CMA as a form of full economic evaluation.” (64). The other said, “A CMA is a costing exercise and not a formal economic evaluation. As such, a CMA is not an appropriate reference case analysis.” (65). CEA is a widely used and well-established method employed by various health technology assessment (HTA) agencies to evaluate the cost-effectiveness of healthcare interventions. Its primary objective is to assess whether the health benefits gained from an intervention justify the associated costs (65-67). Based on the considerations discussed above, the research method employed in this study is CEA.

In 2022, two CEAs were conducted around JUPITER-02 and CAPTAIN-1st, and the results indicated that TorGP was a more cost-effective regimen compared to GP and CamG (47,50). Tian et al. reported ICERs of $6,696/QALY and $55,305/QALY for TorGP and CamGP, respectively, when compared with GP. This analysis suggested that CamGP was not a cost-effective option compared to GP (50). However, another study by Zhu et al. showed that CamGP had favorable economic benefits compared to GP (47). The differences in these findings may be attributed to several factors, including differences in follow-up treatments used in the two studies and a distinct difference in the risk of major AEs in CamGP groups. Additionally, the higher price of camrelizumab in the study by Tian et al. may have led to an overestimation of costs.

The recent price reduction of immunotherapy agents calls for an update to previous cost-effectiveness analyses. In our base-case analysis, TorGP demonstrated the highest cost-effectiveness compared to TisGP, CamGP, and GP for treating R/M NPC with a WTP threshold of $40,354.20/QALY. We propose that several factors may have contributed to the lower costs observed in our study compared to Zhu et al.’s study. First, we suggest that the general price reduction of immunosuppressive drugs following the update of Medicare prices may have had a significant impact on costs. Second, we opted for more cost-effective drugs to manage AEs that were comparable in effectiveness to those employed in other studies. Because the treatment of side effects was standardized, the ICER can be better compared by uniform treatment.

Although TisGP demonstrated greater QALYs and lower AEs management costs, the superior cost-effectiveness of TorGP—driven by toripalimab’s significantly more competitive pricing—ultimately established it as the more economically favorable regimen. These findings underscore the importance of clinicians considering patients’ financial circumstances, clinical status, and treatment tolerance when selecting therapy.

In the management of R/M NPC, current guidelines classify capecitabine, docetaxel, and gemcitabine as Category 2A recommended options for second-line therapy. Recognizing that clinical selection among these regimens often depends more on individual patient characteristics and institutional prescribing patterns than on substantial efficacy differences, our study incorporated all three agents into comprehensive scenario analyses. The results of these analyses provide a more nuanced understanding of cost-effectiveness across diverse practice settings, particularly relevant given the recent expansion of treatment options in this clinical space.

We are focusing on the global expansion of Chinese-manufactured PD-1 inhibitors. On September 19, 2023, Beigene announced that its self-developed anti-PD-1 antibody tislelizumab had been approved by the European Union, becoming China’s first “overseas” product, although it is not yet approved for NPC. On October 27, 2023, toripalimab, an independently developed drug from China, achieved regulatory approval by the FDA for the treatment of NPC and made its entry into the United States market. Currently, toripalimab stands as the first and only drug authorized for comprehensive therapy against advanced NPC in the United States. However, camrelizumab may pose challenges to its international expansion due to adverse reactions such as capillary proliferation. The anticipation is that a greater number of domestic pharmaceuticals will be able to successfully penetrate the international market and gain global recognition.

Limitations

First, because quality-of-life data were not available from the three RCTs, the utility and disutility used in our analysis was drawn from published literature on similar populations. Second, it is worth noting that although the study concept is novel and useful, there were only three RCTs that met the eligibility criteria, and all of them were conducted by the same institution, leading to concerns about potential bias. However, in this case, the I² results were less than 50%, indicating an acceptable level of heterogeneity. Finally, while updated data from the JUPITER-02 trial are now available, we opted for median follow-up times that were expected to be comparable to mitigate the influence of differences in median follow-up times. Our model is likely to become more precise once the final data from CAPTAIN-1st and RATIONALE 309 are released.

Conclusions

In summary, while TisGP shows a clinically favorable trend in efficacy and safety, TorGP demonstrates superior cost-effectiveness with an incremental cost-effectiveness ratio well below the willingness-to-pay threshold. This finding suggests that clinicians should comprehensively evaluate individual patient characteristics when selecting treatment regimens.

Acknowledgments

None.

Footnote

Reporting Checklist: The authors have completed the PRISMA reporting checklist. Available at https://tcr.amegroups.com/article/view/10.21037/tcr-2025-415/rc

Peer Review File: Available at https://tcr.amegroups.com/article/view/10.21037/tcr-2025-415/prf

Funding: This study was supported by Jiangsu Province Entrepreneurship and Innovation Doctoral Talent Program (No. 2019303073386ER19, to S.Z.) and National Natural Science Foundation of China (No. 82003235, to S.Z.).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://tcr.amegroups.com/article/view/10.21037/tcr-2025-415/coif). S.Z. received grants from Jiangsu Province Entrepreneurship and Innovation Doctoral Talent Program (No. 2019303073386ER19) and National Natural Science Foundation of China (No. 82003235). The other authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. As this study involves a retrospective analysis and does not involve the use of individual patient data, it is exempt from ethical review and approval.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Chen YP, Chan ATC, Le QT, et al. Nasopharyngeal carcinoma. Lancet 2019;394:64-80. [Crossref] [PubMed]
2. You R, Liu YP, Huang PY, et al. Efficacy and Safety of Locoregional Radiotherapy With Chemotherapy vs Chemotherapy Alone in De Novo Metastatic Nasopharyngeal Carcinoma: A Multicenter Phase 3 Randomized Clinical Trial. JAMA Oncol 2020;6:1345-52. [Crossref] [PubMed]
3. Tan WL, Tan EH, Lim DW, et al. Advances in systemic treatment for nasopharyngeal carcinoma. Chin Clin Oncol 2016;5:21. [Crossref] [PubMed]
4. Liu Z, Chen Y, Su Y, et al. Nasopharyngeal Carcinoma: Clinical Achievements and Considerations Among Treatment Options. Front Oncol 2021;11:635737. [Crossref] [PubMed]
5. Zhang L, Huang Y, Hong S, et al. Gemcitabine plus cisplatin versus fluorouracil plus cisplatin in recurrent or metastatic nasopharyngeal carcinoma: a multicentre, randomised, open-label, phase 3 trial. Lancet 2016;388:1883-92. [Crossref] [PubMed]
6. Hong S, Zhang Y, Yu G, et al. Gemcitabine Plus Cisplatin Versus Fluorouracil Plus Cisplatin as First-Line Therapy for Recurrent or Metastatic Nasopharyngeal Carcinoma: Final Overall Survival Analysis of GEM20110714 Phase III Study. J Clin Oncol 2021;39:3273-82. [Crossref] [PubMed]
7. Singh S, Hassan D, Aldawsari HM, et al. Immune checkpoint inhibitors: a promising anticancer therapy. Drug Discov Today 2020;25:223-9. [Crossref] [PubMed]
8. Pu Y, Ji Q. Tumor-Associated Macrophages Regulate PD-1/PD-L1 Immunosuppression. Front Immunol 2022;13:874589. [Crossref] [PubMed]
9. Kaushik I, Ramachandran S, Zabel C, et al. The evolutionary legacy of immune checkpoint inhibitors. Semin Cancer Biol 2022;86:491-8. [Crossref] [PubMed]
10. Cohen EEW, Soulières D, Le Tourneau C, et al. Pembrolizumab versus methotrexate, docetaxel, or cetuximab for recurrent or metastatic head-and-neck squamous cell carcinoma (KEYNOTE-040): a randomised, open-label, phase 3 study. Lancet 2019;393:156-67.
11. Burtness B, Harrington KJ, Greil R, et al. Pembrolizumab alone or with chemotherapy versus cetuximab with chemotherapy for recurrent or metastatic squamous cell carcinoma of the head and neck (KEYNOTE-048): a randomised, open-label, phase 3 study. Lancet 2019;394:1915-28.
12. Ferris RL, Haddad R, Even C, et al. Durvalumab with or without tremelimumab in patients with recurrent or metastatic head and neck squamous cell carcinoma: EAGLE, a randomized, open-label phase III study. Ann Oncol 2020;31:942-50.
13. Siu LL, Even C, Mesía R, et al. Safety and Efficacy of Durvalumab With or Without Tremelimumab in Patients With PD-L1-Low/Negative Recurrent or Metastatic HNSCC: The Phase 2 CONDOR Randomized Clinical Trial. JAMA Oncol 2019;5:195-203. [Crossref] [PubMed]
14. Ferris RL, Blumenschein G Jr, Fayette J, et al. Nivolumab vs investigator’s choice in recurrent or metastatic squamous cell carcinoma of the head and neck: 2-year long-term survival update of CheckMate 141 with analyses by tumor PD-L1 expression. Oral Oncol 2018;81:45-51. [Crossref] [PubMed]
15. Yang Y, Qu S, Li J, et al. Camrelizumab versus placebo in combination with gemcitabine and cisplatin as first-line treatment for recurrent or metastatic nasopharyngeal carcinoma (CAPTAIN-1st): a multicentre, randomised, double-blind, phase 3 trial. Lancet Oncol 2021;22:1162-74. [Crossref] [PubMed]
16. Mai HQ, Chen QY, Chen D, et al. Toripalimab or placebo plus chemotherapy as first-line treatment in advanced nasopharyngeal carcinoma: a multicenter randomized phase 3 trial. Nat Med 2021;27:1536-43. [Crossref] [PubMed]
17. Mai HQ, Chen QY, Chen D, et al. Abstract CT226: Final progression-free survival analysis of JUPITER-02, a randomized, double-blind, phase 3 study of toripalimab or placebo plus gemcitabine and cisplatin as first-line treatment for recurrent or metastatic nasopharyngeal carcinoma. Cancer Res 2022;82:CT226.
18. Yang Y, Pan J, Wang H, et al. Tislelizumab plus chemotherapy as first-line treatment for recurrent or metastatic nasopharyngeal cancer: A multicenter phase 3 trial (RATIONALE-309). Cancer Cell 2023;41:1061-1072.e4. [Crossref] [PubMed]
19. Ghate SR, Li Z, Tang J, et al. Economic Burden of Adverse Events Associated with Immunotherapy and Targeted Therapy for Metastatic Melanoma in the Elderly. Am Health Drug Benefits 2018;11:334-43.
20. Dolgin E. Bringing down the cost of cancer treatment. Nature 2018;555:S26-S29.
21. Han J, Zeng N, Tian K, et al. First-line immunotherapy combinations for recurrent or metastatic nasopharyngeal carcinoma: An updated network meta-analysis and cost-effectiveness analysis. Head Neck 2023;45:2246-58. [Crossref] [PubMed]
22. Page MJ, McKenzie JE, Bossuyt PM, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ 2021;372:n71. [Crossref] [PubMed]
23. Sterne JAC, Savović J, Page MJ, et al. RoB 2: a revised tool for assessing risk of bias in randomised trials. BMJ 2019;366:l4898. [Crossref] [PubMed]
24. Takenaka Y, Oya R, Takemoto N, et al. Predictive impact of sarcopenia in solid cancers treated with immune checkpoint inhibitors: a meta-analysis. J Cachexia Sarcopenia Muscle 2021;12:1122-35. [Crossref] [PubMed]
25. Jácome AA, Castro ACG, Vasconcelos JPS, et al. Efficacy and Safety Associated With Immune Checkpoint Inhibitors in Unresectable Hepatocellular Carcinoma: A Meta-analysis. JAMA Netw Open 2021;4:e2136128. [Crossref] [PubMed]
26. Trinquart L, Attiche N, Bafeta A, et al. Uncertainty in Treatment Rankings: Reanalysis of Network Meta-analyses of Randomized Trials. Ann Intern Med 2016;164:666-73. [Crossref] [PubMed]
27. Beis G, Papasotiriou I. Is Network Meta-analysis a Revolutionary Statistical Tool for Improving the Reliability of Clinical Trial Results? A Brief Overview and Emerging Issues Arising. In Vivo 2023;37:972-84. [Crossref] [PubMed]
28. Watt J, Del Giovane C. Network Meta-Analysis. Methods Mol Biol 2022;2345:187-201. [Crossref] [PubMed]
29. Higgins JP, Thompson SG, Deeks JJ, et al. Measuring inconsistency in meta-analyses. BMJ 2003;327:557-60. [Crossref] [PubMed]
30. Husereau D, Drummond M, Augustovski F, et al. Consolidated Health Economic Evaluation Reporting Standards (CHEERS) 2022 Explanation and Elaboration: A Report of the ISPOR CHEERS II Good Practices Task Force. Value Health 2022;25:10-31. [Crossref] [PubMed]
31. She L, Tian K, Han J, et al. Cost-effectiveness analysis of metronomic capecitabine as adjuvant chemotherapy in locoregionally advanced nasopharyngeal carcinoma. Front Oncol 2022;12:904372. [Crossref] [PubMed]
32. Yang L, Wu J. Cost-effectiveness of rivaroxaban compared with enoxaparin plus warfarin for the treatment of hospitalised acute deep vein thrombosis in China. BMJ Open 2020;10:e038433. [Crossref] [PubMed]
33. Zeng H, Wang C, Song LY, et al. Economic evaluation of FLOT and ECF/ECX perioperative chemotherapy in patients with resectable gastric or gastro-oesophageal junction adenocarcinoma. BMJ Open 2022;12:e060983. [Crossref] [PubMed]
34. Statistical Bulletin on the National Economic and Social Development of the People’s Republic of China in 2022. 2023-03-01.
35. Liao M, Jiang Q, Hu H, et al. Cost-effectiveness analysis of utidelone plus capecitabine for metastatic breast cancer in China. J Med Econ 2019;22:584-92. [Crossref] [PubMed]
36. Hollman C, Paulden M, Pechlivanoglou P, et al. A Comparison of Four Software Programs for Implementing Decision Analytic Cost-Effectiveness Models. Pharmacoeconomics 2017;35:817-30. [Crossref] [PubMed]
37. Palmer S, Borget I, Friede T, et al. A Guide to Selecting Flexible Survival Models to Inform Economic Evaluations of Cancer Immunotherapies. Value Health 2023;26:185-92. [Crossref] [PubMed]
38. Latimer NR. Survival analysis for economic evaluations alongside clinical trials--extrapolation with patient-level data: inconsistencies, limitations, and a practical guide. Med Decis Making 2013;33:743-54. [Crossref] [PubMed]
39. Liao JJZ, Liu GF. A flexible parametric survival model for fitting time to event data in clinical trials. Pharm Stat 2019;18:555-67. [Crossref] [PubMed]
40. Baio G. survHE: Survival Analysis for Health Economic Evaluation and Cost-Effectiveness Modeling. Journal of Statistical Software 2020;95:1-47.
41. Wu H, Lin P, Yang S, et al. Cost-utility analysis of palliative care in patients with advanced cancer: a retrospective study. BMC Palliat Care 2021;20:126. [Crossref] [PubMed]
42. Alshamsan B, Gad AM, Alata MK, et al. Capecitabine Monotherapy as Palliative Treatment for Patients with Recurrent/Metastatic Nasopharyngeal Cancer. J Oncol 2022;2022:6860413. [Crossref] [PubMed]
43. Chua D, Wei WI, Sham JS, et al. Capecitabine monotherapy for recurrent and metastatic nasopharyngeal cancer. Jpn J Clin Oncol 2008;38:244-9. [Crossref] [PubMed]
44. Ngeow J, Lim WT, Leong SS, et al. Docetaxel is effective in heavily pretreated patients with disseminated nasopharyngeal carcinoma. Ann Oncol 2011;22:718-22. [Crossref] [PubMed]
45. Zhang L, Zhang Y, Huang PY, et al. Phase II clinical study of gemcitabine in the treatment of patients with advanced nasopharyngeal carcinoma after the failure of platinum-based chemotherapy. Cancer Chemother Pharmacol 2008;61:33-8. [Crossref] [PubMed]
46. Liu GY, Li WZ, Wang DS, et al. Effect of Capecitabine Maintenance Therapy Plus Best Supportive Care vs Best Supportive Care Alone on Progression-Free Survival Among Patients With Newly Diagnosed Metastatic Nasopharyngeal Carcinoma Who Had Received Induction Chemotherapy: A Phase 3 Randomized Clinical Trial. JAMA Oncol 2022;8:553-61. [Crossref] [PubMed]
47. Zhu Y, Liu K, Ding D, et al. Chemo-Immunotherapy Regimes for Recurrent or Metastatic Nasopharyngeal Carcinoma: A Network Meta-Analysis and Cost-Effectiveness Analysis. Front Pharmacol 2022;13:858207. [Crossref] [PubMed]
48. Adalsteinsson E, Toumi M. Benefits of probabilistic sensitivity analysis - a review of NICE decisions. J Mark Access Health Policy 2013;
49. Claxton K, Martin S, Soares M, et al. Methods for the estimation of the National Institute for Health and Care Excellence cost-effectiveness threshold. Health Technol Assess 2015;19:1-503. v-vi. [Crossref] [PubMed]
50. Tian K, Han J, Wang Z, et al. Immune checkpoint inhibition in first-line treatment for recurrent or metastatic nasopharyngeal carcinoma: A CAPTAIN-1st and JUPITER-02 trial-based cost-effectiveness analysis. Oral Oncol 2022;128:105842. [Crossref] [PubMed]
51. Tang YK, Xu Z, Ye ZM, et al. Cost-effectiveness analysis of tislelizumab in combination with chemotherapy for the first-line treatment of patients with metastatic or recurrent nasopharyngeal carcinoma in China. Head Neck 2024;46:5-14. [Crossref] [PubMed]
52. Zhao Z, Chen T, Zhou Z, et al. Cost-effectiveness of camrelizumab combined with chemotherapy in the first-line treatment of recurrent or metastatic nasopharyngeal carcinoma in China. BMJ Open 2023;13:e071832. [Crossref] [PubMed]
53. Chen X, Liang W, Wan N, et al. Cost-effectiveness analysis of gemcitabine plus cisplatin versus fluorouracil plus cisplatin for first-line treatment of recurrent or metastatic nasopharyngeal carcinoma. Oral Oncol 2019;94:80-5. [Crossref] [PubMed]
54. Latimer NR. 2013. NICE Decision Support Unit Technical Support Documents. Survival Analysis For Economic Evaluations Alongside Clinical Trials - Extrapolation with Patient-Level Data. London: National Institute for Health and Care Excellence (NICE); 2013.
55. Fang W, Yang Y, Ma Y, et al. Camrelizumab (SHR-1210) alone or in combination with gemcitabine plus cisplatin for nasopharyngeal carcinoma: results from two single-arm, phase 1 trials. Lancet Oncol 2018;19:1338-50. [Crossref] [PubMed]
56. Lam WS, Wang LZ, Roudi R, et al. Resisting resistance to cancer immunotherapy. Thorac Cancer 2018;9:507-8. [Crossref] [PubMed]
57. Tumeh PC, Harview CL, Yearley JH, et al. PD-1 blockade induces responses by inhibiting adaptive immune resistance. Nature 2014;515:568-71. [Crossref] [PubMed]
58. Sanmamed MF, Chen L. A Paradigm Shift in Cancer Immunotherapy: From Enhancement to Normalization. Cell 2018;175:313-26. [Crossref] [PubMed]
59. Ferguson R, Chat V, Morales L, et al. Germline immunomodulatory expression quantitative trait loci (ieQTLs) associated with immune-related toxicity from checkpoint inhibition. Eur J Cancer 2023;189:112923. [Crossref] [PubMed]
60. Reiners C, Drozd V, Yamashita S. Hypothyroidism after radiation exposure: brief narrative review. J Neural Transm (Vienna) 2020;127:1455-66. [Crossref] [PubMed]
61. Feen Rønjom M. Radiation-induced hypothyroidism after treatment of head and neck cancer. Dan Med J 2016;63:B5213.
62. Higgins AM, Harris AH. Health economic methods: cost-minimization, cost-effectiveness, cost-utility, and cost-benefit evaluations. Crit Care Clin 2012;28:11-24. v. [Crossref] [PubMed]
63. Briggs AH, O’Brien BJ. The death of cost-minimization analysis? Health Econ 2001;10:179-84. [Crossref] [PubMed]
64. Squires BP. Methods for the Economic Evaluation of Health Care Programmes. Canadian Medical Association Journal 1987;138.
65. Guidelines for the Economic Evaluation of Health Technologies: Canada — 4th Edition March 2017. Available online: https://www.cadth.ca/guidelines-economic-evaluation-health-technologies-canada-4th-edition
66. Guide to the methods of technology appraisal: process and methods. National Institute for Health and Care Excellence. Available online: https://www.nice.org.uk/process/pmg9
67. Guidelines for preparing a submission to the Pharmaceutical Benefits Advisory Committee, version 5.0. Department of Health, Australian Government. Available online: https://pbac.pbs.gov.au/content/information/files/pbac-guidelines-version-5.pdf