Gastrointestinal stromal tumors (GIST) are rare mesenchymal tumors that predominantly originate from the gastrointestinal tract, mainly in the stomach (60%) and small intestine (30%) (1). Around 85% of GIST harbor gene mutations in stem cell factor receptor (KIT), and another 5–10% of GIST have a mutation in the gene encoding the platelet-derived growth factor receptors-α (PDGFRA) (2–5). Since the development and application of tyrosine kinase inhibitor (TKI) therapy that inhibits KIT and PDGFRA kinase activity and then intercepted the signal transduction pathways related to tumor proliferation and apoptosis, the therapeutic effects of locally advanced and/or metastatic GIST has achieved a revolutionary breakthrough.

The first TKI imatinib mesylate was approved in February 2002, for the treatment of KIT-positive metastatic and/or locally advanced GIST (6, 7). Treated with initial dose at 400 mg/day of imatinib, patients with metastatic or unresectable GIST reached median progression-free survival (mPFS) at 18 months, median overall survival (mOS) at 55 months (8–10). Other phase III studies have assessed the efficacy of imatinib at two initial dose levels (400 vs. 800 mg daily, given as 400 mg twice a day), showing equivalent response rates and OS for both dose levels (10–12). For resectable GIST patients, imatinib has been used in both pre- and post-operative therapy as several prospective studies have demonstrated the safety and efficacy of preoperative imatinib in patients undergoing surgical resection (13–15), while other studies revealed adjuvant imatinib therapy was associated with longer relapse-free survival (RFS) (16–18) and a longer duration (36- vs. 12-month group) of postoperative imatinib therapy improved RFS and OS for patients with a high risk of recurrence (19, 20).

Resistance to imatinib therapy is categorized into two situations. A small number (<15%) of patients have primary resistance to imatinib therapy (21), which is a disease that cannot be stabilized or progress within 6 months of initiation of treatment. The majority of patients (50%) develop secondary resistance characterized by an initial response or stable disease but subsequent progression, which is the result of acquired mutations generated during the course of treatment (22). For patients with imatinib-resistant or intolerant GIST, sunitinib was approved and recommended in January 2006, as it significantly improved median time to tumor progression (mTTP) (27.3 weeks in patients receiving sunitinib vs. 6.4 weeks in patients on placebo) and estimated OS (23). An recent study suggested that via sunitinib therapy, GIST patients after imatinib failure could reach the mTTP at 8.3 months and median mOS at 16.6 months (24).

In patients with metastatic or unresectable GIST progressing after the failure of imatinib and sunitinib, regorafenib was approved and regarded as the preferred option for third-line therapy, as it provided a significant improvement in PFS compared with placebo (4.8 months for regorafenib vs. 0.9 months for placebo) and higher disease control rate (DCR; 53 vs. 9%) (25).

Concerning rational decision making in health care, a major challenge in pharmacoeconomic evaluation is to make full use of cost-effectiveness data to optimize clinical practice and allocation of healthcare resources. This review was conducted aiming at the cost-effectiveness analysis of TKIs in GIST.

This systematic review was conducted using the Preferred Reporting Items for Systematic Reviews and Meta-Analyses: The PRISMA Statement (26). PICOS criteria (population, intervention, control, outcomes, and study design) was used to guide the development of the search strategy. A thorough literature search of the following online databases was performed: PubMed, Web of Science, and Embase. Medical Subject Heading (MeSH) terms were individually selected using the National Library of Medicine controlled vocabulary thesaurus used for indexing articles: gastrointestinal stromal tumor or GIST, cost, cost-effectiveness, economic evaluation, economics, monetary, reimbursement, insurance. Searches were conducted on December 9, 2020 and all studies published before this date will be investigated.

Eligibility criteria were published studies in English evaluating the cost-effectiveness of any of the TKIs in GIST. Care was taken to ensure that the inclusion criteria were sufficiently broad so that possibly pertinent publications could be assessed by individual screening. Given the heterogeneity of available studies, we were not able to perform a meta-analysis.

Study data extraction was conducted independently by two authors (M.F., Y.Y.) and was extracted using a data extraction form, which included author, published year, country, study population, study design, intervention and comparison, model type, perspective, time horizon, discount rate, sensitivity analysis, threshold, sponsors, cost-effectiveness outcomes, and conclusions. To allow direct comparisons across countries, all costs were converted to US dollars, then inflated to December 2020 using the country-specific Consumer Price Index (CPI) (https://www.bls.gov/data/inflation_calculator.htm).

Completeness of reporting was assessed using the Consolidated Health Economic Evaluation Reporting Standards (CHEERS) checklist, which provides 24 items and accompanying recommendations to optimize reporting of health economic evaluations (27). The quality of the evaluation was assessed using the quality of health economic studies (QHES) instrument, which is designed to discriminate higher-quality cost-effectiveness information to enhance decision making (28). The QHES instrument was a quantitative and weighted scoring approach to appraise health economic evaluations, consisting of 16 items and each of them has a weighted point value ranging from 1 to 9. The sum of the weights of a study ranges between 0 (means extremely poor quality) and 100 (means excellent quality). Both checklists were completed independently by two authors (M.F., Y.Y.), and disagreements were resolved by discussion and arbitration (W.L.) where necessary.

This review has been registered with PROSPERO (CRD42021225253).

Based on the initial searches, a total of 1,440 articles were identified, which were independently screened by two reviewers (M.F., Y.Y.). Of these, 606 were removed as duplicates. Of the 834 publications remaining, 777 records were excluded via reading abstracts and titles with reasons for exclusion: case reports, reviews, and non-original research (e.g., letters or commentaries). Unpublished abstracts and meeting conferences were not included owing to the inability to completely assess quality. Then, 57 full-text articles were assessed for eligibility by two reviewers independently (M.F., Y.Y.). Disagreements were resolved by discussion and arbitration (W.L.) where necessary. Finally, 15 original investigations were found to have sufficient focus and relevance to be incorporated into the systematic review (Figure 1).

Almost every new drug is associated with better clinical benefits in patients and higher costs, posing challenges to cost-effectiveness and affordability, and results of economic evaluations have become increasingly important as criteria for the allocation of health care resources. In our study, there were major differences in the structural assumptions in the identified studies, including in the model types, study perspectives, time horizons, discount rates, assumption of utility, and extrapolation of survival. Therefore, there were large variations in the predicted costs and QALYs associated with each treatment, for example, the predicted QALYs of advanced GIST treated with imatinib varied from 2.96 to 4.85. Variations in QALYs could be explained by the use of utility values derived by different methods, different time horizons, and alternative approaches to survival extrapolation. Variations in total costs could be explained by different healthcare resource use and costs across jurisdictions. Moreover, the different study perspectives would significantly affect total costs. It may also be accounted for by different approaches to capturing costs of post-progression treatment, where some studies assumed no post-progression drug costs while others (35, 36, 42) constructed a series of pragmatic clinical treatment pathways and clearly calculated the costs of each treatment path.

Despite these variations, there was consistency in the conclusions across most of the studies. For patients with advanced/metastatic GIST, all publications agree that TKIs are associated with higher costs and effectiveness than placebo or empirical treatment. Some articles (29–31) concluded that imatinib 400 mg/d in first-line therapy was cost-effective, but these economic analyses were carried on between 2005 and 2008, and some model parameters they used may not be fully standardized. Other studies confirmed the cost-effectiveness of sunitinib in second-line therapy (32–34), and regorafenib was cost-effective compared with imatinib re-challenge in the third-line therapy in Germany (37). Two other studies simulated the most cost-effective medication plan by constructing multiple clinical pathways (35, 36), and based on these results, we suggest for advanced GIST, the treatment of imatinib in first-line, followed by sunitinib in second-line, and regorafenib in third-line was cost-effective.

For patients with resectable GIST, several studies (40, 41), respectively, investigated the 3- vs. 1-year adjuvant imatinib therapy in resected GIST, and both confirmed the cost-effectiveness of the longer-term (3-year) therapy. Another study (43) illustrates the necessity of surgery in rectal GIST patients requiring abdominoperineal resection following neoadjuvant imatinib. Most of the identified studies were conducted in high-income and developed countries, including European and American countries, and most studies had positive conclusions regarding the cost-effectiveness of the interventions except one study (42) taking into account the country's context.

Another two recent economic evaluations carried out by Banerjee et al. (39) and Zuidema et al. (38), respectively, are not limited to a fixed-dose of medication but are concerned about individualized medication methods that guide the use of TKIs in advanced GIST, such as TDM (38) and TGT (39), which are both considered cost-effective. It is known that mutational status has a dramatic impact on response to imatinib or sunitinib in patients with advanced or metastatic GIST. The presence of a KIT exon 11 mutation was associated with better response rates, PFS, and OS compared to KIT exon 9 mutations or wild-type GIST (44–46). In patients whose tumors expressed a KIT exon 9 mutation, high-dose imatinib (800 mg/d) resulted in a significantly superior PFS (44, 45) and increased response rates (46, 47) compared to those treated with imatinib 400 mg/d. And the cost-effectiveness analysis (39) focusing on TGT-guided therapy was performed based on this setting. Another study (38) focused on the TDM-guided dosing imatinib. Therapeutic drug monitoring is a technique used to determine the plasma exposure levels of certain drugs and enable to ensure the GIST patients redistributed with adequate imatinib concentrations in plasma (48, 49). By performing an economic evaluation between TDM-guided and fixed-dose imatinib, the results are a valuable addition to the investigation of the effect of dose optimization. It is foreseeable that with the further development of molecular oncology, there would be more novel economic evaluations.

At the same time, there existed other new TKIs that have been approved by the food and drug administration (FDA) and endorsed by NCCN guidelines, for instance, avapritinib for PDGFRA D842V-mutant GIST as first-line therapy (50), and ripretinib for the progressive disease after imatinib, sunitinib, and regorafenib as fourth-line therapy (51). Nevertheless, sorafenib, nilotinib, dasatinib, and pazopanib have also shown activity in patients with GIST resistant to imatinib and sunitinib. However, much of the data on these TKIs came from phase II studies or retrospective analyses, which lack high-quality clinical evidence. The cost-effectiveness of these TKIs still needs to be measured.

There exist some limitations in this study. First, the QHES instrument employs yes or no responses rather than a continuous scale for each criterion, which would lead to inaccuracy when a study actually partly meets the criteria but is appraised with zero points. Therefore, the CHEERS checklist was applied to cross-evaluate the quality of the literature. But the CHEERS statement is an assessment of reporting, not methodological quality, and failure to follow all the requirements in the CHEERS statement is not indicative of a poor-quality study. Second, our systematic review excluded conference abstracts, unpublished studies (gray literature), and studies that lack full-text resources, which may also introduce some bias.

In conclusion, our systematic review identified 15 economic evaluations of TKIs used in patients with GIST and demonstrated several important findings. First, for patients with advanced GIST, imatinib in the first-line treatment, followed by sunitinib in the second-line treatment was considered cost-effective, and regorafenib was cost-effective compared with imatinib re-challenge in the third-line therapy. Second, for patients with resectable GIST, 3-year adjuvant imatinib therapy represented a cost-effective treatment option compared with 1-year therapy. Third, the precision medicine-assisted imatinib treatment plan represented by TDM- and TGT-guided imatinib therapy was cost-effective compared with empirical fixed-dose treatment.

The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author/s.

MF: material preparation, data acquisition, and wrote the first draft of the manuscript. YY: material preparation and data acquisition. The revised draft of the manuscript was written by MF and YY. All authors contributed to the conception and design of the study, commented on previous versions of the manuscript, read, and approved the final manuscript.

This work was supported by the 1.3.5 Project for Disciplines of Excellence, West China Hospital, Sichuan University (grant No. ZYJC18008 and No. ZYJC18010). The funding source was not involved in the study design, data collection, data analysis, data interpretation, the writing of this article or the decision to submit the paper for publication.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.