Cost-Effectiveness of Treatment Optimisation with Biomarkers for Immunotherapy in Solid Tumours: A Systematic Review

Simple Summary Researchers aim to assess the economic impact of testing predictive biomarkers for immunotherapy in solid tumour treatment using immune checkpoint inhibitors (ICIs). Despite recent advancements, concerns persist regarding the cost-effectiveness and budgetary implications of ICIs. This study systematically reviewed the economic evaluations of biomarker testing from various databases. The team assessed studies from June 2010 to February 2022, evaluating their quality and synthesising findings by tumour type. Understanding the economic implications of these tests could help drive future research, optimise treatment strategies, potentially influencing health care decisions and resource allocation in solid tumour therapy, impacting how we approach and fund immunotherapy for better patient outcomes. Abstract This study investigated the health economic evaluations of predictive biomarker testing in solid tumours treated with immune checkpoint inhibitors (ICIs). Searching PubMed, EMBASE, and Web of Science from June 2010 to February 2022, 58 relevant articles were reviewed out of the 730 screened. The focus was predominantly on non-small cell lung cancer (NSCLC) (65%) and other solid tumours (40%). Among the NSCLC studies, 21 out of 35 demonstrated cost-effectiveness, notably for pembrolizumab as first-line treatment when preceded by PD-L1 assessment, cost-effective at a threshold of $100,000/QALY compared to the standard of care. However, for bladder, cervical, and triple-negative breast cancers (TNBCs), no economic evaluations met the affordability threshold of $100,000/QALY. Overall, the review highlights a certain degree of uncertainty about the cost-effectiveness of ICI. In particular, we found PD-L1 expression associated with ICI treatment to be a cost-effective strategy, particularly in NSCLC, urothelial, and renal cell carcinoma. The findings suggest the potential value of predictive biomarker testing, specifically with pembrolizumab in NSCLC, while indicating challenges in achieving cost-effectiveness for certain other solid tumours.


Introduction
The advent of immunotherapy has significantly changed the therapeutic scenario of cancer patients.Immunotherapy drugs, so-called immune checkpoint inhibitors (ICIs), work by blocking checkpoint proteins from binding to their partners, representing a new weapon for cancer treatment in different settings [1].Moreover, the association of a biomarker test to identify the best treatment choice has revolutionised patient management for many tumour types, driving towards a patient-tailored approach [2].Since the FDA approval of ipilimumab (cytotoxic T-lymphocyte-associated antigen-4 (CTLA-4) inhibitor) in 2011, six more ICIs have been approved for cancer therapy: the programmed cell death-1 (PD-1) inhibitors including nivolumab, pembrolizumab, cemiplimab, and the programmed cell death ligand-1 (PD-L1) inhibitors including atezolizumab, durvalumab, and avelumab.These agents have become the standard of care in many solid tumours [1][2][3].Thus, a predictive biomarker testing approach in this field may represent a virtuous model to invest in for treatment optimisation, with a possible impact also on the patients' quality of life (QoL).Despite the successes achieved in recent years, there are many concerns about the cost-effectiveness (CE) and budget impact of the next wave of ICIs [3].This systematic review, as part of a funded Italian National Research Project, aims to provide a snapshot of the current state-of-the-art regarding the cost-effectiveness, cost-utility, or net-monetary benefits of the use of predictive biomarkers in solid tumours treated with ICIs as tools for customising immunotherapy.

Materials and Methods
A systematic literature review was conducted in accordance with the PRISMA 2020 statement guidelines [4].The review methodology was prospectively registered a priori in PROSPERO (CRD42020201549) as for the entire study protocol [5].

Data Sources and Searches
Searches of relevant studies were conducted across the following electronic bibliographic databases: Ovid MEDLINE, EMBASE, Web of Science, from June 2010 to February 2022 as per the date of the first ICI approval [6].Search strategy and syntaxes are detailed in the Supplementary Box S1 and Supplementary Box S2 in the Supplementary Materials, respectively.References were collected with Reference Manager (Institute for Scientific Information, Berkeley, CA, USA, ver.12).

Study Selection, Data Extraction, and Quality Assessment
Predefined criteria for selecting studies related to economic evaluation (population, interventions and comparisons, outcomes, study design, [PICOS]) were set and published in the study protocol [5].Data were processed by four researchers to identify potentially eligible studies and validated by five clinicians/researchers.The screening process is detailed in Figure 1.The quality of evidence was evaluated by the GRADE approach [7].

Data Synthesis
Country, treatment line, comparators, biomarker test, willingness-to-pay (WTP) threshold, cost-effectiveness ratio, and affordability according to the WTP set were extracted from each study.Information was collected separately for non-small cell lung cancer (NSCLC) and other solid tumour diagnoses.A time-trend analysis was performed.After converting differences in costs to the 2021 US dollar, adjusting for both inflation and purchasing power parities [8] resources, adjusted differences in costs and QALYs were plotted on a CE plane, evaluated with a reference WTP threshold of $100,000/QALY.

Role of the Funding Source
This research was funded by the Ministero dell'Istruzione, dell'Università e della Ricerca (MIUR) within the framework of the PRIN Project 2017, grant number 2017NR7W5K.This funding source had no role in the design, execution, analyses, interpretation of the data, and results of the present study.

Data Synthesis
Country, treatment line, comparators, biomarker test, willingness-to-pay (WTP) threshold, cost-effectiveness ratio, and affordability according to the WTP set were extracted from each study.Information was collected separately for non-small cell lung cancer (NSCLC) and other solid tumour diagnoses.A time-trend analysis was performed.After converting differences in costs to the 2021 US dollar, adjusting for both inflation and purchasing power parities [8] resources, adjusted differences in costs and QALYs were plotted on a CE plane, evaluated with a reference WTP threshold of $100,000/QALY.

Role of the Funding Source
This research was funded by the Ministero dell'Istruzione, dell'Università e della Ricerca (MIUR) within the framework of the PRIN Project 2017, grant number 2017NR7W5K.This funding source had no role in the design, execution, analyses, interpretation of the data, and results of the present study.

Study Identification
A total of 730 articles were identified through the systematic literature search across different bibliographic database queries.As reported in Figure 1, 295 (40.4%) duplicates were removed and 435 abstracts (59.6%) were screened, resulting in 76 studies potentially evaluable.After excluding six papers for which the full-text was not retrieved, a total of 70 full-text articles were examined.Among these papers, 10 systematic reviews and one

Study Identification
A total of 730 articles were identified through the systematic literature search across different bibliographic database queries.As reported in Figure 1, 295 (40.4%) duplicates were removed and 435 abstracts (59.6%) were screened, resulting in 76 studies potentially evaluable.After excluding six papers for which the full-text was not retrieved, a total of 70 full-text articles were examined.Among these papers, 10 systematic reviews and one paper based on a budget impact analysis model were excluded, for a total of 58 studies that met the inclusion criteria.

Study Quality
Quality of evidence and risk of biases were assessed based on the GRADE scale for each study.Overall, 60% of economic evaluation (n = 35) was based on a Markov model and 26% on the partitioned survival model (n = 15).All of the methodological aspects of individual studies are shown in Supplementary Table S1 in the Supplementary Materials.Overall, most economic evaluations (62.1%, n = 36 studies) presented a moderate/high certainty of evidence and strength of recommendations.Next, 25.9% (n = 15) of the studies achieved a moderate certainty rating, 6.9% (n = 4) a high, and 5.2% (n = 3) a moderate/low certainty score (Supplementary Table S2 in the Supplementary Materials).

Study Characteristics
The geographical distribution of papers reporting an economic evaluation of an ICI associated to a biomarker test showed a marked majority of studies conducted in the US (56%; n = 33) and China (27%; n = 16).Figure 2 shows the worldwide distribution of studies included.achieved a moderate certainty rating, 6.9% (n = 4) a high, and 5.2% (n = 3) a moderate/low certainty score (Supplementary Table S2 in the Supplementary Materials).

Study Characteristics
The geographical distribution of papers reporting an economic evaluation of an ICI associated to a biomarker test showed a marked majority of studies conducted in the US (56%; n = 33) and China (27%; n = 16).Figure 2 shows the worldwide distribution of studies included.In the overall analysis, 97% of the included studies (n = 56) assessed the costeffectiveness ratio of an ICI with prior PD-L1 testing, whereas only 3% (n = 2) considered prior PD-1 testing.Among the overall 58 included studies, 60% (n = 35) included patients affected by NSCLC (Table 1) and 40% other solid tumours including 25.9% (n = 6) head and neck cancer squamous cell carcinoma (HNSCC), 15.5% (n = 9) urothelial, bladder cancer, and renal cell carcinoma (RCC), followed by a few studies on triple-negative breast cancer (5.2%; n = 3), melanoma including Merkel cell carcinoma (mMCC) (5.2%; n = 3), cervical, and gastric cancers (3.4%; n = 2) (Table 2).Following a bibliometric analysis, a huge increase in the number of papers on NSCLC from 2016 to 2021 was observed (Figure 3).This was confirmed by the time-trend estimates of the economic evaluations of ICIs associated with biomarker testing in NSCLC being considerably higher than in other solid tumours.In the overall analysis, 97% of the included studies (n = 56) assessed the cost-effectiveness ratio of an ICI with prior PD-L1 testing, whereas only 3% (n = 2) considered prior PD-1 testing.Among the overall 58 included studies, 60% (n = 35) included patients affected by NSCLC (Table 1) and 40% other solid tumours including 25.9% (n = 6) head and neck cancer squamous cell carcinoma (HNSCC), 15.5% (n = 9) urothelial, bladder cancer, and renal cell carcinoma (RCC), followed by a few studies on triple-negative breast cancer (5.2%; n = 3), melanoma including Merkel cell carcinoma (mMCC) (5.2%; n = 3), cervical, and gastric cancers (3.4%; n = 2) (Table 2).Following a bibliometric analysis, a huge increase in the number of papers on NSCLC from 2016 to 2021 was observed (Figure 3).This was confirmed by the time-trend estimates of the economic evaluations of ICIs associated with biomarker testing in NSCLC being considerably higher than in other solid tumours.

Non-Small Cell Lung Cancer (NSCLC)
Among the 35 studies on patients with NSCLC, slightly more than half of those studies (63%, n = 22) [9-14,20-23,29-32,37-39,42,67-69] suggested the cost-effectiveness of ICI treatment with biomarker testing; of these, n = 2 with previous PD-1 assessment and n = 20 with previous PD-L1 assessment.Particularly, when distinguishing between anti PD-1 and anti PD-L1 treatment and considering the multiple analyses performed as part of the selected studies, there were no differences in the proportion of analyses demonstrating the cost-effectiveness of ICIs (p-value = 1), Figure 4.

Non-Small Cell Lung Cancer (NSCLC)
Among the 35 studies on patients with NSCLC, slightly more than half of those studies (63%, n = 22) [9][10][11][12][13][14][20][21][22][23][29][30][31][32][37][38][39]42,[67][68][69] suggested the cost-effectiveness of ICI treatment with biomarker testing; of these, n = 2 with previous PD-1 assessment and n = 20 with previous PD-L1 assessment.Particularly, when distinguishing between anti PD-1 and anti PD-L1 treatment and considering the multiple analyses performed as part of the selected studies, there were no differences in the proportion of analyses demonstrating the cost-effectiveness of ICIs (p-value = 1), Figure 4.  Figures show incremental QALY and incremental costs for different immune checkpoint inhibitors (ICI) among NSCLC patients according to NSCLC type (identified by the type of marker) and immune checkpoint inhibitor (identified by marker colour); the size of the marker identifies first-line (bigger) or second-line (smaller) treatment; the area under the WTP threshold indicates the cost-effective studies and is highlighted in green.Studies for which results are available for more than one country are represented more than once.

Urothelial, Bladder Cancer, and Renal Cell Carcinoma (RCC)
When considering other tumours, statistically significant differences emerged between anti PD-1 and anti PD-L1, with a clear indication of no cost-effectiveness of available evidence for the latter approach in those tumours (Figure 5).
In more detail, five economic evaluations [44][45][46][47]49] were carried out on patients diagnosed with urothelial cancer, demonstrating that avelumab (n = 1) [45] and pembrolizumab (n = 2) [46,47] were cost-effective in first-line treatment when preceded by PD-L1 testing (with thresholds set in the US and Switzerland) with greater impact on the survival benefit of patients.Conversely, first-line treatment with atezolizumab (n = 1) [44] and second-line treatment with pembrolizumab [49] (n = 1) were not cost-effective with a threshold of $150,000/QALY in a US setting.
Regarding renal cell carcinoma from a US perspective, nivolumab was cost-effective compared to everolimus [50], and nivolumab with ipilimumab was cost-effective compared to sunitinib in both cases with a PD-L1 test, and a $150,000/QALY threshold [48] generated a gain of 0.34 QALYs and 0.98 QALYs, respectively.
Figures show incremental QALY and incremental costs for different immune checkpoint inhibitors (ICI) among NSCLC patients according to NSCLC type (identified by the type of marker) and immune checkpoint inhibitor (identified by marker colour); the size of the marker identifies first-line (bigger) or second-line (smaller) treatment; the area under the WTP threshold indicates the cost-effective studies and is highlighted in green.Studies for which results are available for more than one country are represented more than once.When considering other tumours, statistically significant differences emerged between anti PD-1 and anti PD-L1, with a clear indication of no cost-effectiveness of available evidence for the latter approach in those tumours (Figure 5).Figures show incremental QALY and incremental costs for different immune checkpoint inhibitors (ICI) for cancer other than NSCLC according to cancer type (identified by the type of marker) and immune checkpoint inhibitor (identified by marker colour); the size of the marker identifies first-line (bigger) or second-line (smaller) treatment; the area under the WTP threshold indicates the cost-effective studies and is highlighted in green.Studies for which results are available for more than one country are represented more than once.

Discussion
This systematic review examined the cost-effectiveness profile of ICI therapy in solid tumours associated with the biomarker test in oncology.Most of the studies  included focused on non-small cell lung cancer (NSCLC), followed by HNSCC and genitourinary cancers.
This review highlights that there is still no clear evidence regarding the costeffectiveness of anti PD-L1 and PD-1 for treating NSCLC, while at present, the available evidence is generally not in favour of the cost-effectiveness of anti PD-L1 in other solid tumours.
Despite there being a significantly fewer number of studies retrieved for other solid tumours, the PD-L1 testing approach was also cost-effective in some settings for urothelial cancer patients first-line treated with avelumab and pembrolizumab but also for renal cell carcinoma treated with nivolumab.Moreover, this approach was also cost-effective in urothelial cancer and renal cell carcinoma.
Consistent with these findings, the European Society for Medical Oncology (ESMO) NSCLC guidelines emphasise the importance of molecular subtyping in guiding therapeutic decision-making, advocating its execution whenever feasible [70,71].Hence, for patients with advanced NSCLC, ESMO recommends determining PD-L1 expression between others such as EGFR testing, BRAF mutations, and the analysis of ALK, ROS1, and NTRK rearrangements [71].Generally, there is consensus across international guidelines around the need for PD-L1 testing and other biomarkers in advanced NSCLC.These have approved first-line targeted therapies in Europe [72][73][74].
Hence, multiple studies have highlighted the potential of PD-L1 overexpression as a crucial and extensively studied predictive biomarker for assessing the response to PD-L1 In more detail, five economic evaluations [44][45][46][47]49] were carried out on patients diagnosed with urothelial cancer, demonstrating that avelumab (n = 1) [45] and pembrolizumab (n = 2) [46,47] were cost-effective in first-line treatment when preceded by PD-L1 testing (with thresholds set in the US and Switzerland) with greater impact on the survival benefit of patients.Conversely, first-line treatment with atezolizumab (n = 1) [44] and second-line treatment with pembrolizumab [49] (n = 1) were not cost-effective with a threshold of $150,000/QALY in a US setting.
Regarding renal cell carcinoma from a US perspective, nivolumab was cost-effective compared to everolimus [50], and nivolumab with ipilimumab was cost-effective compared to sunitinib in both cases with a PD-L1 test, and a $150,000/QALY threshold [48] generated a gain of 0.34 QALYs and 0.98 QALYs, respectively.
Moreover, the two studies [51,52] conducted on bladder cancer showed that the secondline treatment with pembrolizumab and atezolizumab were not cost-effective compared to chemotherapy in a Canadian, British, and Australian perspective with a WTP less than $100,000/QALY (Figure 5).

Triple-Negative Breast Cancer (TNBC)
The cost-effectiveness analysis of atezolizumab plus nab-paclitaxel versus nab-paclitaxel alone provided similar results.Two Markov models and a multi-country partitioned survival model [59,60] indicated that this combination (atezolizumab plus nab-paclitaxel) was not cost-effective compared to nab-paclitaxel alone, both in China and the US.Similarly, another Markov model-based study [61] corroborated these findings across the overall population (ICER $281,448/QALY; WTP of $200,000/QALY) (Figure 5).

Melanoma including Merkel Cell Carcinoma (mMCC)
Three cost-effectiveness analyses were retrieved on ICIs used for the treatment of melanoma and Merkel cell carcinoma (MCC) [62][63][64].For metastatic MCC, avelumab treatment with previous PD-L1 assessment was cost-effective compared to best supportive care (ICER USD 44,885.06/QALY,WTP USD 53,333.33/QALY)and chemotherapy (ICER USD 42,993.06/QALY)[64] (Figure 5).Regarding patients with advanced melanoma, nivolumab was the most cost-effective treatment option in BRAF wild-type and BRAF mutant patients, as demonstrated by the Markov model developed to estimate the lifetime costs and benefits of nivolumab versus ipilimumab and dacarbazine (for BRAF wild-type) and versus ipilimumab, dabrafenib, and vemurafenib (for BRAF mutant patients) [62].In patients affected by BRAF wild-type advanced melanoma, the discrete simulation event model of Tarhini et al. [63] found that the most cost-effective treatment sequences initiated with anti-PD-1 + anti-CTLA-4 (nivolumab + ipilimumab and chemotherapy through a mix of dacarbazine, temozolomide, paclitaxel, and carboplatin + paclitaxel), followed by either chemotherapy or anti-PD-1 monotherapy (nivolumab and pembrolizumab, assuming an equal share).This provided substantial quality-adjusted survival gains to patients with BRAF wild-type advanced melanoma.

Cervical and Gastric Cancers
Only one cost-effectiveness analysis for the treatment of patients with persistent, recurrent, or metastatic cervical cancers who had not received systemic chemotherapy and were not amenable to curative treatment was found [65].Through a partitioned survival model over a 30-year lifetime horizon, Shi et al. [65] indicated that pembrolizumab was not cost-effective versus the placebo.With respect to patients in second-line therapy with metastatic gastric cancer, the most effective strategy was pembrolizumab for high microsatellite instability (MSI-H) patients and ramucirumab/paclitaxel for all other patients, but the Markov model resulted in a high ICER of $1,074,620/QALY.Among the following strategies, pembrolizumab monotherapy and ramucirumab/paclitaxel combination therapy for all patients and pembrolizumab for patients based on MSI status or PD-L1 expression, the only cost-effective one was paclitaxel monotherapy for all patients, with an ICER of $53,705/QALY (WTP USD 100,000/QALY in the US) [66] (Figure 5).
Figures show incremental QALY and incremental costs for different immune checkpoint inhibitors (ICI) for cancer other than NSCLC according to cancer type (identified by the type of marker) and immune checkpoint inhibitor (identified by marker colour); the size of the marker identifies first-line (bigger) or second-line (smaller) treatment; the area under the WTP threshold indicates the cost-effective studies and is highlighted in green.Studies for which results are available for more than one country are represented more than once.

Discussion
This systematic review examined the cost-effectiveness profile of ICI therapy in solid tumours associated with the biomarker test in oncology.Most of the studies  included focused on non-small cell lung cancer (NSCLC), followed by HNSCC and genitourinary cancers.This review highlights that there is still no clear evidence regarding the cost-effectiveness of anti PD-L1 and PD-1 for treating NSCLC, while at present, the available evidence is generally not in favour of the cost-effectiveness of anti PD-L1 in other solid tumours.
Despite there being a significantly fewer number of studies retrieved for other solid tumours, the PD-L1 testing approach was also cost-effective in some settings for urothelial cancer patients first-line treated with avelumab and pembrolizumab but also for renal cell carcinoma treated with nivolumab.Moreover, this approach was also cost-effective in urothelial cancer and renal cell carcinoma.
Consistent with these findings, the European Society for Medical Oncology (ESMO) NSCLC guidelines emphasise the importance of molecular subtyping in guiding therapeutic decision-making, advocating its execution whenever feasible [70,71].Hence, for patients with advanced NSCLC, ESMO recommends determining PD-L1 expression between others such as EGFR testing, BRAF mutations, and the analysis of ALK, ROS1, and NTRK rearrangements [71].Generally, there is consensus across international guidelines around the need for PD-L1 testing and other biomarkers in advanced NSCLC.These have approved first-line targeted therapies in Europe [72][73][74].
Hence, multiple studies have highlighted the potential of PD-L1 overexpression as a crucial and extensively studied predictive biomarker for assessing the response to PD-L1 antibodies, resulting in improved clinical outcomes [75][76][77][78].However, the implementation of these novel and more effective therapies has been acknowledged as both essential and cost-prohibitive [79].To investigate this, the economic evaluations included in this review [9,10,13,21,24,[29][30][31][32]34,37,42] confirmed that NSCLC patients with high tumour PD-L1 levels with a proportional score ≥ 50% for first-line therapy with pembrolizumab, exhibited superior response rates to immunotherapy and experienced prolonged survival compared to those who underwent conventional chemotherapy.Building upon these widely recognised clinical findings, this systematic review unequivocally demonstrated that the treatment approach in question was not only clinically effective but also cost-effective.The economic evaluations indicated a favourable cost threshold, falling within the range of $100,000/QALY to $150,000/QALY.Contrary to the majority of pharmacoeconomic assessments, the pivotal result of this review revealed that a distinct subset of studies failed to demonstrate the cost-effectiveness of PD-L1 testing and pembrolizumab for the first-line treatment of non-small cell lung cancer (NSCLC).Crucially, this divergence in outcomes emerged exclusively within a specific setting characterised by a willingness-to-pay (WTP) threshold falling below $100,000/QALY.This singular observation underscores the significance of the WTP threshold as a decisive factor in determining the cost-effectiveness of this therapeutic approach for NSCLC [17,23,26,28,35,36,38,40,41].Additionally, of paramount importance is that biomarker testing prior to nivolumab treatment is not deemed costeffective when administered as a second-line treatment [27], in combination with another monoclonal antibody [18], or when the WTP threshold falls below $100,000/QALY [27,43].
Regarding other solid tumours such as in the case of bladder [51,52], cervical [65], and triple-negative breast cancer (TNBC) [59][60][61], no economic evaluation has deemed the costs of the biomarker testing strategy acceptable within a setting where the threshold is less than $100,000/QALY.These findings diverge from the FDA's approval of pembrolizumab, for instance, in patients with recurrent metastatic cervical cancer and platinum failure, but limited to those whose tumours exhibit biomarker test positivity [80].To shed light on these results, we can analyse the outcomes of the clinical trials KEYNOTE-826 [81], designed to assess progression-free survival (PFS) and overall survival (OS) in patients with cervical cancer following genetic testing, and EMPOWER [82], which focused on evaluating OS.While both trials achieved their primary endpoints, subgroup analyses raised discussions on the benefits of ICIs for all cervical cancer patients [83].Despite the limitations of PD-L1 expression as a predictive biomarker, it still informs clinical decision-making and can contribute to pushing advances in immunotherapy research [84].Therefore, it can be concluded that a combination of biomarkers should be employed to identify patients who derive the greatest benefit [83].

Strengths and Limitations
This study has several strengths including a comprehensive literature search and the pioneering nature of the objective, as the economic impact and sustainability of health systems regarding immunotherapy and biomarker testing have not been thoroughly investigated yet.Furthermore, the review was executed using a systematic approach, providing a comprehensive overview of the cost-effectiveness of biomarker testing worldwide and over a wide range of tumours.
Nevertheless, this review did have certain limitations.The primary limitation is related to the inclusion of studies involving prior PD-1 testing constituting only 3% (n = 2) of the overall analysis.These specific studies are confined to pembrolizumab treatment in NSCLC.Due to this limited representation and the statistical insignificance of PD-1 in a broader context, we faced challenges in conducting a separate analysis based on PD-1 versus PD-L1 biomarkers.Another limitation is inherently associated with the methodology employed for conducting cost-effectiveness analyses (CEAs).Indeed, no study can fully encompass all of the potential cost-related factors or account for uncertainties surrounding the factors under investigation such as global economic and market forces, variations in practice and referral patterns, or reimbursements specific to individual insurance companies.Additionally, the set willingness-to-pay (WTP) thresholds do not incorporate other tax-related factors such as the financial impact on patients and supportive care providers, the utilisation of other costly cancer therapies, or indirect cost components such as the ability to return to work and contribute to the workforce and economy [3].

Conclusions
This systematic review highlights that PD-1 and PD-L1 overexpression, when used in combination with immune checkpoint inhibitor (ICI) therapy, could represent a costeffective strategy for treating NSCLC as a first-line treatment with pembrolizumab and with both first-and second-line nivolumab, but also for renal cell and urothelial carcinoma.However, the cost-effectiveness is diminished when the willingness-to-pay (WTP) threshold falls below $100,000/QALY.Therefore, the economic impact of biomarkers upstream of the choice of the specific therapy represents an imperative to validate its effectiveness, the eventual relationship with the quality of life, and economic sustainability.A biomarker testing approach is therefore a virtuous model to invest in, providing the patient with a greater chance of receiving increasingly effective therapy and minimising adverse events due to the administration of untargeted therapies, resulting in an undoubted improvement in quality of life, while also optimising the management of health care resources.

Figure 1 .
Figure 1.PRISMA diagram of the review's systematic searches.

Figure 1 .
Figure 1.PRISMA diagram of the review's systematic searches.

Figure 2 .
Figure 2. Global distribution of economic evaluations of an ICI associated with a biomarker test.

Figure 2 .
Figure 2. Global distribution of economic evaluations of an ICI associated with a biomarker test.

Figure 3 .
Figure 3. Economic evaluations of an ICI associated with a biomarker test per year focused on solid tumour diagnoses.Other tumours: Urothelial, bladded and renal cell cancer, cervical cancer, gastric cancer, head and neck cancer squamous cell carcinoma; melanoma; Merkel cell carcinoma, triplenegative breast cancer.Abbreviations: Non-small cell lung cancer (NSCLC).Notes: Dotted lines are estimates of linear predictions on the yearly trend of economic evaluations of ICIs associated with biomarker testing in NSCLC (blue dotted line) and other solid tumours (red dotted line).

Figure 3 .
Figure 3. Economic evaluations of an ICI associated with a biomarker test per year focused on solid tumour diagnoses.Other tumours: Urothelial, bladded and renal cell cancer, cervical cancer, gastric cancer, head and neck cancer squamous cell carcinoma; melanoma; Merkel cell carcinoma, triple-negative breast cancer.Abbreviations: Non-small cell lung cancer (NSCLC).Notes: Dotted lines are estimates of linear predictions on the yearly trend of economic evaluations of ICIs associated with biomarker testing in NSCLC (blue dotted line) and other solid tumours (red dotted line).

Figure 4 .
Figure 4. Results from the different studies on a cost-effectiveness plane for NSCLC patients according to anti PD-1 (a) or anti PD-L1 therapy (b).Note: Studies in the figure are identified by reference number.

Figure 4 .
Figure 4. Results from the different studies on a cost-effectiveness plane for NSCLC patients according to anti PD-1 (a) or anti PD-L1 therapy (b).Note: Studies in the figure are identified by reference number.

Cancers 2024 , 22 Figure 5 .
Figure 5. Results from the different studies on a cost-effectiveness plane for patients with cancer other than NSCLC according to anti PD-1 (a) or anti PD-L1 therapy (b).Note: Studies in the figure are identified by reference number.

Figure 5 .
Figure 5. Results from the different studies on a cost-effectiveness plane for patients with cancer other than NSCLC according to anti PD-1 (a) or anti PD-L1 therapy (b).Note: Studies in the figure are identified by reference number.

Table 1 .
Characteristics of studies focusing on non-small cell lung cancer (NSCLC).

et al. Argentina, Brazil, Peru 2018 [41]
* Charlson score was calculated based on comorbidities present in the year before cancer diagnosis using the Comorbidity SAS Macro provided by the National Cancer Institute.Comorbidity burden level was divided into three groups-Charlson score equal to 0 (Charlson 0), Charlson score equal to 1 (Charlson 1), and Charlson score equal to 2+ (Charlson 2+).

Table 2 .
Characteristics of studies stratified by other solid tumours.