A Systematic Review of the Economic Burden of Proton Therapy in Head and Neck Cancer

Background: Radiation therapy is used to treat head and neck cancer (HNC) patients. Proton beam therapy (PBT) is one of the newer treatment options. This systematic review will describe the cost and cost-effectiveness of PBT compared with other first-line treatment options based on available literature and provide a better understanding of its usage in HNC in the future. Methods: This systematic review was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines. Systematic searches were conducted in PUBMED, EMBASE and SCOPUS till February 2022. Original pharmacoeconomic articles written in English that considered PBT for HNC were included; the title, abstract and full text of the search items were screened. The included studies were critically appraised using the Drummond Checklist followed by data extraction. Results: Eight of the ten included studies were of good quality; most were cost-effectiveness or cost comparison studies and used the Markov model and lifetime horizon. The dominant comparator was intensity-modulated radiotherapy. The willingness to pay threshold ranged from $30,828 to $150,000 per QALY. The incremental cost-effectiveness ratio (ICER) was between $4,436.1 and $695,000 per QALY. In HNC patients with human papillomavirus infection, the ICER was lower ($288,000/QALY) from the payer’s perspective, but much higher ($390,000/QALY) from the societal perspective. Conclusion: Our systematic review showed that appropriate patient selection can make PBT cost-effective. HPV-associated tumors can be cost-effectively treated with PBT. From the payer’s perspective, PBT is a cost-effective treatment option. In younger patients, PBT can result in lesser incidence of adverse effects, and hence, can reduce the subsequent need for long-term supportive care. Lower fractionation schedules can also make PBT a cost-effective treatment.


Introduction
Head and neck cancer (HNC) is the seventh most common malignancy in the world with an annual incidence of 800,000 cases and 300,000 deaths (Kawakita et al., 2022).South and Southeast Asia have the highest malignancy rates (Argirion et al., 2019); India currently accounts for roughly 60% of all HNC cases worldwide., and this number is projected to double by 2030 (Prabhash et al., 2020).In the United States (US), 53,000 new cases of HNC and 10,860 fatalities were reported in 2019 (Siegel et al., 2019).
Currently, whether used as the main treatment or as an adjuvant after surgical resection, radiation helps around 75% of patients with HNC (Alfouzan 2021).However, developments in newer targeted therapies over the past two decades coupled with increasing experience in the use of radiation therapy have made more treatment options available for an individual patient and increased the complexity of decision-making.Radiation exposure REVIEW A Systematic Review of the Economic Burden of Proton Therapy in Head and Neck Cancer to vital bodily structures is decreased by intensitymodulated radiation treatment (IMRT), which offers a precisely targeted dose distribution.It was demonstrated in a randomized trial that IMRT performed better than conventional therapy in improving xerostomia (Grutters et al., 2010;Nutting et al., 2011).Protons rather than photons are used to administer the dose in intensity-modulated proton therapy (IMPT), a more contemporary delivery method for proton beam treatment (PBT) (Lukens et al., 2015).
PBT was established for use in radiotherapy to target tumors with proximity to vital anatomical structures.This method enables the radiation energy to be accurately focused at a given depth, resulting in dosage reduction.It differs from external photon (x-ray) radiotherapy in that it allows the radiation dose delivered to the tumour area to be increased without a corresponding increase in exposure to the surrounding healthy tissues (Alfouzan 2021).In patients with nasal cavity, paranasal sinus, and nasopharyngeal cancer, dosimetric comparisons of IMPT with IMRT indicate better sparing of the parotid glands, oral cavity, oesophagus, and larynx (Holliday et al., 2016;Jeremic et al., 2021).Therefore, IMPT's dosimetric advantages can lessen the radiation toxicity, both acute and long-term, in patients with HNC (Moreno et al., 2019).PBT delivery, on the other hand, necessitates sophisticated equipment and is far more expensive than photon-based therapy.
Over the past decade, a large number of centres equipped to offer IMPT have been established, with more than 70 now in operation and another 40 in various stages of development globally (Jones et al., 2019).PBT is now more widely available, and as a result, more clinical evidence demonstrating its significance and efficacy has been produced (Verma et al., 2016).With the available resources and cost associated with proton therapy, information regarding cost and cost-effectiveness is potentially crucial for healthcare decision-makers The available evidence is limited and a thorough systematic review and assessment of available studies including studies based on modeling could potentially help to identify areas for further research.Though the benefits of proton therapy have been shown in the treatment of HNC, there is still doubt and debate over the potential eventual role of PBT in disease management given the paucity of level 1 evidence for PBT in HNC (Gunn et al., 2016;Mody et al., 2021).The debate appears to have centered on cost-effectiveness and cost-competitiveness in recent years.From the standpoint of global health, PBT's cost is not outrageous and shouldn't be used as an excuse to deny our patient's potentially curative and less toxic therapies (Lievens and Van den Bogaert 2005).Even though PBT is currently expensive, it may become far less expensive if the initial built-in cost recovery is not considered.The additional cost of PBT can be justified only if it provides a significant clinical benefit.This needs an understanding of the outcomes and dose distributions to specific cancer sites (Goitein and Jermann 2003).Extensive pharmacoeconomic evaluations on proton therapy were carried out for prioritized sites, basically pediatric tumors and base of skull cancers, but not for head and neck cancers which is one of the important sites (Ontario Health 2021;Austin et al., 2019).This systematic review is an attempt to describe the cost and cost-effectiveness of IMPT compared with other treatment modalities based on data from published studies in HNC.

Materials and Methods
The protocol was entered into PROSPERO under the registration number CRD42022306597 (Accessible from: https://www.crd.york.ac.uk/prospero/display_record.php?ID=CRD42022306597).We used a particular search strategy (https://sites.google.com/view/searchstragegypbt/home) to conduct systematic searches in PUBMED, EMBASE, and SCOPUS, three electronic bibliographic databases, from January 2011 to February 2022.
Original articles for cost studies that only consider PBT for HNC treatment, pharmacoeconomic studies (costeffectiveness, cost-minimization, cost-benefit analyses and cost-utility), and cost comparison evaluations that take into account both proton therapy and comparators, regardless of the comparator or comparator(s), for HNC treatment were incorporated in the review.Articles published in the aforementioned time period in the English language from any nation were taken into consideration.
The following studies were excluded: editorials, commentaries and systematic reviews; preclinical and other research that did not fall under the aforementioned categories; studies whose objectives did not include economic evaluation cost comparison or economic evaluation, such as budget impact analyses, the burden of disease or cost of illness.
First, the article titles and abstracts were reviewed separately by two authors; any discrepancies in the screening were settled by the third author.The entire studies that were qualified were also analyzed for relevance.A Microsoft Excel sheet was used to define and record the coding for the inclusion criteria and the exclusion criteria for each step.Finally, references of all the shortlisted studies were screened for any additional studies that met the inclusion/exclusion criteria.
The Drummond Checklist was used by two authors to critically evaluate the included research impartially (Drummond 2005).The primary outcomes, including incremental cost-effectiveness ratios (ICERs), cost per treatment and cost/quality-adjusted life year (QALY), were extracted into a pre-piloted, standard format by two authors independently, and any discrepancies were settled by the third author.Additional outcomes like cost per treatment, methods/ source utilized to calculate the cost, efficacy/ effect of therapy, methods/ source used to estimate effectiveness, benefits, discount rate, time horizon; standard outcomes for cost-utility and costeffectiveness analysis, analysis of sensitivity, willingness to pay (WTP) threshold or uncertainty measures, expected reported value of perfect value, the pharmacoeconomic evaluation type, analysis type and utilities were also obtained in a similar manner.

Results
The preferred reporting items for systematic reviews and meta-analyses (PRISMA) flow chart is illustrated in Figure 1.

Characteristic features of pharmacoeconomic literature included in the systematic review
Characteristic features of the pharmacoeconomic literature are presented in table 1 and table 2. The included studies were hosted in the USA, the Netherlands, China and Sweden.All the studies used the Markov model as the study design except for two studies that used case control.Eight studies compared proton therapy (IMPT, PBT) with photon therapy (IMRT).One study compared PBTwith both photon and combined therapy and one study compared PBT with conventional radiation.
Four evaluations were conducted using the payer's perspective, two used the provider's perspective, two used the societal perspective and one used both the payer and the societal perspective.Two studies from USA done from a payer perspective showed comparable  $100,000 per QALY and $150,000per QALY Assumed symmetric triangular distributions (from 0% -50%) 25% mode for I-year reduction dysgeusia, PEG, and xerostomia.The impressive improvement (50%) in this distribution was considered a "best-case scenario," to intentionally favor the more expensive therapy.
For 65-year-old patients, the mean net monetary benefit of perfect information was $0 in both perspectives at a societal WTP of $100,000/QALY.Over a 30-year period, at a societal WTP of $150,000/ QALY, the total value of information was $247,000 from the payer perspective and $0 dollars from the societal perspective.
CRT (first 3 mo), Base NED (after 3 mo), Base NED with PEG (first 2 y), Base NED with PEG, (after 2 y), Dysgeusia disutility, Xerostomia disutility, DM, Dead Ramaekers et al., 2013 Price/ unit and use of resources depended on guidelines, the cross-sectional survey, or expert opinions.IMPT treatment cost was calculated by multiplying IMRT treatment cost with a 2.1 cost ratio.To consider IMPT-if efficient, IMRT and IMPT plans were compared for each dose distribution and decide upon the most efficient treatment per patient.Therefore, the costs of an extra treatment plan (€88) were added for this strategy.A half-cycle correction was applied for QALYs and costs.
The occurrence of xerostomia and/or dysphagia was estimated according to 2 available NTCP models.The proportion of patients who had both xerostomia and dysphagia was calculated using conditional toxicity probabilities from a cross-sectional survey.Scoring for the utility was sourced from cross-sectional research (with n = 396) by using a Dutch Euroqol 5D questionnaire in HNC patients.Utility scoring were added with the life expectancy to derive QALY.
Lifetime horizon used.The future QALY and the costs were further discounted by 1.5 and 4.0 %, respectively.Costs were converted to 2010 money value.Expected mean costs, toxicity occurrence, disease and toxicity free lifeyears, QALYs, ICER.

Partial economic evaluations assessing the cost
The main reason why IMPT is more expensive than IMRT on average is because of higher equipment expenses, although a subset of IMRT patients have expenditures that are comparable to IMPT patients because they require supportive care resources more frequently.Sher et al., 2018 In terms of payer's and societal perspectives, the ICER for proton therapy in HPV-positive patients is $ 288,000/QALY and $ 390,000/ QALY, respectively.From the payer's and societal perspectives, HPV-negative patients ICER is $ 516,000/QALY and $ 695,000/QALY, respectively.
IMRT $23,137 (payer) and $27,192 (societal) IMPT $45,457 (payer) and $56,659 (societal) CEA IMPT is cost-effective only from Payer's perspective if it significantly reduces the long-term morbidity in the younger population; not cost-effective from the societal perspective

Ramaekerset al., 2013
When IMPT if efficient (expected to be cost-effective) was compared with IMRT for all the patients an ICER of € 60,278 / QALY was calculated.When IMPT for every patient was compared with the IMPT if found efficient, an ICER of € 127,946 / QALY was calculated.

Brodin et al., 2021
Considerable Patient-to-patient variation in the estimated ICERs, with a median of $361,405/QALY (IQR, $45,453-$1,556,948) for the whole cohort.Patients under 65 years old have a median ICER of $341,081/QALY as opposed to patients over 65, who have a median ICER of $399,533/QALY.According to the p16 status, the median ICER for patients with p16 negative tumors was $ 516,297 / QALY while the median ICER for those with p16 positive tumors was $ 234,201 / QALY.$ 20,257 and $ 36,659, as the initial price of 33 IMRT and IMPT fractions respectively.

Discussion
Our systematic review identified 10 studies conducted between 2009 to 2021.The study quality assessment based on 10-item Drummond's checklist showed that eight of the ten included studies received a score ≥7.The studies by Cheng et al., (2016) and Ning et al., (2020) scored 3 and 4, respectively; costs and consequences were not accurately measured in the right physical units in either of these studies, their values were not modified for differing timelines, no incremental analysis of alternatives was carried out, and neither the presentation nor the test results addressed all user-relevant concerns.The quality of the source research affects the internal validity of the synthesis produced by the systematic review.Cheng et al., (2016) suggest that proton therapy would be cost-effective in 35% of HNC patients at a WTP of €80,000/QALY without analyzing the ICER.Ning et al., (2020) suggest that though the direct costs of proton therapy are higher upfront, the patients can still obtain benefits from lesser medical costs over their lifetime.Since these costs were not measured in appropriate physical units, not adjusted for differential timing, and no allowance was made for uncertainties in cost estimation, the findings of these studies need to be considered with caution.
Except for two case-control studies, all other studies used the Markov model which simulates the development of chronic conditions, such as a tumour, through numerous cycles of operation, and therefore, is useful to assess the long-term cost-effectiveness of cancer therapy (Russell et al., 1996).Eight studies compared IMPT with IMRT.The head and neck area is an excellent target for IMRT, thereby making it the lead comparator (Lee et al., 2007).A horizon of a lifetime has been used in most studies even though there is a lack of long-term evidence following proton therapy.The premise is that there is QALY) when compared to both HPV positive and HPV negative status in societal perspective ($ 390,000/QALY, $ 695,000/QALY) where PBT was not found cost-effective.All other evaluations included suggested that a particular subset of patients is cost-effective.(refer to Table 1) A lifetime horizon was used as a time horizon in majority of the studies.In one study, the time horizon was from the first hospital visit to 30 days post-radiation therapy; in another, the time horizon was considered from one month before to six months post-treatment.The sources of cost were literature, medicare payment schedule, institutional database, cross-sectional survey and value-based analysis.The populations in included studies were head and neck patients, oropharyngeal cancers patients, nasopharyngeal carcinoma patients and paranasal sinus and nasal cavity cancer patients.
Willing to pay (WTP) threshold is stated in five studies.Three studies mentioned the WTP threshold in US dollars ranging from 30,828 to 1,50,000 per QALY.Two studies mentioned the WTP threshold in euro currency as 80,000 per QALY.

Quality of pharmacoeconomic literature included in the systematic review
The study quality is illustrated in the 10-item Drummond's checklist in Table 3. Eight of the 10 included studies received a score ≥7.The items with low scores were regarding comparator, effect, cost and consequences credible valuation, incremental analysis and incomplete results.

Principal findings of pharmacoeconomic literature included in the systematic review
Full economic evaluation and cost-effectiveness analysis was conducted in all except three studies (refer to Table 4) where partial economic evaluation was carried on to primarily focus on the costs.This was done to analyze and compare the costs of two or more alternatives without taking into account the effects.All these studies which assessed only cost used specific analysis methods.Thaker et al. (2021) used time-driven activity-based costing (TDAB), a tool that helps providers study alternative payment models.Peeters et al. (2010) conducted a cost analysis comparing per fraction and per treatment costs.Ning et al. (2020) utilized value-based analysis that addressed both employer and payer's potential concerns.
The efficacy of the proton therapy and comparators in the studies were analyzed in the cost-effectiveness analysis.It includes QALY, survival rates, and tissue complications.Among the 7 studies that conducted cost-effectiveness analysis, only 5 calculated ICER; ICER ranged from $4436.1 to $695000.This wide variation in ICER can be explained based on the country where the study was conducted, HPV p16-status, the pharmacoeconomic perspective used, age of the patient, efficiency and reduction in complications.The upper limit of the ICER range in China ($102,684/QALY) is much less than the lower limit in the USA ($234,201/ QALY).The ICER upper limit in studies from the Netherlands (€127,946) is also lesser than that from the USA.These variations reflect the varied healthcare costs in each no substantial difference between the two therapies in terms of delayed recurrence risk or toxicity risk despite their potential improvements (Sher et al., 2018).This time horizon is used because delayed side effects of radiation, which generally manifest after six months, can adversely affect the quality of life during the entire life span.A shorter follow-up will account for only the acute side effects and will not reflect the actual QALY.While majority of published studies focused on all HNC subsites, three of them focused on oropharyngeal malignancies which have a high propensity for HPV positivity, with associated superior treatment response and survival.The cost-effectiveness threshold, which reflects society's WTP for an additional unit of benefit, is a significant issue connected to the generalizability of research results.Studies from the Netherlands displayed a highly uniform pattern with respect to the threshold, which was €80,000/QALY gained.Studies from the United States displayed an extended range of this threshold, from $100,000 to $150,000/QALY gained.However, the Chinese studies had a consistently lower threshold of $30,000/QALY gained.WTP for healthcare services is based on the thematic domains of sociodemographic traits; perceived barriers, benefits, and threats; study design and setting.This explains the variation across the nations (Steigenberger et al. 2022).This predetermined WTP is used to compare with ICER to determine what will constitute a cost-effective approach in that country (McDougall et al., 2020).
Cost-effectiveness analysis was conducted in seven included studies; only five of these determined ICER.The valuable measurement ICER is the ratio of the variation in costs of two medical procedures to the variations in their outcome.Hence, ICER represents extra costs per increased units of treatment outcome received by changing between one medical treatment and the other.Despite having certain drawbacks, ICER is one of the crucial measures required to assist and guide decisions about allocating limited resources among competing healthcare programs (Bambha and Kim, 2004).In the current review, ICER ranged from $4,436.1 to $695,000.HPV virus infection/ p16 status, the pharmacoeconomic perspective used, patient age, and incidence of adverse effects accounted for the variation in ICER.Sher et al., (2018) concluded the cost-effective superiority of proton therapy in younger HPV-positive patients; ICERs for proton therapy were $288,000/ QALY and $516,000/QALY in favorable HPV-positive and negative patients, respectively.Brodin et al., (2021) inferred that, when compared with p16 negative tumors ($516,297 per QALY), p16 positive tumors ($234,201 per QALY) were cost-effectively treated with the proton therapy.Both these studies which were done in the USA from the payer's perspective give us a comparable ICER for both HPV positive and negative status which makes this inference highly relevant.Contrasting with p16 negative tumors, ≥ 50% of p16 positive tumors were cost-effectively treated with proton therapy at $500,000 per QALY (Brodin et al., 2021).But it is worth to mention that a very high ICER is still observed in these tumors.
The various vantage points from which health costs and benefits might be evaluated are known as perspectives (Tonis et al., 2021).Societal perspective is the most thorough since it incorporates the perspectives of all healthcare stakeholders and aims to reflect the complete spectrum of societal opportunity costs related to various interventions.This specifically refers to productivity losses brought on by patients' incapability to work and modifications to these losses brought on by new technologies (McIntosh and Luengo-Fernandez, 2006).Patient mortality and morbidity as well as the total cost of providing and receiving medical care are considered from a societal perspective.Societal perspectives are more prevalent in nations with nationalized healthcare.When insurance firms and employers collaborate together to choose medical coverage for their employees, the payer perspective is used.
According to the current research, even with a proton facility that costs $25 million per gantry, there is no chance that PBT would indeed be cost-effective from the societal perspective under given favorable assumptions.However, the United States and other countries already have a large number of proton facilities.Compared to the up-front costs of building a proton center, the marginal cost for each patient is important from the payer's perspective.Payer's perspective in the pharmacoeconomic evaluation is of value only when PBT is cost-effective when compared to its remuneration.In the study done by Sher et al., (2018), the payer's perspective found IMPT to be cost-effective, while a societal perspective did not.
Current policy recommendations state that a clinically substantial decrease in toxicities (xerostomia, oral mucositis, dysphagia, and percutaneous endoscopic gastrostomy tube implantation) is necessary for patients to be cost-effective candidates for PBT (Health Care Insurance Board 2011; Sherry et al., 2021).Lunkdvist et al., (2005) indicated that, if the right risk categories are selected as proton therapy targets, proton therapy might be a cost-effective therapy and that a proton treatment facility investment could be better cost-effective relative to employing conventional radiation.Individuals who have an increased risk of adverse effects should be identified and treated with proton therapy in practice advantages of this treatment.
The study by Li et al., (2022) shows that, according to the present WTP in China ($33558 per QALY), NTCP of xerostomia, dysphagia and sensorineural hearing loss should be reduced by ≥17%, ≥19%, ≥20%, ≥24%, ≥28% and ≥39% in patients aged >10, >20, >30, >40, >50, and >60 years appropriately, for proton therapy to be cost-effective; proton therapy is not cost-effective for the ≥70-year-old patients as NTCP reduction of ≥90% needs to be achieved.According to Brodin et al., (2021), patients less than 65 years of age had a median ICER of $ 341,081 per QALY, whilst patients equal and greater than 65 years of age had a median ICER of $ 399,533 per QALY.As per the study by Li et al., (2020) of the cost-effectiveness of IMPT in patients with varying ages, the ICERs ranged from $14,999.4 per QALY at age 0 to $74,440.1 per QALY at 70 years, respectively.Therefore, utilizing the existing WTP threshold of China, IMPT was deemed to be cost-effective in all the patients with age ≤56 years.Hence, it can be inferred that head and neck proton therapy is cost-effective in patients of younger age groups who are at risk of treatment-related complications that warrant additional supportive care following treatment.
In three of the included studies, only partial economic evaluation was done to analyze and compare the costs of two or more alternatives without taking into account the outcomes (Thaker et al., 2021;Peeters et al., 2010;Ning et al., 2020).Thaker et al., (2021) determined that IMPT expenses are, on average, greater than IMRT because of the increased equipment expenditures.They identified that 28% of the total costs of IMRT and IMPT patients were overlapping.Higher expenses of IMRT resulted in part from a larger usage of supporting resources (such as emergency, gastroenterology, and inpatient care), which may not have been needed with IMPT because of its improved dosimetry.Ning et al., (2020) found that despite having greater direct costs, PBT may provide long-term indirect benefits in terms of productivity and disability.The overall cost of medical care did not increase in the setting of ancillary service.Both studies point out that PBT reduces the cost of treatment by reducing the need for supportive care.Peeters et al., (2010) suggested that it may be possible to administer particle treatment in the future using much lower fractions than now, thereby reducing the cost of particle treatment.
While most of the included studies estimated the total cost, Thaker et al., (2021) and Ning et al. (2020) estimated incremental cost differences in their study as they were unable to reveal absolute costs for each step in the workflow due to the sensitive nature of internal expenditures but instead offered an overview of relative costs.Lunkdvist et al., (2005) in their study indicated that investing in the proton setting might be deemed cost-effective based on the simulation's assumptions.It should be mentioned that the study findings were predicated on the idea that the proton facility exclusively treated patients with the four forms of cancer-medulloblastoma, breast cancer, prostate cancer and HNC.It won't be the case in practice, since identifying other patients who would be more cost-effective might potentially boost the cost-effectiveness of the proton therapy (Lunkdvist et al., 2005).Only articles published in English were included; so there is a chance that all the relevant articles were not included in this systematic review.The quantitative description cannot be done, since the cost per treatment and outcome was not measured using uniform criteria in the included evaluations.The majority of the studies have not taken into consideration the initial investment costs and machine throughput.Also, with increasing adoption of the technology, the cost of machinery is bound to reduce, and thereby increase cost-effectiveness; this has not been considered in the studies included in this systematic review.
To conclude, our systematic review found that proton therapy in HNC was cost-effective from the payer's perspective when compared to the societal perspective.There was a significant variation in WTP amongst the countries, with lower WTP countries demonstrating greater cost-effectiveness.Appropriate patient selection is required.The proton therapy will be cost-effective in HPV-associated tumors, in young patients due to lesser incidence of adverse effects by decreasing the need for supportive care and in patients with lower fractionation schedules.

Figure 1 .
Figure 1.The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) Flow Chart is Illustrated

Table 2 .
Features of Incorporated Pharmacoeconomic Evaluations Asian Pacific Journal of Cancer Prevention, Vol 24 3647

Table 3 .
Quality of Incorporated Pharmacoeconomic Evaluations levels to evaluate.