Cost–effectiveness of concurrent radiation with cetuximab or chemotherapy in older patients with oropharyngeal cancer

Squamous cell carcinoma of the oropharynx is a distinct subset of head and neck cancers (HNC) that involve the base of the tongue, tonsil, soft palate, vallecula and pharyngeal wall. While historically many oropharyngeal cancer (OPC) cases were related to the risk factors of tobacco and alcohol use, the incidence of oncogenic human papillomavirus (HPV) infection as an etiologic factor has increased over time. The incidence of OPC in the USA is projected to increase from 20,124 cases in 2016 to about 30,629 cases in 2029 [1]. Previous studies have indicated that there is an ongoing shift toward an increased incidence of OPC in the USA among older age groups, in particular, those over 65 years of age, which is likely attributed to HPV-related OPC [1,2]. Trends of increasing HPV-related OPC incidence have been observed in the USA and other economically developed countries [3].

Data from the Surveillance, Epidemiology and End Results (SEER) registry indicate that about 77% of the oral cavity and pharynx cancer cases are diagnosed at a localized or regional stage [4]. While the increasing availability of transoral robotic surgery (TORS) has allowed for surgical resection to be offered as a curative approach to more patients [5], definitive cytotoxic chemotherapy plus radiation (CRT) is still widely employed as a curative treatment option. The added benefit of CRT compared with radiation therapy (RT) alone was established based on multiple clinical trials, the results of which were disseminated between 1999 and 2004 [6–9]. However, the addition of chemotherapy to radiation can be associated with significant toxicity, especially in older individuals, and there are limited prospective data with regard to alternative chemotherapy dosing schedules. Thus, treatment with RT alone remained a relevant option for older patients with comorbidities, who might not tolerate chemotherapy.

In 2006, Bonner et al. presented randomized data showing a survival benefit with the monoclonal antibody EGFR inhibitor cetuximab plus RT (cetuximab-RT) over RT alone [10]. The study was met with widespread interest among head and neck oncologists as it presented a new treatment option, especially for patients at higher risk of increased toxicity from platinum-based chemotherapy. Over time, a debate developed with regard to the efficacy of cisplatin plus RT (cisplatin-RT) versus cetuximab-RT, in part supported by a retrospective analysis that indicated the survival outcomes were inferior with cetuximab-RT compared to cisplatin-RT [11]. In addition, subsequent studies indicated that the toxicity of cetuximab-RT may be more pronounced than originally anticipated [12]. In a randomized clinical trial (RTOG 1016) among patients with HPV-associated OPC, cisplatin-RT outperformed cetuximab-RT in locoregional control, progression-free survival and overall survival (OS) [13]. In another randomized clinical trial (De-ESCALaTE HPV) in patients with HPV-positive low-risk OPC, cetuximab-RT did not demonstrate reduced toxicity compared with cisplatin-RT, but did show inferior tumor control, with a lower rate of two-year OS and a higher rate of two-year recurrence [14].

Thus, while this data did not support the use of cetuximab for HPV-positive OPC, its use for HPV-negative disease has not specifically been studied. There remain questions about efficacy in this subset, coupled with the known increased cost associated with cetuximab. A previously published study demonstrated that treatment with cetuximab-RT is associated with substantial incremental costs compared with treatment with CRT or RT alone [15]. Therefore, it is important to evaluate the clinical and economic outcomes of OPC treatment options, establish their cost–effectiveness and compare their results as the treatment landscape changes over time.

Although there are a few cost–effectiveness studies of cetuximab for locally advanced HNC [16–18], there is a need for studies that focus on OPC given the etiologic nature of this subsite and its common association with HPV. In addition, previous cost–effectiveness studies were conducted by modeling data from clinical trials and this evidence must be complemented by real-world data that reflect actual spending by payers. Moreover, the shifting age demographics and HPV prevalence over time as well as developments in radiation treatment in recent years warrant evaluating how the value of OPC treatments has evolved over a long interval of time (2000–2013).

The objective of this study was to conduct cost–effectiveness analyses to assess the value of definitive therapeutic options for OPC between 2000 and 2013. To facilitate the analysis of cost–effectiveness over time, the study period was divided into two based on the year of diagnosis (2000–2005 and 2006–2011). In the early time period, the survival and costs outcomes of CRT versus RT alone were compared. In the later time period, comparisons included CRT versus RT alone, cetuximab-RT versus RT alone and CRT versus cetuximab-RT.

This study examined the cost–effectiveness of definitive treatments for OPC using 13 years of data for individuals newly diagnosed with nonmetastatic OPC as their first cancer and represented in the linked SEER-Medicare data. Leveraging the long time period, cost–effectiveness estimates were compared over time periods defined by the availability of cetuximab. The comparisons provided novel information regarding the cost–effectiveness of treatments for OPC before and after a significant change in the treatment landscape.

The first comparison of CRT versus RT alone in the earlier time period (2000–2005) showed that the ICER of CRT fell within acceptable ranges of cost–effectiveness, which was anticipated given the expected survival benefit of adding platinum-based chemotherapy [25] at a relatively low cost. The second comparison of CRT versus RT alone in the post-cetuximab era (2006–2011) showed that CRT was dominated by RT because it was associated with higher costs without evidence of improved RMST compared with RT. It was expected that CRT would offer a survival advantage over RT in the post-cetuximab era. However, there are several factors to consider in reviewing these results. While CRT was associated with a numerically improved RMST, the limited follow-up time and the small sample size may have prevented capturing a survival difference. Similar trends were observed in the Kaplan–Meier functions of both groups (Figure 1B) and in the cost–effectiveness scatter plot (Figure 2B). Another factor to consider is the increasing prevalence of HPV among older patients with OPC in more recent time periods. For instance, analysis of HPV prevalence in OPC has shown a significant increase from 64% between 2000 and 2004 to 72% between 2005 and 2009 [26]. HPV-positive OPC has a better prognosis compared with HPV-negative OPC [27], which may be reflected by the observed improvements in survival with both RT and CRT in Cohort II compared with Cohort I (Figure 1A & B). The improved prognosis of HPV-positive OPC and the concerns over morbidity and mortality related to chemotherapy have promoted discussion of treatment de-escalation in this patient group with the goal of reducing toxicity while maintaining tumor control [14]. Several strategies for de-escalating the treatment of HPV-positive OPC have been previously tested and continue to be tested in ongoing trials [28], and this topic may be of special importance in the case of older patients. Additionally, in a meta-analysis that evaluated the benefit of adding chemotherapy to locoregional treatment, the benefit of chemotherapy was found to decrease with older age, potentially due to higher toxicity and reduced adherence to treatment [29]. Nonetheless, future confirmatory analyses are warranted to investigate these findings.

Third, the comparison of cetuximab-RT versus RT alone in Cohort II showed that cetuximab-RT was dominated by RT alone, due to incremental costs of US$46,557 with no incremental survival benefit (the incremental RMST point estimate was negative). These survival results were consistent with an exploratory subgroup analysis from the landmark randomized controlled trial of cetuximab, which suggested no additional benefit of cetuximab among patients 65 years or older [30]. In the fourth comparison, cetuximab-RT demonstrated incremental costs of US$34,422 coupled with inferior survival compared with CRT. The survival results were consistent with two clinical trials and multiple retrospective analyses [11,13,14,31]. These findings suggest that other treatment options are preferred over cetuximab-RT from both survival and cost perspectives.

In HNC, in general, several studies found that cetuximab plus radiation therapy was cost-effective compared with radiation alone [16–18]. These studies utilized modeling approaches (e.g., Markov models or decision trees) and data from the cetuximab landmark clinical trial. However, the current study provides important insights to supplement available evidence because it is based entirely on observed survival and healthcare resource utilization in real-world practice. It also included individuals who are older and with comorbidities and thus, less likely to be enrolled in trials. The current study also provides a novel examination of the value of CRT over time. While the results indicate that CRT may have offered good value compared with RT alone in the pre-cetuximab era, the findings suggest that CRT had similar survival benefits compared with RT alone but at a higher cost.

The study results must be considered in the context of study limitations. First, the observational nature of the study may result in patient selection into treatment groups that may impact the estimated outcomes. Although we aimed to reduce this patient selection bias by matching treatment groups, unobserved characteristics, factors such as performance status and patient preference, could not be controlled for and might have led to residual confounding. Second, in order to reduce the amount of censored cost data, the time horizon of the study was reduced, which may have limited our ability to observe differences between the comparison groups in the long run. Future studies with longer follow-up data are needed to complement the evidence from this study. Third, while the type of cytotoxic chemotherapy (i.e., carboplatin vs low-dose cisplatin vs high-dose cisplatin) may have an impact on patient outcomes, specific types of cytotoxic chemotherapy could not be studied due to sample size limitations. Therefore, as the first step for a cost–effectiveness analysis using real-world data, the family of cytotoxic chemotherapy agents that were utilized in the real-world setting were evaluated as one group. Future studies with larger sample sizes are warranted to evaluate the cost–effectiveness of specific cytotoxic chemotherapy agents. Fourth, this study assessed the value of concomitant RT with cytotoxic chemotherapy or cetuximab as treatment options in nonmetastatic OPC, which included localized (10–16% of the patients in the matched cohorts) and regional tumors (84–90% of the patients in the matched cohorts). In future studies using larger samples, stage-specific subgroup analyses may be warranted to evaluate the cost and survival outcomes of such treatments within each stage group. Fifth, the study reports an additional year of survival as the effectiveness measure and did not adjust the estimates based on quality of life. Additional research on health-related quality of life will be needed to provide reliable inputs for population-based cost–effectiveness studies of OPC treatments based on real-world or modeled data. The study results also may not generalize to individuals diagnosed with OPC who are younger than 66 years old. Lastly, the results may not generalize to the source SEER-Medicare OPC population because all comparisons were conducted using matched samples. While matching is appropriate to maintain the internal validity of the group comparisons, there is potential that the results are not externally valid. Despite these limitations, the results provide important, novel information regarding the value of therapies across a changing treatment landscape for nonmetastatic OPC.

The value of definitive CRT compared with RT has changed over time for older patients diagnosed with OPC. In the earlier time period, the ICER for CRT was US$56,650 per additional year of survival, which may fall within acceptable cost–effectiveness thresholds. In the later time period, CRT was associated with similar survival as RT alone, but at a higher cost. Finally, cetuximab-RT was associated with substantial costs compared with both RT and CRT despite the lack of a survival benefit compared with RT and the inferior survival compared with CRT. Based on these findings, the value of cetuximab for older patients with OPC was not favorable.

H Albarmawi: Conception and design; data analysis and interpretation; drafting the manuscript; final approval of the manuscript; agreement to be accountable for all aspects of the work; KJ Kullen: Conception and design; data interpretation; revising the manuscript critically for intellectual content; final approval of the manuscript; agreement to be accountable for all aspects of the work; R Mehra: Study design; data interpretation; drafting and revising the manuscript; final approval of the manuscript; agreement to be accountable for all aspects of the work; E Onukwugha: Conception and design; data analysis and interpretation; drafting and revising the manuscript; final approval of the manuscript; agreement to be accountable for all aspects of the work; O Goloubeva: Conception and design; data acquisition, analysis and interpretation; drafting and revising the manuscript; final approval of the manuscript; agreement to be accountable for all aspects of the work.

This study used the linked SEER-Medicare database. The interpretation and reporting of these data are the sole responsibility of the authors. The authors acknowledge the efforts of the National Cancer Institute; the Office of Research, Development and Information, CMS; Information Management Services (IMS), Inc.; and the Surveillance, Epidemiology and End Results (SEER) Program tumor registries in the creation of the SEER-Medicare database. The collection of cancer incidence data used in this study was supported by the California Department of Public Health as part of the statewide cancer reporting program mandated by California Health and Safety Code Section 103885; the National Cancer Institute's Surveillance, Epidemiology and End Results Program under contract HHSN261201000140C awarded to the Cancer Prevention Institute of California, contract HHSN261201000035C awarded to the University of Southern California and contract HHSN261201000034C awarded to the Public Health Institute; and the Centers for Disease Control and Prevention's National Program of Cancer Registries, under agreement # U58DP003862-01 awarded to the California Department of Public Health. The ideas and opinions expressed herein are those of the author(s) and endorsement by the State of California Department of Public Health, the National Cancer Institute and the Centers for Disease Control and Prevention or their Contractors and Subcontractors is not intended nor should be inferred. The authors acknowledge the efforts of the National Cancer Institute; the Office of Research, Development and Information, CMS; Information Management Services (IMS), Inc.; and the Surveillance, Epidemiology and End Results (SEER) Program tumor registries in the creation of the SEER-Medicare database.

No external funding was received for the conduct of this study. H Albarmawi reports employment by Genentech, Inc. after the completion of the analysis. E Onukwugha reports research funding from Pfizer, Inc. and BeiGene, Ltd. unrelated to the current work. R Mehra reports research funding from Merck & Co., Inc. and Astra Zeneca Plc. as well as consulting fees (advisory board) from Rakuten unrelated to the current work. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

No writing assistance was utilized in the production of this manuscript.

This study was approved by the Institutional Review Board at the University of Maryland, Baltimore (HP-00067264).