Costs Associated With Imaging Surveillance After Treatment for Head and Neck Cancer

Key Points

Question

What are the costs associated with imaging surveillance after treatment for head and neck cancer (HNC)?

Findings

In this economic evaluation study including 136 patients, 14 imaging studies were performed for patients with HNC during a mean surveillance period of 3.2 years, with a total cost of $36 800. In patients without HNC recurrence, a mean of 13 studies were performed during surveillance, with a total cost of $35 000, yet only half the patients who developed recurrence had imaging surveillance in accordance with guidelines.

Meaning

The findings of this economic evaluation suggest that the cost burden of imaging in asymptomatic patients with HNC should be considered in this current health care environment.

Abstract

Importance

The National Comprehensive Cancer Network recommends imaging within 6 months after treatment for head and neck cancer (HNC). Further imaging is recommended only if the patient has symptoms or abnormal findings on physical examination. However, in many instances, asymptomatic patients continue to have imaging evaluations.

Objectives

To assess practice patterns in surveillance imaging in patients with HNC and evaluate the costs associated with these imaging practices.

Design, Setting, and Participants

This single-institution retrospective economic evaluation study screened 435 patients to identify patients newly diagnosed with head and neck mucosal and salivary gland malignant tumors between January 1, 2010, and December 31, 2016. Data analyses were performed from October 25, 2018, to November 24, 2020.

Exposure

Imaging practice patterns.

Main Outcomes and Measures

Number and costs of imaging studies during the surveillance period for all patients, patients who remained disease free, and patients who developed recurrence.

Results

A total of 136 patients (mean [SD] age at diagnosis, 62 [14] years; 84 [61.8%] male; 106 [77.9%] White) with HNC were included in the study. The oropharynx was the most common subsite (64 [47.1%]), most HNCs were stage IVA (62 [45.6%]), and most patients received definitive radiation-based treatment (71 [52.2%]). During the median surveillance period of 3.2 years (range, 0.3-6.8 years), a mean (SD) of 14 (10) imaging studies were performed for all patients, with a mean (SD) total cost of $36 800 ($24 500). In patients who remained disease free, a mean (SD) of 13 (10) imaging studies were performed during the surveillance period, with a mean (SD) total cost of $35 000 ($21 700). Patients who lacked symptoms had a mean (SD) of 4 (3) studies performed per year, resulting in a mean cost of $9600 ($5900) per year. Patients who developed recurrence had more studies per year of follow-up (mean difference, 5.0; 95% CI, 3.4-6.6) and higher associated mean costs (mean difference, $10 600; 95% CI, $6100-$15 000) than patients who remained disease free.

Conclusions and Relevance

In this economic evaluation study, many patients treated for HNCs received imaging studies beyond what is recommended by National Comprehensive Cancer Network guidelines. These findings suggest that the cost burden of imaging in the asymptomatic patient needs to be considered against the value obtained from routine imaging in this current health care environment.

This economic evaluation assesses practice patterns in surveillance imaging in patients with head and neck cancer and evaluates the costs associated with these imaging practices.

Introduction

Head and neck cancer (HNC) accounts for 3% of all cancers in the US, with approximately 48 000 new cases annually.¹ Traditionally, most of these malignant tumors were alcohol and tobacco related. However, during the past 2 decades, human papillomavirus has emerged as a newly recognized risk factor for the development of HNC.^2,3 Human papillomavirus–related HNC tends to affect a younger patient population and have a better response to treatment, leading to an increase in HNC survival.⁴ Given this larger population, issues related to long-term follow-up and surveillance care have become increasingly apparent.

An estimated $4.2 billion is spent annually on HNC,⁵ and this figure is expected to increase each year. The most significant factors associated with the costs include hospitalizations, surgery, radiotherapy, and diagnostic imaging. When considering these costs, it is important to identify factors that are associated with increased patient survival and quality of life but can reduce overall financial burden. Although it is difficult to alter costs associated with primary cancer treatment (surgery, radiotherapy, and chemotherapy), an area for improvement involves the use of posttreatment imaging, which is aimed at identifying locoregional recurrence and distant metastases. Research thus far has failed to support the notion that increased surveillance imaging leads to prolonged survival or increased patient well-being. With regard to posttreatment imaging, the National Comprehensive Cancer Network (NCCN) guidelines recommend 1 baseline image within 6 months of treatment completion. The baseline image impacts decision-making, assists with prognosis, and is a cost-effective strategy for decreasing further interventions.^6,7 The NCCN guidelines expressly recommend against further use of routine surveillance imaging in otherwise asymptomatic patients because there is no proven survival benefit and associated costs (financial, radiation exposure to the patient, and use of limited hospital resources).⁸ Despite the NCCN guidelines’ recommendation against routine surveillance imaging in asymptomatic survivors of HNC, a recent survey⁹ of physicians who treat HNC found that up to 79% of physicians continued to obtain these follow-up images in their asymptomatic patients, with up to 39% of physicians using them in more than half of their follow-up patients.

As the health care environment shifts to an increased focus on value and long-term patient-centered outcomes, it is essential to ensure that our tests are providing an important health benefit. The goals of this study are to assess practice patterns in surveillance imaging in patients with HNC and evaluate the costs associated with these imaging practices.

Methods

We performed a retrospective review of adult patients newly diagnosed with and treated for mucosal and salivary gland malignant tumors at a single institution between January 1, 2010, and December 31, 2016. Study data were collected and managed using REDCap (Research Electronic Data Capture) tools hosted at NorthShore University HealthSystem.^10,11 REDCap is a secure, web-based software platform designed to support data capture for research studies, providing (1) an intuitive interface for validated data capture, (2) audit trails for tracking data manipulation and export procedures, (3) automated export procedures for seamless data downloads to common statistical packages, and (4) procedures for data integration and interoperability with external sources. This study was approved by the institutional review board at NorthShore University HealthSystem with a waiver of informed consent. All data were deidentified. This study followed the Consolidated Health Economic Evaluation Reporting Standards (CHEERS) reporting guideline.

A total of 435 patients were screened. Patients were excluded if they were lost to follow-up in less than 2 years, did not have complete medical records available for review, or had metastatic disease at presentation, leaving 136 patients. Basic patient demographic and clinical variables, including tumor subsite, staging, and treatment details, were collected for analysis. Current Procedural Terminology (CPT) codes for neck ultrasonography, neck computed tomography (CT), chest CT, maxillofacial CT, neck magnetic resonance imaging, chest radiography, and positron emission tomography (PET) were used to determine the number of imaging procedures obtained per patient and were then cross-referenced with the medical record to determine whether symptoms or abnormal physical examination findings were present at the time of the imaging study. Symptoms included pain, dysphagia, odynophagia, otalgia, sore throat, shortness of breath, voice change, and neck mass.

Professional and hospital charges linked to the CPT codes associated with each imaging study encounter were used as a proxy for costs. Imaging studies associated with emergency department visits or inpatient stays not associated with the patient’s HNC diagnosis were excluded. Primary outcome measures were overall number and costs of imaging for all patients, in patients who developed recurrence, and in patients who remained disease free.

Study variables were summarized using number (percentage), mean (SD), or median (range). Comparisons between groups were summarized with mean differences and 95% CIs. Statistical analysis was performed using SAS software, version 9.3 (SAS Institute Inc).

Results

We identified 136 patients who met criteria for inclusion in this study (mean [SD] age at diagnosis, 62 [14] years; 84 [61.8%] male; 106 [77.9%] White). The oropharynx was the most common subsite (64 [47.1%]) followed by the oral cavity (24 [17.7%]). On the basis of American Joint Committee on Cancer staging system, 62 patients (45.6%) had stage IVA disease followed by 34 (25.0%) with stage III disease. Most patients received definitive radiation-based treatment (71 [52.2%]). The median follow-up period was 3.2 years (range, 0.3-6.8 years). Patient characteristics are listed in Table 1.

Table 1. Characteristics of the 136 Study Patients^a.

Characteristic	Finding
Age at diagnosis, mean (SD), y	62 (14)
Follow-up, median (range), y	3.2 (0.3-6.8)
Sex
Female	52 (38.2)
Male	84 (61.8)
Race
White	106 (77.9)
African American	5 (3.7)
Asian	4 (2.9)
Other races or not provided	21 (15.4)
Subsite
Oral cavity	24 (17.7)
Oropharynx	64 (47.1)
Larynx	10 (7.4)
Hypopharynx	10 (7.4)
Nasopharynx	7 (5.2)
Sinonasal	4 (2.9)
Salivary	15 (11.0)
Unknown	2 (1.5)
T classification
T0	2 (1.5)
T1	43 (31.6)
T2	43 (31.6)
T3	28 (20.6)
T4a	12 (8.8)
T4b	6 (4.4)
Unknown	2 (2.9)
N classification
N0	41 (30.1)
N1	28 (20.6)
N2a	6 (4.4)
N2b	38 (27.9)
N2c	16 (11.8)
N3	3 (2.2)
Unknown	4 (2.9)
Overall stage
I	21 (15.4)
II	10 (7.4)
III	34 (25.0)
IVA	62 (45.6)
IVB	6 (4.4)
Unknown	3 (2.2)
Treatment
Primary surgery	64 (47.1)
Primary radiation	71 (52.2)
Concurrent chemotherapy	61 (44.9)
Other or unknown	3 (2.2)
Recurrences (n = 27)
Local	9 (33.3)
Regional	9 (33.3)
Distant	9 (33.3)

^aData are presented as number (percentage) of patients unless otherwise indicated.

During the median surveillance period of 3.2 years, a mean (SD) of 14 (10) imaging studies were performed for all patients, with a mean (SD) total cost of $36 800 ($24 500) (Table 2). Of a total 1914 imaging studies during the study period, 1268 (66.2%) were CT, 381 (19.9%) were chest radiography, 129 (6.7%) were PET, 96 (5.0%) were magnetic resonance imaging, and 40 (2.1%) were ultrasonography. These numbers translated into a mean (SD) of 5 (4) imaging studies per year of follow-up, with a mean (SD) cost of $11 800 ($8200), with CT the most common modality (Table 2). Imaging studies were slightly more common in the first year (mean [SD], 4.5 [3.3]) of follow-up compared with the second year (mean [SD], 4.0 [2.6]; mean difference, 0.5; 95% CI, −0.1 to 1.1). Twenty-seven patients developed recurrence after treatment, of whom 13 had documented symptoms or abnormal examination findings to prompt further imaging workup. The most common subsites for the 14 patients without a prompt for further imaging workup were salivary (n = 6), oropharynx (n = 3), and hypopharynx (n = 3). The remaining 109 patients remained disease free throughout the study period, 92 of whom were asymptomatic or lacked abnormal examination findings.

Table 2. Number and Total Charges for All Imaging Studies.

Group	No. of imaging studies	Total charges (in thousands), $
Entire surveillance period
All patients (N = 136)
Mean (SD)	14.1 (10.2)	36.8 (24.5)
Median (range)	11 (0-49)	34.5 (0.0-110.9)
No recurrence (n = 109)
Mean (SD)	13.4 (9.9)	35.0 (21.7)
Median (range)	11 (0-49)	34.5 (0.0-93.5)
Recurrence (n = 27)
Mean (SD)	17.0 (11.2)	44.3 (33.1)
Median (range)	14 (3-38)	34.7 (5.3-110.9)
Mean difference (95% CI)	3.7 (−0.7 to 8.0)	9.3 (−4.3 to 23.0)
Per year of surveillance
All patients (N = 136)
Mean (SD)	4.8 (4.2)	11.8 (8.2)
Median (range)	4 (0-18)	10.0 (0.0-40.0)
No recurrence (n = 109)
Mean (SD)	3.8 (2.5)	9.7 (5.8)
Median (range)	3 (0-17)	9.3 (0.0-25.2)
Recurrence (n = 27)
Mean (SD)	8.7 (6.6)	20.3 (10.9)
Median (range)	8 (1-32)	22.3 (1.5-40.0)
Mean difference (95% CI)	5.0 (3.4-6.6)	10.6 (6.1-15.0)
Per year of surveillance in disease-free patients
No symptoms (n = 92)
Mean (SD)	3.6 (2.5)	9.6 (5.9)
Median (range)	3 (0-17)	9.4 (0.0-25.2)
Symptoms (n = 17)
Mean (SD)	4.4 (2.9)	10.4 (5.3)
Median (range)	4 (1-14)	9.3 (2.5-22.9)
Mean difference (95% CI)	0.8 (−0.5 to 2.1)	0.8 (−2.2 to 3.9)
Per year of surveillance in patients with recurrence
No symptoms (n = 14)
Mean (SD)	10.3 (7.5)	23.8 (9.3)
Median (range)	8 (2-17)	23.4 (8.3-40.0)
Symptoms (n = 13)
Mean (SD)	7.1 (5.3)	16.5 (11.6)
Median (range)	8 (1-18)	19.2 (1.5-32.8)
Mean difference (95% CI)	−3.2 (−8.4 to 2.0)	−7.4 (−15.7 to 0.9)

In patients who remained disease free, a mean (SD) of 13 (10) imaging studies were performed during the surveillance period (4 [3] imaging studies per year of follow-up). This number translated into overall costs of $35 000 ($21 700) for these imaging studies. On a per-year analysis, disease-free patients were charged $9700 ($5800) for imaging studies. Patients who remained asymptomatic and those who developed concerning symptoms or examination findings had a mean (SD) of 4 (3) imaging studies performed per year, resulting in a mean (SD) cost of $9600 ($5900) for asymptomatic patients and $10 400 ($5300) for patients with concerning symptoms or examination findings per year (mean difference, $800; 95% CI, −$2200 to $3900) (Table 2).

As expected, the 27 patients who developed recurrence had more imaging studies per year of follow-up (mean difference, 5.0; 95% CI, 3.4-6.6) and higher associated costs (mean difference, $10 600; 95% CI, $6100-$15 000) than patients who remained disease free. Among the patients with disease recurrence, there were fewer studies (mean difference, −3.2; 95% CI, −8.4 to 2.0) and lower costs (mean difference, −$7400; 95% CI, −$15 700 to $900) in the 13 patients who were symptomatic or had abnormal examination findings and developed recurrence compared with those who were asymptomatic with no abnormal examination findings (Table 2).

Among all patients, 14 (10.3%) had a recurrence detected after routine imaging, which translated into a mean (SD) cost of $23 800 ($9300). Among the 13 (9.6%) who presented with abnormal signs or symptoms that prompted further workup and eventual confirmation of recurrence, the mean (SD) cost was $16 500 ($11 600). Among the 109 patients (80.1%) remained disease free, a mean (SD) cost of $35 000 ($21 700) was incurred during the surveillance period, which translated into $9700 ($5800) per year of follow-up.

Discussion

This economic evaluation found a mean (SD) cost of $36 800 ($25 800) for surveillance imaging after patients were treated for HNC, which translated into a mean (SD) of 5 (4) imaging studies and $11 800 ($8200) per year of follow-up. These costs included not only the ordered imaging examinations but also the associated expenses of ordering and interpreting the examination findings, such as laboratory charges for required blood tests, hospital charges, and professional charges of the radiologists who read the images. To the best of our knowledge, this is the first study of the direct costs associated with surveillance imaging in this population, and it sheds light on the potential for cost savings in the current health care environment.

The economic burden of HNC is substantial, and attempts to quantify its comprehensive costs have been successful, if not a bit disconcerting. In a comparative analysis of different estimation methods used to evaluate annual direct medical expenditures for patient with HNC, Coughlan et al¹² calculated estimates ranging from $1.09 to $3.18 billion. A recent incremental expenditure analysis⁹ of costs associated with HNC care in the US estimated a total annual cost of $4.2 billion, with annual expenditures per individual comparable to other more common cancers, such as lung and colon. The costs of HNC treatment far exceed the costs of other aspects of HNC care, with 1 study¹³ estimating total treatment costs comprising up to 90% of the total cost of care. In a systematic literature review of costs associated with all aspects of HNC care, Wissinger et al¹⁴ highlight the enormous direct and indirect costs of living with HNC in the US and abroad. With such extensive costs only expected to increase,¹⁵ it is important to adopt cost-effective strategies to care for patients with HNC.

Clinical surveillance of patients with HNC is intended to identify recurrences and second primary tumors and monitor late effects of treatment and toxic effects. Data on the direct costs associated with posttreatment imaging surveillance of HNC are limited. An older literature search by Virgo et al¹⁶ identified a wide variety of published posttreatment surveillance strategies for patients with HNC, which translated into Medicare-based charges that ranged from $1198 to $22 807 during a 5-year follow-up period. These costs included all follow-up charges, including outpatient clinic visits, laboratory tests, and imaging examinations.

In the current study, only 10.3% of all patients developed recurrence that was first identified on routine imaging in the absence of symptoms or an abnormal examination finding. When only those patients who remained disease free and never had concerning signs or symptoms were evaluated, the mean (SD) cost of imaging was $9600 ($5900) per year. One could argue that such examinations were unnecessary and clearly not cost-effective. The amount of money spent on imaging in patients who ultimately proved to be disease free is substantial and should be carefully considered in the context of the current health care financial setting. The infrequency of asymptomatic recurrences is supported in the literature. Ng et al¹⁷ performed a retrospective study of patients with HNC treated definitively with chemoradiation and found that approximately two-thirds of disease recurrences presented with an adverse symptom or examination finding, with most presenting within the first 2 years. They demonstrated a 3% yield of detecting a salvageable recurrence with routine imaging after 2 years in an asymptomatic patient with no abnormal clinical finding. Their findings are consistent with other studies that found that most patients with salvageable disease recurrences presented with abnormal symptoms or examination findings within the first 2 years after treatment^18,19,20,21 and that, although routine imaging overall detected recurrences earlier, asymptomatic patients with image-detected recurrences were more likely to have distant disease, limiting salvage treatment options. This factor could add to the emotional distress of patients with incurable disease who could otherwise be observed.

The NCCN guidelines provide a clear recommended clinical surveillance schedule of regularly scheduled visits.²² Regarding imaging surveillance, the NCCN recommends a baseline posttreatment imaging examination within 6 months of treatment completion but reserves further imaging as indicated based on abnormal signs or symptoms. This approach is shared by other professional organizations, including the American Cancer Society,²³ American Society of Clinical Oncology,²⁴ and several European professional organizations.^25,26 However, surveillance strategies are often determined by departmental or institutional policy or treating physician preference, which has resulted in marked variation in physician-reported practices during the last 30 years.^9,16,27 Indeed, a survey administered as recently as 2013 to members of the American Head and Neck Society and American Society for Radiation Oncology revealed wide variation and high levels of PET-CT use in asymptomatic patients even among physicians who reported familiarity with the NCCN guidelines.⁹

The NorthShore University HealthSystem has no standardized care pathway for imaging surveillance. In general, patients will undergo imaging approximately 3 months after treatment completion, but each treating physician determines the subsequent surveillance schedule for his or her patient. Although this certainly limits the current study’s generalizability, it reflects the widespread variation in practices that are documented in the survey study by Roman et al.⁹ That study had an uneven sample of radiation oncologists (67%) and head and neck surgeons (33%), but the authors did not find that physician specialty was associated with imaging use. In the current study, the patients were roughly evenly divided between primary surgery and primary radiation therapy, but specialty was not associated with imaging orders to determine whether a particular specialty was driving imaging practices.

The NCCN and other guidelines are informed by many studies^18,28,29,30 that have found no associations between survival and more rigorous radiographic and even clinical surveillance. Such studies^18,28,29,30 suggest that routine office follow-up in advanced HNC is sufficient and should focus on providing emotional support and managing late treatment complications. A recent study³¹ found that a symptom-based telephone questionnaire can help identify patients at high risk for oral cancer recurrence, thereby reducing even clinical visits and improving health care resource allocation.

Limitations

There are important limitations to this study. As a retrospective medical record review, the findings may be limited by inaccurate documentation of indications for imaging examinations, both in the clinic note and in the International Classification of Diseases, Ninth Revision codes associated with the orders. Such inaccurate documentation could lead to overestimation of costs if imaging tests that were not specifically related to HNC posttreatment surveillance were included. The current study used CPT code–associated hospital and professional charges as a proxy for the actual costs of imaging studies and therefore may not accurately reflect the true costs of surveillance imaging. The study does not take into account the actual payments by insurers, the variability of negotiated rates between hospitals and insurers, or the differences in charges among different hospital systems. Although a more nuanced analysis with these variables is certainly required for more generalizable results, this study is an important first step in trying to quantify the cost of a potentially modifiable variable in HNC care.

The scenario of surveillance imaging ordered in patients who are difficult to examine clinically was also disregarded, often because of the adverse effects of their treatment, such as severe trismus, neck fibrosis, or bulky flaps used for reconstruction. Furthermore, patient-driven imaging or imaging tests that were ordered because of patient anxiety were not taken into account. Finally, survival outcomes were not evaluated in the study. Future analysis could examine whether routine imaging that identified recurrences in asymptomatic patients had an actual association with survival or merely caught a recurrence that eventually would have been clinically detectable, without any association with survival.

Despite these shortcomings, this study is the first contemporary attempt, to our knowledge, to estimate head and neck imaging surveillance costs, and it highlights the potential to improve health care resource use. On the basis of the findings, NCCN guidelines to obtain imaging within 6 months of treatment, with further tests based on patient symptom or abnormal examination findings, should be followed. Physicians should rely on their clinical judgment and tailor the surveillance to the individual patient. Therefore, it may be reasonable to perform annual imaging in selected high-risk patients, such as patients who continue to smoke or who are difficult to clinically examine. Institutions should also focus on incorporating defined survivorship plans with imaging guidelines as part of their comprehensive, multidisciplinary HNC treatment.

Conclusions

This study found that many patients who were treated for HNC receive imaging studies despite NCCN guideline recommendations against it. These findings suggest that the necessity of surveillance imaging in the asymptomatic patient needs to be balanced with the potential clinical benefit in an effort to contain costs without oncologic compromise to the patient.