Chest CT and Whole-Body ¹⁸F-FDG PET Are Cost-Effective in Screening for Distant Metastases in Head and Neck Cancer Patients

Imaging Techniques

All patients underwent chest CT and ¹⁸F-FDG PET, in random order as dictated by logistics. Spiral CT scans were obtained with a fourth-generation Somaton Plus (Siemens AG) after the intravenous administration of contrast medium (Ultravist; Schering AG). ¹⁸F-FDG PET was performed after patients had fasted for 6 h with ample access to water. At 60–90 min after the intravenous administration of 250–370 MBq of ¹⁸F-FDG, the imaging of trajectory knee–skull base was performed using dedicated full-ring bismuth germinate PET scanners (in Amsterdam/Groningen, ECAT HR+ [CTI/Siemens]; in Nijmegen, ECAT EXACT [CTI/Siemens]). Any focal abnormality suggestive of malignancy was reported (11).

Data Analysis

The result of the clinical diagnostic work-up between presentation and a follow-up of 12 mo was used as reference standard, and patients were classified as positive or negative with respect to the presence of distant metastases. Follow-up was performed every 6 wk in the first year and consisted of visits to the outpatient clinic. During follow-up, the dates of the detection of distant metastases, second primary tumors, or death were recorded. Although the primary goal was screening on distant metastases, second primary tumors were also registered.

Cost Analysis

The hospital's perspective was considered. The cost analysis focused on direct medical costs. The base year was 2008. The costs of diagnosis and treatment were based on the total clinical consumption of all evaluable patients. For the most important items, unit costs were determined because these were a better estimator of the theoretic opportunity costs (13,14). These costs include not only the measurable costs of an intervention (e.g., radiotherapy, surgery, and imaging) but also the services that are not directly allocated to patient care, such as hospital overhead and administrative personnel. Therefore, all hospital costs can be assigned to the interventions given in the hospital. For the determination of these unit costs, the microcosting approach was used (15).

Table 1 shows the most important unit costs used in this analysis. The costs of ¹⁸F-FDG PET scanning, hospital days, outpatient visits, and day-care treatments are composed of variable and overhead costs. The variable costs consisted of manpower (e.g., doctors, nurses) and materials (e.g., medication, supportive patient care, meals). The overhead costs were related to general hospital services and housing. The costs of radiotherapy covered the entire process, including preparation. When patients were subjected to chemotherapy, the costs of the chemotherapeutic agent were derived from the Pharmaceutical Compass (16) and included in the cost analysis; costs of administration are covered by hospitalizations and day-care treatment (15). For most laboratory and diagnostic tests, the Dutch tariff system was used as an approximation of unit costs.

Allocation of Resource Use and Details of Cost Analysis

Data on resource use were collected from the hospital information system, patient files, and the case report forms. Data on the numbers of hospital days, outpatient visits, day-care treatments, diagnostic activities, laboratory testing, radiation therapy sessions, surgical procedures, and medication were collected.

As in all clinical studies concerned with diagnostic techniques, there are several possible diagnostic outcomes: true-positive, false-positive, true-negative, and false-negative. A positive test outcome (distant metastases and [incurable] second primary tumor) results in palliative treatment, and a negative outcome (no distant metastases and [incurable] second primary tumor) results in curative treatment. In Table 2, all different test outcomes and treatment possibilities, together with the translation and consequence for the cost analysis, are presented. However, for the clinical decision making toward either curative or palliative treatment, other aspects are also of interest; these should be incorporated into the cost analysis as well. These aspects relate to the general condition of the patient, the patient's preferences, the cost implications of the clinical approach, and the possible use of second-line diagnostics to confirm or reject an initial test outcome. In addition, ¹⁸F-FDG PET has proven its value in previous studies, and it would, therefore, be unethical to leave its test outcome out of the clinical decision. This decision to include ¹⁸F-FDG PET resulted in several additional test outcome treatment combinations that had to be adjusted for in the cost analysis to be able to judge the added value of ¹⁸F-FDG PET in screening for distant metastasis and synchronous primary tumors in HNSCC patients.

In the cost analysis, the following 3 diagnostic strategies are compared: CT alone, ¹⁸F-FDG PET alone, and the combination of the 2 visually correlated. On the basis of the test outcomes and the aspects already mentioned, clinical experts determined the appropriateness of the clinical approach for each patient in the 3 diagnostic scenarios; subsequently, the consequences for the cost analysis were determined. These consequences were based on resource use in patients undergoing comparable interventions and on information from clinical experts who indicated what resource use was incorrectly withheld or used based on clinical characteristics of the patients. This resource use includes the costs of hospitalization, operations, radiotherapy, and revalidation and also includes the number of ancillary imaging techniques, hospitalizations, and interventions that would not have been applied in the absence of the ¹⁸F-FDG PET scan. It was decided to exclude the impact on laboratory testing from the base case analysis because these are patient-specific and independent of further treatment.

Sensitivity Analysis

The estimation of resource use is associated with substantial uncertainty; because of patient variation, these are tested in multiple sensitivity analyses to determine the impact on the cost-effectiveness outcomes. Additionally, we performed a sensitivity analysis for which we included an estimation of the costs of laboratory testing associated with hospitalization for extensive treatment that was futile or that should have been performed based on the diagnostic test outcome.

Patient Characteristics

One hundred forty-five patients were entered in the study (11). After the exclusion of patients who were incorrectly included or had logistical problems, 111 patients remained. Because the reference standard for further data analysis was the detection of distant metastases or negative follow-up of 12 mo, we excluded 19 patients who died without distant metastases within this 12-mo follow-up. Therefore, we obtained evaluable data for 92 patients. For the cost-effectiveness analysis, the complete data for 80 patients were available. Of these 80 patients, 63 were men, and the mean age was 60 y (range, 40–81 y). Primary tumor sites were the oral cavity (n = 17), oropharynx (n = 24), hypopharynx (n = 16), larynx (n = 14), cervical esophagus (n = 3), and lymph node metastases of an unknown primary tumor (n = 18). Ten patients had more than 1 synchronous primary tumor. The patient characteristics in this study were comparable with the patient characteristics included in the clinical study.

Clinical Study

Pretreatment screening identified distant metastases in 17 of 80 patients (21%; 95% confidence interval [CI], 15%−28%) and second primary tumors in 6 of 80 (8%; 95% CI, 3%−12%). All patients with distant metastases were treated palliatively. Half of the patients with a second primary tumor had disseminated lung cancer (lung or brain metastases), and they also received palliative treatment. The other 3 patients appeared to have limited-stage disease of their secondary primary and were treated with curative intent for both primary tumors. In 32 of 80 of the total group of patients (41%; 95% CI, 33%−50%), distant metastasis (33%; 95% CI, 25%−41%) or a second primary tumor (9%; 95% CI, 5%−15%) was detected during screening or within the 12 mo of follow-up.

¹⁸F-FDG PET had a higher sensitivity (53% vs. 37%) and positive predictive value (80% vs. 75%) than did chest CT. The combination of CT and ¹⁸F-FDG PET had the highest sensitivity (63%).

Details of the clinical study are presented elsewhere (11).

Cost Analysis

The distribution of test outcomes and treatment options is presented in Figure 1. From this figure, it appears that in the CT-only scenario 11.25% of the patients were not treated in accordance with the test outcome. However, in 2.50% of these patients the positive test outcome related to an operable second primary, and therefore the curative operations were appropriate. In another 2.50%, a palliative approach was chosen, despite a negative test outcome, because of patient preference or clinical condition. As a consequence, 6.25% of these patients remained for whom a correction must be made for inappropriately withheld (3.75%) or given curative treatment (2.50%).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1. 

Distribution of patients over test outcome and treatment approach.

In the ¹⁸F-FDG PET–only scenario, 16.25% of the patients were not treated according to their test outcome. Of these patients, 14% were curatively treated for an operable second primary tumor, and 7% were treated palliatively despite a negative test because of the patient's condition. In another 7% of these patients, the ¹⁸F-FDG PET result raised questions that were answered with additional diagnostic testing; the ¹⁸F-FDG PET result was, therefore, considered to be false-positive but this did not influence the treatment decision. Finally, 5% of the patients with a false-negative test outcome were treated palliatively; this had no impact on the cost analysis because the clinicians indicated there would have been no difference in curative and palliative treatments. Thus, 11.25% of the patients remained for whom a correction must be made for inappropriately withheld or given curative treatment.

In the third scenario, the combination of CT and ¹⁸F-FDG PET, 13.75% received a treatment that was not in line with the test outcome. Of these patients, 16.75% were curatively treated for an operable second primary tumor, and 8.25% were treated palliatively, despite a negative test result because of the patient's condition. This scenario also included a patient for whom the ¹⁸F-FDG PET result raised questions (these questions were solved with additional diagnostic testing) and a patient for whom the clinicians indicated that there was no difference in resource use between palliative and curative treatments. This scenario resulted in an adjustment for inappropriate resource use of the patients who were treated curatively despite a positive test outcome.

In 3 patients, the ¹⁸F-FDG PET scan resulted in ancillary imaging or hospitalizations and interventions that would not have been applied in the absence of the ¹⁸F-FDG PET scan; these resources were also defined and subtracted from the total resource use in the CT-only scenario. The results of the cost analysis for all diagnostic strategies are presented in Table 3. In this table, the difference in costs between treatment strategies and study-related diagnostic procedures (CT and ¹⁸F-FDG PET) were included. A curative treatment cost approximately €41,369 (≈$61,912) and a palliative treatment €26,328 (≈$39,402).

Comparing the diagnostic strategies, the most important changes are seen in the costs of hospital days, surgery, and radiotherapy. These changes are caused by the reduction of futile operations and curative radiotherapy in both the ¹⁸F-FDG PET and the CT + ¹⁸F-FDG PET scenarios versus the CT scenario (Table 3).

These reductions countered the introduction of additional diagnostic costs by ¹⁸F-FDG PET testing and resulted in an average cost per patient of €38,558 (≈$57,705) in the CT-only scenario, €38,355 (≈$57,402) in the ¹⁸F-FDG PET–only scenario, and €37,954 (≈$56,801) in the CT + ¹⁸F-FDG PET scenario. The differences between these scenarios are small; the introduction of ¹⁸F-FDG PET led to a cost reduction between €203 (≈$303) and €604 (≈$903).

Sensitivity Analysis

The costs of hospitalizations, operations, radiotherapy, and diagnostic imaging are the main cost drivers in this patient population. The results of several sensitivity analyses on these main cost drivers are presented in Table 4.

The cost differences now ranged between €110 (≈$164) and €697 (≈$1,043). Variation in resource use and costs affected the differences between the strategies. Furthermore, the results still remained robust in favor of the CT + ¹⁸F-FDG PET scenario.