Added value of cost-utility analysis in simple diagnostic studies of accuracy: ¹⁸F-fluoromethylcholine PET/CT in prostate cancer staging

Abstract

Diagnostic studies of accuracy targeting sensitivity and specificity are commonly done in a paired design in which all modalities are applied in each patient, whereas cost-effectiveness and cost-utility analyses are usually assessed either directly alongside to or indirectly by means of stochastic modeling based on larger randomized controlled trials (RCTs). However the conduct of RCTs is hampered in an environment such as ours, in which technology is rapidly evolving. As such, there is a relatively limited number of RCTs. Therefore, we investigated as to which extent paired diagnostic studies of accuracy can be also used to shed light on economic implications when considering a new diagnostic test. We propose a simple decision tree model-based cost-utility analysis of a diagnostic test when compared to the current standard procedure and exemplify this approach with published data from lymph node staging of prostate cancer. Average procedure costs were taken from the Danish Diagnosis Related Groups Tariff in 2013 and life expectancy was estimated for an ideal 60 year old patient based on prostate cancer stage and prostatectomy or radiation and chemotherapy. Quality-adjusted life-years (QALYs) were deduced from the literature, and an incremental cost-effectiveness ratio (ICER) was used to compare lymph node dissection with respective histopathological examination (reference standard) and ¹⁸F-fluoromethylcholine positron emission tomography/computed tomography (FCH-PET/CT). Lower bounds of sensitivity and specificity of FCH-PET/CT were established at which the replacement of the reference standard by FCH-PET/CT comes with a trade-off between worse effectiveness and lower costs. Compared to the reference standard in a diagnostic accuracy study, any imperfections in accuracy of a diagnostic test imply that replacing the reference standard generates a loss in effectiveness and utility. We conclude that diagnostic studies of accuracy can be put to a more extensive use, over and above a mere indication of sensitivity and specificity of an imaging test, and that health economic considerations should be undertaken when planning a prospective diagnostic accuracy study. These endeavors will prove especially fruitful when comparing several imaging techniques with one another, or the same imaging technique using different tracers, with an independent reference standard for the evaluation of results.

Introduction

In diagnostic research, imaging techniques are required to detect and localize disease and, thereby, to discriminate between diseased and disease-free (or metastasized and metastasis-free) patients by means of sensitivity and specificity. Diagnostic research in recent decades has been affected by the introduction and increased clinical use of rapidly evolving imaging techniques like single-photon emission computed tomography (SPECT), magnetic resonance imaging (MRI), positron emission tomography (PET), PET/computed tomography (PET/CT), and, most recently, PET/MRI. Sufficient sensitivity and specificity of a test is, though, not considered appropriate as surrogate endpoint for clinical benefit which preferably has to be demonstrated by means of randomized controlled trials (RCT) in which mortality, morbidity, symptoms, and quality of life are measured [1,2]. Ideally, analyses examining societal costs and benefits of a given diagnostic imaging technology complete the picture [3].

Cost-effectiveness analysis (CEA) compares alternative interventions using costs and a common effectiveness measure (e.g., correct staging or life-years gained). The results of such comparisons may be stated in terms of costs per unit of effectiveness (e.g. dollars spent per life-year gained) or effectiveness per unit of cost (life-years gained per dollar spent). In this context, the relative cost-effectiveness of alternative tests can be assessed as long as the alternatives under consideration are not of an exceptionally different scale. In cost-utility analysis (CUA), the incremental cost of an intervention is compared to the incremental health improvement attributable to the intervention, where health improvement is for instance measured in quality-adjusted life-years (QALYs) gained [4-6].

Studies demonstrating clinical benefit by improvement of long-term patient outcome as well as studies on cost-effectiveness or cost-utility do require RCTs in order to compare imaging techniques head-to-head. However, studies of this type are rare [7-9] and most of the studies in diagnostic research today are still diagnostic accuracy studies. An alternative approach to RCTs is the application of economic modeling techniques.

Economic modeling is a relatively cheap and effective way of synthesizing existing data and evidence available on the costs and outcomes of alternative interventions [10-13]. In intervention trials, for instance, intervention thresholds were introduced as the absolute threshold of disease risk at which intervention becomes acceptable in terms of both efficacy and cost-effectiveness [14,15].

We propose to utilize simple diagnostic studies of accuracy to assess the incremental cost-effectiveness ratio (ICER) for the comparison of a (new) diagnostic imaging technique and a reference standard diagnostic procedure. Compared to the reference standard, any imperfections in accuracy of a new diagnostic imaging technique will indeed imply that replacing the reference standard generates a loss in effectiveness. However, lower boundaries of sensitivity and specificity of a new imaging test can be assessed at which the replacement of a standard procedure by the new test comes with a trade-off between worse effectiveness and lower costs. We demonstrate our approach by using data from a recent investigation at our institution in which the value of ¹⁸F-fluoromethylcholine (FCH) PET/CT for lymph node staging in prostate cancer was assessed using pelvic lymph node dissection (LND) with subsequent histological examination as the reference standard [16].

Methods

Background on clinical study

The methodology of the clinical study was published earlier [16]. In brief, between January 2008 and December 2010 FCH-PET/CT was performed in 210 intermediate or high risk patients prior to regional LND and subsequent histological examination (reference standard). The surgical technique comprised an open retroperitoneal bilateral pelvic LND that was undertaken through a midline incision, either as part of the radical prostatectomy or as an individual operation. The LND was performed along the medial side of the external iliac vessels from the femoral canal up to the bifurcation of the internal and external iliac vessels, including the obturator fossa. The LND included most of the lymph nodes of the external iliac, obturator, and hypogastric nodes. The specimens were prospectively mapped according to their anatomical location and processed according to standard protocols for the subsequent histological examination. The result of the histological examination of the lymph nodes was compared with the result of FCH-PET/CT, as obtained by blinded review. Sensitivity and specificity of FCH-PET/CT for the detection of lymph node metastases were estimated to be 0.73 (Wilson score-based 95% confidence interval (95% CI): 0.58-0.84) and 0.88 (95% CI: 0.82-0.92), respectively.

Cost data collection

Cost data were collected post hoc in addition to the clinical trial. These comprised costs for lymph node dissection, histological examination of lymph nodes, prostatectomy, radiation and chemotherapy as per the Danish Diagnosis Related Groups Tariff in 2013. Details can be found in Table 1.

Table 1.

Calculation of average cost and QALYs, depending on the outcome of the imaging procedure. One-hundred DKK correspond to 17.11 US$ (daily exchange rate on October 29, 2014)

	Dissection of lymph nodes and histopathological examination (reference standard)		FCH-PET/CT
Result of diagnostic procedure	Positive	Negative	True positive	False negative	False positive	True negative
Lymph node dissection	32,859 DKK	32,859 DKK
Histopathological examination	1,500 DKK	2,250 DKK
FCH-PET/CT scan			11,110 DKK	11,110 DKK	11,110 DKK	11,110 DKK
Radiation and chemotherapy	276,413 DKK		276,413 DKK	276,413 DKK	389,586 DKK
Prostatectomy		60,026 DKK		60,026 DKK		60,026 DKK
Sum	310,772 DKK (53,171 $)	95,135 DKK (16,277 $)	287,523 DKK (49,193 $)	347,549 DKK (59,463 $)	400,696 DKK (68,556 $)	71,136 DKK (12,171 $)
QALYs	4.2	19.5	4.2	3.8	9.0	19.5

Life-expectancy and QALYs

The expected number of life years remaining was estimated for a 60 year old patient and depended on the stage of disease as indicated by the diagnostic imaging procedure. The following numbers were not explicitly available and were deduced from the literature to the best of our knowledge [17-20]:

● True positive (metastatic prostate cancer, radiation and chemotherapy): 8 years.

● False negative (metastatic prostate cancer, prostatectomy): 8 years.

● False positive (localized prostate cancer, radiation and chemotherapy): 14 years.

● True negative (localized prostate cancer, prostatectomy): 20.4 years.

QALYs, i.e., (score per year) x (expected number of life years), for these four outcomes of the diagnostic staging procedure were also not explicitly available and instead deduced from the literature to the best of our knowledge [21-24]. Scores were estimated using values between one (good health) and zero (death):

● True positive: 0.7 × 5 + 0.4 + 0.2 + 0.1 = 4.2 QALYs. (Five years of mild disease including consequences of castration, one year of moderate disease, one year of heavy disease, one year near death.)

● False negative: 0.8 × 2 + 0.5 × 3 + 0.4 + 0.2 + 0.1 = 3.8 QALYs. (Two years of disease including consequences of treatment and disease, three years of mild disease including further treatment and consequences of castration, one year of moderate disease, one year of heavy disease, one year near death.)

● False positive: 0.7 × 12 + 0.4 + 0.2 = 9.0 QALYs. (Twelve years of mild disease including belief of metastatic disease and consequences of castration, one year of moderate other disease, one year of heavy other disease.)

● True negative: 0.9 × 2 + 1 × 17 + 0.5 × 1.4 = 19.5 QALYs. (Two years of discomfort due to treatment, but knowledge of being cured, seventeen life years in good health, 1.4 years of other disease).

ICER

We observed a prevalence for progressed disease (malignant metastases in the lymph nodes) in 41 (19.5%) of the 210 patients in our clinical study [16]. We modeled the alternative pathways by means of a decision tree (Figure 1) and evaluated the expected cost and expected utility for both FCH-PET/CT and the reference standard accordingly (Table 2), given the estimated sensitivity and specificity of FCH-PET/CT of 0.73 and 0.88, respectively. The ICER is then the ratio between the difference in expected cost and the difference in expected utility (FCH-PET/CT minus reference standard).

Figure 1.

Decision tree model for the detection of lymph node metastases. The estimated prevalence of 0.195 as well as estimates of sensitivity and specificity of FCH-PET/CT stem from the clinical study [16].

Table 2.

Calculation of expected cost and expected utility for FCH-PET/CT and the reference standard, according to the decision tree model in Figure 1. One-hundred DKK correspond to 17.11 US$ (daily exchange rate on October 29, 2014)

Pathway	Probability	Cost (DKK)	Expected cost (DKK)	Utility	Expected utility
A	0.195	310,772	60,675	4.2	0.82
B	0	371,548	0	3.8	0
C	0.805	95,135	76,561	19.5	15.69
D	0	423,945	0	9	0
Total	1		137,236		16.51
E	0.143	287,523	40,979	4.2	0.60
F	0.053	347,549	18,321	3.8	0.20
G	0.708	71,136	50,378	19.5	13.81
H	0.097	400,696	38,696	9	0.87
Total	1		148,374		15.48

Apart from abovementioned primary analysis in which the point estimates for sensitivity and specificity of FCH-PET/CT from our clinical study were used, we assessed lower boundaries of sensitivity and specificity of FCH-PET/CT at which the replacement of the reference standard comes with a trade-off between worse effectiveness and lower costs. This was done by varying the levels of sensitivity and specificity of FCH-PET/CT in the ICER calculations. These boundaries were then compared to the 95% CI of sensitivity and specificity from our clinical study.

All analyses were performed using SAS 9.1.3 (SAS Institute Inc., Cary, NC, USA) and Stata/MP 13.1 (StataCorp, College Station, TX, USA).

Results

The ICER was -10,760 DKK (-1,841 $) per QALY as the difference in expected cost and the difference in expected utility (FCH-PET/CT minus reference standard) was 11,138 DKK (1,906 $) and -1.0351 QALY, respectively (see also Figure 1 and Table 2). This means that FCH-PET/CT was, given the estimated values of sensitivity and specificity of 0.73 and 0.88, respectively, both more costly and less effective in terms of QALYs than the reference standard. Figure 2 shows the cost-effectiveness plane, in which the abovementioned ICER value of -10,760 DKK (-1,841 $) per QALY corresponds to the point P1 in quadrant II. Points in the four quadrants have the following interpretation:

Figure 2.

Cost-effectiveness plane (adapted from [25] (p.212) and [26] (p.12)). The blue line K divides the plane into cost-effective (lower right: I A, III B, IV) and non-cost-effective areas (upper left: I B, II, III A). Here, line K is based on a hypothetical benchmark value of 30,000 $ per QALY.

● I: FCH-PET/CT is more effective (in terms of QALYs), but also more costly than LND staging.

● II: FCH-PET/CT is less effective and more costly than LND staging, i.e. LND staging dominates FCH-PET/CT.

● III: FCH-PET/CT is less effective, but also less costly than LND staging.

● IV: FCH-PET/CT is more effective and less costly than LND staging, i.e. FCH-PET/CT dominates LND staging.

The blue line K indicates the maximum acceptable ICER for which a hypothetical value of 30,000 $ per QALY was assumed here. Line K divides the plane into cost-effective (lower right) and non-cost-effective outcomes (upper left).

When varying the levels of sensitivity and specificity of FCH-PET/CT, it was found that the specificity of FCH-PET/CT must be at least 0.92 in order to change the sign of the ICER to be positive, meaning here that losses in terms of QALYs were accompanied by monetary savings when moving from the reference standard to FCH-PET/CT. For specificities of 0.92, 0.93, 0.94, and 0.95, the sensitivities needed to exceed 0.78, 0.55, 0.33, and 0.10, respectively, in order to get a positive number as ICER (see filled circles and squares in Figure 3). For instance, the pair (sensitivity, specificity) = (0.78, 0.92) was associated with a difference in expected cost and a difference in expected utility of -57 DKK (-9.75 $) and -0.693 QALY, respectively, resulting in an ICER of 82 DKK (14.03 $) per QALY. This value is indicated by point P2 in Figure 2 and still reflects a non-cost-effective result at an ICER benchmark value of 30,000 $ per QALY.

Figure 3.

Indication of pairs (sensitivity, specificity) for which the expected cost and the expected utility of FCH-PET/CT were greater than and less than those of the reference standard, respectively, thereby implying a negative ICER (unfilled circles and squares) and for which both the expected cost and the expected utility of FCH-PET/CT were less than those of the reference standard, yielding a positive ICER (filled circles and squares). The estimated values of sensitivity and specificity of FCH-PET/CT from our clinical study and the respective 95% confidence intervals are shown (squares). The prevalence of malignant lymph node metastases was 0.195 [16].

The abovementioned lower boundaries of sensitivity and specificity of FCH-PET/CT imply a trade-off between worse effectiveness and lower costs when replacing the reference standard with FCH-PET/CT. Comparing these boundaries to the 95% CIs of sensitivity and specificity of PET/CT from our clinical study showed that only some values from the upper range of these 95% CIs (i.e. a specificity of 0.92 and a sensitivity between 0.78 and 0.84; see filled squares in Figure 3) were accompanied by worse effectiveness at lower costs when replacing the reference standard with FCH-PET/CT. Figure 4 shows a contour plot which illustrates an increasing ICER with increasing sensitivity and specificity of FCH-PET/CT. The orange area indicates combinations of sensitivity and specificity at which the ICER is negative, hence indicating both lower costs and superior effectiveness of the reference standard over FCH-PET/CT. This area accords to the area of unfilled circles and squares in Figure 3. The remaining, i.e. non-orange, area in Figure 4 shows combinations of sensitivity and specificity at which the ICER was positive, thereby indicating lower costs of FCH-PET/CT, but favorable utility for the reference standard. Due to the prevalence of progressed disease (19.5%) and its impact on cost calculations (Figure 1, Table 2), even pairs with relatively low sensitivity, but high specificity, were associated with a positive ICER. Only the dark blue area indicates combinations of sensitivity and specificity at which the replacement of LND staging by FCH-PET/CT would be cost-effective, given an ICER benchmark value of 30,000 $ per QALY (see, for instance, point P3 in Figure 2).

Figure 4.

ICER of FCH-PET/CT in prostate cancer patients depending on varying levels of sensitivity and specificity. One-hundred DKK correspond to 17.11 US$ (daily exchange rate on October 29, 2014). The prevalence of malignant lymph node metastases was 0.195 [16].

Discussion

The assessment of the cost-effectiveness or the cost-utility of diagnostic imaging techniques like PET/CT does, in principle, require larger scale clinical trials. Such trials are difficult to conduct in an environment in which technology is rapidly evolving and costly. RCTs that also evaluate cost-effectiveness or cost-utility represent a study design scarcely found in clinical trials that evaluate non-invasive imaging modalities [27,28]. Moreover, not many cost-effectiveness analyses for PET/CT were undertaken in cancer imaging prior to 2010 [29,30]. Recent systematic reviews on the cost-effectiveness of PET and PET/CT in cancer and non-cancer indications and of ¹⁸F-Fluorodeoxyglucose PET in tumors other than lung cancer comprised only 47 and 16 studies, respectively [28,31]. Modeling techniques remain the way to proceed in generating and synthesizing the economic evidence efficiently in order to inform decision makers regarding PET [32].

Our approach is a simple, applied example of a cohort model-based cost-utility analysis of a diagnostic test for lymph node staging of prostate cancer in which we take the perspective of the hospital. The model is essentially a decision tree, with crude life expectancy estimates for each endpoint. It does not replace the assessment of cost-effectiveness or cost-utility in clinical trials by means of more sophisticated stochastic modeling or, whenever possible, in addition to RCTs. However, facing scarce evaluative resources, we propose an extended use of simple diagnostic studies of accuracy. An evaluation of lower boundaries of sensitivity and specificity at which the replacement of the reference standard comes with a trade-off between worse effectiveness and lower costs is easily performed and gives valuable insight into the potential economic burden caused by new diagnostic procedures. Obviously, these lower economic thresholds of sensitivity and specificity must also be assessed from a clinical point of view, not just an economical one, as, for instance, only relatively large values of sensitivity and specificity of FCH-PET/CT will support the replacement of lymph node dissection by FCH-PET/CT scanning in the future. In our example, the former clinical study results indicated that from a clinical point of view FCH-PET/CT was not sufficiently accurate to replace the reference procedure. This post hoc cost-effectiveness assessment pointed to values at the upper range of the 95% CI of accuracy measures (i.e. a specificity of 0.92 and a sensitivity between 0.78 and 0.84) at which worse effectiveness at least comes along with lower costs when replacing the reference standard with FCH-PET/CT.

The histological examination of the lymph nodes in our clinical study [16] was done according to standard protocols. Later, fellow colleagues of ours reinvestigated the lymph nodes from all 169 patients of whom the lymph nodes were judged to be negative in the standard examination by means of an extended pathological examination, which is a costly and time-consuming procedure [33]. The extended pathological examination included a 100-μm deep haematoxylin and eosin (HE) section followed by a slide stained with cytokeratin AE1/AE3 and then by four HE sections at 0.5-mm intervals. The standard pathological examination detected 41 patients with malignant lymph node involvement, whereas the extended pathological examination revealed 5 additional patients with lymph node metastases. This means that the sensitivity of the reference procedure in our example could be assessed as low as 89.1%, whereas it is still reasonable to assume a specificity of the reference procedure of 100%. Then, the lower boundaries of sensitivity and specificity at which the ICER becomes positive decrease slightly to the pairs (sensitivity, specificity) = (0.90, 0.91), (0.67, 0.92), (0.44, 0.93), and (0.22, 0.94), thereby comprising additional pairs from the estimated 95% CI of sensitivity and specificity of FCH-PET/CT as compared to assuming the reference standard to be perfect. Besides, sensitivity and specificity of FCH-PET/CT need also to be reassessed on these grounds (here, a reclassification of these 5 patients would actually lead to a slightly decreased sensitivity of 0.67 (95% CI: 0.53-0.79), whereas specificity remains unchanged). The consequence of assuming an imperfect reference procedure is a larger number of sensitivity/specificity pairs for which switching to a new imaging test could be of interest from an economical point of view and emphasizes the outstretched reach of the validity of the reference standard on both the accuracy and the cost-effectiveness assessment of competing tests.

PET/CT is a technology that is rapidly evolving and continually improving, most likely resulting in improved sensitivity and specificity of any given tracer. As a consequence, any prediction of the economic burden of PET/CT will either hold (when sensitivity and specificity are not subjected to significant change despite technological advancement) or will be overestimated, i.e., a conservative estimation, when re-visited later (when sensitivity and specificity actually are subjected to significant improvements due to technological advancement). According to this, any economic assessment today can be considered a lower boundary of the actual economic burden tomorrow when technology has improved since the economic assessment.

Limitations of our study comprise its retrospective nature, working with a base-case scenario of a typical 60 year old patient only, and the extent to which costs and QALYs were assessed post hoc, failing to comprise ranges for both costs and QALYs and merely using point estimates instead, thereby failing to provide supplementary sensitivity analyses. Using fixed values does, indeed, influence the calculated numbers, but not necessarily the principles applied. Further limitations of our study are the relatively short time horizon, the disregard of some form of discounting of costs and benefits in order to consider their present value, and the hospital perspective that we take rather than a more societal point of view. We did not consider side effects of LND, radiation and chemotherapy, or prostatectomy which entails an over-simplification of the clinical setting. All in all, the ICER and, thereby, the concrete lower boundaries of sensitivity and specificity derived depend heavily on our assumptions made. However, we would like to stress the added value of our approach in principle and do not claim robustness of the derived lower boundaries of sensitivity and specificity against model variations.

Future research on the added value of cost-utility analysis in diagnostic studies of accuracy in other clinical settings will improve our approach by bearing the following issues in mind:

● Tariff-based cost estimation is far from actual costs; costs should be, whenever possible, estimated using claims data or accounted costs on a per-patient basis in order to enable the assessment of cost distributions across patients and patient groups. However, tariff-based cost data may be the only available source in countries like Denmark.

● Instead of using fixed values of point estimates for costs, life years, or utility index, ranges should be applied. Costs differ greatly between different age groups/cancer stage groups, those who received different diagnosis/treatment options, or those with and without operative complications. Also life-years vary greatly. At least, 95% CI for the mean value of costs/life years should be applied to the cost-utility sensitivity analysis.

● Quality of life scores in a healthy population typically average around 0.85 on a scale of 0 to 1. Even an advanced stage group of patients rarely score, on average, less than 0.3. If suitable literature on the long-term quality of life of patients’ post-diagnostic tests is not available, looking beyond the cancer form under consideration to other cancers may prove beneficial in order to get an idea of average quality of life scores. Expert elicitation can then be used to adjust these scores depending on how the cancer form under consideration is felt to differ.

● The derivation of a utility index by measuring patients’ preferences is, generally speaking, a challenging endeavour and methods used are various (e.g. Visual Analogue Scale, time trade-off, or standard gamble). Cost-effectiveness analysis using clinical endpoints such as survival or time-to-progression as effectiveness measures is probably easier to convey to clinicians and policy makers as it makes use of more direct measures of patient-benefit, being independent of patient preferences and the measurement thereof.

● Probabilistic sensitivity analysis sheds light on the implications of the uncertainty of the model by investigating the consequences of choosing alternative, plausible values for quality of life scores, survival, cost of treatment, starting age of the population in the model, and the results of the diagnostic imaging test.

● A cost-effectiveness threshold by means of a benchmark ICER value is needed to get context of what QALY benefit would have to be seen to make a diagnostic imaging test cost-effective. The ICER is of little value without it and the risk is that international thresholds are then considered that might not be representative for the country to which the respective study results are supposed to apply.

In 2013, Consolidated Health Economic Evaluation Reporting Standards (CHEERS) were published in order to consolidate and update former health economic evaluation guidelines into one current, useful reporting guidance document, including a 24-item checklist [34].

Conclusions

Simple diagnostic studies of accuracy can be put to a more extensive use, over and above a mere indication of sensitivity and specificity of an imaging test. An evaluation of lower boundaries of sensitivity and specificity can be performed and gives valuable insight into the economic burden of new diagnostic procedures or tracers. Compared to the reference standard in a diagnostic accuracy study, any imperfections in accuracy of a diagnostic test imply that replacing the reference standard generates a loss in effectiveness and utility. In our example, substituting the conventional staging procedure with FCH-PET/CT would produce a loss, and not a minor one, due to the clinically inefficient sensitivity and specificity of FCH-PET/CT. Giving our study’s limitations, we cannot claim cost-effectiveness of FCH-PET/CT in staging prostate cancer, but we have hopefully stimulated further perspectives to the planning and evaluation of simple diagnostic studies of accuracy.

We focused on the comparison of a non-invasive imaging technique and the current, invasive reference procedure in that respective indication. Analogously, cost-utility and cost-effectiveness analyses using ICERs can and should be done especially in paired diagnostic studies of accuracy when comparing several imaging techniques with one another, or the same imaging technique using different tracers, with an independent reference standard for the evaluation of results. This will enable to rank order several different imaging techniques from an economical point of view, supplemented by a graphical display by means of a cost-effectiveness plane, which completes the picture on top of the imaging techniques’ clinical assessments. Health economic considerations should be undertaken when planning any prospective diagnostic accuracy study in order to investigate if adding cost-utility or cost-effectiveness analysis to sensitivity and specificity determination of diagnostic tests improves the ability to identify the better performing alternative between competing diagnostic modalities when each are referenced to a gold standard.

Cost-utility and cost-effectiveness analyses involve various challenges. However, modalities (or tracers) that also seem to be economically defensible will be easier to argue for when it comes to regular use in daily clinical practice (such as in Denmark) or reimbursement (such as in the US or Germany).