all other possible potential strategies to stage lung cancer. However, PET scanning after CT
positive results is a less clear cut strategy with some studies reporting PET to be cost effective
and others reporting the opposite to be true. Whilst, studies not taking into account the
difference between CT negative and CT positive patients may not be picking up this effect.
Despite the international evidence that PET is likely to be cost effective, especially in
CT negative patients the results are not entirely transferable to the Australian context, that is,
excluding the Viney et al trial, which only looked at stage I and II candidates. This is mainly
because of the different costs structures and reimbursement rates in overseas healthcare
systems, which makes the translation of them to the Australian context inappropriate.
Model Structure
There is great variability in the various national clinical guidelines for treating NSCLC
and therefore there is some difference in how NSCLC is treated both nationally and even
locally.25,26 The model presented here is, therefore, based on a simplification of the potential
outcomes and pathways that on average most likely to used to treat NSCLC in a cost effective
manner. The model uses a societal perspective.
As with the HTBS and NICE economic evaluation, the model has been based upon the
decision pathway used in the Dietlein study. This is considered to be one of the best studies
conducted on PET. However, there are three main problems with this study and its
transferability to other decision making contexts has been highlighted.
First, it does not contain all the plausible alternatives for PET (see the HTBS study for
other potential strategies).27 However, this is unlikely to matter greatly as the other strategies
suggested are unlikely to ever be used in a clinical situation because they deny patients
histological confirmation of their lung cancer.
The second problem is that they take a simple approach to combing data from the
literature on the sensitivity and specificity of tests independently, which is inappropriate.
Therefore, the sensitivity and specificity for detecting mediastinal metastasis in this study are
taken from the meta-analysis conducted by the HTBS study and sensitivity for detecting distant
metastasis is taken from the meta-analysis used in the NICE guidelines.
The third problem is that the costs are based upon the German reimbursement rates and
charges and are therefore are inappropriate for translation to other healthcare systems. This has
bee rectified by inserting comparative costs for tests, surgery and palliative care from
Australia. Nevertheless, the Dietlein model is still useful for giving the outline of the decision
problem and some of the parameters are the best estimates available to the model for
evaluating the cost and benefits of PET in NSCLC. The model inputs can be found in table 4.
Strategies
The model takes the two most cost effective strategies that have been identified from
the literature above. These strategies also offer a histological confirmation of NSCLC and they
have been applied to both CT positive and CT negative patients. The first strategy is a normal
CWU without the use of PET. After the NSCLC has been diagnosed patients either proceed to
25 Clinical Practice Guidelines for the Prevention, Diagnosis and Management of Lung Cancer. Australian
Government; National Health and Research Council. 2004.
26 National Institute of Health and Clinical Effectiveness, “The Diagnosis and treatment of Lung Cancer; Methods,
Evidence and Guidance. National Collaboration Centre of Acute Care. 2005.
27 Bradbury et al. Health Technology Board Scotland, “Positron Emission Tomography (PET) Imaging in Cancer
Management” October 2002.
11