Adjuvant cyclophosphamide, methotrexate and fluorouracil for node-positive breast cancer: a lifetime cost-utility...

The Q-TWIST (time spent without symptoms and toxicity) method [1-4] and the Gompertz extrapolation method [5-11] are two techniques that have been proposed for evaluating survival in cancer patients. The mathematical basis of the Q-TWIST method relies on estimating the area under the survival curve and partitioning its value into three components with different levels of quality of life (presence of toxicity, presence of symptoms, absence of symptoms and toxicity). The Gompertz approach utilises a curve-fitting procedure to extrapolate the survival curves to infinity. A recent report [12] has described a combined application of the Q-TWIST method and the Gompertz approach ("extrapolated Q-TWIST" method), which allows one to conduct a cost-utility analysis with the calculation of the cost per QALY (quality-adjusted life year) gained. In this paper, we describe a reappraisal of an earlier cost-effectiveness study [7] by application of this extrapolated Q-TWIST method [12]. Our cost-effectiveness study [7] evaluated the pharmacoeconomic profile of adjuvant cyclophosphamide, methotrexate and fluorouracil (CMF) in patients with node-positive breast cancer [13] and utilised a Gompertz analysis to estimate lifetime overall survival (OS), which was 862 and 756 discounted years per 100 patients in the CMF and the control groups, respectively. In applying the extrapolated Q-TWIST method to this data set, a second Gompertz analysis is carried out on the disease-free survival (DFS) curve of the two patient groups. Lifetime DFS of Bonadonna's patients can thus be estimated as 741 years in the CMF group and 572 years in the controls (discounted values normalised to 100 patients). Figure 1 shows the two curves of OS and DFS for the CMF group and the partition of the area under the OS curve into its three components. To introduce an assessment of quality of life into these data, the Q-TWIST method partitions the value of OS into the three components called TWIST (absence of symptoms and toxicity), TOX (time spent with toxicity) and REL (survival after relapse). Hence, OS = TWIST + REL + TOX, where REL = OS - DFS. In the control group, TOX = O and TWIST = DFS because no treatment-related toxicity is present.