Chemotherapy alongside endocrine treatment in ER +ve breast cancer patients post resection of a primary tumour has been estimated to reduce mortality rates by up to 30%. However, the high cost of the therapies, heterogeneous nature of the disease and adverse event profile implies that not all patients should receive the treatment. Many existing prognostic tools such as the NPI, PREDICT, and Adjuvant! Online may not definitively estimate the risk profile of patients, resulting in an indeterminate risk classification. In such cases gene expression profiling tests such as EndoPredict can aid the treatment decision. It is important to examine if the test represents a cost-effective use of limited NHS resources in such intermediate risk patients.
This small (n = 151) multi-centre, two-stage study evaluated the cost-effectiveness of EndoPredict in patients with no clear treatment based on current prognostic criteria. The primary analysis examined whether EndoPredict test results increased or decreased the use and intensity of chemotherapy and the associated direct cost implications. Secondly, a mathematical model was constructed to determine how the change in treatment decisions impacted the long term health of the population, and the future cost implications to the NHS.
A cost increase per patient treated with chemotherapy was identified when EndoPredict test results were available (£149), alongside no significant change in the total number being prescribed chemotherapy. However, chemotherapy was offered to a very different patient population, with 36.9% of patients having a change in treatment decision. The long term analysis found the use of EndoPredict to be associated with greater total costs but a potential increase in population health, resulting in an incremental cost-effectiveness ratio of £26,836 per quality adjusted life year.
While EndoPredict was found to be more expensive overall, the ability of the EPClin score to affect a more optimal allocation of chemotherapy, resulted in long term health gains. However, this result was on the margin of what is conventionally considered a cost-effective use of limited NHS resources and subject to significant uncertainty.
Clinical trial identification
Legal entity responsible for the study
Sussex Health Outcomes Research and Education in Cancer
S. Hinde C. Theriou, S. May, L. Matthews, A. Arbon, L. Fallowfield, D. Bloomfield: This research was funded through an unrestricted educational grant from Myriad.