Economic evaluation of personalised versus conventional risk assessment for women who have undergone testing for hereditary breast and ovarian cancer genes: a modelling study.

Abstract

The management of women with germline pathogenic variants (GPVs) in breast (BC) and ovarian cancer (OC) susceptibility genes is focused on surveillance and risk-reducing surgery/medication. Most women are assigned an average range of risk and treated accordingly, but it is possible to personalise this. Here, we explore the economic impact of risk personalisation. We compared two strategies for risk stratification for female participants: conventional risk assessment (CRA), which only involves information from genetic testing and personalised risk assessment (PRA), using genetic and non-genetic risk modifiers. Three different versions of PRA were compared, which were combinations of polygenic risk score and questionnaire-based factors. A patient-level Markov model was designed to estimate the overall National Health Service cost and quality-adjusted life years (QALYs) after risk assessment. Results were given for 20 different groups of women based on their GPV status and family history. Across the 20 scenarios, the results showed that PRA was cost-effective compared with CRA using a £20 000 per QALY threshold in women with a GPV in <i>PALB2</i> who have OC or BC+OC family history, and women with a GPV in <i>ATM</i>, <i>CHEK2</i>, <i>RAD51C</i> or <i>RAD51D</i>. For women with a GPV in <i>BRCA1</i> or <i>BRCA2</i>, women with no pathogenic variant and women with a GPV in <i>PALB2</i> who have unknown family history or BC family history, CRA was more cost-effective. PRA was cost-effective compared with CRA in specific situations predominantly associated with moderate-risk BC GPVs (<i>RAD51C</i>/<i>RAD51D</i>/<i>CHEK2</i>/<i>ATM</i>), while CRA was cost-effective compared with PRA predominantly with high-risk BC GPVs (<i>BRCA1</i>/<i>BRCA2</i>/<i>PALB2</i>). PRA was cost-effective in specific situations compared with CRA in the UK for assessment of women with or without GPVs in BC and OC susceptibility genes.