Search

PURPOSE Genomic tests improve accuracy of risk prediction for early breast cancers but these are expensive. This study evaluated the clinical utility of EndoPredict®, in terms of impact on adjuvant therapy recommendations and identification of parameters to guide selective application. METHODS Patients with ER-positive, HER2-negative, and early-stage invasive breast cancer were tested with EndoPredict®. Two cohorts were recruited: one consecutively and another at clinical team discretion. Systemic treatment recommendations were recorded before and after EndoPredict® results were revealed to the multidisciplinary team. RESULTS 233 patients were recruited across five sites: 123 consecutive and 110 at clinical team discretion. In the consecutive cohort 50.6% (62/123) cases were classified high risk of recurrence by EndoPredict®, compared with 62.7% (69/110) in the selective cohort. A change in treatment recommendation was significantly more likely (p

Genomic tests improve accuracy of risk prediction for early breast cancers but these are expensive. This study evaluated the clinical utility of EndoPredict®, in terms of impact on adjuvant therapy recommendations and identification of parameters to guide selective application.Patients with ER-positive, HER2-negative, and early-stage invasive breast cancer were tested with EndoPredict®. Two cohorts were recruited: one consecutively and another at clinical team discretion. Systemic treatment recommendations were recorded before and after EndoPredict® results were revealed to the multidisciplinary team.233 patients were recruited across five sites: 123 consecutive and 110 at clinical team discretion. In the consecutive cohort 50.6% (62/123) cases were classified high risk of recurrence by EndoPredict®, compared with 62.7% (69/110) in the selective cohort. A change in treatment recommendation was significantly more likely (p 0.0001) in the selective cohort (43/110, 39.1%) compared to the consecutive group (11/123, 8.9%). The strongest driver of selective recruitment was intermediate grade histology, whilst logistic regression modelling demonstrated that nodal status (p 0.001), proliferative rate (p = 0.001), and progesterone receptor positivity (p 0.001) were the strongest discriminators of risk.Whilst molecular risk can be predicted by traditional variables in a high proportion of cases, EndoPredict® had a greater impact on treatment decisions in those cases selected for testing at team discretion. This is indicative of the robust ability of the clinical team to identify cases most likely to benefit from testing, underscoring the value of genomic tests in the oncologists' tool kit.