INNV-19. International multi-centre clinical performance evaluation of a brain cancer liquid biopsy
Delayed diagnosis of brain cancer leads to two-thirds of patients receiving a diagnosis only after presenting to the emergency department with more severe symptoms or neurological deficits. A simple, rapid, liquid biopsy implemented in primary care could enable more efficient triage of patients with non-specific symptoms potentially related to brain cancer, prioritising patients for urgent brain imaging, accelerating diagnosis. We present the international, multi-centre, observational Early and tiMely detection of BRAin CancEr (EMBRACE) study. Patients were prospectively recruited across seven sites in the United Kingdom, Belgium, Sweden and Switzerland. Target population was patients with symptoms potentially associated with brain cancer. Blood serum samples were analyzed by the Dxcover® Brain Cancer Liquid Biopsy Platform. Test performance was assessed by comparison of the liquid biopsy result to diagnostic imaging. The primary objective was to determine clinical performance of the test for patients with suspected brain cancer, reporting diagnostic sensitivity and specificity. 2554 patients were enrolled across the seven collection sites; 2345 were deemed eligible and taken forward for test assessment. There were 718 brain tumours in total, of which 399 were malignant, and 1627 non-tumour diagnoses. Overall diagnostic performance for the primary objective was 86% sensitivity for brain cancer detection with 99% negative predictive value (NPV). Sensitivity for all brain tumours combined (malignant and benign) was 76%. Notably, for the most prevalent and most aggressive brain cancer, glioblastoma, 86% of cases were successfully identified. Additionally, 94% of central nervous system lymphoma, 90% of brain metastases were predicted correctly. Existing symptom-based referral pathways are ineffective for the detection of brain cancer. There is an urgent need for new tests, like this liquid biopsy, to help with clinical decision making and efficiently stratify patients towards diagnostic imaging.