DIAGnosis using NOvel technology for Subtypes In Stroke (DIAGNOSIS study): performance evaluation protocol.

INTRODUCTION

Timely and accurate identification of stroke subtypes, particularly intracerebral haemorrhage (ICH) and large vessel occlusion (LVO), is essential for improving stroke outcomes. Biomarkers have shown potential in identifying these subtypes. Glial fibrillary acidic protein (GFAP), which is elevated in ICH, and D-dimer, which is associated with LVO, are promising for rapid, point-of-care identification. This study evaluates whether combining GFAP and D-dimer with clinical features improves the diagnostic accuracy for ICH and LVO.

METHODS AND ANALYSIS: This is a multicentre, observational diagnostic accuracy study recruiting a minimum of 257 patients with suspected stroke within 6 hours of symptom onset from two UK hyperacute stroke centres. The study will integrate clinical features and point-of-care biomarker results using the LVOne test (GFAP and D-dimer) to develop diagnostic models that differentiate ICH and LVO from other stroke presentations. Patients will be identified on hospital arrival, and clinical variables and blood samples collected immediately, with the final diagnosis recorded prior to discharge. Statistical analyses will assess discrimination, calibration and clinical utility, with sensitivity, specificity and predictive values reported. Model development and reporting will follow the Transparent Reporting of a Multivariable Prediction Model for Individual Prognosis or Diagnosis+Artificial Intelligence and the Standards for Reporting of Diagnostic Accuracy Studies guidelines.

ETHICS AND DISSEMINATION: The study is registered with ISRCTN (ISRCTN35174477). Ethical approval has been obtained from the Yorkshire & The Humber - Leeds West Research Ethics Committee (25/YH/0102) and the Health Research Authority. Results will be disseminated via open-access peer-reviewed journals, presentations at national and international conferences and summaries for patient and public contributors.