PT - JOURNAL ARTICLE AU - C I Clement AU - I Goldberg AU - P R Healey AU - S Graham TI - Humphrey matrix frequency doubling perimetry for detection of visual-field defects in open-angle glaucoma AID - 10.1136/bjo.2007.119909 DP - 2009 May 01 TA - British Journal of Ophthalmology PG - 582--588 VI - 93 IP - 5 4099 - http://bjo.bmj.com/content/93/5/582.short 4100 - http://bjo.bmj.com/content/93/5/582.full SO - Br J Ophthalmol2009 May 01; 93 AB - Aim: Matrix perimetry uses frequency-doubling technology (FDT) incorporated into a 5° test target. This permits testing of the same number of locations within a defined visual field as standard automated perimetry (SAP) and may improve performance compared with original FDT perimetry. This study investigates the performance of Humphrey Matrix perimetry for detecting glaucomatous visual-field loss.Design: Prospective case control study.Methods: We recruited 115 participants with glaucomatous visual-field loss and 33 normal controls from an urban glaucoma practice. Each participant performed SITA 24-2 SAP then threshold 24-2 Matrix perimetry. Severity of visual-field loss was defined using SAP mean deviation (MD) as early (MD >−6 dB), moderate (MD −6 to −12 dB) or advanced (MD <−12 dB). The sensitivity and specificity of Humphrey Matrix perimetry were calculated for different automated indices.Results: The matrix perimetry sensitivity and specificity were up to 100% for moderate and advanced glaucomatous visual-field loss. A receiver operator characteristic area under the curve (AUC) analysis revealed MD to be slightly better than pattern standard deviation (PSD) for defining moderate (AUC: MD 0.997; PSD 0.987) and advanced defects (AUC: MD 1.000; PSD 0.987). Matrix was less sensitive (up to 87.3%) for detecting early glaucomatous visual-field loss compared with SITA 24-2 SAP (AUC: PSD 0.948; MD 0.910).Conclusions: Matrix perimetry is excellent for detection of moderate to advanced glaucomatous visual-field loss but may miss some early defects. It may be well suited to following progression of early to moderate field loss because of a smaller target size compared with original FDT perimetry.