Da Pozzo Stefano, Iacono Pierluigi, Marchesan Roberta, Fantin Anna, Ravalico Giuseppe
Eye Clinic, Ospedale Maggiore, University of Trieste, Piazza dell'Ospedale 1, 34129 Trieste, Italy.
Graefes Arch Clin Exp Ophthalmol. 2005 Aug;243(8):774-9. doi: 10.1007/s00417-004-1118-1. Epub 2005 Mar 9.
The aim of this study was to evaluate the ability of scanning laser polarimetry (SLP) parameters provided by commercially available GDx with variable corneal compensator (VCC) to discriminate between healthy and glaucomatous eyes.
Sixty-five healthy and 59 glaucomatous age-matched patients underwent a complete ophthalmological evaluation, an achromatic automated perimetry (AAP), and SLP with GDx-VCC. One randomly selected eye from each subject was considered. All glaucomatous eyes had reproducible visual field defects. Mean values (+/- SD) of all SLP-VCC parameters measured in the two groups were compared. Area under receiver operating characteristics (AUROC) curve and sensitivities at predetermined specificities of >or=80% and >or=95% for each single parameter were calculated. Moreover, the nerve fiber indicator (NFI) diagnostic accuracy was evaluated calculating positive, negative, and interval likelihood ratios (LRs) at different cutoff values.
All SLP parameters were significantly different between the two groups (p<0.001). The NFI showed the best AUROC curve (0.938, SE 0.02) whereas temporal, superior, nasal, inferior, temporal (TSNIT) average was second best (0.897, SE 0.03), and normalized superior area was third (0.879, SE 0.04). At fixed specificity >or=95%, sensitivities ranged from 22% to 79.7% whereas for values >or=80%, sensitivities were in the 44.1-89.8% range. At a cutoff NFI value of 30, positive LR was 17.6 (95% CI: 5.8-53.6) and negative LR was 0.19 (95% CI: 0.11-0.33). Interval LRs for NFI showed that values <or=20 or >40 were associated with large effects on posttest probability.
SLP-VCC allows good discrimination between healthy and glaucomatous eyes. New software-provided parameters NFI, TSNIT average, and normalized superior and inferior areas appear to be reliable in the evaluation of glaucomatous disease. In particular, after evaluation on interval LRs, the NFI showed a high diagnostic accuracy for values <or=20 or >40.
本研究旨在评估配备可变角膜补偿器(VCC)的商用GDx提供的扫描激光偏振仪(SLP)参数区分健康眼和青光眼眼的能力。
65名年龄匹配的健康受试者和59名青光眼患者接受了全面的眼科评估、消色差自动视野计检查(AAP)以及使用GDx-VCC的SLP检查。从每个受试者中随机选择一只眼睛进行分析。所有青光眼患者的视野缺损均具有可重复性。比较两组测量的所有SLP-VCC参数的平均值(±标准差)。计算每个参数在预定特异性≥80%和≥95%时的受试者操作特征曲线下面积(AUROC)和敏感度。此外,通过计算不同临界值下的阳性、阴性和区间似然比(LRs)来评估神经纤维指数(NFI)的诊断准确性。
两组之间所有SLP参数均有显著差异(p<0.001)。NFI的AUROC曲线最佳(0.938,标准误0.02),而颞上鼻下平均厚度(TSNIT)次之(0.897,标准误0.03),归一化上半面积排第三(0.879,标准误0.04)。在固定特异性≥95%时,敏感度范围为22%至79.7%,而在≥80%时,敏感度在44.1%-89.8%范围内。当NFI临界值为30时,阳性似然比为17.6(95%置信区间:5.8-53.6),阴性似然比为0.19(95%置信区间:0.11-0.33)。NFI的区间似然比显示,≤20或>40的值与对检验后概率有较大影响相关。
SLP-VCC能够较好地区分健康眼和青光眼眼。新软件提供的参数NFI、TSNIT平均厚度以及归一化上半和下半面积在青光眼疾病评估中似乎是可靠的。特别是,在评估区间似然比后,NFI在≤20或>40时显示出较高的诊断准确性。