目的：通过节细胞内网状层(GCIPL)评估中度和重度青光眼的损伤程度并比较其与视盘周围视网膜神经纤维层(PRNFL)的诊断效能，包括敏感性与特异性及ROC曲线下面积(AUC)。

方法：前瞻性研究。共210眼(包括中度青光眼患者30例54眼，重度青光眼患者34例59眼和正常人50例97眼)纳入本研究。所有参与者均接受全面眼科检查，包括视野检查、3D-OCT视盘检查和3D-OCT黄斑部垂直扫描。记录并比较所有参与者的GCIPL和PRNFL的AUC、敏感性与特异性。

结果：在中度和重度青光眼组中，上部、下部及整体GCIPL和PRNFL厚度明显变薄(P<0.001)。在中度青光眼组中，GCIPL的敏感度与特异性高于PRNFL(仅上半部分敏感性相同)。重度青光眼组，上部、下部及整体GCIPL的敏感度均低于PRNFL。整体GCIPL的特异性低于PRNFL。上部GCIPL的特异性高于PRNFL。下部GCIPL的特异性与PRNFL相同。

结论：对于区分中度和重度青光眼，黄斑GCIPL参数的功能远高于PRNFL。两者联合在病情分析中效果最优，能够提供更准确的损伤程度评估。

AIM:To evaluate the extent of damage caused by moderate and severe glaucoma on ganglion cell inner plexiform layer(GCIPL)and to evaluate the diagnostic performance of this layer compared to the peripapillary retinal nerve fiber layer(PRNFL). This was performed by comparing their area under the curve(AUC)sensitivity and specificity.

METHODS:This study is a prospective study. Two hundred ten eyes(54 eyes of 30 moderate glaucoma subjects, 59 eyes of 34 severe glaucoma subjects and 97 eyes of 50 normal subjects)were enrolled in this study. Patients underwent complete ophthalmic examination, visual field(VF)examination and also 3D optical coherence tomography(OCT)of the disc and 3D vertical(V)OCT of the macula were performed. The GCIPL and PRNFL AUC, sensitivity and specificity were performed and compared.

RESULTS:A significantly thinner superior, inferior and total GCIPL and PRNFL thickness in moderate and severe glaucoma groups was detected(all P<0.001). In moderate glaucoma, GCIPL showed higher sensitivity and specificity than PRNFL(only the superior half shows equal sensitivity). As regard severe glaucoma the total, superior and inferior sensitivities of the GCIPL were lower than the PRNFL. The total GCIPL specificity was lower than the PRNFL. The superior GCIPL specificity was higher than the PRNFL. The lower GCIPL specificity was equal to that of the PRNFL.

CONCLUSION:The ability of the macular GC/IPL parameters to discriminate moderate and severe glaucoma is high and comparable to that of the PRNFL. A combination of both in the baseline evaluation is optimal and provides more accurate assessment of the extent of damage.