Background/aims Previously, we found a strong structure–function relationship between Octopus visual field cluster sensitivity and corresponding sector retinal nerve fibre layer thickness (RNFLT) values, using normal strategy and 16 custom-made cluster–sector pairs. However, Octopus commercially offers 10 clusters, correction for diffuse defect (corrected clusters) and tendency-oriented perimetry (TOP) strategy. In the current investigation, we evaluated structure–function relationship for these conditions, respectively.

Methods Twenty-four healthy, 45 ocular hypertensive and 63 glaucoma eyes underwent Octopus perimetry (normal and TOP strategies) and RNFLT measurements on the same day. Pearson's correlation between corresponding cluster defect and sector RNFLT values was investigated.

Results Correlation between the corresponding cluster defect and sector RNFLT values was significant (p<0.0001) for all clusters, with both test strategies. The r values ranged from −0.3529 and −0.3877 to −0.5798 and −0.5893, respectively. The highest r values were found for the inferotemporal RNFLT sector superior and superior paracentral cluster pairs with both strategies. No difference in correlation was seen between the strategies (p≥0.183). For corrected clusters, significant correlation (p<0.01) was seen only for two inferotemporal and one superotemporal RNFLT sectors, with both strategies (r values: −0.2446, −0.2375; −0.3526, −0.3119; −0.2392, −0.2340, respectively).

Conclusions Octopus normal and TOP strategies provide similar, moderate-to-strong negative correlation between the corresponding cluster defect and sector RNFLT values for all 10 manufacturer-provided clusters. For corrected clusters the relationship is poor, which suggests that by correcting for diffuse sensitivity loss some local glaucomatous defect is removed. For structure-function based clinical decisions, use of TOP is possible but use of corrected clusters is not recommended.