To create a method to identify the edge of visual field (VF) scotoma and to investigate the relationship between the steepness of the ‘border’ of scotoma and the prediction error of VF sensitivity.Methods
Twenty-two glaucomatous patients with reliable VFs were recruited. Then, using the latest VFs (Humphrey 24-2) of these patients, the gradient of the plane on the hill of vision is calculated from the sensitivity of adjacent four/three points, so that the ‘border’ of VF damage is identified. Next, VF measurement was carried out, adding ten test points at the centre of adjacent four/three points where the gradient of the plane is largest (Full threshold, Custom mode). The VF measurements were performed using two approaches; random target presentation of 62 test points together and showing ten added points following or prior (randomly selected) to 52 (24-2) points. Each measurement with each approach was repeated twice in a same visit. Then, the absolute value of the difference between the measured sensitivity of the added ten test points and the average of the sensitivities of surrounding three/four test points were calculated. Finally, the relationship between the gradient of sensitivity plane and the absolute difference was investigated using the multilevel modeling (MLM).Results
MLM revealed significant positive relationship between the absolute difference and the gradient of plane with all of the four measurements (p<0.01).Conclusion
It may be advantageous to increase the spatial information by carrying out additional measurement at the gap where the border of scotoma is steep.