Specificity of the program threshold noiseless trend for perimetric progression analysis

To evaluate the specificity of the Threshold Noiseless Trend program (TNT), designed to measure progression in visual fields, using four procedures.

A. In 63 eyes with ocular hypertension, whose last examination showed no perimetric or morphological defects, we performed a mean of 7.70 +/- 1.71 follow-up examinations during 2.2 +/- 0.6 years. B. In 81 glaucomatous eyes examined twice with a bracketing strategy (Octopus 1-2-3 perimeter), we calculated mean threshold value and long-term fluctuation. We simulated 12 different visual fields, adding a random component to simulate an equivalent fluctuation of amplitude. C. Seventy-two glaucomatous eyes, with and without progression, were examined 7.76 +/- 1.25 times during 4.88 +/- 1.39 years using the Humphrey-Sita Standard strategy. Visual field tests were randomly disordered and analyzed using TNT. D. 1221 eyes were examined 7.19 +/- 3.5 times during 3.50 +/- 1.45 years (10,407 visual fields) using TOP-G1 program. We detected progression in 204 eyes using TNT. They were re-evaluated after random disordering of visual field tests.

The four procedures indicated specificity > or = 95% as from the seventh examination; this reduced to 90% in experiment C with six examinations, and in experiments A and C with five examinations.

The specificity of TNT may be considered to be over 95% with a large number of examinations, and 90-100% with fewer examinations. At least five examinations are required for a basic interpretation of progression, and preferably more than six to guarantee the specificity of the result.