The standard test, test B, for vision impairment in children is 65% sensitive and 80% specific. It is anticipated that a new, more sophisticated test, test A, will be more specific (95%). In order to proceed with further development of the test, the study investigators have decided that they would need to draw the conclusion that test A is at least as test B. With significance level 0.05 and power 0.90, how can I calculate the required sample size for patients with vision impairment.

(Upper 0.025, 0.05, 0.1 quantiles of the standard normal distribution are 1.96, 1.64 and 1.28)