Sample size calculation and re-estimation based on the prevalence in a single-arm confirmatory diagnostic accuracy study