发表在23年N Engl J Med的一篇鼻咽癌筛查队列研究，其癌症样本估计基于敏感性统计学假设采用了两个率比值的非劣效检验，而对照组样本估计基于特异性McNemar Test，以下是原文：

Sample size estimation The sample size was estimated by PASS 2011 (NCSS, Kaysville, Utah, USA). The estimated sensitivity and specificity of P85-Ab was 90% and 98.0%, respectively. The sensitivity of EBNA1-IgA and VCA-IgA was between 75% to 95%, while the specificity was between 80% to 92%.

Non-inferiority test for the ratio of two correlated proportions was used to estimate the sample size for sensitivity comparison, in which α and power was set as 0.025 and 0.8, non-inferiority ratio (R0) and actual ratio (R1) was set as 0.8 and 1.0, respectively, **standard proportion (Ps) was set as 0.75 to 0.95,** while **nuisance (sensitivity) was set as 0.90 to 0.95**, respectively. According to the estimation, at least 33 to 57 samples from NPC cases should be collected.

Test for two correlated proportions (McNemar Test) was used to estimate the sample size for higher specificity of P85-Ab, in which α and power was set as 0.05 and 0.8, **difference (P10- P01) was set as 0.006 to 0.18, and Proportion discordant (P10+P01) was set as 0.10 to 0.22**. According to the estimation, at least 54 to 231 samples from healthy controls should be collected.

请教各位专家：

1. 为什么癌症样本量和对照组样本量估计采用不同的检验方法？

2. 在估计癌症组样本量时，参数standard proportion (Ps)和nuisance (sensitivity) 分别是什么意思？其赋值应如何确定？

3. 在估计对照组样本量时，参数difference (P10- P01) 和Proportion discordant (P10+P01) 分别是什么意思？其赋值应如何确定？

期待解答，非常感谢!