我记得中国从核酸检验确诊改为CT确诊是因为等待核酸结果太久了（好几天），而且试剂盒数量也远远不够。还没出结果，很多人已经不行了，出结果之前不隔离，几天内可以传很多人，尤其是医护人员，所以还不如CT看出了疑似就隔离，那也是不得已的措施。那时候，远观者（包括我）还没明白有很多无症状或轻微症状的感染者继续传播病毒，但一线医生已经知道了。所以二月中旬开会的时候UF Emerging Pathogens 研究所的教授很肯定地说有 community transmission，靠隔离是堵不住的。
Nature Biotechnology 上这篇报道对比 PCR 和抗体测试，很详细：
Fast, portable tests come online to curb coronavirus pandemic
by Cormac Sheridan
23 MARCH 2020
还有用 CRISPR 技术的测试方法在研究中。Neither PCR nor immunoassay technologies are ideal, however. PCR tests are highly accurate and can be developed at speed, but they are complex to use and slow to deliver a result. Immunoassays are less accurate and take longer to develop, but they are easy to use and deliver results in 20–60 minutes. And because immunoassays detect patient antibodies to a pathogen, these inevitably must contend with the inherent variability of the polyclonal human antibody response. “That antibody response takes time to characterize,” says Ranga Sampath, chief scientific officer at FIND. Results obtained during validation testing, therefore, may not necessarily be replicated in clinical settings. “These tests tend not to be definitive — that’s the biggest challenge with immunoassays,” he says.
Mammoth Biosciences and Sherlock Biosciences, companies cofounded by CRISPR pioneers Jennifer Doudna and Feng Zhang, respectively, both aim to match its accuracy while achieving genuine POC testing speeds by employing the specificity and sensitivity of CRISPR–Cas editing to develop rapid POC SARS-CoV-2 tests.