In situhybridization is the only DNA- or RNA-based molecular biology based test that allows for the direct correlation of the results with the histologic and cytologic features of the sample. The DNA/RNA extraction that precedes filter hybridization (slot blot, hybrid capture, or Southern blot) and polymerase chain reaction (PCR) precludes this analysis. The relative sensitivities of the three different assays are presented in Table 1 . It is evident thatin situhybridization is a relatively insensitive test. A reflection of this relative insensitivity is seen in occult or latent infection by a virus where the copy number is low. In such situations, the virus is rarely detected byin situhybridization even though it was detected by either PCR or filter hybridization (1 -6 ). This is not to say that the technique ofin situhybridization has remained static. The detection threshold of this assay has improved substantially in the last 10 years. Another point worth emphasizing aboutin situhybridization is that one does not need to use radiolabeled probes (usually 35 S or 3 H) to maximize its sensitivity. Although true 10 years ago, recent and dramatic advances in nonisotopic labeling and, more importantly, detection systems has greatly enhanced the sensitivity when using such common labels as biotin and digoxigenin (7 -12 ). Still, only the most aggressive salesman would claim (and incorrectly at that) that any givenin situsystem can routinely detect one DNA or RNA copy per cell. In my experience, this statement applies to the newer generation of posthybridization “signal amplification systems” (such as the cascade amplification system) which are not able to routinely detect one copy per cell.
