Gene Expression Detection Assay for Cancer Clinical Use

Narrandes, Shavira; Xu, Wayne

Although qRT-PCR tests have shown consistent results across replicate assays, they are not 100% reproducible as errors may be amplified along with the exponential amplification of the original nucleotide strand and the PCR can result in some loss of information during amplification. 82 Expression microarrays have also been shown to be highly reproducible, depending on the ability of a probe to anneal to the same number of transcripts across replicate experiments. The protocols used for RNA extraction and sample purification, data acquisition and normalization, type and construction of probes, and the methods of labelling and hybridization used greatly affect the reproducibility of DNA microarrays. 83 , 84 Numerous studies have confirmed the reproducibility of the NanoString nCounter platform due to its requirement of little input RNA and lack of a reverse transcription step. For example, its reproducibility has been demonstrated using diffuse B-cell lymphoma (DLBCL), gastric, pancreas, xenograft, lung, breast, blood, and melanoma samples, while a separate study demonstrated inter-laboratory (Canada, Switzerland, and the United Stated) reproducibility of nCounter with medulloblastoma samples. 42, 85 , 86 Similarly, studies have used techniques such as PCR, qRT-PCR, and microarrays to show the MiSeq platform and RNA-Seq to be reproducible. 87 , 88 , 89 , 90 For example, the SEQC/MAQC-III Consortium have demonstrated RNA-Seq to have a >80% validation with quantitative PCR (qPCR) when sequencing unannotated exon-exon junctions, yet microarrays do not provide reliable reproducibility data. 89 One of the greatest drawbacks of TMAs is the tumor inter-heterogeneitythe fact that a single spot punched from a tumor block is not representative of the rest of the tumor. Regardless, TMAs have a high degree of reproducibility due to the maximum 1000 spots on a single slide simultaneously being sampled and subjected to an experimental test using the same reagents. Multiple punches can be sampled from one tumor block, allowing for the same tumor block to be used in multiple experiments; for example, TMAs from one recipient block can be distributed between labs. Immunohistochemical and FISH tests can then be standardized for reproducible results. 91