A team of researchers in Germany has developed a new assay to help determine which non-small cell lung cancers are likely to respond to certain targeted therapies.
As a screen for gene rearrangements, the assay uses quantitative reverse transcription-PCR to show altered ratios of 3' to 5' ends of the aplastic lymphoma tyrosine kinase, or ALK gene, rather than specific mutations. It is also optimized to work on formalin-fixed paraffin-embedded tissue samples.
The researchers, hailing from the Dr. Margarete Fischer-Bosch Institute of Clinical Pharmacology (IKP), Robert Bosch Hospital, and other German institutes, published details of the assay this month in The Journal of Thoracic Oncology. In the study they tested the assay against fluorescent in situ hybridization, the gold standard test for ALK rearrangements, and found that it accurately typed 97 percent of tumors and also identified ALK deregulation in two cases where FISH was insufficient.
Crizotinib, an ALK inhibitor marketed by Pfizer as Xalkori, is highly effective on lung cancers containing ALK rearrangements. As these constitute less than five percent of cases, according to a 2012 study in Frontiers in Oncology, reliably determining which patients should receive this therapy is crucial.
As recently as 2013, testing guidelines published in The Journal of Thoracic Oncology recommended ALK testing by FISH assay evaluated by a pathologist. The authors wrote "RT-PCR is not recommended as an alternative to FISH for selecting patients for ALK inhibitor therapy." This recommendation was due to "concerns for a higher failure rate of an RNA-based assay in routine FFPE pathology material, and the risk of false negatives, owing to variability in the EML4-ALK fusion structure and the existence of other ALK fusion partners," according to the paper.
However, a few recent studies have revealed some patients' cancers responded to crizotinib despite negative FISH or IHC, and others have show heretofore unknown rearrangement patterns as reported in PGx Reporter.
Although RT-PCR can be complicated by the large number of fusion variants and inability to detect previously unknown partners, the authors of the current study cited three recent assays that detect unbalanced ALK transcript expression, rather than a set of specific fusion transcripts, allowing detection of ALK rearrangements regardless of a specific fusion partner. However, they said these were either not suited for FFPE, or relied on exon array or NanoString's nCounter platform, which are not commonly available technologies in pathology labs.
The German group's new assay is similar to one published in 2012 in Clinical Cancer Research, in which researchers at Fudan University in Shanghai developed a qRT-PCR assay for the 5' and 3' portions of ALK transcripts. This assay was for use on frozen tissue samples, however, not FFPE.
To validate the qRT-PCR method in FFPE, the German researchers screened specimens from 652 NSCLC patients. Eighty percent of these had RNA of sufficient quality based on amplification of the control gene PGK1. Of these, about 5 percent showed ALK 3' expression and non-expression of the 5' amplicon, 1 percent expressed the full transcript, and 94 percent, or 493 patient samples, did not significantly express ALK. This is in line with the 5 percent expected to have ALK alterations.
IKP's Claudia Kalla, lead author on the study, told PCR Insider via email that while there are many commercially available FFPE nucleic acid sample prep kits, her group tested three different RNA isolation methods to find the one that worked most efficiently for their sample set.
She said, the AllPrep DNA/RNA FFPE kit from Qiagen "turned out to be the most efficient with respect to the quality and amount of tumor material needed for positive amplification results." To determine this, she said they tested a series of FFPE samples and amplicons of increasing size. The simultaneous isolation of RNA and DNA was a strong argument for the Qiagen kit, she said, and "it opens up the possibility to combine DNA mutation and RNA expression analysis and, thus, to perform complex investigations even of small biopsies," Kalla said.
Kalla added the caveat that not all FFPE conditions are identical in all pathology labs. Furthermore, expanding this test to a clinical diagnostic will require some collaborative effort, since the Qiagen kit is currently for research use only.
In the assay, PGK1 gene expression was used to normalize the data, and an ALK-expressing rhabdomyosarcoma cell line served as the positive control. Another recent study reviewed commonly used internal control genes in lung cancer PCR studies, concluding PGK1 was too varied in its expression. But Kalla pointed out that study was of squamous cell carcinoma, not NSCLC. She said her group chose their control gene carefully; they assessed five genes to arrive at the best one, since "one reference gene is rather [more] practicable than three in a compact test assay that should work with low cDNA amounts." Kalla said, "Although I am aware that PGK1 will not be the perfect control for each NSCLC sample it seems to be a good compromise: qRT-PCR using this control accurately typed 97 percent of 19 rearranged and 158 nonrearranged tumors in comparison with FISH."
Insight Genetics has commercialized a qRT-PCR assay for ALK which "detects both known and unknown ALK fusions as well as over-expression of full-length ALK" and is specific for the kinase region, according to the company's website. It also measures RNA from FFPE. As reported in PCR Insider, this test was initially licensed from St. Jude's hospital, and is now being performed in the company's recently certified CLIA lab.
Kalla said she recently became aware of this test, but couldn't say how similar her group's assay is to Insight's. "I don't know the technical details of [their] test … I assume that the general principle will be the same and speculate that possible differences could be the use of another control system, [such as the] reference gene to control RNA quality, RH30 as positive PCR control."
Kalla said her group is currently analyzing a prospective series of samples, and "If the results are satisfying, we will decide whether to use it clinically."