SAN DIEGO — St. Jude Children's Research Hospital is developing a clinical sequencing pipeline in which it plans to enroll every incoming pediatric oncology patient. James Downing, scientific director of St. Jude, discussed the hospital's plans at the American Association for Cancer Research's annual meeting in San Diego, Calif., this week.
The clinical pipeline will include whole-genome sequencing, exome sequencing, and transcriptome sequencing of the patient's tumor genome to look for somatic alterations, as well as whole-genome and exome sequencing of normal tissue to look for germline mutations. The researchers will return data on somatic alterations from 567 cancer-related genes as well as on germline mutations from 31 cancer predisposition genes and 22 non-cancer disease genes that will be put into patients' electronic medical records.
The development of the clinical sequencing pipeline stems from the Pediatric Cancer Genome Project, a $65 million project that began in 2010 with the aim of sequencing tumor/normal genome pairs of pediatric cancer patients. St. Jude collaborated with the Genome Institute at Washington University and the project was completed in February 2013. It included whole-genome sequencing of 700 matched tumor/normal pairs as well as 2,000 exomes from 21 different types of pediatric cancers.
"Every pediatric cancer we've sequenced has made fundamental insights into the genetics that are driving that cancer," Downing said.
The second phase of the Pediatric Cancer Genome Project is a $30 million, two-year initiative that includes two parts: a discovery component focused on the epigenome and a clinical component that will include whole-genome, exome, and transcriptome sequencing of the tumors, as well as whole-genome and exome sequencing of the germline of every pediatric cancer patient that is seen at St. Jude, about 400 per year.
Downing said the hospital would launch the pipeline in the summer of this year. Already, it has completed an 80-patient pilot study to validate the tumor sequencing protocol as well as a 1,120-patient pilot study to validate the germline analysis.
In the first, 80-patient pilot study, researchers sequenced patients' tumor and normal genomes, and 53 of these patients had also been sequenced as part of the initial Pediatric Cancer Genome Project, which allowed the researchers to "compare to the original sequencing."
Additionally, the fist pilot allowed the team to "assess our ability to identify the diagnostic lesions currently screened for in the molecular diagnostic laboratory," Downing said, such as BCR-ABL1, TEL-AML1, E2A-PBX1, and MLL rearrangements for leukemia risk stratification. This gave the team confidence that next-gen sequencing approaches could "allow us to replace many of the diagnostic approaches we use to stratify patients." Moving forward, Downing said the group will continue to study whether it makes sense to replace some current tests with sequencing.
Following that pilot, the team conducted another pilot to test their ability to analyze germline mutations. From 1,120 pediatric cancer cases they sequenced the whole genomes of 700 patients and the whole exomes of 420 patients, to "determine how we will handle this kind of data," Downing said.
The researchers also assessed the frequency of germline mutations in cancer predisposition genes in their cohort and found "somewhat surprisingly low numbers," Downing said. Looking at the 31 cancer predisposition genes, nearly 8 percent, or 89 of the 1,120 cases, had a pathogenic or likely pathogenic lesion in one of those genes. Not surprisingly, the most commonly mutated gene was TP53, which was found in just under 50 patients.
St. Jude has also developed infrastructure to enable some of the data to be put into patients' electronic medical reports, including known pathogenic somatic alterations from 567 cancer genes and known pathogenic germline mutations in 31 cancer predisposition genes and 22 non-cancer disease genes. "We are only reporting [germline] alterations of genes where there are therapeutic or preventative interventions," Downing said. He said that the team would also be assessing additional genes on a quarterly basis. "As new insights are gained on the clinical relevance of specific gene mutations and this becomes incorporated into clinical practice, we will add any results on these genes to patient’s medical record by issuing addendums," he said, for any patient that has consented to receiving the new information.
He added that in developing the pipeline, the researchers quickly realized that many public databases of gene variants are "riddled with inaccuracies and errors," and as such St. Jude's has developed its own database that includes data from the Pediatric Cancer Genome Project as well as data from other databases. "We now have an automated way to classify mutations," he said.
While much of the variant classification process will be automated, Downing said that there will still be an analyst that looks at the classifications, and there will be a review of those findings to determine how to act on them. Downing also noted that there will be separate reports for somatic and germline mutations.
For the clinical sequencing pipeline, Downing said St. Jude now has the ability to do next-generation sequencing in its CLIA certified laboratory. He said they have six Illumina instruments — three HiSeq 2500s, two HiSeq 2000s and one MiSeq. The laboratory will be purchasing an additional two Illumina systems in the next several weeks and is considering purchasing the NextSeq 500 as one of its next systems.
Initially, the sequence data will not be used for clinical decision making, but to study the impact of sequencing in terms of how physicians handle the data; whether comprehensive genomics uncovers valid data; and the best ways to incorporate sequence data into the clinical management of pediatric cancer. Downing said it would likely be about a year before the data would be used on a regular basis to guide treatment. Although, he said that in the cases of refractory or relapse patients, the data could be used to guide treatment.
The clinical sequencing project will be another layer on top of the other research-directed protocols that children enter into when they are admitted at St. Jude, Downing said. Patients will be consented upon admission to the sequencing, the reporting of pathogenic variants in both the germline and tumor genomes, as well as the use of their genomic data for research at St. Jude.
Downing also stressed the importance of doing a comprehensive approach of whole-genome, exome, and transcriptome analysis. "It requires that in order to be accurate," he said. "We need that triple analysis." Even though the mutations that will be reported and put into the medical records will primarily be coding mutations, "our discovery efforts are focused on non-coding regions," he said. "We're moving this into the clinic, but always translating back into the biology."